Neurophysiological Basis of

Motor Control

Third Edition

Mark L. Latash, PhD

The Pennsylvania State University

Tarkeshwar Singh, PhD

The Pennsylvania State University
 Library of Congress Cataloging-in-Publication Data
Names: Latash, Mark L., 1953- author. | Singh, Tarkeshwar, 1981- author.
Title: Neurophysiological basis of motor control / Mark L. Latash, PhD, The
Pennsylvania State University, Tarkeshwar Singh, PhD, The Pennsylvania State
University.
Description: Third edition. | Champaign : Human Kinetics, [2024] |
Neurophysiological basis of movement / Mark L. Latash. 2nd ed. c2008. |
Includes bibliographical references and index.
Identifiers: LCCN 2022022940 (print) | LCCN 2022022941 (ebook) | ISBN
9781718209527 (paperback) | ISBN 9781718209534 (epub) | ISBN
9781718209541 (pdf)
Subjects: LCSH: Locomotion. | Neurophysiology. | Motor ability. | Movement
disorders.
Classification: LCC QP301 .L364 2023 (print) | LCC QP301 (ebook) | DDC
612.7/6--dc23/eng/20220720
LC record available at https://lccn.loc.gov/2022022940
LC ebook record available at https://lccn.loc.gov/2022022941
ISBN: 978-1-7182-0952-7 (print)
Copyright © 2024 by Mark L. Latash and Tarkeshwar Singh
Copyright © 2008, 1998 by Mark L. Latash
Human Kinetics supports copyright. Copyright fuels scientific and artistic endeavor,
encourages authors to create new works, and promotes free speech. Thank you
for buying an authorized edition of this work and for complying with copyright laws
by not reproducing, scanning, or distributing any part of it in any form without
written permission from the publisher. You are supporting authors and allowing
Human Kinetics to continue to publish works that increase the knowledge,
enhance the performance, and improve the lives of people all over the world.
To report suspected copyright infringement of content published by Human
Kinetics, contact us at permissions@hkusa.com. To request permission to
legally reuse content published by Human Kinetics, please refer to the information
at https://US.HumanKinetics.com/pages/permissions-information.
The web addresses cited in this text were current as of June 2022, unless
otherwise noted.
Acquisitions Editor: Diana Vincer; Managing Editor: Melissa J. Zavala;
Copyeditor: Kevin Campbell; Proofreader: Pamela Johnson; Permissions
Manager: Laurel Mitchell; Graphic Designer: Dawn Sills; Cover Designer: Keri
Evans; Cover Design Specialist: Susan Rothermel Allen; Photograph (cover):
SCIEPRO/Getty Images; Photographs (interior): © Human Kinetics, unless
otherwise noted; Photo Asset Manager: Laura Fitch; Photo Production
Manager: Jason Allen; Senior Art Manager: Kelly Hendren; Illustrator: Mark L.
Latash, unless otherwise noted; Printer: Versa Press
Printed in the United States of America 10 9 8 7 6 5 4 3 2 1
The paper in this book is certified under a sustainable forestry program.
Human Kinetics
1607 N. Market Street
Champaign, IL 61820
USA
United States and International
Website: US.HumanKinetics.com
Email: info@hkusa.com
Phone: 1-800-747-4457
Canada
Website: Canada.HumanKinetics.com
Email: info@hkcanada.com
E8444
 Contents
Preface

Introduction
Chapter 1 History, Evolution, and Motor Control
1.1 Brief History of
Movement Studies
1.2 Evolution of
Movements and Nikolai
Bernstein’s Theory
1.3 Motor Control and
Laws of Nature

Part I Excitable Cells and Their

Communication
Chapter 2 Membranes, Particles, and Equilibrium
Potentials
2.1 The Biological
Membrane
2.2 Movement in a
Solution
2.3 Concentration of
Water: Osmosis
2.4 Movement of Ions:
The Nernst Equation
Chapter 3 Action Potential
3.1 Creation of a
Membrane Potential
3.2 Basic Features of
the Action Potential
3.3 Mechanisms of
Generating an Action
Potential
Chapter 4 Information Conduction and
Transmission
4.1 Conduction of an
Action Potential
4.2 Myelinated Fibers
4.3 Structure of a
Neuron
4.4 Information Coding
in the Nervous System
4.5 Synaptic
Transmission
4.6 Neurotransmitters
4.7 Temporal and Spatial
Summation
Chapter 5 Skeletal Muscle
5.1 Skeletal Muscle
Structure
 5.2 Myofilaments
5.3 Neuromuscular
Synapse
5.4 Mechanisms of
Contraction
5.5 Types of Muscle
Contractions
5.6 Elements of
Mechanics
5.7 Force–Length and
Force–Velocity
Relations
5.8 External Regimes of
Muscle Contraction
Chapter 6 Peripheral Receptors
6.1 General
Classification and
Properties of
Receptors
6.2 Muscle Spindles
6.3 The Gamma-System
6.4 Golgi Tendon Organs
6.5 Other Muscle
Receptors
6.6 Articular Receptors
6.7 Cutaneous Receptors
 6.8 Signals From
Peripheral Receptors
Chapter 7 Motor Units and Electromyography
7.1 The Motor Unit
7.2 Fast and Slow Motor
Units
7.3 The Henneman
Principle
7.4 Functional Roles of
Different Motor Units
7.5 Electromyography
7.6 Processing
Electromyographic
Signals
Problems for Part I

Part II Neuroanatomical Foundations

of Motor Control
Chapter 8 Cerebral Cortex
8.1 Structure of the
Cerebral Cortex
8.2 Cells in the
Cerebral Cortex
8.3 Premotor Cortex and
Supplementary Motor
 Areas
8.4 Primary Motor Cortex
8.5 Efferent Output From
the Cortical Motor
Areas
8.6 Afferent Input Into
the Cortical Motor
Areas
8.7 Hemispheric
Lateralization in the
Cortical Motor Areas
8.8 Preparation for a
Voluntary Movement
8.9 Neuronal Population
Vectors
8.10 Encoding Movement
Parameters in the
M1
8.11 Brain–Machine
Interfaces
Chapter 9 Basal Ganglia
9.1 Anatomy of the Basal
Ganglia
9.2 Inputs and Outputs
of the Basal Ganglia
 9.3 Direct and Indirect
Pathways Within the
Basal Ganglia
9.4 Dopamine Modulation
of Basal Ganglia
Circuits
9.5 Motor Circuits
Involving the Basal
Ganglia
9.6 Activity of the
Basal Ganglia During
Movements
9.7 Movement Disorders
Associated With the
Basal Ganglia
9.8 Other Functions of
the Basal Ganglia
Chapter 10 Cerebellum
10.1 Overall Structure
of the Cerebellum
10.2 Inputs and Outputs
of the Cerebellum
10.3 Pathways Within the
Cerebellum
10.4 Distinct Cerebellar
Regions Control
 Discrete Motor
Functions
10.5 Cerebellar Control
of Movement
10.6 Consequences of
Cerebellar Lesions
on Movements
10.7 Cerebellar
Contribution to
Motor Learning
10.8 Cerebellar
Interactions With
the Basal Ganglia
and Cortex
Chapter 11 Brainstem and Extrapyramidal Tracts
11.1 Brainstem Anatomy
11.2 Reticular Formation
11.3 Superior Colliculus
11.4 Red Nucleus
11.5 Vestibular Nuclei
11.6 Cranial Nerves
11.7 Descending Tracts
Problems for Part II
Part III Sensory Basis of Motor
Control
Chapter 12 Central Processing of Somatosensory
Information
12.1 First-Order Neurons
12.2 Second-Order
Neurons
12.3 Third-Order Neurons
12.4 Proprioceptive
System
12.5 Primary and
Secondary
Somatosensory
Cortex
12.6 Integration of
Somatosensory Input
With Other Sensory
Modalities
12.7 Injuries to
Somatosensory
Pathways
Chapter 13 Vestibular and Auditory Systems
13.1 Transduction in the
Vestibular System
 13.2 Vestibular
Afferents Respond
to Head Motion
13.3 Central Projections
From the Otolith
Organs and
Semicircular Canals
13.4 Central Pathways
That Stabilize
Gaze, Posture, and
Head Movements
13.5 Peripheral Auditory
System
13.6 Central Auditory
Projections From
the Cochlea
13.7 Auditory
Integration
13.8 Auditory Thalamus
and Cortex
13.9 Auditory Cortex and
Limb Motor Control
Chapter 14 Visual System
14.1 Structure of the
Eye
 14.2 Structure of the
Retina
14.3 Rods and Cones
14.4 Optic Nerve,
Tracts, and
Radiations
14.5 Striate Cortex
14.6 Retinotopic
Organization of V1
14.7 Extrastriate Cortex
14.8 Neurons of the Two
Visual Streams
14.9 Visual Deficits Due
to Area-Specific
Visual System
Damage
14.10 Ocular Movements
Problems for Part III

Part IV Reflexes and Reflex-Like

Movements
Chapter 15 Reflexes
15.1 Definition of a
Reflex
 15.2 Reflex Arc, Gain,
and Latency
15.3 Reflex
Classifications
15.4 Conditioned
Reflexes
Chapter 16 Excitation and Inhibition Within the
Spinal Cord
16.1 The Spinal Cord
16.2 Excitation Within
the Central Nervous
System
16.3 Postsynaptic
Inhibition
16.4 Recurrent
Inhibition and
Renshaw Cells
16.5 Reciprocal
Inhibition
16.6 Presynaptic
Inhibition
16.7 Persistent Inward
Currents
Chapter 17 Monosynaptic Reflexes
 17.1 H-Reflex and M-
Response
17.2 Tendon Tap Reflex
(T-Reflex)
17.3 Effects of
Voluntary Muscle
Activation on
Monosynaptic
Reflexes
17.4 F-Wave
Chapter 18 Oligosynaptic and Polysynaptic
Reflexes
18.1 Oligosynaptic
Reflexes
18.2 Polysynaptic
Reflexes
18.3 Flexor Reflex
18.4 Tonic Stretch
Reflex
18.5 Tonic Vibration
Reflex
18.6 Interaction Among
Reflex Pathways
18.7 Interjoint and
Interlimb Reflexes
Chapter 19 Long-Loop Reflexes and Reflex-Like
Reactions
19.1 Preprogrammed
Reactions
19.2 Preprogrammed
Reactions Versus
Stretch Reflexes
19.3 Afferent Sources of
Preprogrammed
Reactions
19.4 Preprogrammed
Reactions During
Movement
Perturbations
19.5 Basic Features of
Preprogrammed
Reactions
19.6 Preprogrammed
Corrections of
Vertical Posture
19.7 Corrective
Stumbling Reactions
Problems for Part IV

Part V Control and Coordination of

Goal-Oriented Movements
Chapter 20 Voluntary Control of a Single Muscle
20.1 What Is Voluntary
Movement?
20.2 Feedforward and
Feedback Control
20.3 Servo Control
20.4 Servo Hypothesis
20.5 α-γ Coactivation
20.6 Voluntary
Activation of
Muscles
20.7 Equilibrium-Point
Hypothesis
Chapter 21 General Issues of Motor Control
21.1 Force Control
21.2 Engrams and the
Generalized Motor
Program
21.3 Internal Models
21.4 Equilibrium-Point
Hypothesis: Main
Ideas
21.5 Equilibrium-Point
Hypothesis: Subtle
Details
 21.6 Dynamic Systems
Approach
Chapter 22 Motor Synergies
22.1 The Problem of
Motor Redundancy
22.2 Optimization
Approaches
22.3 Bernstein’s Level
of Synergies
22.4 Uniting Muscles
Into Groups
22.5 Principle of
Abundance
22.6 Ensuring Stability
of Movements
22.7 Uncontrolled
Manifold Hypothesis
Chapter 23 Patterns of Single-Joint Movements
23.1 Isotonic Movements
and Isometric
Contractions
23.2 Task Parameters and
Performance
Variables
 23.3 Kinematic Patterns
During Single-Joint
Isotonic Movements
23.4 EMG Patterns During
Single-Joint
Isotonic Movements
23.5 EMG Patterns During
Single-Joint
Isometric
Contractions
23.6 Dual-Strategy
Hypothesis
23.7 Single-Joint
Movements Within
the Equilibrium-
Point Hypothesis
Chapter 24 Multijoint Movement
24.1 Two Issues With
Controlling Natural
Reaching Movements
24.2 Interjoint Reflexes
24.3 Multijoint
Coordination by the
Spinal Cord
24.4 Supraspinal
Mechanisms
 24.5 Neural Control
Variables for
Multijoint
Movements
24.6 Equilibrium-
Trajectory
Hypothesis
24.7 Hierarchical
Control With
Spatial Referent
Coordinates
24.8 Multijoint
Synergies
Chapter 25 Postural Control
25.1 Vertical Posture
25.2 Postural Sway
25.3 Role of the
Vestibular System
25.4 Role of Vision
25.5 Role of
Proprioception
25.6 Anticipatory
Postural
Adjustments
25.7 Corrective Postural
Reactions
 25.8 Postural Synergies
Chapter 26 Locomotion
26.1 Two Approaches to
Locomotion
26.2 Central Pattern
Generator
26.3 Locomotor Centers
26.4 Spinal Locomotion
26.5 Spinal Control of
Locomotion in
Humans
26.6 Gait Patterns
26.7 Dynamic Pattern
Generation
26.8 Step Initiation
26.9 Corrective
Stumbling Reaction
Chapter 27 Prehension
27.1 Hand Joints and
Muscles
27.2 Cortical
Representations of
the Hand
27.3 Indices of Finger
Interaction
 27.4 Multifinger
Synergies in
Pressing Tasks
27.5 Grasping
27.6 Prehension
Synergies and
Principle of
Superposition
Problems for Part V

Part VI Sensorimotor Integration for

Perception and Action
Chapter 28 Kinesthetic Perception
28.1 Sensation and
Perception
28.2 Weber-Fechner Law
28.3 Ambiguity of
Sensory Information
28.4 Afferent and
Efferent Components
of Perception
28.5 Vibration-Induced
Kinesthetic
Illusions
28.6 Distorted Efferent
Copy and
 Preconceptions
28.7 Sense of Effort
28.8 Stability of
Percepts
28.9 Perception–Action
Coupling
Chapter 29 Multisensory Integration
29.1 Spatial
Multisensory
Integration for
Limb Motor Control
29.2 Temporal
Multisensory
Integration for
Limb Motor Control
29.3 Coordinate Frames
for Limb Motor
Control
29.4 Postural Balance
and Motion
Perception
29.5 Neural Correlates
of Multisensory
Integration
Chapter 30 Visual Perception and Action
30.1 Two Visual Streams
30.2 Magnocellular and
Parvocellular
Ganglion Cells and
Streams
30.3 Motion Processing
in the Cortex
30.4 Color, Object, and
Face Recognition in
the Ventral Stream
30.5 Feedforward and
Feedback Control
for Reach-to-Grasp
Movements
30.6 Neural Structures
Involved in
Oculomotor Control
30.7 Roles of Two Visual
Streams in Eye–Hand
Coordination
30.8 Eye and Hand
Coordination for
Movements Starting
From Rest
30.9 Eye and Hand
Coordination During
Movement
 30.10 Eye and Hand
Coordination While
Intercepting
Moving Targets
Problems for Part VI

Part VII Emerging, Evolving, and

Adapting Movements
Chapter 31 Fatigue
31.1 Fatigue and Its
Contributors
31.2 Muscular Mechanisms
of Fatigue
31.3 Spinal Mechanisms
of Fatigue
31.4 Supraspinal
Mechanisms of
Fatigue
31.5 Adaptive Changes
During Fatigue
31.6 Abnormal Fatigue
Chapter 32 Effects of Aging
32.1 General Features of
Movements in
Elderly Persons
 32.2 Changes in Muscles
and Motor Units
32.3 Muscle Reflexes in
Elderly Persons
32.4 Changes in Sensory
Function
32.5 Muscle Activation
Patterns During
Fast Movements
32.6 Changes in Posture
and Gait
32.7 Hand Function in
Elderly Persons
32.8 Changes in Motor
Synergies
32.9 Adaptive Changes in
Motor Patterns
32.10 Effects of
Training
Chapter 33 Typical and Atypical Development
33.1 Humans at Birth
33.2 Motor Milestones
During Typical
Development
33.3 Exploration and
Emergent Motor
 Patterns
33.4 Development of
Motor Synergies
33.5 Down Syndrome
33.6 Effects of Practice
in Persons with
Down Syndrome
33.7 Autism
33.8 Developmental
Coordination
Disorder
Chapter 34 Motor Learning
34.1 Adaptation,
Learning, and
Memory
34.2 Muscle Memory
34.3 Habituation of
Reflexes
34.4 Conditioned
Reflexes
34.5 Operant
Conditioning and
Learning Spinal
Reflexes
34.6 Short-Term and
Long-Term Memory
 34.7 Adaptation to
Unusual Force
Fields
34.8 Motor Skills
34.9 Learning Motor
Synergies
34.10 Stages in Motor
Learning
34.11 Effects of
Practice on
Cortical
Representations
Problems for Part VII

Part VIII Motor Disorders

Chapter 35 Peripheral Muscular and Neurological
Disorders
35.1 Myopathies and
Neuropathies
35.2 Muscular
Dystrophies
35.3 Continuous Muscle
Fiber Activity
Syndromes
35.4 Myasthenia Gravis
35.5 Mononeuropathies
 35.6 Multiple
Mononeuropathies
35.7 Polyneuropathies
35.8 Radiculopathies
Chapter 36 Spinal Cord Injury and Spasticity
36.1 Consequences of
Spinal Cord Injury
36.2 Signs and Symptoms
of Spasticity
36.3 Possible Mechanisms
of Spasticity
36.4 Defining Muscle
Tone
36.5 Treatment of
Spasticity
Chapter 37 Disorders Involving the Basal Ganglia
37.1 Clinical Features
of Parkinson’s
Disease
37.2 Voluntary Movements
in Parkinson’s
Disease
37.3 Vertical Posture
and Locomotion in
Parkinson’s Disease
 37.4 Motor Synergies in
Parkinson’s Disease
37.5 Treatment of
Parkinson’s Disease
37.6 Huntington’s Chorea
37.7 Hemiballismus
37.8 Dystonia
37.9 Tardive Dyskinesia
Chapter 38 Cerebellar Disorders
38.1 Consequences of
Cerebellar Injuries
in Animals
38.2 Consequences of
Cerebellar
Disorders in Humans
38.3 Abnormalities of
Stance and Gait
38.4 Voluntary Movements
in Cerebellar
Disorders
38.5 Cerebellar Tremor
38.6 Ataxias
38.7 Changes in Motor
Synergies
38.8 Cerebellar
Cognitive Affective
 Syndrome
Chapter 39 Cortical Disorders
39.1 Consequences of
Lesions of
Different Cortical
Lobes
39.2 Stroke
39.3 Myoclonus
39.4 Tics
39.5 Tourette Syndrome
39.6 Williams Syndrome
Chapter 40 Systemic Disorders
40.1 Amyotrophic Lateral
Sclerosis
40.2 Multiple Sclerosis
40.3 Multisystem Atrophy
40.4 Essential Tremor
40.5 Cerebral Palsy
40.6 Wilson’s Disease
Chapter 41 Motor Rehabilitation
41.1 Do “Normal
Movements” Exist?
41.2 Changes in CNS
Priorities
41.3 Neural Plasticity
 41.4 Adaptive Changes in
Motor Patterns
41.5 Consequences of
Amputation
41.6 Functional
Electrical
Stimulation
41.7 Constraint-Induced
and Discomfort-
Induced Therapies
41.8 Brain–Computer
Interface
41.9 Practical
Considerations
Problems for Part VIII
Glossary
References
Index
About the Authors
 Preface
Since the original publication of the first edition of Neurophysiological
Basis of Movement in 1998, major changes have happened in the
field of the neural control of movements or, briefly, motor control.
Over the past 25 years, the International Society of Motor Control
(ISMC) has been formed, which now has its official journal Motor
Control, runs a series of biennial conferences Progress in Motor
Control, awards every two years the Bernstein Prize for outstanding
contributions to motor control, and helps organize the annual Motor
Control Summer School. There have been major developments in
the theoretical foundations of motor control accompanied by an
increase in the number of experimental studies, including studies of
animals and humans over the life span, from neonates to the elderly,
as well as various patient populations, aiming to discover the
neurophysiological mechanisms involved in the production of
functional movement patterns. There are courses in motor control at
both undergraduate and graduate levels, textbooks, reference
books, and national and international conferences on motor control.
At this time, motor control is striving to become a field of natural
science comparable in its rigor and exactness to established fields
such as classical physics. This requires bringing more exactness into
the terminology and rigor into the logics of the textbook. In particular,
we strive to offer clear definitions to such frequently used but rarely
defined concepts as synergy, motor program, motor command, and
muscle tone. Neurophysiology forms the foundation of motor control,
and therefore we consider an understanding of the basic
neurophysiological structures and processes to be vital for progress
in the field.
The two previous editions of this book have been used to teach
both upper-level undergraduate and entry-level graduate classes. In
particular, at Penn State, the textbook has been used to teach a 400-
level undergraduate course, Movement Disorders, and a 500-level
graduate course, Neurophysiological Basis of Movement. The former
course targeted those students in the department of kinesiology who
planned to continue their education in areas related to motor
disorders and rehabilitation. Most students who have taken this
course so far planned to continue their education in a medical
school, a physical therapy or occupational therapy school, a
chiropractor’s school, a physician’s assistant program, or other
related program. For that class, the course was tailored by selecting
only the basic information on neurophysiology and focusing more on
chapters related to changes in movements and neurophysiological
mechanisms associated with fatigue, development, aging, and
various movement disorders.
In contrast, the graduate class was targeting students who joined
the graduate program in kinesiology after completing undergraduate
studies without taking any classes related to the central nervous
system. These students commonly came from areas such as
engineering, physics, mathematics, philosophy, and sociology. Those
more mature students were exposed to material on basic
neurophysiology, neurophysiological mechanisms of various
behaviors, and more theoretical issues such as the ideas of
parametric control and the control of action stability.
This third edition has been expanded to allow more flexibility in
tailoring the material to courses targeting different audiences and
course contents. In particular, we added chapters on motor learning
and sensorimotor integration, and we expanded significantly the
sections on the role of different sensory modalities in motor control,
on kinesthetic perception, and on action–perception interactions. We
also took this opportunity to add more one-minute questions and
self-test problems, which have proven to be highly useful (based on
the student feedback).
We are addressing rather explicitly many of the controversial
issues in the area of motor control and coordination. Originally, the
textbook was somewhat sugarcoated, presenting only well-
established facts in the most noncontroversial manner. A number of
chapters in the third edition deal directly with current theories of
motor control and coordination and present different opinions on the
very basic issues.
 The preparation of this edition was helped a lot by the very useful
(and frequently harsh) feedback from the students who took the two
aforementioned courses at Penn State and from colleagues who
took their time to tell us how the textbook could be improved. We are
particularly grateful to Vladimir Zatsiorsky for the innumerable helpful
comments over the past 30 years, to Robert Sainburg for the many
fruitful discussions of how to teach motor control and
neurophysiology, and to Karl Newell for the encouragement to
develop the two courses at Penn State.
Mark Latash and Tarkeshwar Singh
Introduction
 Chapter 1

History, Evolution, and Motor

Control

KEY TERMS AND TOPICS

history of movement studies
Nikolai Bernstein
evolution of movements
motor control
laws of nature
physical approach

What is the origin of purposeful movements by animals, including

humans? Why is it easy for an observer to distinguish movement by
an animal from that of an inanimate object, for example a stone?
These questions have been in the center of attention of researchers,
philosophers, and clinicians for literally thousands of years (for
reviews on the history of movement science, see Latash and
Zatsiorsky 2001; Meijer 2002). Classical Greek philosophers
formulated the first of these questions somewhat differently: How
does the soul control the body? And what is the origin of soul action?
1.1 Brief History of Movement
Studies
One of the founders of geometry, Pythagoras (571?-497? B.C.),
viewed soul as an entity (a number) that can move by itself following
the motion of heavenly spheres. Further, motion of the soul was
somehow transmitted to body parts. Somewhat later, Democritus
(460-370? B.C.) introduced a theory that all objects in the world
consisted of very small basic objects, atoms, and concluded that
movement of the soul was transmitted to the body by the movement
of atoms. This was a brilliant insight preceding the current
understanding of the role of moving ions in the communication
between the central nervous system and muscles.
Animals, according to Plato (428-347 B.C.), possessed a unique
ability to show self-motion (i.e., they can move without being pushed
by other objects). He viewed this ability as a reflection of the
immortal soul, which was present in animals but not in inanimate
objects. The soul was assumed to move body parts similarly to how
fingers of the puppeteer pull on strings and move the marionette.
The development of this view over the past 2,000 years is reflected
in the idea that the brain prescribes spatial trajectories of body parts
during voluntary movements. Aristotle (384-322 B.C.) was arguably
the first to emphasize the importance of coordination during
voluntary movements, an undefined feature prescribed by the
Creator that makes biological movements look harmonious.

PROBLEM 1.1
What makes the soul command the body according to classics of
Greek philosophy? And what could be the origin of commands to
the soul?
 Until the second century A.D., philosophers discussed only very
general features of biological movements without paying much
attention to organs within the body that were crucial for movements
to occur. The great Roman physician Galen (129-201) was arguably
the first to emphasize two important features of animal movements.
The first is the involvement of muscle pairs with opposing actions
(agonists and antagonists, for example flexors-extensors) in
voluntary movements of body segments. And the second was the
role of nerves that delivered animal spirits to muscles that led to their
force production.
The formulation of the main question of movement as that of soul-
body interaction persisted until relatively recently. A great
philosopher of the Renaissance, René Descartes (1596-1650), is
credited for introducing the basics of dualism, a branch of philosophy
that views two independent entities forming every human being, the
soul and the body (not very different from the ancient Greeks).
Following Galen, Descartes thought that the soul used animal spirits
to move the body, although some movements, for example the
beating of the heart, were viewed as independent of the soul.
Descartes also emphasized the importance of senses and the
central nervous system for some movements. Descartes and a
British anatomist Thomas Willis (1628-1678) viewed quick motor
reactions to sensory stimuli as building blocks for voluntary
movements—a view developed in the 20th century by Sherrington.
Later, Jean Astruc (1684-1766) introduced the term reflex for such
actions.
At about the same time, the science of biomechanics was born
with particularly impressive contributions by Giovanni Alfonso Borelli
(1608-1679), a disciple of Galileo. Borelli viewed muscles as elastic
structures controlled by the soul with droplets of nerve juice
delivered by nerves—a great insight into the role of chemical
processes and neuromuscular mediator (acetylcholine) in the
production of biological movements.

PROBLEM 1.2
 Present examples of biological movements dependent and
independent of the soul according to Descartes.

The importance of electricity for biological movements was

discovered just over 200 yrs ago by Luigi Galvani (1737-1798). This
field of research developed rapidly in the 19th century, leading to the
discoveries of the role of electricity within muscles by Carlo
Matteucci (1811-1868) and Etienne DuBois-Reymond (1818-1896).
In the middle of the century, studies of muscle reflexes by a German
scientist Eduard Friedrich Wilhelm Pflüger (1829-1910)
demonstrated the role of the spinal cord in reflex movements and its
ability to produce activations of different muscles to the same
stimulus depending on the initial state of the body.
At about the same time, studies of biomechanics were moved
forward by the invention of photography, and two great researchers,
Etienne-Jules Marey (1830-1904) and Eadweard J. Muybridge
(1830-1904), developed photographic methods specifically for the
analysis of natural movements. Mechanical movement analysis was
also refined by the Weber brothers (Ernst Heinrich, 1795-1878;
Wilhelm Eduard, 1804-1891; and Eduard Friedrich Wilhelm, 1806-
1871), who contributed significantly to the understanding of the
mechanics of locomotion and other movements involving multijoint
effectors.
Neurophysiological studies of movements got very important
contributions from studies of neuroanatomy by the great Italian
neuroanatomist Camillo Golgi (1843-1926), who developed the
method of silver-staining of neurons, and the Spanish
neuroanatomist Santiago Ramon y Cajal (1852-1934), who used this
method to visualize a variety of cells. His amazing drawings of
neurons in different parts of the brain remained unsurpassed in
quality for many years. One of the greatest neurophysiologists of the
20th century, Sir Charles Sherrington (1852-1952), and his colleague
Sir Michael Foster (1826-1907) coined the term synapsis (later
transformed into synapse) for sites of interaction between neurons.
The contributions of Sherrington were many and varied, and his
name will be featured in several chapters, particularly those
dedicated to description of muscle reflexes and the neural control of
locomotion. In particular, Sherrington introduced the notion of active
inhibition within the central nervous system and emphasized the
importance of reflex connections from peripheral receptors to
motoneurons in the spinal cord. He viewed muscle reflexes not as
hardwired stereotypical responses to stimuli but rather as flexible
mechanisms that formed the basis of motor behavior.
A student of Sherrington, Thomas Graham Brown (1882-1965),
performed studies of animals without sensory input into segments of
the spinal cord from peripheral receptors. He documented
locomotion-like movements in these animal preparations and
introduced the notion of networks within the spinal cord (and,
possibly, other parts of the central nervous system) able to generate
rhythmic patterns of neural activity. Such neural networks were later
termed central pattern generators.
Further important steps in movement studies were helped by the
development of new methods of analysis of electrical processes in
the muscles (electromyography, EMG) and in the brain
(electroencephalography, EEG). One of the pioneers of
electromyography, Kurt Wachholder (1893-1961), and his student
Hans Altenburger published a series of studies of EMG patterns
during voluntary movements and described important regularities in
muscle activation, including the triphasic pattern of muscle
activation.
The name of Nikolai Bernstein (1896-1966) will be featured in
many chapters of this textbook. Bernstein is viewed by contemporary
researchers as the father of motor control and physiology of activity.
His research method was based on Darwin’s ideas of evolution: He
viewed anatomical, physiological, and behavioral phenomena as
reflections of the evolutionary process and fight for survival. Based
on these views, he developed a hierarchical scheme for the control
of biological movements (Bernstein 1947/2020), which remains one
of the best developed schemes deeply rooted in neurophysiology.
1.2 Evolution of Movements and
Nikolai Bernstein’s Theory
Bernstein assumed that the process of evolution led to gradual
development of the central nervous system based on new tasks
faced by animals. He used the expression “new tasks formed the
brain” in a very direct meaning. Furthermore, he viewed each
qualitatively new evolutionary step as associated with the
construction of a new level of control on the foundation of pre-
existent levels. In other words, in higher animals up to humans, one
can see reflections of the whole process of evolution. Based on
these views, Bernstein classified electrical phenomena in the body
into paleokinetic and neokinetic. The first term was used to address
evolutionarily older, relatively slow processes, which are going to be
discussed further in this textbook under the label of postsynaptic
potentials and local currents, and the second term described
evolutionarily more recent, fast processes such as action potential
generation and conduction.
In his most detailed book, Bernstein (1947/2020) introduced
several levels within the central nervous system participating in the
construction of movements. His classification of levels was based on
two principles. The first associated certain classes of motor tasks
and task components with specific levels. The second linked the
levels to different neurophysiological structures and pathways. The
main levels identified by Bernstein are illustrated in table 1.1.
The lowest level, Level A, also known as the paleokinetic level, or
the rubro-spinal level, was viewed as the level where muscle tone
(sometimes addressed as muscle tonus) was defined. We will return
to this notion later in the textbook, in particular in chapters dedicated
to movement disorders in neurological patients. This level was
supposed to play auxiliary roles, for example, providing necessary
postural backgrounds for actions led by higher levels. Bernstein
emphasized the importance of spinal reflexes, in particular
reciprocal inhibition, in the functioning of this level. According to
Bernstein, descending control of this level from the brain was
associated with the output of the red nucleus, a nucleus in the
brain, the origin of the rubrospinal tract, which mediates, in particular,
effects of processes in the cerebellum on limb and body
movements.

Table 1.1 Hierarchical Scheme of the Neural Control of

Movements
Level Alias Physiology Functions
A Paleokinetic level Rubrospinal Muscle tone, posture
B Level of synergies Thalamic-pallidar Synergies, stability
C Level of spatial field Pyramidal-striatal Targets, affordances
C1 Striatal Trajectory
C2 Pyramidal Target
D Level of actions Parietal-premotor Topology, skills
E Level of symbolic actions Higher cortical Language, music

Adapted from Bernstein (1947/2020).

The next level, Level B, which had two names, the level of
synergies and patterns and the thalamic-pallidar level, was
responsible for uniting numerous muscles that take part in most
functional actions, for example locomotion, into groups. Muscles
within each such group were expected to show parallel changes in
their activation levels, which is currently viewed as one of the
distinguishing features of multi-muscle synergies. Bernstein realized
the importance of movement stability in the continuously changing,
unpredictable world, and associated this function with Level B. It is
rather amazing that Bernstein linked this level to the thalamus and
globus pallidus—two structures within the brain involved in neural
loops that define movement-related output of cortical areas. As we
will see later, this insight maps well on the current understanding of
the role of the thalamus and basal ganglia (globus pallidus is part of
this system of subcortical structures) in the control and coordination
of movements.
 Bernstein associated the next level, Level C, the level of the
spatial field, or the pyramidal-striatal level, with the concept of spatial
field, a portion of the external space accessible for specific actions,
with its geometry and metrics. This notion included not only motor
elements but also sensory signals relevant to that portion of the
surrounding space and its perception with the help of various
sensory modalities, including proprioception, touch, vision, and
vestibular system. The notion of spatial field is a close relative to the
notion of affordances introduced later by the great American
psychologist James Gibson (1904-1979) to describe possibilities for
action existing in the environment given the properties of the body,
location of targets, obstacles, and external forces (Gibson 1979).
The level of spatial field was viewed as responsible for actions
toward targets in the external space. Two sublevels were responsible
for identification of a target and ultimately reaching it (C2, pyramidal
sublevel) and for ensuring specific trajectories from the initial state to
the target (C1, striatal sublevel). Note that proper functioning of this
level always has to take into consideration actual and expected
forces acting from the environment. Since such expectations are
never 100% perfect, behaviors controlled by Level C are always
associated with motor variability and can be characterized with
accuracy—consider, for example, catching a ball. Bernstein
introduced one of his famous expressions with respect to actions
controlled at Level C: “Repetition without repetition.” He implied that,
when a person tried to perform the same task multiple times, the
process of solving this motor problem was repeated but actual
trajectories were not—a great insight considered in detail later.

PROBLEM 1.3
Can spatial field be described with nonspatial variables? Present
examples.

Actions at the next level are seen nearly exclusively in humans.

This level, Level D, the level of actions, or the parietal-premotor
level, is responsible for object-oriented meaningful actions. Actions
performed at this level not only reach for an object or transport it but
do so for a reason; the main characteristic of these movements is
their topology (i.e., higher-level spatial characteristic). One of the
examples discussed by Bernstein is handwriting: One can write any
given letter, for example the letter r, in many different ways, using
different implements and holding them differently, even with different
body parts, but they will all be recognized as r and will carry
respective meaning. This is the level where motor skills are
developed. As the name of this level suggests, Bernstein associated
it with processes in parietal and premotor areas of the cortex.
The highest level, Level E, the level of symbolic, highly
coordinated actions such as speech and writing, was not associated
by Bernstein with any specific brain structures. He described this
level as based on “highest cortical control.” Using the example of
handwriting, this level took responsibility not only for writing the letter
r in a legible way but writing it in a proper position within a word.
Using a different example, a musician has to produce not only an
action that leads to a specific note when playing a musical
instrument, but playing it at a proper time and level of sound within
the piece.
We will return to Bernstein’s multilevel scheme for the
construction of movements later when we discuss specific
neurophysiological structures, circuits, reflexes, and behaviors.

1.3 Motor Control and Laws of

Nature
Although this textbook is not explicitly about motor control, we will
always view specific neurophysiological structures and circuits as
contributors to the control and coordination of functional movements.
There is no agreed upon definition of the commonly used expression
“motor control,” and this situation is a major source of
misunderstanding in the field studying the neural control of
movements. Two attitudes to motor control dominate the field. One
of them is based on the impressive success of control theory and
engineering during the 20th century in the design of control systems
for artificial objects, from ballistic missiles to artificial satellites, to
self-driving cars, and to robots. According to this approach, the
central nervous system of a moving animal performs computational
operations similar to those performed by the control systems in
human-built objects with the purpose of predicting future states of
the animal, changes in the environment, and interactions between
the two, and generating neural signals needed to produce requisite
force profiles. The goal of motor control, within this theoretical
framework, is to understand the software underlying those
computations and the neurophysiological systems involved in
delivering relevant signals to and from structures performing the
computations.
According to principles of engineering, animal bodies are built
rather poorly. As we will see further in the textbook, our bodies have
sluggish and not very predictable motors (muscles) that produce
forces dependent on muscle length and velocity; they are equipped
with rather ambiguous and fuzzy sensors (receptors) and have very
long delays in the conduction of information along neural fibers from
the receptors to the central nervous system, within the central
nervous system, and back to the muscles. In contrast, artificial
systems, such as robots, are equipped with powerful actuators that
can produce patterns of force and torque independently of the
kinematics; their sensors are accurate and dedicated to specific
salient variables, and conduction delays in the electrical circuits are
very short. Overall, if we were designed by a 21st century engineer,
this person would have not been praised by his or her employers.
That is why we need an omnipotent computer (the brain) to handle
all the difficulties introduced by this apparently faulty design.
The alternative view is: Evolution makes no mistakes or, at the
very least, it makes fewer mistakes when compared to contemporary
engineers. (Indeed, look at the best robots trying to play soccer!
They are amazingly clumsy and inept.) If our bodies have specific
features, these are the best to handle everyday motor problems and
be successful at solving them. This approach views animals and
their body parts, including the brain, as natural objects that behave
according to laws of nature and perform no computational
operations. Indeed, very coordinated and agile animals cannot be
taught to add two numbers. Within this approach, motor control is
defined as an area of natural science exploring laws of nature that
define how body parts interact with each other and with the
environment during natural movements. This view of biological
movements is going to be addressed further as the physical
approach.
To introduce the physical approach, one has to start with a few
very basic definitions. Laws of nature are concise descriptions of
regularities in the behavior of different classes of objects observed
by researchers, commonly in the form of equations. Consider, for
example, two of the best-known laws of nature from the field of
classical mechanics, Newton’s second law and Hooke’s law.
Newton’s second law describes how forces change the motion of
material objects:

where F is force, a is acceleration, and m is mass (figure 1.1a). Note

that, here and later, italics will be used to denote variables and
parameters characterized by magnitude only, and bold-italic fonts will
be used for vectors (variables characterized by both magnitude and
direction). Hooke’s law described how forces lead to the deformation
of certain classes of objects, which we will address as springs (figure
1.1b):

where ΔF stands for a change in external force, Δx for a change in

the object’s dimension, and k for stiffness.
Figure 1.1 An illustration of two basic laws of nature from classical physics, (a)
Newton’s second law and (b) Hooke’s law. Both laws involve variables: F, a, ΔF,
and ΔX and parameters: m and k. F = force; a = acceleration; m = mass; ΔF =
change in force; X = coordinate; X0 = coordinate corresponding to zero length of
the spring; ΔX = change in coordinate; k = stiffness.

Note that each of the two presented equations involves two types
of symbols, variables (F, a, ΔF, and Δx) and parameters (m and k).
Variables are constrained by each law, while parameters are not.
Indeed, mathematically it does not matter whether you write F = ma
or F = am. However, if you apply force to an object, it would lead to
proportional acceleration, not a change in mass. The same is true for
Hooke’s law: Changing force acting on a spring will lead to its
deformation, not a change in its stiffness. This is true even for
objects that are characterized by quickly changing parameters. For
example, m remains a parameter in Newton’s second law even if it
changes quickly as, for example, in a rocket burning lots of its fuel.
Classical laws of nature described in physics textbooks are
applicable to all objects, biological and inanimate. However, there is
a qualitative difference. Inanimate objects behave in a well-
predictable way if we know all the variables and parameters of the
relevant laws of nature. Biological objects do not violate those laws,
but their behavior is not prescribed, only constrained, by them.
Indeed, animals frequently show behaviors that are not expected
from inanimate objects. Examples include running uphill and
swimming against the current. These behaviors suggest that some of
the relevant laws of nature have not been discovered yet, which
makes the field of motor control very exciting. Further, we will try to
link specific features of neurophysiological systems and circuits in
the body to possible biology-specific laws of nature.
Figure 1.2 uses the well-known law of gravity to illustrate the
difference between the two main approaches, based on the control
theory (left panel) and on laws of nature (right panel). The law of
gravity was introduced by Isaac Newton in the following form:

where FG stands for the gravity force acting between two bodies, M1
and M2 for the masses of the two bodies, R for the distance between
the bodies, and γ is a constant. Imagine now that there is a
computational device on the Sun that receives signals from sensors
on the planets informing it of their masses and distances from the
Sun. This computational device also has information on the mass of
the Sun itself. Based on those data, it computes the requisite force
using equation 1.3 and sends signals to actuators that produce the
precomputed force on the planets.
Figure 1.2 An illustration of two approaches to the origin of gravity force acting
between the Sun and a planet. (a) The Sun gets information from sensors on the
mass (m) of and distance (R) to the planet, computes planned force (F) based also
on its own mass (M), and sends a signal to hypothetical actuators that put this
force into action. (b) The Sun creates gravity field G. Any object in this field
experiences force proportional to the field and the mass of the object. γ = constant.

The alternative is to introduce the notion of a gravitational field

created by any object with mass and formalizing the force it creates
on other objects with mass in the form of equation 1.3. No sensors,
computational devices, or actuators need to be assumed: Celestial
bodies move as dictated by the law of gravity, which is one of the
basic laws of nature.
The fact that human bodies are equipped with apparent
anatomical “actuators” (muscles) and “sensors” (receptors) does not
by itself prove that the brain performs computations similar to those
that are used to control robots and satellites. Note that the
functioning of other organs such as the liver, the heart, and the
stomach is not based on assumed computations but on physiological
mechanisms that are based on laws of nature. In this book, we will
try not to invoke computations performed by our objects of study,
including the brain, but to understand how they participate in the
production of natural movements based on the known physiological
mechanisms and laws of nature.
According to one of the influential theories of motor control
(reviewed in Feldman 2015), which we will describe in more detail
later, the neural control of biological actions is based on prescribing
parameters within the relevant laws of nature. This type of control—
addressed as parametric control—is qualitatively different from
movements in inanimate nature. Indeed, there is only one method to
induce movement (or to change movement) of an inanimate material
object: to apply force—a variable—to that object, which will induce
acceleration according to Newton’s second law. Biological objects,
however, change parameters to initiate or modify movements.
Figure 1.3 A simple physical system: an ideal pendulum (a point mass on a
massless, rigid cord). Its motion can be controlled by changing parameters of the
pendulum: L (length) and {x0, y0, z0}—coordinates of its suspension point. Forces
acting on the pendulum emerge with its motion: FG = force of gravity, FC = force
from the cord, and FRES = resultant force.

These two methods of control can be illustrated with the example

of a simple pendulum consisting of a ball attached by a massless
rigid cord to the point of suspension (figure 1.3). Natural movement
of the pendulum is defined by its parameters. In particular, the
natural frequency of its oscillations is defined by the length of the
cord (L), and the portion of the external space where it oscillates is
defined by the coordinates of the point of its suspension in space (x0,
y0, z0; we assume that the direction of gravity, g, is not changing). Its
motion is accompanied by changes in the resultant force acting on
the ball (FRES in figure 1.3), but this force is not prescribed by a
controller. It emerges during the motion of the pendulum given the
external field of gravity and parameters of the pendulum.
How can one change the movement of this simple pendulum?
There are two methods. First, one can push the ball during its motion
(i.e., apply an external force). Second, one can change some of the
parameters that define its motion; for example, one can shorten the
cord (and it will start to oscillate faster) or move the point of
suspension (and it will oscillate in a different area of space). The
second method is an example of parametric control. So, the idea that
biological movements are produced by parametric control implies
that living objects can change their relevant parameters. Imagine
that a pendulum has learned (as a result of “pendulum evolution”)
how to change—at will!—its length and coordinates of suspension.
This would lead to any desired movement of the ball without any
computation of forces.

CHAPTER 1 IN A NUTSHELL
Philosophers and researchers have been
thinking about problems of interaction
between the human mind (soul,
intention) and body literally for
millennia. Development of new methods
of observation, measurement, and
analysis played a major role in
leading the evolution of thinking
about the interactions between the
central nervous system and the rest of
the body. Nikolai Bernstein introduced
a multilevel scheme of the
construction of movements based on the
evolutionary approach; many of his
guesses have been confirmed in later
studies. There are two approaches to
motor control. One of them is based on
concepts from control theory and
engineering; the other is based on
ideas from natural science. According
to the latter approach, motor control
is a field of study trying to discover
laws of nature that define
interactions among body structures,
including the central nervous system,
and between the body and the
environment during natural movements.
 Part I

Excitable Cells and Their

Communication
 Chapter 2

Membranes, Particles, and

Equilibrium Potentials

KEY TERMS AND TOPICS

biological membranes
movement of particles in solutions
osmosis
movement of ions
Nernst equation
equilibrium potential

Nikolai Bernstein singled out two events in biological evolution that

he viewed as crucial for the development of life on Earth (Bernstein
1947, 1996; Latash 2020). One of them is easy to guess: This is the
emergence of a stable molecule that can replicate itself. Otherwise, it
would be impossible to pass information to future generations. The
second event is less obvious. It is the emergence of the biological
membrane, which separates the contents of the simplest organism,
the cell, from the environment. We are not going to discuss viruses
here. They may be viewed as even simpler organisms able to
survive and succeed in the evolutionary process without the
membrane.
 Further, Bernstein described a feasible chain of events leading to
more and more complex organisms up to humans. He emphasized
the following steps, which are going to be discussed in more detail in
future chapters:
1. Cells form groups. Within such groups, cells on the surface
become sensitive to salient external stimuli, and cells inside
the groups develop an ability to change shape (i.e., to produce
movements). This leads to the emergence and specialization
of receptors and muscles.
2. One of the ends of such a group of cells, its “mouth end,”
becomes its active end, searching for food, defending from
predators, and providing information on objects that are not in
direct contact with the body with the help of specialized
sensors—telereceptors.
3. Specialization of groups of cells for fast information
transmission. Emergence of groups of cells (neural ganglia,
the spinal cord, the medulla, and the brain) dedicated to the
control of certain tasks.
4. Searching for more food and avoiding predators leads to new
motor tasks, which require more and more sophisticated
problem-solving abilities. Bernstein emphasized, in particular,
the role of locomotion and postural tasks for the development
of the central nervous system.
5. The process of encephalization, when larger and larger
groups of tasks and associated problems are delegated to the
brain.
We will begin with one of the two earliest “inventions” of evolution,
the cellular membrane. Other steps will be clarified later in the book.

2.1 The Biological Membrane

The cellular membrane (figure 2.1) isolates information within the cell
from the external world and thus allows for its storage, it protects the
contents of the cell, and it defines its boundary, thus making it a unit
separate from the environment. If the membrane were absolutely
impermeable to any substances, the cell would not be able to
interact with the world, to extract the necessary information and
substances (e.g., sources of energy), and to get rid of products of
metabolism. This would make it an alien structure rather than a part
of the environment. If the membrane were permeable to everything,
its function would be completely lost. So, one of the most important
functions of the cellular membrane is its partial permeability that
allows exchange of information with the environment while protecting
the contents of the cell.

Figure 2.1 Cellular membrane is a sophisticated structure that is permeable to

some substances but not to others. Its selective permeability makes the
membrane a unique structure that allows the cell both to interact with the
environment and to be separate from it.

The movement of substances across the membrane is a central

theme of an area of biology called “membrane physiology.”
Membranes are commonly very thin (about 40 Ångstrom or about 4
nm = 4 · 10−9 m), but they control the movement of substances much
more effectively than the cells themselves, despite their relatively
large volume. There are three major groups of substances that can
travel across the membrane and whose properties we are going to
consider:
1. Solvents. The most common solvent is water; however, some
substances are soluble in lipids, which allows them to pass
through cellular membranes more easily because the
membranes are built mostly with lipid molecules.
2. Electrolytes. These are ions (fragments of molecules) that
have a nonzero electric charge.
3. Nonelectrolytes. These are molecules or fragments of
molecules without a net electric charge. Many products of
cellular metabolism are nonelectrolytes.
Movements of electrolytes will play a particularly important role in
this course, because they create electric current through the
membrane. Most of the urgent information is transmitted within the
nervous system (as well as within other systems of our body) with
the help of electricity. So, electric currents created by movements of
electrolytes are vital for information transmission that, in turn,
underlies all the processes of the generation of commands to
muscles and the execution of these commands by the muscles.

2.2 Movement in a Solution

Water is a very good solvent because of the polar nature of its
molecule, H2O, because this molecule has local positive and
negative charges that sum up to zero. As a result, an electrolyte
(e.g., molecules of salt, NaCl) dissociates in water, creating ions.
Nonelectrolytes that, like water, are polar also dissolve rather well
but without breaking down to ions.
Movement of water commonly occurs because of a difference in
hydrostatic pressure; this bulk flow or convection is proportional to
pressure difference (figure 2.2). Bulk flow carries water with all the
dissolved particles.
Figure 2.2 Convection is movement of a solvent (for example, water) and
solutes from an area of high pressure to an area of low pressure.

The concentration of particles of a certain kind in water defines

another type of movement, which is called diffusion. If the
concentrations in two areas of a solution are different, random
motion of particles (molecules or ions) in different directions, the
Brownian motion, leads to a net movement in a direction from the
site with a higher concentration to the site with a lower concentration
(figure 2.3).
Figure 2.3 Diffusion is movement of particles dissolved in a solvent from an area
of their high concentration to an area of their low concentration.

As a result, diffusion changes the concentration of particles

leading to a decrease in the difference in concentrations at different
sites. Note that relative change in concentration depends on several
factors, including actual difference in the number of particles and
total volume of each site (compartment). The bigger the
compartment, the smaller the change. When discussing diffusion of
particles across cellular membrane, the extracellular space is
typically considered to be much larger than the intracellular space,
so diffusion processes lead to a change in the concentration inside
the cell but not outside it. Note also that diffusion takes time,
particularly when it occurs across large distances. So, our body uses
other means of transporting solutes over large distances, in
particular, convection with the help of the circulatory system. The
rate of diffusion from or into a cell depends on the surface/volume
(S/V) ratio for the cell. Small cells have large S/V ratios, and diffusion
occurs quickly, while large cells have low S/V ratios and diffusion is
slow. For a spherical cell:
where r is radius.
Electrolytes and nonelectrolytes both move with convection and
diffusion. Electrolytes, however, can also move under the action of
an electric field. In this case, the movement of electrolytes obeys
Ohm’s law:

where I is current or change in electric charge (I = dQ/dt), V is

voltage or difference of electric potentials, and R is a coefficient
termed resistance (figure 2.4).

Figure 2.4 An electric field creates a difference of potentials (U), which induces a
flow of charged particles (current, I). The current is proportional to the difference of
potentials. The inverse of the coefficient of proportionality is termed resistance, R.

Convection, diffusion, and movement under the action of an

electric field occur in a solution irrespectively of the presence or
absence of membranes. Let us now turn to movements of these
substances across biological membranes. Membranes typically are
built of lipid layers that are very permeable to water, quite
impermeable to virtually any particles, but particularly hate to let ions
through. So, almost all movement of substances across membranes
occurs at special sites called membrane channels. At these sites,
specialized macromolecules let certain substances across the
membrane. For example, sodium channels use a polypeptide with an
enormous molecular weight of about 260,000 daltons. The rate of
movement of a substance through the membrane depends on the
concentration gradient (like in diffusion) and on voltage gradient (if
we are dealing with ions). There are substances that can cross
membranes in substantial quantities without the help of channels;
these are solutes that can dissolve in lipids, examples being
anesthetics and some other drugs.

2.3 Concentration of Water:

Osmosis
To measure concentration of all the particles in a volume of a
solvent, one needs to know the total number of different particles in
that volume. For this purpose, it is useful to borrow a special unit
from electrochemistry, namely a mole. A mole is the amount of a
substance for which the weight in grams is equal to the substance’s
molecular weight. For example, molecular hydrogen has a molecular
weight of 2 (1 for each hydrogen atom); thus, one mole of hydrogen
weighs 2 g; similarly, one mole of oxygen weighs 32 g (16 for each
atom). Note that one mole of any chemical substance, for example,
atom, molecule, or ion, always contains 6.02 times 10 to the power
of 23 particles (Avogadro’s number).
The concentration of water is measured as the total concentration
of all particles. Thus, the osmolarity of a solution with a
nondissociating substance (e.g., sucrose), will correspond to the
number of molecules of this substance. So, a 1 millimolar (mM,
remember that “milli” means divided by 10 to the power of 3) solution
has an osmolarity of 1 mOs (milliosmole). If the substance is one
that can dissociate, for example, salt, each molecule of this
substance (NaCl) will produce two particles, Na+ and Cl−, so that a 1
mM solution of NaCl has an osmolarity of 2 mOs. Note that the
concentration of a substance can change without changing its
amount (i.e., if the total volume of the cell changes). This can
happen, for example, if you put a red blood cell (erythrocyte) into a
solution with a smaller or higher concentration of salt than in blood
plasma.
 PROBLEM 2.1
What will happen with a red blood cell in these solutions? Note
that the membrane surface cannot change much.

A solution is called isoosmotic if it has the same concentration of

solute as the reference solution (plasma), hypoosmotic if it has a
lower concentration, and hyperosmotic if it has a higher
concentration of solute. These are nearly synonyms to the commonly
used terms isotonic, hypotonic, and hypertonic.
Osmosis is a process of movement of the solvent (for example,
water), rather than the solute, across the membrane, in order to
obtain osmotic equilibrium. Remember that motion of ions and other
particles through the membrane is typically restricted, while water
can travel freely.
It is important to understand that osmotic equilibrium (when water
does not move from one side of the membrane to the other) is
achieved only if the osmolarity of the solution on either side of the
membrane is equal. So, the concentration of particles inside the cell
(Si) should be equal to the concentration outside the cell (So). Note
that concentration equals the number of particles (A) divided by
volume of the site (V),

So, if you take a cell from a solution with the concentration of

particles S1 and place it into a new solution with the concentration of
particles S2, cell volume will change so that osmotic equilibrium is
reached (figure 2.5).
Initially, Si1 = So1. From equation 2.3,
A1/V1 = So1
In the new solution, similarly we get
A2/V2 = So2
A simple transformation gives:
Figure 2.5 If you take a cell from a solution with a concentration of particles S1
and place it into a new solution with a concentration of particles S2 (S2 > S1), cell
volume will change (decrease) until a new osmotic equilibrium is achieved.

So, in order to know how cell volume will change in a new

solution, we need to know the concentration of the solute outside
and the amount of the solute inside the cell.

PROBLEM 2.2
What will happen with the cell if it is placed in a solution containing
only permeable substances?

2.4 Movement of Ions: The

Nernst Equation
Ions move by both diffusion and voltage gradient (figure 2.6).
As mentioned earlier, diffusion is driven by a concentration
difference. The chemical force driving diffusion is termed chemical
potential (Fc):
Figure 2.6 Ions move under the influence of two forces. The first is related to the
concentration gradient (Fc), while the second is related to the difference of
potentials (Fe). C = concentration, V = voltage, T = temperature (Kelvin scale), z =
valence, R and Φ = coefficients (Avogadro’s number and Faraday number).

where R is the gas constant, T is absolute temperature (Kelvin

scale), and C is concentration.
If there is an external electrical field, the electrical force (Fe) acting
on a charged particle can be defined from:

where z is the valence (don’t forget that valence can be positive or

negative!), Φ is the Faraday constant, and V is voltage. So, the total
electrochemical force acting on an ion is

If an ion is in equilibrium, forces acting on particles of the ion on

the two sides of the membrane must be equal. These forces are
sometimes addressed as electrochemical potentials:

From this equation, one can calculate the equilibrium potential

(Veq) inside the membrane with respect to the potential outside (Veq
= Vin − Vout)—that is, the potential at which there is no net movement
of the ions through the membrane:

This is the Nernst equation. So equilibrium potential, by

definition, is an electric potential that induces the movement of an
ion across the membrane equal and in the opposite direction to
movement of the ion due to the difference in concentrations (note
that in figure 2.6, Fc, force due to the difference in concentrations,
and Fe, force due to the difference in the electric potentials, are
acting in opposite directions). Electric force acting on an ion is
directly proportional to its charge (i.e., it is twice as high in the case
of Ca++ as in the case of Na+ or K+). It is equal in magnitude and acts
in the opposite direction in the case of Cl− as compared to Na+. At
body temperature, RT/Φ is a constant (62 mV), and so:

Note the following properties of the equilibrium potential:

1. It is a measure of the concentration ratio for an ion that has
the meaning of energy available for diffusion;
2. it is a potential when there is no net passive movement of an
ion across the membrane; and
3. it is actual voltage on the membrane, but only if just one ion
species can move through it (e.g., if there is just one kind of
channel, as in squid axon membrane, which is permeable at
rest only to K+).

PROBLEM 2.3
In which direction will electric current flow across a membrane if
the potential inside the membrane is higher than Veq? Solve it for
Na+ and for Cl−.
 The direction of the current is defined by the potential on the
membrane, while its magnitude is defined by Ohm’s law (see
equation 2.2). So, for example, electric current due to the movement
of K+ ions will be:

where I is current, gk is the conductance for K+, V is voltage, and Vk

is equilibrium voltage for K+. Note that gk is not a constant and may
change quickly. Note also that the concentration gradients do not
change much during brief events such as an action potential. So,
virtually all ion movements through the membrane will be defined by
equation 2.11.

CHAPTER 2 IN A NUTSHELL
Biological membranes are unique
structures that allow cells to
interact with the environment and be
separate from it. Particles in
solutions can move among compartments
under the influence of differences in
pressure, differences in
concentration, and electrical field.
Osmosis is a process of movement of
the solvent, rather than solute,
across the membrane, in order to
equilibrate concentrations of all
particles. Equilibrium potential is a
potential on a membrane that creates
an electrical force acting on charged
particles, which is equal and opposing
the force due to the difference in
particle concentrations. It is defined
by the Nernst equation.
 Chapter 3

Action Potential

KEY TERMS AND TOPICS

membrane potential
ion conductance
sodium–potassium pump
generation of action potential

The action potential is the most important unit of information

transmission in the bodies of higher animals. Its importance is
immense. In lower animals, information within the body is transmitted
mostly by diffusion and by convection—that is, by bulk flow of liquids
containing important chemical substances. This mechanism of
information transmission is called humoral, and its speed is limited
by the rate of liquid flow under the difference in pressure. In the
process of evolution, the emergence of the ion mechanism leading to
the generation and transmission of action potentials signified a
many-fold increase in the speed of information processing and
conduction, giving species who possessed this “novelty” a significant
advantage in the everlasting competition of life. The humoral
mechanism of information transmission is still present in the higher
animals, but all the processes that require quick decision-making
and quick action take advantage of the much faster electrochemical
mechanism. Some of the properties, limitations, and disorders of
movements are rather directly linked to the mechanism of generation
and transmission of action potentials.

Figure 3.1 A membrane is separating two areas, with and without ions of Na+
and Cl−. Diffusion of these ions may occur at different speeds. As a result, a new
equilibrium will be reached with different ion concentrations to the right and to the
left, when the electric force will exactly compensate for the concentration gradient
force.

3.1 Creation of a Membrane

Potential
Consider a membrane separating a volume into two halves (figure
3.1). Initially, there is no NaCl to the right of the membrane, and
there is some to the left. Note that there is no voltage across the
membrane because the number of ions of Na+ to the left of the
membrane is exactly the same as the number of ions of Cl−.
Diffusion will begin because of the concentration gradient. However,
different ions may move at different velocities. Imagine that in our
case, Cl− moves faster. Then, when the concentration of ions on both
sides is equal, there will be a little bit more of Cl− to the right and a
little bit more of Na+ to the left. Thus, an electric potential will emerge
across the membrane, or, more precisely, a difference of potentials
known as the membrane potential. Note that the potential is being
created not by all the ions but only by a tiny fraction that is not
balanced. For example, the extra amount of ions needed to create a
potential of 100 mV (a typical value for membrane potentials) is only
10-12 M (one picomole) for the area of membrane of 1 cm2. All
biological potentials are created by tiny amounts of unbalanced ions.
So, to a good approximation, you can always consider the total
concentration of positive ions in a solution to be equal to the total
concentration of negative ions.

PROBLEM 3.1
Find an error (an imprecise statement) in the previous paragraph.

Note that the membrane also behaves like an electric capacitor, a

physical structure able to store electrical charge in the presence of
an external electric field. In particular, like in regular electric
capacitors, its ability to store electric charge depends on its surface
but not on the volume of the solution. The net (unbalanced) charge
(Q) on a membrane equals its capacitance (C) multiplied by voltage
(V) across the membrane:

This is a version of Coulomb’s law, which is also the definition for

capacitance as a coefficient of proportionality between the difference
of potentials and stored charge.
When the voltage changes, the charge changes as well. By
definition, capacitative current is the change of charge (Ic = dQ/dt).
So,
 Note that capacitative current is different from the current created
by the movement of ions through a membrane. Capacitative current
is created by changing an electric field, and it does not require any
carriers or channels. It may play a significant role in some cases of
small changes in membrane potential.
There are several important ions that play special roles in the
electric phenomena in neurons of mammals. We are now going to
consider sodium (Na+), potassium (K+), and chlorine (Cl−). Their
concentrations inside and outside a membrane are rather different
(figure 3.2). Using the Nernst equation, we can calculate equilibrium
potentials for these ions. Later, we will turn our attention to one more
important ion, Ca++, which has very low concentration inside the
typical membrane.

Figure 3.2 A membrane may be considered a capacitor. Its charge (Q) is

proportional to the difference of potentials (V) across the membrane with a
coefficient termed capacitance (C).

PROBLEM 3.2
Calculate (approximately) equilibrium potentials for Na+, K+, and
Cl−.
 Figure 3.3 shows typical concentrations of the three ions K+, Na+,
and Cl− inside and outside the membrane and the corresponding
equilibrium potentials. The difference in ion concentrations inside
and outside the membrane is maintained actively and requires
energy. This mechanism is commonly called the sodium–
potassium pump. Ion pumps involve the action of proteins that
carry ions against their concentration gradient. The sodium–
potassium pump involves proteins that carry Na+ outside the
membrane and carry K+ inside the membrane. The number of carried
ions is not balanced: The protein carries two K+ ions inside for every
three Na+ ions carried outside. As a result, the pump generates a net
negative potential inside the membrane or, in other words,
contributes to its hyperpolarization. Figure 3.4 shows schematically
how the pump works, receiving energy from adenosine triphosphate
(ATP) stored in mitochondria.

Figure 3.3 The differences in the concentration of the three most important ions
across the membrane. Equilibrium potentials for each ion are shown in
parentheses.
Figure 3.4 Maintaining the ion concentration gradients across the membrane
requires energy, which is provided by a chemical process that transforms ATP
(stored in mitochondria) into ADP. This mechanism is called the sodium–potassium
pump.

Let us imagine that a number of ions, for example K+, Na+, and
Cl−, can cross a membrane through the same channels and thus are
in a competition. Membrane potential will be defined according to
equation 3.3:

where P is permeability of the membrane to an ion and C is

concentration of an ion (note the subscripts K, Na, and Cl) in and out
of the cell. This is the Goldman-Hodgkin-Katz equation.
If the channels are perfectly selective, the equation will look like
this:

where g is conductance and E is equilibrium potential for a given ion.

You may consider g as a reflection of the number of open channels
for a particular ion. Then, the more channels that are open, the
bigger is the contribution of the equilibrium potential of this ion to the
actual resting potential on the membrane. In particular, if only one
ion can cross the membrane, its resting membrane potential will be
equal to the equilibrium potential for that particular ion. Equation 3.4
is a decent approximation since membrane channels are rather ion-
specific. However, note that it is not applicable during fast changes in
membrane potential, as for example, during the action potential.

PROBLEM 3.3
Why is equation 3.4 inadequate during fast changes in membrane
potential? What has not been taken into account? Why, in
equation 3.3, are members for Cl− represented differently than
those for K+ and for Na+?

Actual resting (or equilibrium) potential on the membrane depends

on the values of conductance (g) for different ions and on their
equilibrium potentials. The resting membrane potential can be quite
different in different neurons; for example, it is about −40 mV in rods
and cones in the retina at rest and about −75 mV in the pyramidal
neurons in the cortex. The resting membrane potential can also
show changes within the sleep-wakefulness cycle (Hirsch et al.
1983). For simplicity, we are now going to assume that the resting
potential equals −70 mV and is mostly defined by the higher g for
ions of Cl− as compared to the conductance values for K+ and Na+
ions. This value is not far from the membrane potential on the squid
giant axon, which has a diameter of about 0.5 mm and, as a result,
can be pierced by relatively large microelectrodes. Studies on the
squid giant axon by Alan Hodgkin and Andrew Huxley in the 1930s
documented a potential inside its membrane of about 60 mV.

3.2 Basic Features of the

Action Potential
The word “potential” has quite a few meanings. We are going to
speak now about a process, a time function of the transmembrane
voltage that we are going to call action potential. Don’t confuse it
with membrane potential, which describes a membrane’s state at a
particular instant of time.

Figure 3.5 If you stimulate a membrane with relatively small electrical stimuli, its
resting potential will change somewhat in response to each stimulus, and then will
return to its resting level.

One of the most interesting features of an action potential is its

threshold nature. Imagine that you start to stimulate a membrane by
applying short pulses of electric current through the membrane with
an external stimulator (figure 3.5). At low values of the stimulating
current, the membrane will respond with a small change in its
potential that will rather quickly return to its equilibrium (or resting)
value. Your stimulus will certainly spread because an electric field
spreads, but it will not spread far because the electric field drops
quickly with distance from the source of stimulation. So, the maximal
deviation of the membrane potential from its resting value will be
seen close to the site of stimulation only.
If you start to increase the current, the deviation of the membrane
potential will also increase (figure 3.6), and at a certain value of the
stimulus something amazing will happen: The membrane will
respond with a disproportionally huge change in its potential. The
value of membrane potential at which this qualitative change occurs
is termed the membrane threshold or stimulation threshold. If you
continue to increase the strength of the stimulus, surprisingly, no
further change will occur. The membrane will react with exactly the
same action potential. This feature of the action potential to either be
of a standard height or not to be at all without any intermediate
behavior is termed the all-or-none law.

Figure 3.6 An increase in the stimulation current will lead, at low values, to a
gradual increase in the deviation of the membrane potential from its resting level.
At some value of the stimulus, when it reaches a certain threshold magnitude, an
action potential will be generated. Further increase in the strength of the
stimulation will not lead to a change in the membrane response.

PROBLEM 3.4
Suggest examples of the all-or-none law from everyday life.

Please note that we are addressing now transmembrane

potentials (i.e., the difference between the potential inside the
membrane and the potential outside the membrane). If you put a
couple of measuring electrodes outside the membrane, you can
record a difference of potentials between the electrodes but not
across the membrane. Extracellular potentials are typically much
smaller than the amplitude of the action potential (by a factor of one
thousand!). Similarly, if you try to stimulate a membrane without
penetrating it, you will need rather high currents because the
extracellular solution and the membrane effectively shield the inside
of the cell from the effects of externally applied currents.
Let us perform a mental experiment and insert a very thin
stimulating electrode through the membrane into a cell so that the
integrity of the membrane is not violated (figure 3.7). If we now apply
current that makes the voltage in the cell more negative, the change
in the membrane potential will be called hyperpolarization. A
current in the opposite direction will induce a change in the
transmembrane potential called depolarization. Both
hyperpolarization and depolarization spread electrotonically (i.e.,
they affect neighboring areas of the membrane due to the
membrane’s electrical properties), and typically they quickly become
smaller and disappear with distance.

Figure 3.7 A thin electrode is inserted into the cell without breaking the
membrane. Now we can apply electrical current to change the membrane resting
potential either toward lower (depolarization) or higher (hyperpolarization)
magnitudes of membrane polarization.
3.3 Mechanisms of Generating an
Action Potential
First, it is important to realize that an action potential emerges
because of the dependence of membrane permeability for certain
ions upon the membrane potential.
Let us consider an example (figure 3.8). There is only one ion that
can move through special channels in a membrane. Each channel is
being guarded by a demon who sometimes falls asleep. The
probability of the demon falling asleep depends on membrane
potential so that, at rest, all the demons are awake and do not let the
ions cross the membrane. We can apply short-lasting pulses of
stimulation to change the membrane potential. A depolarizing pulse
puts some of the demons to sleep so that some ions can cross the
membrane. The more demons that are asleep, the bigger is the
current created by the ions. Note, however, that the current itself will
change the membrane potential.
Figure 3.8 A nonscientific illustration. A demon is guarding each channel for Na+
in the membrane. Membrane depolarization makes some of the demons fall
asleep, so that their channels become open. Ions will cross the membrane and will
increase the depolarization, putting more demons to sleep.

Now, we have two major possibilities:

1. The current hyperpolarizes the membrane and, therefore,
wakes up some of the demons, who quickly start closing the
gates and restoring the resting potential.
2. The current further depolarizes the membrane (i.e., works in
the same direction as the stimulus). Then, the current puts to
sleep more demons, thus opening more channels, thus
increasing the current, thus putting to sleep more demons,
thus … and so on.
The process described in the second example is called positive
feedback (figure 3.9), while the first possibility corresponds to
negative feedback. Apparently, systems with positive feedback are
capable of generating large signals very quickly, while systems with
negative feedback generally tend to bring any “perturbing” signal
down to zero.
 A very similar mechanism gives rise to the “all or none” signal that
emerges on the membrane when it is excited to the threshold:
depolarization increases membrane permeability to a certain ion,
while increased permeability induces membrane current that
increases depolarization.

Figure 3.9 A positive feedback process where (a) leads to a rapid amplification
of the effect, while a negative feedback process (b) quickly restores the original
state.

One can study the mechanisms involved in the process of

generating an action potential with the voltage clamp technique. This
technique is used to keep membrane potential at a certain level with
the help of external electronics, which add electrical charges to or
remove them from the membrane, thus keeping the potential
constant (like a thermostat keeping room temperature constant by
adding or removing heat). These conditions do not allow the positive
feedback mechanism to generate an action potential, but they allow
the experimenter to study the dependence of conductance in
specialized ion channels upon membrane potential.
Figure 3.10 A constant depolarization is applied to the membrane. Note that Na+
conductance (gNa) turns on and off while K+ conductance (gK) changes slowly and
stays at a new level. Note also that higher stimuli lead to higher values of gNa
achieved over shorter times.

Figure 3.10 shows the dependence of the sodium conductance

(gNa) upon voltage after a depolarizing voltage step is applied to the
membrane. Note that gNa turns off spontaneously—that is, it goes
down to its original, very low value without an obvious additional
external stimulus, even when the membrane voltage is kept constant
artificially (shown by the “Stim” line in figure 3.10). At higher
stimulation voltages, peak values of gNa are much higher, they are
reached more quickly, and gNa drops faster. Note also that the time it
takes gNa to reach peak value is shorter for larger stimuli while for gK
it is almost unchanged. When the conductance for both major ions,
Na+ and K+, is increased, one can say that all the channels are open,
the sodium–potassium pump becomes functionally disabled, and
membrane potential changes are primarily defined by ion movement
through the open channels.
If you turn the voltage off and let the membrane potential return to
its resting value, gNa gets down close to zero if it is not zero already.
There is an important phenomenon of channel inactivation,
which means that after a spontaneous turn-off, gNa cannot be
increased immediately even if you apply a very strong voltage (figure
3.11). It needs some time to recover. When the conductance cannot
be increased by any external voltage, the membrane is said to be in
an absolute refractory period. When you need higher than usual
voltage to increase the conductance, but you can do it, the nerve is
said to be in a relative refractory period.

Figure 3.11 After a stimulus (St1) leading to an increase in gNa, another stimulus
is less able to turn it on for some time. For a short period of time this inactivation is
absolute—that is, gNa will not respond even to a very strong stimulus (St2, absolute
refractory period). Then, a stronger-than-usual stimulus can turn gNa on (St3,
relative refractory period).
 So, the channels can close in response to a change of membrane
potential and also spontaneously, when they require some time to
recover. In the first case, you can open the channels with an external
depolarizing stimulus; in the second case, the only available method
is to wait.
Figure 3.10 also shows the dependence of potassium
conductance (gK) on membrane potential. Note that gK starts to
increase with depolarization but does not turn off spontaneously. It
goes down to its original value only when the membrane potential
returns to its resting level. This means that there is no channel
inactivation, and no refractory period for potassium channels. Note
also that gK increases more slowly than gNa, which means that early
in the process of membrane depolarization, the open sodium
channels will play a bigger role.
As can be seen from the figures already shown, both gNa and gK
behave “smoothly” with membrane voltage (i.e., they do not show
any threshold effects). In order to understand the mechanisms giving
rise to the all-or-none action potential, we need to remove the
voltage clamp and allow the potential to change.
Note that opening channels for sodium and for potassium leads to
different consequences for the membrane potential because of the
difference in the concentrations of Na+ and K+ ions inside and
outside the cell. An increase in gK, for example, induced by a short
depolarizing pulse, leads to a flow of K+ out of the cell. The loss of
positive ions leads to a drop in the membrane potential (remember,
membrane potential is measured inside the cell with respect to the
outside!)—that is, to a decrease in depolarization or to
hyperpolarization. This, in turn, will lead to a drop in gK. So we have
a system with a negative feedback that will quickly restore the
original resting potential. An increase in gNa, however, will lead to an
inflow of Na+ inside the cell (i.e., to further depolarization). Here we
deal with a system with a positive feedback which, as is well known,
loves to go berserk. Different dependencies of gNa and gK on
membrane potential, together with the property of sodium channels
to inactivate, lead to the generation of the action potential.
 Remember that the direction of flow of an ion depends on its
equilibrium potential such that the difference between actual
membrane potential and the equilibrium potential of an ion defines
the direction in which the ion will flow. On the other hand, an ion with
a higher permeability plays a bigger role in defining the overall
membrane potential than ions with smaller permeabilities. This
means that changes in gNa and gK can lead to changes in the resting
membrane potential.
Figure 3.12 shows an action potential and the changes in gNa and
gK in different phases of the potential. The sequence of the events is
as follows:
1. The initial depolarization (created by an external stimulus)
increases gNa so that the membrane potential tries to reach
sodium equilibrium potential.
2. The sodium channels are quickly inactivated, leading to a drop
in gNa; as a result, the membrane potential has no time to
reach the sodium equilibrium potential.
3. After a brief delay, gK increases and draws membrane
potential to its equilibrium potential (i.e., repolarizes the
membrane).
4. There is a rather long period of hyperpolarization (the
afterpotential) after which membrane potential returns to its
resting value.
Figure 3.12 Changes in Na+ and K+ conductance during an action potential.
Note that the peak of the action potential is positive, and after the action potential
the membrane remains hyperpolarized for some time.

PROBLEM 3.5
Why does the membrane potential drop below the resting level?
Can you imagine a situation in which the afterpotential would be
higher than the resting potential?

Neurons within the central nervous system can exhibit a variety of

electrophysiological properties reflected in different shapes and
features of their action potentials (figure 3.12). For example, the
large Purkinje cells in the cerebellum can generate very high
frequency trains of action potentials (over 200 Hz) interrupted by
effects from Ca++ channels. These cells can also generate unusual,
long-lasting action potentials (complex spikes) discussed later in
the chapter on the cerebellum. Broad action potentials at lower
frequencies can be seen in other neurons in the brain due to an
interaction among conductance changes for Ca++, Na+, and K+. Many
neurons in the central nervous system can generate rhythmic bursts
—clusters of action potentials at a very high frequency within each
cluster.
Until now, we have only discussed transient, relatively short-
lasting ion currents through the membrane. There are also longer-
lasting, persistent ion currents. In particular, persistent Na+ currents
can be seen in response to relatively low levels of membrane
depolarization, below the level needed for the generation of an
action potential. These currents can help the membrane to generate
an action potential in response to relatively small depolarizing
stimuli. They can also contribute to the plateau potentials—relatively
long-lasting (up to a few seconds) membrane depolarization.
Persistent Ca++-dependent currents can be seen in relatively thin
neural fibers—dendrites. These will be discussed later in more detail.

CHAPTER 3 IN A NUTSHELL
The action potential is the unit of
information transmission within the
bodies of higher animals. Membrane
potential is created by a small number
of unbalanced ions. Movement of ions
through the membrane occurs at special
sites called ion channels. An active
molecular mechanism, a sodium–
potassium pump, maintains the
difference in the concentrations of
sodium and potassium ions across the
membrane. The dependence of the sodium
ions’ conductance on the membrane
potential leads to the generation of
an action potential when membrane
depolarization reaches its threshold.
After an action potential, the
membrane stays in a short-lasting
state of insensitivity due to the
inactivation of sodium channels.
 Chapter 4

Information Conduction and

Transmission

KEY TERMS AND TOPICS

conduction of action potentials
neuron
myelinated and nonmyelinated axons
information transmission in the central
nervous system
synapse
synaptic transmission
neurotransmitters
persistent inward currents
temporal and spatial summation

As mentioned earlier, the action potential is probably the most

important process or event in our body because it is used for
information transmission over considerable distances within the
neuromuscular system. An important feature of the action potential is
its propagation—that is, an action potential never stays at one place;
it travels along nerve or muscular fibers. One of the main
characteristics of action potential propagation is its velocity. Since an
action potential is a process, a time function measuring its velocity
requires identification of a well-defined point on the action potential
curve (e.g., the time of its peak) and a measure of the time interval
from the moment when this point occurs at one location to the
moment when it occurs at another location: ΔT = T2 − T1. If a
researcher stimulates a neural fiber with an electrical stimulator at a
certain point, an action potential may occur at this point. The
membrane potential recorded further down the nerve fiber will show
a similar time function—that is, a similar action potential that will
occur, however, at a time delay after the first one (figure 4.1). If one
knows the distance between the two points (ΔS), the average
velocity of transmission (V) can be computed as:

Figure 4.1 Action potential travels along a neural fiber. For calculation of the
velocity of its propagation (V), the action potential has to be recorded at different
times (t1 and t2) at different coordinates (s1 and s2).
 PROBLEM 4.1
If a neural fiber is stimulated strongly at some point in the middle,
in what direction will the action potential propagate?

4.1 Conduction of an Action

Potential
When an action potential occurs in a certain segment of a cellular
membrane, it sets up local current circuits that flow to the
neighboring segments largely according to Ohm’s law (i.e., without
any help from pumps, channels, and other sophisticated
mechanisms). The charge leaks through the membrane capacitance
according to equation 3.1 in chapter 3. Figure 4.2 shows
schematically a simplified electrical system of a membrane and the
currents that flow within this system. The local currents depolarize
the membrane and, if the depolarization is strong enough, another
action potential(s) can occur. So, strictly speaking, the action
potential does not travel along a membrane of a neural fiber but
rather emerges at different spots and disappears, giving rise to new
potentials. However, since all the potentials look alike (because of
the all-or-none law), the process looks as if one potential were
traveling along the fiber.
Figure 4.2 A simple electrical scheme of a membrane and the direction of local
currents (shown by arrows).

Two factors are very important for the process of propagation of

the action potential:
1. inactivation of sodium channels leading to the absolute
refractory period within an area of the membrane just after an
action potential, and
2. different densities of sodium channels at different sections of
the membrane.
The first factor does not allow an action potential to “backfire” during
its natural propagation along a fiber. That is, if an action potential
appears at point 1 in figure 4.3 at time t1 and then disappears, giving
rise to an action potential at a neighboring point 2 at t2, the
membrane at point 1 stays refractory for some time and cannot be
excited by local currents created by the second action potential. So,
the second action potential at point 2 can excite the membrane at
point 3 but not back at point 1.
Figure 4.3 The phenomenon of inactivation of sodium channels does not allow
an action potential to “backfire.” If an action potential comes to point 2 from point 1,
it cannot go back; it travels only forward to point 3.

PROBLEM 4.2
What would happen in response to a single strong stimulus
applied to the membrane of a long neural fiber if there were no
sodium inactivation mechanism?

The second factor makes some areas of the membrane more

readily excitable and, therefore, favors the generation of an action
potential in those particular areas.
The process of generation of a single action potential is brief, but
nevertheless it takes time on the order of 1 ms, while local currents
spread almost instantaneously. Let us consider two neural fibers. An
action potential has just emerged in both at point A in figure 4.4. The
local currents generated by the action potential spread in the
surrounding tissue and decrease very quickly with distance from
point A. However, the current spreads more easily, and decreases at
larger distances, in thick fibers as compared to thin fibers. So the
next most distant action potential will be generated in fiber 1 at point
B and in fiber 2 at point C. Note that it will take the same time for the
potential to jump over distance (B–A) in fiber 1 and (C–A) in fiber B.
We have come to the conclusion that thick fibers conduct action
potentials at higher velocities.
Although action potentials typically have the same shape, their
width can change under certain special circumstances. Wider action
potentials induce larger local currents and can bring more distant
areas of the membrane to its threshold. Figure 4.5 illustrates the
dependence between the pulse’s duration and the strength it needs
to just reach the threshold at a certain point on a membrane and
between pulse duration and maximal distance at which membrane
potential can reach the threshold for a constant amplitude of the
pulse. Note that there is a value in the strength of a stimulus
(rheobase) below which an increase in the duration is unable to
induce an action potential.

Figure 4.4 Action potentials emerge simultaneously at point A in fibers 1 and 2.

Local currents decrease with distance more slowly in the thicker fiber 2. Thus, they
will bring the membrane to the threshold at a more distant point. So the next action
potential will emerge at the same time at point B in fiber 1 and at point C in fiber 2.
Hence, the speed of conduction is higher in the thicker fiber.
Figure 4.5 The dependence between stimulus duration and the amplitude
needed to just reach the threshold of a membrane.

4.2 Myelinated Fibers

Some neural fibers, particularly the thicker ones, are covered with a
sheath made of myelin that is built of specialized, nonneuronal
Schwann cells. Schwann cells wrap around such neural fibers in
several layers, forming the myelin sheath. The sheath has breaks
that are called Ranvier nodes (figure 4.6). This design allows the
action potential to travel at much higher speeds. There are two
important features of myelinated fibers. First, myelin sheath
increases the distance at which local currents from an action
potential are able to reach the threshold of membrane depolarization
for the generation of another action potential. Second, the sodium
channels are concentrated in Ranvier nodes so that their density
there is much higher than average, and their density under the
myelin sheath is much lower than average. As a result, if an action
potential occurs at a Ranvier node, it gives rise to local currents that
bring the membrane to the threshold at the neighboring node, and
the action potential kind of jumps from one node to another. This
phenomenon is addressed as saltatory conduction, which is
characterized by a considerable increase in conduction velocity.
Thicker fibers have larger intervals between neighboring Ranvier
nodes, and therefore, they conduct action potentials at higher
velocities. Actually, there is a very simple equation relating the
diameter of a myelinated fiber and the conduction velocity of action
potential:

where V is velocity in m/s and d is fiber diameter in microns. Note

that this equation is not applicable to nonmyelinated fibers.

Figure 4.6 A myelinated fiber is enclosed in a sheath made of nonneural cells

(glial cells). The myelin sheath has breaks (Ranvier nodes) where action potentials
are generated.

PROBLEM 4.3
What will happen if a myelinated fiber suddenly loses its sheath?
What can you expect from such a fiber in a hot bath and in a cold
bath? Note that ion diffusion proceeds much more quickly at high
temperatures.
 Table 4.1 compares the velocities of different processes. Note that
speeds of conduction in our body are rather high, but not extremely
high. Certainly, they are not comparable to the speed of light, which
is equal to the speed of propagation of an electromagnetic field.

Table 4.1 Characteristic Velocities of Different Processes

Process Velocity (m/s)
Slow nerve conduction 0.5
Sprinting (humans) 10
Driving at 90 km/h (65 mph) 25
Fast nerve conduction 120
Sound traveling in air 330
Light (electromagnetic field) 300,000,000

PROBLEM 4.4
The nature of an action potential is electric. Why is the speed of
its propagation so much lower that the speed of electric events
like electric current?

PROBLEM 4.5
The speed of neural conduction is comparable to the highest
movement velocities observed in athletes. Does this mean that
there is an upper limit for movement velocity set by action
potential conduction speed?

Knowing the speeds of conduction is very important for

understanding the characteristics of many neurophysiological
processes that include conduction of information from one place to
another within our bodies. In some situations, these conduction
delays dominate in the total time delay between a stimulus and a
response—for example, in the case of spinal reflexes described in a
later chapter. In other situations—for example, when a person is
asked to perform a simple action (e.g., press a button) as quickly as
possible following a signal—conduction delays contribute
significantly to the reaction time, although they may not dominate it.
We are going to encounter a few classifications in this textbook.
One of the most helpful and commonly used is the classification of
neural fibers according to their diameter and function, sensory
versus motor, or, using a different pair of terms accepted in
neurophysiology, afferent versus efferent. It was suggested by a
great physiologist, David Lloyd, and is illustrated in table 4.2.

4.3 Structure of a Neuron

Before moving further, we need to introduce certain basic notions
related to the structure of a single neural cell (a neuron). Figure 4.7
shows an illustration of a “typical neuron”; the quotation marks reflect
the fact that many neurons within the human body look significantly
different from this cartoon. The neuron consists of three major parts:
soma, axon, and dendrites.
The soma, or body of the neuron, contains the nucleus (or several
nuclei) and other important small structures (organelles).
Mitochondria are major storage places and sources for the release of
molecules whose chemical transformations generate energy for the
processes inside the cell, in particular for the sodium–potassium
pump.

Table 4.2 Types of Neural Fibers (Axons) and Conduction Speeds

of Action Potentials
Innervated Fiber diameter Conduction
Type structure (microns) velocity (m/s)

Afferent or sensory muscle nerves*

Ia Muscle spindle, 13-20 80-120
(Aα) primary endings
Ib Golgi tendon organ 13-20 80-120
(Aα)
 Innervated Fiber diameter Conduction
Type structure (microns) velocity (m/s)
II Muscle spindle, 6-12 40-80
(Aβ) secondary endings
III Muscle deep pressure 1-5 5-30
(Aδ) endings
IV (C) Nociceptors (pain) 0.2-1.5 0.5-2

Efferent or motor nerves

Aα Skeletal muscles 18 100
Aβ Muscles and spindles 8 50
Aγ Muscle spindle 5 20

*Classifications for cutaneous nerves are shown in parentheses.

Figure 4.7 A neuron schematic.

The axon is typically a long, rather thick branch that carries the
output signals generated by the cell. At its end, the axon splits into a
whole bunch of smaller, thin branches (terminal branches) that make
contacts with other cells and transduce information to these cells.
These branches are commonly much shorter than the main part of
the axon. Axons can be very long, up to 1 meter, as in the case of
the axon of a motoneuron with its soma located in the spinal cord
that sends signals to a muscle in a foot. There are long axons of
neurons within the central nervous system as well, including axons
of neurons in the cortex of the large hemispheres that send their
signals to neurons in the lower parts of the spinal cord. The place
where the axon exits the soma is called the axon hillock. At this
place, the density of sodium channels is very high, and so this is the
place where action potentials are typically being generated. The
axons of groups of neural cells, typically united by a functional or
anatomical feature, commonly run together over relatively large
distances. In such cases, the groups of axons are addressed as
neural tracts (if they travel from one place to another within the
central nervous system) or as nerves (if they connect the central
nervous system with peripheral structures such as muscles and
sensory organs).
Dendrites form a tree around the soma and serve as sites of
inputs into the cell. Terminal branches of the axons of other cells
make connections (synapses) on the dendrites as well as on the
soma itself.
To summarize, dendrites and soma serve as sites where
information comes to the neuron from other neurons and is
integrated (assessed, compared, and put together); the axon hillock
is the place where action potentials are generated in response to the
incoming information, and the axon serves to conduct action
potentials to distant places and to transmit information to other cells.
Dendrites’ properties have attracted much attention in relation to
the ability of the dendrite membranes to show steady depolarization.
In other words, dendrites have been shown both theoretically
(Gutman 1991) and experimentally (Schwindt and Crill 1977, 1981;
Heckman et al. 2003, 2005) to have long-lasting changes in the
membrane potential induced by persistent inward currents. A
persistent inward current is a depolarizing current produced by
voltage-sensitive channels that do not show the phenomenon of
inactivation (this is why they can be long-lasting or persistent). These
channels on dendrites are specialized for Ca++ ions. Effectively,
persistent inward currents may be seen as the means of reducing
the threshold for action potential generation and, hence, reducing the
distance from the membrane resting potential to this threshold. So
these currents seem to facilitate the process of generating action
potentials. On the other hand, persistent inward currents tend to
sharply limit the efficacy of additional synaptic inputs into the neuron,
so their overall effects on the excitability of the neuron may not be
easily predictable.
Figure 4.8 illustrates the dependence between the voltage and
current over a membrane in the absence of persistent inward
currents (thin dashed curve) and when these currents are present
(thick solid curve). Note that when the current is zero, the
membrane, by definition, is at its resting potential. The thin line
shows only one value of the membrane potential when the current is
zero (i.e., only one value of the resting potential). The thick line
shows three crossings of the abscissa axis. The first and third
crossings are stable, while the intermediate one is not. An amazing
feature of the second resting potential (point 3 in figure 4.8) is that it
can be above the threshold for the generation of an action potential.
If the resting potential of a dendrite membrane is at that second
state, the dendrite starts generating action potentials and would
continue doing so without any external stimuli as long as the
membrane potential stays above the threshold value and the neuron
does not run out of energy. We will discuss the implications of this
phenomenon for voluntary movements in later chapters.
Figure 4.8 Current–voltage characteristics of a dendrite membrane in the
absence (thin lines) and in the presence (thick lines) of persistent inward currents
(PIC). In the second case, potentially there may be two stable values of the
membrane potential (points 1 and 3). Point 3 may be over the membrane threshold
for the generation of action potentials.

4.4 Information Coding in the

Nervous System
Because of the all-or-none law, individual neurons can only generate
single action potentials of a relatively constant duration and
amplitude. Thus, an action potential by itself transmits only limited
amounts of information: It either occurs or it does not occur at any
point in time. The only way a neuron can encode significant amounts
of information is by generating sequences of action potentials. In
other words, information is encoded by changing the frequency of
firing. Note that here we talk about instantaneous frequency of
neuronal firing—that is, an inverse of the time interval between two
successive action potentials:
where T2 and T1 are times of occurrence of two successive action
potentials. Intervals between successive action potentials fluctuate
all the time, even if the apparent input to the neuron stays constant.
So, neurons never fire at a constant frequency, and they should be
characterized either by instantaneous frequency or by average
frequency of firing over some time period. As mentioned earlier,
some neurons can demonstrate bursts of action potentials at a
relatively high frequency separated by intervals of silence; in such
cases, one number is apparently not enough to describe the
behavior.
This type of information transmission is called frequency coding or
frequency modulation. However, if one considers groups of neurons,
frequency coding loses its exclusive right as the only method of
information transmission in the central nervous system because of
possible changes in the number of neurons within the group that is
generating action potentials within any time interval. Hence,
frequency modulation is supplemented with amplitude (or
magnitude) modulation. The ability of neurons to integrate incoming
information (see the discussion of spatial and temporal summation in
section 4.7) allows them to take into account both the timing of
incoming action potentials and their number. So, the firing rate
(instantaneous frequency of firing) of a neuron depends on both the
frequency and the magnitude of its input.

4.5 Synaptic Transmission

A very important feature of neurons is the ability both to conduct
information from one place to another and to transmit it to other cells.
In this textbook, we focus primarily on synaptic transmission that
does not involve electrical contact between the membranes of two
cells. Electrical transmission via a mechanism addressed as
ephaptic was demonstrated in the brains of vertebrates (Bennett et
al. 1959), including mammals (Baker and Llinas 1971; Korn et al.
1973). In contrast to the more typical synaptic transmission involving
neuromediators, such electrotonic synapses (sometimes called
ephapses) are primarily excitatory, can transmit signals in both
directions (from cell A to cell B and from cell B to cell A), and are
less able to show plastic changes.
More typically, transmission of information from one cell to another
occurs at specialized sites of the membranes of the two cells and do
not involve direct contact. At these sites, the membranes come very
close to each other and form synapses. A synapse consists of three
major components: presynaptic membrane, postsynaptic membrane,
and synaptic cleft (figure 4.9). The presynaptic membrane belongs to
the cell that transmits information (coded as a sequence of action
potentials), while the postsynaptic membrane belongs to the cell that
receives the information.

Figure 4.9 A synapse consists of a presynaptic membrane, a synaptic cleft, and

a postsynaptic membrane. An action potential in the presynaptic fiber makes
synaptic vesicles move to the membrane, fuse with it, and release molecules of a
neurotransmitter into the cleft. The neurotransmitter acts at the postsynaptic
membrane and changes its potential.
 There are two major groups of synapses, obligatory and non-
obligatory. If an action potential on the presynaptic membrane
always gives rise to an action potential on the postsynaptic
membrane, such a synapse is called obligatory. Typical examples of
obligatory synapses are those between neural cells and muscle
cells. Non-obligatory synapses are much more common within the
central nervous system: A single action potential on the presynaptic
membrane is typically unable to induce an action potential on the
postsynaptic membrane.
Neuron-to-neuron synaptic transmission uses various chemical
substances called neurotransmitters or synaptic mediators.
Neurotransmitters are normally synthesized by the presynaptic
neuron and stored in special reservoirs (vesicles) close to the
presynaptic membrane. The typical scheme of synaptic transmission
is as follows (figure 4.9):
1. An action potential arrives to the presynaptic membrane.
2. It induces chemical changes in the membrane properties (with
an important role played by Ca++ ions), which lead to
movement of vesicles with a neurotransmitter to the
presynaptic membrane, their fusion with the membrane, and
release of the neurotransmitter molecules into the synaptic
cleft. This process is called exocytosis.
3. The molecules of the neurotransmitter travel across the
synaptic cleft (its typical width is on the order of 100 nm) by
passive diffusion.
4. These molecules act at special sites (receptors) on the
postsynaptic membrane and change its potential.
5. These molecules are quickly removed from the synaptic cleft
by a special chemical substance (an enzyme) or are taken
back into the presynaptic membrane and recycled.
Molecules of neurotransmitter bind to receptor sites on the
postsynaptic membrane and induce one of two basic effects: They
can depolarize the membrane or they can hyperpolarize it (figure
4.10). In the first case, a depolarizing potential will appear that is
called an excitatory postsynaptic potential (EPSP). In the second
case, a hyperpolarizing potential will emerge that is called an
inhibitory postsynaptic potential (IPSP). The peak magnitude of
EPSPs and IPSPs is relatively small in neuro-neural synapses and
can be very large (up to several tens of millivolts) in neuromuscular
synapses. In response to a single presynaptic action potential,
postsynaptic potentials last for about 15 ms and then disappear.
When a number of action potentials come to presynaptic membranes
that make synapses with the same postsynaptic membrane, the
balance of EPSPs and IPSPs on the postsynaptic membrane will
define whether the potential reaches the threshold and whether an
action potential is generated.

Figure 4.10 A presynaptic action potential can induce either a depolarization or a

hyperpolarization of the postsynaptic membrane. These effects are called EPSP
and IPSP, respectively.

Note that typical synapses transmit signals in one direction only, in

particular because there are no vesicles with neurotransmitters in the
postsynaptic membrane and no receptors sensitive to
neurotransmitters on the presynaptic membrane. The efficacy of
neuro-neural synapses can be modified by previous activity of the
presynaptic neuron, which leads to plasticity and activity-dependent
changes in neural transmission—a very important mechanism for
memory and learning.
4.6 Neurotransmitters
Neurotransmitter release is quantal in nature. It can happen
spontaneously, leading to small, stereotypical potentials on the
postsynaptic membrane called miniature excitatory or inhibitory
postsynaptic potentials (MPSPs). There are three major groups of
neurotransmitters: amino acids, biogenic amines, and
neuropeptides.
Amino acids are building blocks for all proteins and are very
common in our bodies, in the nervous system in particular. Not all
amino acids act as neurotransmitters. One of the most frequently
encountered neurotransmitters is gamma-aminobutyric acid
(GABA). It can be found in a significant proportion of all the
synapses (about 25% to 40%). Among dominant excitatory
neurotransmitters let us mention glutamic acid and leucine. They
depolarize the postsynaptic membrane and thus bring it closer to the
threshold for the generation of an action potential. Glycine is an
inhibitory mediator found, in particular, in the spinal cord.
Biogenic amines are found in smaller quantities than amino acids.
There are several biogenic amines whose role as neurotransmitters
is particularly important. These are acetylcholine, serotonin,
dopamine, and norepinephrine. Their action on the postsynaptic
membrane is not as unambiguous as that of GABA and glutamic
acid. In particular, acetylcholine commonly exerts inhibitory effects
on postsynaptic neurons within the central nervous system, but it is
also the most important excitatory mediator in the transmission of
signals from neurons to muscle fibers. The phenomenon of
persistent inward currents mentioned in an earlier section involves
complex intracellular mechanisms. It can be put into action by
several neurotransmitters, including serotonin and norepinephrine.
Neuropeptides were viewed as occurring only in small quantities
within the central nervous system and therefore were overlooked for
many years. They generally modulate the synaptic efficacy of other
neurotransmitters. Typical examples are endorphins and enkephalins
that act at specific receptor sites that can also be taken by certain
drugs, such as opiates. One of the widely encountered
neuropeptides is neurotensin, which has been linked to burst-like
activity in neurons using other neurotransmitters, such as dopamine.

4.7 Temporal and Spatial

Summation
It has already been mentioned that neuron-neuronal synapses are
mostly non-obligatory. This means that one presynaptic action
potential cannot force the postsynaptic membrane to generate an
action potential. Such stimuli are called subthreshold. So, in order to
generate an action potential, the postsynaptic membrane must
somehow sum up the effects of a number of presynaptic signals.
There are two basic ways of doing this.
The first way is based on the fact that postsynaptic excitatory
potentials (EPSPs) are of a relatively long duration (about 15 ms).
So, if another action potential comes to the same synapse at a delay
smaller than the typical duration of the EPSP, its postsynaptic effects
will superimpose on the effects on the previous signal and lead to a
larger EPSP (figure 4.11). This mechanism is called temporal
summation. A sequence of presynaptic action potentials may be
able to bring the postsynaptic potential to the membrane threshold
when a single potential is unable to do this.
Figure 4.11 Temporal summation occurs when several action potentials arrive at
a presynaptic membrane at intervals that do not allow individual EPSPs to
disappear. Their effects can sum up and induce an action potential.

PROBLEM 4.6
What is the minimum frequency of presynaptic action potentials
that can theoretically lead to temporal summation and to a
postsynaptic action potential?

Another mechanism is based on the fact that a postsynaptic

membrane can receive many presynaptic inputs located close to
each other and on the existence of local currents through the
membrane and in the surrounding media. When a presynaptic action
potential induces a subthreshold EPSP (figure 4.12), membrane
depolarizes in an area that is in direct contact with the
neurotransmitters released by the presynaptic membrane. Local
currents spread this depolarization to neighboring areas of the
postsynaptic membrane, certainly with a quick decrement in its
amplitude. So, if other synapses are located nearby (synapses 2 and
3 in figure 4.12), the postsynaptic membrane may “feel” the effects of
the local currents from all three synapses. If three action potentials
come simultaneously to the three synapses, the depolarization of the
postsynaptic membrane in all three synapses will be bigger than that
which occurs in response to only its own action potential. This effect
is called spatial summation.

Figure 4.12 Spatial summation occurs when several action potentials AP1, AP2,
and AP3 arrive simultaneously at different synapses on the same presynaptic
membrane so that their individual EPSPs sum up and can induce an action
potential. The insert drawing shows possible locations of the three synapses on
the target neuron.

PROBLEM 4.7
What will happen if action potentials to synapses 1, 2, and 3 come
not simultaneously but at a delay?

The mechanisms of temporal and spatial summation are

examples of how postsynaptic membrane can integrate information
coming from presynaptic cells. They make it possible to transfer
signals through non-obligatory synapses.
 PROBLEM 4.8
Imagine that two groups of neurons (A and B) send their signals to
another group of neurons (C). Action potentials generated
simultaneously by all the neurons of group A lead to a response
C1 (the number of activated neurons in group C); action potentials
generated simultaneously by all the neurons of group B lead to a
response C2. What can be the magnitude of the response to
action potentials in all the neurons in both groups A and B? Can it
be bigger than, smaller than, or equal to (C1 + C2)? Why?

CHAPTER 4 IN A NUTSHELL
Passive spread of local currents from
an action potential leads to
depolarization of adjacent segments of
the membrane to the threshold and to
the generation of a new action
potential. Speed of conduction is
higher along thicker neural fibers.
Some fibers are covered with a special
substance, myelin, which increases the
speed of conduction of action
potentials. Information exchange among
cells occurs at special sites called
synapses. The mechanism of synaptic
transmission involves special chemical
substances, mediators that can
depolarize or hyperpolarize the
postsynaptic membrane. Neural cells
integrate the incoming information and
generate action potentials when the
effects of several synapses or several
action potentials coming at a high
rate are summed up.
 Chapter 5

Skeletal Muscle

KEY TERMS AND TOPICS

skeletal muscle
myofibril
neuromuscular synapse
excitation–contraction coupling
twitch and tetanic contractions
elements of mechanics
length and velocity dependence of muscle
force
external regimes of muscle contraction

Skeletal muscle is a machine (a “motor”) that converts chemical

energy to mechanical work and heat. It is probably the most amazing
motor known to humans. Its ability to quickly generate power is
superior to virtually any human-designed motor of approximately the
same size. It has numerous features that may seem weird to an
external observer. These include relative slowness of force
generation, its dependence on muscle length and velocity, and a few
others reviewed later. Some of them look rather suboptimal or even
bizarre. There are two ways to look at these unique features. The
first is to ask oneself: How does the central nervous system cope
with (or compensate for) all the “weirdness” of muscle design, its
nonlinearities (you will learn what these are), time delays, and other
features that seem terrible when they are viewed through the eyes of
a 21st century engineer? The alternative is to ask oneself: How are
the unusual features of skeletal muscle used by the central nervous
system to produce the unique properties of human movements that
make them far superior to any robot? These properties include, in
particular, flexibility, dynamic stability, and the ability to handle fragile
objects. As suggested in chapter 1, we will look at skeletal muscle
optimistically—as a unique design developed by evolution, and not
as a blunder of nature.
When people talk about muscles, they sometimes mean different
things. For example, when a person flexes the knee, it is commonly
said that the quadriceps muscle is being stretched. In this case, the
word “muscle” is used to imply the whole complex of structures that
includes muscle fibers, tendons, and ligaments. However, depending
on a number of factors, the muscle–tendon complex may be
stretching while the muscle fibers are shortening. This can happen,
in particular, when the muscle generates active force while being
elongated by an external force (eccentric contraction; see later
discussion)—for example, by the force of gravity while landing after a
jump. In the example of the quadriceps, it is also implied that the four
anatomically distinct muscles behave in a qualitatively similar way.
Some muscles have less obvious separation into portions (called
compartments) that have different actions and show a degree of
mechanical and physiological independence (Jeneson et al. 1990;
Danion et al. 2003a). Typical examples are the extrinsic finger
flexors: flexor digitorum profundus and flexor digitorum superficialis.
These muscles have single proximal tendons and four distal tendons
directed at their points of attachment at the four fingers. Their
mechanical behavior is rather complex due to the transmission of
forces across the compartments.
In this chapter, we will first discuss the “naked muscle”—that is,
the properties of muscle fibers irrespective of the way in which the
force generated by the processes in muscle fibers is being
transferred by tendons to create torques in joints. The role of
tendons will be considered at the end of the chapter.

5.1 Skeletal Muscle Structure

Whole muscle is composed of parallel fibers (muscle cells or
myofibrils). Each fiber is a rather large cell; the fibers may be several
centimeters long and from 10 to 100 μm in diameter. Muscle fiber,
like any cell, has a membrane (sarcolemma), inside which there is
sarcoplasma containing myofilaments and sarcoplasmic reticulum.
The structure of a muscle fiber is shown in figure 5.1. The
sarcolemma has many invaginations that are called T-tubules. They
plunge deeply into the interior of the fiber, thus increasing the
surface area by a factor of 3 to 10. T-tubules come very close to
cisternae in the sarcoplasmic reticulum. The gap there is very
narrow, about 300 Å (i.e., even smaller than the typical synaptic
cleft).
Figure 5.1 (a) The structure of a muscle fiber. (b) A schematic representation of
a cross-section of a fiber.

A major role in the mechanism of muscle contraction is played by

Ca++ ions. We have already mentioned major physiological roles for
Ca++ in the persistent inward currents through dendritic membranes
and in synaptic transmission; the same ion is vital for proper muscle
functioning. At rest, ion pumps pump Ca++ ions from sarcoplasma
into the sarcoplasmic reticulum (similar to the sodium–potassium
pump). This process, similar to other ion pumps, involves specialized
macromolecules and requires energy. Sarcoplasmic reticulum
contains a special protein that binds Ca++ and does not let it escape.
As a result, the concentration of Ca++ in sarcoplasma is very low
(less than 10-7 M). We will examine the role of calcium in muscle
contraction in the next section.

5.2 Myofilaments
Myofilaments are major force-producing elements of muscle cells,
consisting of two major molecules, myosin and actin (figure 5.2).
Thick filaments contain mostly myosin, while thin filaments contain
mostly actin. Actually, thin fibers contain two actin molecules that
form a structure that looks like a double helix (resembling the famous
DNA structure). In order to develop force, the two sets of filaments,
actin and myosin, must make connections to each other. These
connections are called cross-bridges. Actin and myosin filaments
are organized so that regularly arrayed myosin filaments are
surrounded by six actin filaments (like a kitchen floor mosaic), while
each actin molecule is in contact with three molecules of myosin. At
one end, actin filaments are attached to a structure called a Z-line.
Two sets of actin filaments and one set of myosin filaments between
two Z-lines constitute a sarcomere. Sarcomeres are the most
important functional units of myofibrils in the production of muscle
force. In order to characterize the state of a myofibril, two more
terms are used: The length of the myosin filaments within a
sarcomere is called the A-band, while the length of actin filaments
that do not overlap with myosin filaments is called the I-band. These
bands are clearly seen under a strong microscope as alternating
dark (A-band) and light (I-band) zones (figure 5.2).
Figure 5.2 The structure of a myofibril. The lower figure shows the sequence of
dark and light bands (A-bands and I-bands). The upper drawing shows the typical
configuration of actin and myosin molecules within a myofibril.

There are several more important protein molecules that play

major roles in the mechanism of muscle contraction. The first is
tropomyosin. Its long molecules lie along the actin molecules in thin
filaments (figure 5.3). Another important molecule is troponin.
Troponin molecules are attached at regular intervals to tropomyosin
molecules, forming a complex that can change its configuration
under the action of calcium ions. These changes play an important
role in the formation of cross-bridges, which are the units of active
muscle contraction (see later discussion).
There is a third filament system inside the myofibrils that acts in
parallel to the actin–myosin complex. This system has elastic
properties. A major macromolecule forming this system is titin.
Studies of titin have suggested its major role not only in defining the
elastic properties of muscle fibers at rest but also during active
muscle contractions (Granzier and Labeit 2004, 2005).
Figure 5.3 The structure of the thin filament (actin). Note long tropomyosin
molecules in parallel with the actin strands. Troponin attaches to tropomyosin at
regular intervals.

Figure 5.4 Neuromuscular synapse. A presynaptic nerve action potential

induces the movement of vesicles with acetylcholine (ACh) to the presynaptic
membrane, their fusion, and the release of ACh into the cleft. ACh diffuses to the
postsynaptic muscle membrane, depolarizes it, and induces an action potential.
Remaining ACh molecules are either re-uptaken into the presynaptic membrane or
removed by an enzyme, acetylcholinesterase.
5.3 Neuromuscular Synapse
The neuromuscular synapse (or neuromuscular junction) is a
region of contact between a single presynaptic nerve fiber (recollect
that an axon branches, giving rise to many presynaptic fibers) and a
muscle fiber. These two fibers come very close to each other so that
the synaptic cleft is only about 500 Å wide, but they do not make
direct contact (figure 5.4). The presynaptic axonal membrane has
active zones that contain many synaptic vesicles with a
neurotransmitter (acetylcholine) and also a high concentration of
mitochondria that store and supply molecules that are metabolized to
get energy.
When a decision is made by the central nervous system to induce
a muscular contraction, signals eventually go to neurons—known as
α-motoneurons—in the spinal cord (or in the brainstem, for head and
neck muscles) that send their long axons to appropriate muscles or,
in other words, innervate muscles. Action potentials travel at a high
speed along these thick efferent fibers and arrive at the point of
branching. There, the action potential excites each of the branches,
so that each of them delivers an action potential to the presynaptic
membrane at about the same time. Neuromuscular synapses are
obligatory—that is, a presynaptic action potential always induces a
postsynaptic action potential and initiates the process of muscle
contraction. This is achieved by an amplification of the incoming
signal with chemical mechanisms. Let us consider the most
important steps involved in this process:
Step 1: An action potential arrives at the presynaptic membrane
and opens voltage-dependent Ca++ channels. Normally, the
intracellular concentration of Ca++ is very low. However, after an
action potential arrives, the concentration of Ca++ increases
dramatically (by a factor of 20). Intracellular Ca++ activates processes
leading to movement of synaptic vesicles to the presynaptic
membrane. The vesicles fuse with the membrane and let their
contents out, into the synaptic cleft (exocytosis).
 Step 2: The neurotransmitter (acetylcholine, ACh) released into
the cleft diffuses across the short distance to the postsynaptic
membrane and binds to specific molecular receptors on the
postsynaptic membrane. There is very high density of ACh-sensitive
receptors on the postsynaptic membrane (up to 10,000/μm2). ACh in
the synaptic cleft is quickly broken down by a special enzyme,
acetylcholinesterase, into two substances, acetate and choline. The
presence of this enzyme makes the duration of the postsynaptic
effects very brief. Relatively large amounts of ACh are taken back by
the presynaptic membrane and recycled.

PROBLEM 5.1
Imagine that you have a muscle without acetylcholinesterase.
What can you expect in response to a single presynaptic action
potential?

Step 3: ACh acts on the postsynaptic membrane and induces

changes in its ion permeability, leading to a depolarizing potential
(excitatory postsynaptic potential, EPSP). In healthy muscles,
EPSPs induced by presynaptic signals are always suprathreshold,
leading to the generation of an action potential on the postsynaptic
membrane. Note that subthreshold depolarizing potentials may
emerge spontaneously (i.e., without an apparent stimulus) at the
postsynaptic muscle membrane (endplate region of muscle fiber).
These potentials are about 1 mV in peak amplitude, and their
functional meaning is unclear. They are called miniature endplate
potentials or MEPPs (see figure 5.5).
Figure 5.5 Miniature excitatory postsynaptic potentials (endplate potentials,
MEPPs) spontaneously occur on the muscle postsynaptic membrane. A
presynaptic nerve action potential always makes the postsynaptic membrane
reach the depolarization threshold and induces a muscle action potential.

PROBLEM 5.2
A presynaptic action potential induces contraction of 20 myofibrils.
Can a sequence of action potentials induce the contraction of
more than 20, less than 20, or exactly 20 myofibrils?

The next steps involve events already happening in the muscle

fibers, and we will address them as the mechanisms of muscle
contraction.

5.4 Mechanisms of Contraction

Step 4: Postsynaptic action potential travels along the muscle cell
membrane (sarcolemma) entering the T-tubules (figure 5.6). There it
opens Ca++ channels. At rest, virtually all Ca++ ions are stored in
sarcoplasmic reticulum where they are “captured” by a special
protein. Opening calcium channels leads to a massive influx of Ca++
ions into the sarcoplasma, increasing the concentration of these ions
by a factor of 100. This increase is transient, and Ca++ is quickly
pumped back into the sarcoplasmic reticulum.

Figure 5.6 Muscle action potential travels along the sarcolemma, enters T-
tubules, and leads to a release of Ca++ ions from the sarcoplasmic reticulum.

Step 5: The sliding filament theory (figure 5.7). Calcium ions in

sarcoplasma act on the troponin–tropomyosin complex. At rest,
tropomyosin blocks the myosin-binding site on actin—that is, a
special place on an actin molecule that would eagerly attach to a
special place on a myosin molecule. Calcium ions make the binding
site available. If there is energy available (normally, ATP molecules),
the myosin head binds to a site on actin and uses the energy to
ratchet the filaments with respect to each other. As mentioned
earlier, these attachments between actin and myosin molecules are
called cross-bridges. Then myosin releases itself from the actin site
and springs back, getting ready to attach to the next available site
and to repeat the cycle (certainly, if Ca++ ions and energy sources
are still available). Note that the interaction between myosin and
actin filaments occurs in the three-dimensional space so that each
myosin molecule simultaneously makes and breaks cross-bridges
with six actin molecules. This means, in particular, that when some
of the cross-bridges break, others maintain force of contraction.
Force developed by a muscle fiber may be considered approximately
proportional to the average number of simultaneously engaged
cross-bridges.

Figure 5.7 The sliding filament theory. Ca++ ions remove troponin and free a site
for myosin to bind to actin (this process uses the energy from ATP). A ratchet
motion occurs, moving the filaments with respect to each other.

PROBLEM 5.3
What would happen if the troponin–tropomyosin complex were
permanently inactivated?

Steps 4 and 5 are commonly known as excitation–contraction

coupling.
Step 6: After excitation stops (action potentials do not arrive
anymore), Ca++ is actively pumped from sarcoplasm into the
sarcoplasmic reticulum, the troponin–tropomyosin complex takes
over all the sites of myosin binding, and the filaments slide back
along each other (relaxation). Note that muscle relaxation is a
passive process: One cannot command a muscle to relax, only to
stop sending excitatory action potentials.
The sliding filament mechanism provides an explanation for some
of the features of the dependence of muscle force upon muscle
length that are going to be considered in the next section.
 PROBLEM 5.4
Imagine that a muscle is developing force under a constant level
of stimulation. Draw a cartoon depicting the dependence of
muscle force upon muscle length based on what you already
know.

5.5 Types of Muscle

Contractions
Muscular contraction leads to the generation of force that is always
directed to shortening the muscle fibers. Muscles cannot actively
elongate. Later in this chapter, we will see that a muscle can develop
active force while its length is increasing (eccentric contraction).
However, in such cases, muscle length is always changing under the
action of another force, produced either by other muscles or by the
environment, or because of the inertia of the muscle and body parts
to which it is attached.
When a single action potential comes to a muscle fiber, it
responds with a unitary contraction called a twitch contraction or
simply a twitch (figure 5.8). Depending on the properties of the fiber,
twitch contractions last from a few tens of milliseconds to a couple of
hundred milliseconds. Note for comparison that a muscle action
potential has a duration of about 10 ms. This means, in particular,
that the mechanical consequences of an action potential last much
longer than the action potential itself.
Figure 5.8 A typical twitch contraction of a muscle in response to a single
stimulus.

PROBLEM 5.5
Why does the twitch contraction outlast the action potential?

If several fibers are stimulated simultaneously, their twitch

contractions superimpose. This superposition can lead both to an
increase in the peak amplitude of the twitch contraction and to its
prolongation if a muscle fiber with a longer-lasting twitch is added to
a group of fibers with short-lasting twitches.
If two action potentials come to the same fiber at a short interval,
their mechanical effects may superimpose, as shown in figure 5.9,
so the peak force of the contraction induced by the second action
potential will increase. If many action potentials come at a frequency
that allows superposition of their mechanical effects, a sustained
level of contraction is observed, called tetanus or tetanic contraction
(figure 5.10). Tetanus may display local peaks of contraction at
relatively low frequencies of action potentials (sawtooth tetanus) or
may lead to total fusion of individual twitches, in which case it is
called smooth tetanus.
Figure 5.9 Two action potentials come at a short interval and induce two twitch
contractions. Their mechanical effects superimpose, leading to a higher level of
muscle force.

Figure 5.10 A sequence of action potentials may lead to a tetanus (a sustained

contraction). At a high frequency of action potentials, individual contractions may
fuse, leading to a smooth tetanus.

PROBLEM 5.6
Smooth tetanus is rarely observed in individual fibers in real life.
Why are our muscle contractions normally smooth?
Figure 5.11 A simple mechanical model of a muscle. It contains a force
generator (F), a damping element (B), and two elastic elements, a parallel spring
(K1) and a series spring (K2).

5.6 Elements of Mechanics

Let us now turn to real life and remind the reader that muscles do not
exist by themselves, but their action at joints is affected by the
mechanical properties of tendons and ligaments, as well as by the
geometry of the tendon attachment to the bones. Typical models of
muscles involve at least four components (figure 5.11): a contractile
element (force generator) about which we have been speaking until
now, a damping element (dashpot), and two elastic elements, one
parallel and one serial. Unfortunately for investigators, most of these
elements are essentially nonlinear, which means that their
mechanical behavior differs from that of ideal springs and dampers
and may be characterized by abrupt changes in their characteristics.
It seems to be the proper time to introduce a little bit of
mechanics. Elastic elements or springs are physical objects that
resist external attempts at changing their length by developing force
acting against the imposed deformation. During deformation, springs
accumulate potential energy that can be released. In the simplest
case of a linear spring, the force developed by the spring is
described with Hooke’s law:

where Fe is elastic force, x is spring length, x0 is zero length (i.e., the

length at which elastic forces are zero), and k is a coefficient termed
stiffness. Note that the minus sign implies that force acts against the
change in length with respect to x0. The notion of stiffness can only
be applied to springs. One cannot measure a change in force and a
change in length (or displacement) of an arbitrary object and claim
that the ratio of the two is stiffness (Latash and Zatsiorsky 1993). In
particular, “joint stiffness” (also “limb stiffness,” or “body stiffness,”) is
not a very meaningful expression because joints do not deform, they
move. Tissues crossing joints, such as muscles and tendons, do
deform, but characterizing their overall time-varying behavior by a
single parameter, stiffness, is very questionable.
Damping is the ability of a system to generate force against the
vector of velocity:

where Fv is damping force, V is velocity of length change, and b is a

coefficient. Note again that the minus sign implies that force acts
against the velocity vector; as a result, damping always acts to
reduce the kinetic energy of a moving object. Sometimes damping is
imprecisely referred to as viscosity, which is a property of liquids,
whereas equation 5.2 can be applied to various objects.
All material objects also have inertia, which is a coefficient
between applied force and acceleration:

where Fi is inertial force, a is acceleration, and m is a coefficient

termed mass.
 Equations 5.1 through 5.3 describe what are called linear
elements. Such elements produce outputs in proportion to input
signals. For example, if a force F1 acts on a spring and induces a
displacement x1, and force F2 produces a displacement x2, the
combined action of two forces (F1 + F2) will produce displacement (x1
+ x2). The same rule of simple summation can be applied to damping
and inertial forces. Such systems are relatively easy to analyze, and
equations describing their behavior can often be solved analytically.
Linear systems are commonly studied in textbooks of elementary
physics; however, in real life they are rare. It does not take much to
make a system nonlinear. For example, if stiffness depends on
spring length, this element is already nonlinear, and consequently, a
system with such an element is very likely to be nonlinear. The
situation becomes even more complicated if a characteristic of an
element changes in a step-like fashion, which is typical for intact
muscles as described later in the book.

5.7 Force–Length and Force–

Velocity Relations
A typical example of the nonlinear behavior of a whole muscle is its
force–length relationship. Such a relationship can be obtained in an
experiment in which muscle length is fixed at a certain value
(isometric conditions), a standard stimulation is applied to the
muscle nerve with the help of an external electrical stimulator, and
peak muscle force is measured. Then, muscle length is fixed at a
different value, the same stimulation is applied, and the force is
measured again. And so forth. As a result, a force–length curve is
observed similar to those shown in figure 5.12. It is important to
remember that the measurements are performed when muscle
length is not changing, which means that inertial and viscous
properties do not play a major role.
Figure 5.12 Force–length curves measured in a muscle for different levels of
external stimulation (S1, S2, and S3). Note that the muscle behaves like a
nonlinear spring. Changing the strength of the stimulation modifies the zero length
of the spring and the shape of the characteristic. Force drop with stretch does not
happen during natural movements because the anatomical range of muscle length
changes is limited.

PROBLEM 5.7
The last statement is not 100% correct. Why?

If you change the parameters of the stimulation (its amplitude or

frequency) and perform the same experiment, the curve will shift to
the right or to the left nearly parallel to the length axis. So the muscle
behaves like a nonlinear spring (note that its stiffness, the slope of
the curve, changes with length as well as with the level of excitation)
whose zero length changes in response to a change in an incoming
activation signal (stimulation).

PROBLEM 5.8
In which direction will the curve shift if you increase either the
frequency or the amplitude of the stimulation?
 One can measure the force–length relationship in a single
sarcomere (i.e., inside the contractile element shown as F in figure
5.11). The active force developed by the sarcomere will show a
dependence on the sarcomere length that is somewhat similar to the
one for the whole muscle: At low values of sarcomere length, cross-
bridges cannot develop force because of the lack of space for new
attachments; at intermediate values of length, the force is maximal;
while at high values of sarcomere length, there are only a few cross-
bridges that can generate force, and it drops again.
Another important relation for a whole muscle is the force–velocity
curve. Such curves are usually studied in experiments when a
muscle performs a twitch contraction under different loads and the
peak velocity of muscle shortening is measured. The curve typically
looks hyperbolic (figure 5.13) and can be well approximated with the
famous Hill equation:

Figure 5.13 A typical force–velocity curve for a whole muscle. The x-axis
represents velocity of muscle stretch (frequently, this curve is drawn using velocity
of muscle shortening). Note that the muscle develops higher forces when it is
lengthening (positive velocity) than when it is shortening (negative velocity).
Compare this figure with the Hill equation.

where F is force, F0 is force at zero velocity (in isometric conditions),

V is peak velocity of shortening (i.e., it is negative for a stretched
muscle), and a and b are constants specific for a given muscle.
5.8 External Regimes of Muscle
Contraction
Muscle contraction in conditions that prevent changes in muscle
length is termed isometric, and a load leading to such conditions is
termed “isometric load.” When muscle contracts acting against a
constant external force, such contraction and load are termed
isotonic. If a muscle is acting against a springlike load, the load is
termed elastic. Examples of different loading conditions are shown
in figure 5.14.

Figure 5.14 A muscle always works against a load. Three types of loads are
illustrated. Isometric load prevents changes in the length of the “muscle plus
tendon” complex; an isotonic load does not change; an elastic load acts like a
spring. A typical muscle characteristic is shown for comparison (the thin curve).

The problem with these terms is that they are misleading.

Consider, for example, what will happen if movement in a joint is
prevented (isometric conditions). If a muscle acting at this joint is
activated, it develops contractile force that acts on all the elements,
including the parallel and serial elastic elements. Depending on the
relative properties of these elements, the muscle will induce a
change in their relative length even if the length of the whole
complex “muscle plus tendon” is kept constant. Note that a relaxed
muscle is usually less stiff than its tendon, while activated muscle is
typically more stiff than the tendon. Thus, muscle fiber length is
going to change in isometric conditions as well.

PROBLEM 5.9
What will happen with the relative length of muscle fibers and with
the tendon under activation in isometric conditions?

Isotonic conditions are also not associated with a constant

external load if we try to look at it from the point of view of the
muscle, not of an external observer. Joint movement is associated
with changes in joint geometry, including changes in the distance
from the center of joint rotation to the line of muscle action. As a
result, to produce a certain value of the moment of force required to
balance the action of the external load, muscle force will have to
change with joint angle.

CHAPTER 5 IN A NUTSHELL
Muscle contractions are produced by an
interaction of two types of molecules,
actin and myosin, inside muscle cells.
Muscle cells are excited through
neuromuscular synapses with the help
of a mediator, acetylcholine. Action
potentials lead to the release of
calcium ions, which make cross-bridge
formation between actin and myosin
molecules possible. In response to a
single stimulus, muscle fibers
generate a single twitch contraction.
A number of stimuli coming at a high
frequency lead to the summation of
individual twitches and the generation
of a tetanic contraction. Muscle force
increases with muscle length and
decreases with the velocity of muscle
shortening. Muscles always act against
external loads, typical examples being
isometric, isotonic, and elastic
loads.
 Chapter 6

Peripheral Receptors

KEY TERMS AND TOPICS

Classification of receptors
Weber-Fechner law
muscle spindles
fusimotor innervation
Golgi tendon organs
articular receptors
skin and subcutaneous receptors

Receptors are specialized cells or subcellular structures that change

their properties in response to stimuli (sources of energy) of a
special type or modality. Thus, different receptor systems enable
humans and animals to differentiate among different sources of
energy (for example, light, sound, and mechanical energy) that are
being absorbed by the body. Receptors of a certain type are typically
rather specific (i.e., they ignore alien stimuli), although many of them
can be forced to fire with an electrical stimulation or even with a
strong mechanical stimulus. For example, a hard hit to an eye can
lead to a whole bunch of sparks, that is, to visual images induced by
the activity of visual receptors in the eye responding to mechanical
deformation. Interestingly, although many of our receptors react to
electrical stimulation and transduce information using electrical
phenomena, we do not have a developed system for sensing an
electromagnetic field outside the visible light range.

6.1 General Classification and

Properties of Receptors
The obvious function of receptors is to make information about
particular types of stimuli available to other neurons within the
central nervous system. Some of this information is related to the
environment, while other information is related to the state of the
body itself. There are four major groups of receptors:
1. Interoceptors transduce information from within the body;
2. Exteroceptors transduce information from the environment;
3. Proprioceptors transduce information about the relative
configuration of body segments; and
4. Haptic receptors transduce information about the direct
mechanical contacts with the environment.
Sometimes, receptors associated with pain are classified as a
special group of nociceptors. These receptors will be discussed in
more detail later in the chapter on kinesthetic perception.
Receptors within each group can be sensitive to stimuli of different
modalities; on the other hand, receptors belonging to different
groups may react to energy of the same kind. For example, there are
mechanoreceptors (i.e., receptors that react to mechanical stimuli) in
each of the four groups. Receptors sensitive to certain chemicals
(chemoreceptors) are rather widespread. Some of them are located
on membranes and are sensitive to certain neurotransmitters. As
already discussed, these receptors play a major role in synaptic
transmission of information. On the other hand, the activity of
chemoreceptors in the mouth and nose plays an extremely important
role in one’s life, creating the senses of taste and smell.
Before moving to the mechanisms and function of specific groups
of receptors, let us mention a law that is applicable to the conscious
perception of signals from many of the receptor systems. This law
states that perceived sensation is related to stimulus magnitude by a
logarithmic function. It can be studied by changing the magnitude of
a stimulus of a certain modality and asking the subject to report how
strong the stimulus feels on a scale from 0 to 10. It is called the
Weber-Fechner law:

where P is the magnitude of perception, M is the magnitude of the

stimulus, M0 is the magnitude of the stimulus when the subject just
feels it (the threshold stimulus), and k is a constant. This law is an
example of psychophysical functions that relate dimensions
measured directly by an experimenter to perceived dimensions
reported by the subject.

PROBLEM 6.1
Give an example of a receptor system that does not obey the
Weber-Fechner law.

In this chapter, we will consider proprioceptors and some of the

haptic receptors whose activity is closely tied to the motor function.
A typical proprioceptor is a specialized neural cell with the body
located in a special place, a ganglion, close to the spinal cord
(figure 6.1). These neurons are rather unusual in their structure; in
particular, they do not receive inputs from other neurons and do not
have a typical dendritic tree. Their axons serve both for inputs into
the neuron and for its output. The axon of such a neuron is termed
an afferent axon or afferent fiber; it has a characteristic T-shape. An
afferent axon splits into two branches close to the neuron body. One
of the branches goes to a peripheral site in the body, where it ends
with a specialized ending (sensory ending); the membrane of the
ending can be depolarized to the threshold by a stimulus of a certain
strength and modality. The other branch goes through the back
(dorsal) portion of the spinal canal (through the dorsal roots) into
the spinal cord, where it can make connections with many different
neurons, leading to various motor and sensory effects.

Figure 6.1 The body of a sensory neuron is located in a ganglion near the spinal
cord. One branch of its T-shaped axon goes to the peripheral sensory ending,
while the other branch goes through the dorsal roots into the spinal cord.

Proprioceptor neurons differ from most of the neurons within the

central nervous system in the way they generate and transmit action
potentials. While most neurons generate action potentials on their
body membrane (in particular, at the axonal hillock) in response to
excitatory synaptic effects, proprioceptors generate action potentials
in the periphery, in the sensory ending excited by an external
stimulus. Most neurons integrate information from many other
neurons to generate an action potential. Proprioceptor neurons have
only one source of excitation, at the peripheral sensory ending. Most
neurons transmit action potentials from the body down the axon; this
type of transmission is called orthodromic. Action potentials
generated by a sensory ending travel in the opposite direction,
toward the body; this type of transmission is called antidromic.
Further, the action potential is conducted orthodromically along the
central branch of the T-shaped axon.

6.2 Muscle Spindles

The muscle spindle is one of the most amazing inventions of nature
(figure 6.2). These structures with a very sophisticated design supply
other neurons within the central nervous system with information
related to the length and velocity of muscle fibers. Muscle spindles
have an elongated shape (they are commonly up to 1 cm long) with
a thicker area in the middle that makes them look like regular
spindles. They are scattered among muscle fibers in large quantities.
Each spindle contains specialized muscle fibers, intrafusal
fibers, that are oriented parallel to the regular power-producing
extrafusal fibers. The middle part of a spindle is covered with a
capsule made of connective tissue. At both ends, intrafusal muscle
fibers are connected either to extrafusal fibers or to tendinous
ligaments. So, when extrafusal fibers change their length, intrafusal
fibers are stretched or shortened correspondingly.
Figure 6.2 (a) A muscle spindle is oriented parallel to extrafusal muscle fibers. It
is covered with a capsule connected to the extrafusal fibers by strings of
connective tissue (b). Spindles contain two types of intrafusal muscle fibers, the
bag fibers (BF) and the chain fibers (CF). Two types of sensory endings can be
found in muscle spindles, primary (Ia) and secondary (II). Primary endings are
typically seen in virtually all intrafusal fibers, while secondary endings are seen in
CF and static BF, but not dynamic BF.
 Spindles contain two major types of intrafusal fibers, called bag
fibers and chain fibers. These names reflect the distribution of nuclei
within a fiber, clustered as in a bag, or spread along the fiber in a
chain-like fashion. In turn, bag fibers are of two subtypes, static and
dynamic, reflecting their role in the responses to muscle length and
velocity of its changes. Two types of sensory endings can be found
in muscle spindles. They are located mostly in the middle (equator)
portion of the spindle. Endings of the first type, called primary
spindle endings, are seen on virtually all intrafusal fibers, including
both bag and chain fibers, while endings of the second type, called
secondary spindle endings, are rarely seen on dynamic bag fibers,
but are common in static bag and in chain fibers. A spindle ending,
like any sensory ending of a proprioceptive neuron, is located at the
end of the axon of a neuron whose body is in a spinal ganglion.
Axons of primary endings belong to group Ia afferent fibers, while
axons of secondary endings belong to afferent fibers of group II.
Primary sensory endings are sensitive to both muscle length and
velocity. Figure 6.3 shows a typical response of a primary ending to
an externally imposed muscle stretch at different velocities. Note that
the frequency of firing of the ending is higher after the stretch than
before, which means that the ending is sensitive to muscle length. In
addition, during stretching, the frequency of firing of the ending is
higher for faster stretches, meaning that the ending is sensitive to
velocity as well. Note that velocity sensitivity of primary spindle
endings leads to an increase in the frequency of their firing during
muscle elongation and also to a decrease in the frequency of firing
during muscle shortening. Primary afferent axons are among the
fastest neural fibers. They are myelinated, with the diameter ranging
from 12 to 20 μm, which corresponds to action potential velocity of
up to 120 m/s.
Figure 6.3 Typical responses of a primary spindle ending to an externally
imposed muscle stretch at different velocities. Note that the response increases
with muscle length and with the velocity of stretch. The top drawing shows the
length changes and individual action potentials. The bottom drawing shows the
changes in firing frequency (ƒ) starting from some nonzero level (ƒ0).

PROBLEM 6.2
What can you say about the state of the muscle (its length and
velocity) if you know the frequency of firing of a primary ending in
this muscle?

Secondary endings are sensitive only to muscle length but not to

velocity. Figure 6.4 shows the response of a typical secondary
ending to muscle stretch and shortening. The proprioceptor neurons
innervating secondary endings have smaller axons, and the speed of
conduction is also lower, ranging from 20 to 60 m/s.

Figure 6.4 A typical response of a secondary spindle ending to an externally

imposed muscle stretch and shortening. Note that the response increases with
muscle length and does not depend on velocity. The figure is organized similarly to
figure 6.3.

PROBLEM 6.3
Draw a graph of time changes in the firing frequency for a primary
ending and for a secondary ending if muscle length changes as a
sine function.

PROBLEM 6.4
Draw a graph of length changes for a muscle whose typical
primary spindle ending shows a smooth, ramp-like increase in the
firing frequency with time from a certain steady level to another,
higher steady level followed by a ramp-like decline in the firing
frequency to a level somewhat higher than the original one.
 Endings in muscle spindles have very high sensitivity to low-
amplitude changes in muscle length, particularly if these changes
occur at a high frequency. This is particularly true for primary spindle
endings, which can be forced to generate an action potential or even
several action potentials in response to every cycle of high-
frequency vibration (on the order of 100 Hz) when the amplitude of
the vibration is 1 mm and the vibration is applied to the skin over the
muscle belly or over the tendon. If a vibrator is attached directly to
muscle fibers, a few μm of vibration amplitude are enough to drive
primary spindle endings at the frequency of the vibration.

6.3 The Gamma-System

Primary and secondary spindle endings are unique among
proprioceptors in their ability to change their sensitivity to the
respective physical variables, muscle length and velocity, in
response to signals from special groups of neurons that form the
gamma-system (γ-system).
Intrafusal muscle fibers receive signals from the efferent axons of
neurons of a particular type with the bodies located in the spinal
cord. These neurons belong to the class of motoneurons together
with spinal neurons that send their signals to power-generating,
extrafusal fibers. Motoneurons innervating muscle fibers inside
muscle spindles belong to the gamma system and are called γ-
motoneurons. They are much smaller than the other group of
motoneurons, α-motoneurons; their axons are of approximately the
same length since their target is located in the same muscles, but
they are much thinner and, correspondingly, conduct action
potentials at much lower velocities (about 20 m/s).
There are two types of γ-motoneurons (figure 6.5), dynamic and
static. Dynamic γ-axons innervate dynamic bag muscle fibers and,
thus, affect the sensitivity of primary spindle endings that are located
in these fibers to muscle velocity. Static γ-motoneurons send their
axons to static bag and chain fibers. They can change the sensitivity
of both primary and secondary endings to muscle length.
Figure 6.5 There are two types of small motoneurons (γ-motoneurons)
innervating intrafusal fibers of muscle spindles. Dynamic γ-motoneurons innervate
dynamic bag fibers and change the sensitivity of primary endings. Static γ-
motoneurons innervate static bag and chain fibers. They change the sensitivity of
both primary and secondary endings.

Figure 6.6 shows how electrical stimulation of dynamic γ-axons

can change the response of a primary ending to muscle stretch and
shortening. Note the increase in the effects of stretch on the spindle
response. Static γ-axons innervate intrafusal fibers containing both
primary and secondary sensory endings. So, the effect of their
stimulation can be seen in the response of both groups of sensory
endings to muscle length as an increase in the frequency of firing.
Figure 6.6 The effects of activation of dynamic γ-motoneurons on the response
of a primary spindle ending to muscle stretch and shortening. Thin line shows
frequency changes (ƒ) without gamma stimulation. Thick dashed line shows the
effects of γ-dynamic stimulation.

PROBLEM 6.5
How will a secondary ending react to an increase in the activity of
dynamic γ-motoneurons innervating the spindle?

PROBLEM 6.6
When we voluntarily contract a muscle, its length decreases.
However, the frequency of firing of the spindle endings in the
muscle may remain constant. How can this happen?

6.4 Golgi Tendon Organs

Another group of proprioceptors are located close to the junction
between tendons and muscle fibers (figure 6.7). These receptors are
sensitive to mechanical deformation and are called Golgi tendon
organs. Note that tendons may be viewed as nearly perfect elastic
structures (springs). This means that mechanical deformation of a
tendon increases with muscle force, and so Golgi tendon organs
appear to be nearly perfect force sensors. Golgi tendon organs do
not receive any additional innervation (that would be similar to the
gamma-system for muscle spindles); they are also not responsive to
the rate of force change. Thus, their response to muscle force is
relatively independent of other factors. The fact that tendons are
nonlinear springs makes Golgi tendon organs nonlinear sensors;
however, this is unlikely to be a major problem as long as their
properties do not change.

Figure 6.7 Golgi tendon organs are located in series with extrafusal muscle
fibers at their junction with the tendon. They are innervated with fast-conducting Ib-
group axons of sensory neurons in spinal ganglia.

Figure 6.8 illustrates a typical response of a Golgi tendon organ to

force generated by muscle fibers that are in series with the tendon
organ. Note, however, that Golgi tendon organs are rather selective
—that is, they respond to force generated by “their” muscle fibers. If
muscle force is generated by fibers that do not act at the area where
a Golgi organ is located, it will not increase its firing frequency, and it
may even show a drop in activity. This may result from an unloading
of a particular “braid” of the tendinous structures where the Golgi
organ is located.
Figure 6.8 A response of a Golgi tendon organ to muscle force. Note that it is
similar to the response of secondary spindle endings to muscle length.

PROBLEM 6.7
Imagine that you prevent movement of your right elbow joint and
then quickly activate the biceps to a rather large force (isometric
conditions). Draw a graph of the change in the firing frequency for
a primary and secondary muscle spindle ending and for a Golgi
tendon organ. Assume that all endings showed a steady-state
firing level before the force increase.

PROBLEM 6.8
Now, do the same for a very fast elbow flexion movement against
a constant external force.

Axons originating from Golgi tendon organs are nearly as large as

axons of primary spindle endings, and their speed of conduction is of
the same order of magnitude; it may reach about 80 m/s.

6.5 Other Muscle Receptors

A few other types of sensory endings can be found in a muscle. The
first is paciniform corpuscles that are similar in their structure to
Pacinian corpuscles found in the skin (see section 6.7), although
smaller. Paciniform corpuscles are commonly located at the area of
musculotendinous junction, and they are quite sensitive to high-
frequency vibration. Not much is known about their functional
significance and central connections.
There are also free sensory endings scattered all around the
muscle. They are sensitive to strong mechanical stimuli (like those
occurring during pinching) as well as to certain chemicals. These
receptors are likely to play an important role in the sense of pain and
in certain reflex responses that are described later. They are
innervated with small, nonmyelinated axons with a relatively slow
conduction speed (under 10 m/s).

6.6 Articular Receptors

Another group of proprioceptors reside in the joints; these are called
articular receptors. There are actually several different types of
articular receptors, including sensory endings similar to Golgi tendon
organs, as well as free nerve endings. They are innervated by axons
of variable size, from rather thin ones lacking myelin sheath to large
ones with a diameter of over 10 μm that belong to group I fibers and
are characterized by fast speeds of conduction (up to 80 m/s).
These receptors were thought for a long time to be perfect angle
transducers that inform the central nervous system about the
position of the joints. However, a closer inspection of their behavior
has revealed that individual articular receptors fire within a rather
small range of joint angles (figure 6.9). Moreover, many active
receptors are found when joint position is close to one of its
anatomical extremes, while only a few of them are active in the
middle range of joint motion. Thus, articular receptors are unlikely to
provide reliable information about joint position during natural
movements.
Figure 6.9 Most articular receptors fire in rather narrow ranges of joint angles,
mostly close to the anatomical limits. An increase in muscle force leads to an
increase in joint capsule tension, and articular receptors increase their response
(bold lines).

Articular receptors are also sensitive to changes in joint capsule

tension. Typically, a receptor increases the frequency of its firing
over the whole range of its activity in response to an increase in joint
capsule tension (figure 6.9). Note that joint capsule tension
increases with muscle force. All this makes signals from articular
receptors even less attractive as a major source of information about
joint position because a receptor may demonstrate similar firing
frequencies at different positions if joint capsule tension magnitudes
are different—for example, as a result of coactivation of agonist and
antagonist muscles acting at the joint. These receptors appear to be
better suited for detecting and signaling potentially damaging states
of the joints, such as those associated with high joint capsule
tension, joint angles close to one of the anatomical limits, and joint
inflammation.

6.7 Cutaneous Receptors

Human skin houses cutaneous receptors sensitive to different
sensory modalities. Among them are thermoreceptors sensitive to
temperature, nociceptors sensitive to potentially damaging stimuli
and giving rise to the sense of pain, and mechanoreceptors sensitive
to pressure. The last group of receptors plays a particularly important
role in the control of human movements, especially those involving
tactile discrimination (haptic perception) and manipulation of objects.
Figure 6.10 shows the major types of cutaneous
mechanoreceptors in the glabrous skin of the hand. Meissner
corpuscles and Merkel disks are located closest to the skin surface,
on the border of epidermis and dermis. Deeper down, in dermis,
there are Ruffini endings, and even deeper, in subcutaneous tissue,
there are Pacinian corpuscles.

Figure 6.10 Major types of cutaneous and subcutaneous mechanoreceptors in

the glabrous skin of the hand.

Merkel disks and Meissner’s corpuscles are closer to the surface

of the skin, in the epidermal sweat ridges and dermal papillae
respectively, and have much smaller receptive fields than the Ruffini
endings and Pacinian corpuscles that reside in the dermis and
deeper tissue. Merkel disks respond to vertical pressure on the skin
surface but not to lateral displacements. A group of Merkel disks is
commonly innervated by one afferent axon, which effectively
integrates information from a skin area. Meissner corpuscles are
sensitive to quickly changing pressure on a small area of skin. They
quickly adapt and stop responding if the pressure does not change.
Each Meissner corpuscle is innervated by two or more axons. Ruffini
endings can be activated from much larger skin areas, from as far as
5 cm away from the location of an ending. They are slowly adapting
and continue firing in response to stable deformation of the skin.
Pacinian corpuscles are the biggest of them all; their size ranges
from 1 to 5 mm. Pacinian corpuscles react to very quickly changing
mechanical deformation (for example, to vibration).
Sensory afferents are further distinguished by the temporal
dynamics of their responses to mechanical stimulation. A particular
subset of afferents fires rapidly when the rate of the application of
mechanical stimulation changes but falls silent during sustained
constant stimulation. These afferents are called rapidly adapting
afferents, and they are very effective in communicating to the
nervous system about dynamic changes in the stimulation. Meissner
corpuscles and Pacinian corpuscles are rapidly adapting fibers. In
contrast, other afferents produce a constant rate of discharge in the
presence of sustained stimulus. These afferents are called slowly
adapting afferents, and they relay to the nervous system properties
of the stimulus such as shape, size, and dimensions. Merkel disks
and Ruffini endings are slowly adapting fibers.
Each of these mechanoreceptors adapts differently to mechanical
pressure and serves different purposes for stereognosis.
Stereognosis is the ability to perceive and recognize the shape and
form of objects with tactile manipulation in the absence of visual or
auditory stimuli. Stereognosis relies on a symphonic interplay
between responses of the four sensory afferents to mechanical
stimulation. A detailed list of their properties and functions appears in
table 6.1.

Table 6.1 Afferent System and Properties

Receptor Merkel Meissner Pacinian Ruffini
Receptive field Small Small Large Large
Receptor Merkel Meissner Pacinian Ruffini
Receptive field 9 mm2 22 mm2 Entire hand or 60 mm2
area finger
Axon diameter 7-11 mm 6-12 mm 6-12 mm 6-12 mm
Conduction velocity 40-65 35-70 m/s 35-70 m/s 35-70 m/s
m/s
Innervation density 100/cm2 150/cm2 20/cm2 10/cm2
Spatial acuity 0.5 mm 3 mm 10+ mm 7+ mm

6.8 Signals From Peripheral

Receptors
Where does the information go after activating the receptors? Three
major effects of proprioceptor activity are of considerable importance
for us at this point.
First, proprioceptors induce changes in muscle activity, bypassing
our consciousness. Some of these effects are called reflexes, while
others are termed triggered reactions or preprogrammed
reactions. For now, let us avoid linguistic debates on the
appropriateness of the term reflex. There is considerable
disagreement on whether this term is informative or ambiguous and
misleading (e.g., Prochazka et al. 2000). It will be considered in
more detail in a later chapter. This term has been used for “more or
less automatic” and “more or less stereotypical” reactions to external
stimuli that do not involve conscious participation by the subject. As
described in other chapters, reflexes may be not so automatic and
not so standard; they may even occur in different muscles in
response to the same stimulus. However, they do reflect the
stimulus, albeit in a more complex way than was thought in the first
half of the 20th century.
Second, proprioceptors inform the central nervous system about
the body’s configuration, its orientation, and its interaction with
external objects. In particular, they let us know where our arms and
legs are, and how heavy or light, how rough or soft, the objects we
handle are.
Third, proprioceptors play a major role in creating an internal
system of coordinates that is used by the brain to plan and execute
movements. Problems with the functions of the proprioceptive
system result in major disruption of the neural control of movements,
making some habitual actions impossible—for example, standing
with one’s eyes closed or pointing to a memorized target without the
help of vision. This can be seen, for example, in “deafferented
persons”—an imprecise term that reflects severe cases of peripheral
large-fiber neuropathy, which prevents signals generated by the
sensory endings from reaching the central nervous system, and
resulting in major problems with joint coordination (Sainburg et al.
1993, 1995).
Signals from proprioceptors travel along afferent axons into the
spinal cord (figure 6.11). There, they make connections with different
kinds of neurons. Primary spindle afferents are the only known ones
to make direct projection on spinal α-motoneurons. The majority of
afferent axons make synapses on interneurons that are typically
smaller cells processing incoming information and making
projections onto other neurons, including motoneurons. Careful
tracing of spinal projections from different proprioceptors revealed a
rather complex picture (Jankowska 1979; Jankowska and McCrea
1983): Different afferents project onto the same interneurons, so the
original information about length, velocity, pressure, force, and joint
angles gets totally and seemingly irrecoverably mixed. Some afferent
fibers go as far as the brain without making intermediate
connections, which makes them probably the longest neural fibers in
the body—up to 2 m long from the sensory ending to the projection
in the brain. After they get into the brain, they apparently participate
in such processes as perception of the body and limb position and
movement planning.
Figure 6.11 Afferent nerves from peripheral receptors enter the spinal cord
through the dorsal roots. There they make synapses on interneurons and
motoneurons and send signals to the brain. Note that the same interneurons may
receive signals from afferents of different modalities.

The axons of mechanosensory afferents (both tactile and

proprioceptive) enter the spinal cord through the dorsal roots and
divide into ascending and descending branches. For the tactile
system, both branches project onto the gray matter of the spinal cord
and extend across several segments. The ascending branches of
the first-order neurons, called the dorsal column, ascend ipsilaterally
to the level of the caudal medulla, where they synapse on the
neurons of the dorsal column nuclei. Thus, the axons of these
neurons are quite long and extend all the way from the receptors,
through the spinal cord, to the brainstem.
The proprioceptive afferent axons also enter the dorsal roots and
divide into ascending and descending systems. But more
importantly, some neurons entering the dorsal roots synapse on
motor neurons in the ventral horn. This circuit forms the basis of the
stretch reflex. The stretch reflex or the myotatic reflex refers to a
muscle contraction in response to passive stretching of muscle
fibers. The proprioceptive afferent axons in the ascending branches
travel along with axons of tactile afferents in the dorsal column. From
there they follow the same route to the primary somatosensory
cortex as the tactile afferents. More information about the reflex
effects of proprioceptive afferent axons can be found in Chapters 17
through 19. Perceptual effects are described in Chapters 12 and 28.

CHAPTER 6 IN A NUTSHELL
Receptors are specialized cells or
parts of cells that can respond to
external stimuli of certain types.
Signals from receptors lead to a
perception of stimuli that is related
to the strength of the stimuli by a
nonlinear law. Proprioceptors produce
information about the relative
configuration of body segments. Muscle
spindles contain sensory endings of
two types, sensitive to muscle length
and sensitive to both muscle length
and velocity. The sensitivity of
spindle endings is modulated by a
special system of neurons, the
fusimotor or gamma-system. Golgi
tendon organs are sensitive to muscle
force. Articular receptors are
sensitive to both joint angle
(typically, close to the anatomical
limits of joint rotation) and joint
capsule tension. Skin and subcutaneous
receptors measure pressure on the
skin. Sensory endings send signals
along the peripheral branch of the T-
shaped axon to spinal ganglia, where
the bodies of sensory neurons are
located. Then, signals travel along
the central branch of the T-shaped
axon into the central nervous system.
 Chapter 7

Motor Units and

Electromyography

KEY TERMS AND TOPICS

motor units
Henneman principle
recruitment patterns
electromyography

It is time to make a step from considering the properties of single

cells to the next functionally important level of complexity. It would
certainly be unwise to expect the central nervous system to control
the level of activity of each and every neural and muscular cell
separately. Such an approach would impose a computational load
beyond human imagination. It is comparable to calculating the
trajectory of each individual elementary particle within a baseball in
order to ensure a desired trajectory of the ball. The central nervous
system simplifies the task and decreases the computational load by
uniting small elements of the neuromuscular system into functional
units that are controlled with just one or two parameters. The
smallest functional unit of the neuromotor system is termed a motor
unit.
Figure 7.1 Alpha-motoneurons in the spinal cord send their axons through the
ventral roots. Each axon branches in the target muscle and innervates several
muscle fibers. A motoneuron and the muscle fibers it innervates are called a motor
unit.

7.1 The Motor Unit

Figure 7.1 shows a couple of neural cells in the spinal cord
innervating a muscle (i.e., α-motoneurons). Their axons branch at
the end and innervate several muscle fibers each. Since each
neuron obeys the law of all-or-none, such an arrangement leads to
synchronized contraction of all the muscle fibers innervated by one
α-motoneuron in response to each action potential delivered by the
axon of the motoneuron. So, all these muscle fibers also behave
according to the law of all-or-none. The motoneuron and the muscle
fibers it innervates are called a motor unit. Typically, each muscle
fiber is innervated by only one axon branch, although during
development and during recuperation following nerve injury, muscle
fibers may receive inputs from a number of axons. With time,
however, such “redundant” inputs disappear, and each muscle fiber
ends up being innervated by one axon.
Motor units differ in size, which relates to the size of the
motoneuron body, the diameter of its axon, and the number of
muscle fibers innervated by the motoneuron. These parameters are
closely correlated, so large motoneurons (with a large body cell and
large axon) innervate more muscle fibers than smaller motoneurons.
The number of muscle fibers innervated by a single motoneuron (the
innervation ratio) varies in a wide range, from under 10 in muscles
controlling eye movements to over 1,000 in large muscles
participating in postural control during standing. With age, the
number of motoneurons decreases, and a process of reinnervation
of orphan muscle fibers takes place that leads to an increase in the
size of individual motor units and a corresponding increase in the
innervation ratio. These issues will be considered in more detail in
chapter 32 on the effects of aging.

PROBLEM 7.1
Can one motor unit produce different levels of muscle force?
Why?

PROBLEM 7.2
You want a large motor unit and a small motor unit to contract
simultaneously. How would you time neural commands to the two
motoneurons?

Up to now, we have been discussing the most common muscle

fibers that generate action potentials in response to local membrane
depolarization at the neuromuscular synapse, conduct these
potentials, and generate twitch contractions in response to each
action potential. Another type of muscle fiber is relatively rare and
has been seen in muscles involved in eye, throat, and ear
movements. These are sometimes called tonic fibers. Each tonic
fiber is typically innervated at a number of places. It does not
generate action potentials on its membrane in response to
presynaptic inputs but rather spreads postsynaptic depolarization
with local currents. As a result, the response of a tonic fiber to
presynaptic stimuli does not obey the all-or-none law and can be
graded. However, we are not going to consider these unusual fibers
further, but instead we return to the more common twitch fibers that
are involved in movements of the human body and limbs.

7.2 Fast and Slow Motor Units

Each muscle consists of a number of motor units. This number
ranges from less than a hundred for small muscles (e.g., those
controlling eye movements), to thousands for large muscles
controlling the movements of large body segments. Within a muscle,
one can see motor units that differ not only by their relative size but
also by their contractile properties.
Figure 7.2 (a) Twitch contractions and (b) tetanic contractions of three motor
units. Note that the fastest and strongest motor unit (motor unit 1) shows the
largest drop in force with time (fatigue), while the smallest and slowest motor unit
(motor unit 3) does not show fatigue at all.

There are two basic tests that are used to describe the functional
properties of motor units. One is twitch contraction, while the other
is fatigue. Figure 7.2a shows twitch contractions of three motor
units. Note that these motor units generate different levels of peak
force and take different times to reach the peak force level. In
particular, motor unit 3 takes the longest time for its twitch
contraction and generates the lowest peak force, while motor unit 1
is the first to reach the peak force level and has the highest peak
force. If the same motor units are stimulated at a rather high
frequency, they produce tetanic contractions. This can be achieved
by stimulating the axons of the motor units with an external electrical
stimulator. If the frequency of the stimulation is the same for all three
motor units, the peak force will once again be the highest for motor
unit 1 and the lowest for motor unit 3 (figure 7.2b). If the stimulation
continues for a rather long time, changes in the level of the
contraction force will be observed (figure 7.2b), induced by fatigue.
The mechanisms of fatigue will be discussed in chapter 30. Here, it
is important to note that the changes in the force levels of motor unit
2 and motor unit 3 are small, while the force level of motor unit 1
drops significantly.
Thus, there seem to be three types of motor units. Motor units of
the motor unit 1 type are typically called fast-twitch fatigable, motor
units of the motor unit 2 type are called fast-twitch fatigue resistant,
and motor units of the motor unit 3 type are called slow-twitch fatigue
resistant. These groups are sometimes referred to as FF, FR, and S
motor units. Slow motor units (S) typically have fewer muscle fibers,
smaller motoneurons, and thinner axons. Correspondingly, the
speed of conduction of action potentials along the axons of slow
motor units is the lowest (although it is rather high since, as you
remember, the axons of α-motoneurons are thick, myelinated, group
I neural fibers). FF motor units are characterized by the highest
conduction velocity. The difference in conduction velocity may be
more than twofold (from 40 m/s in some S motor units to 100 m/s in
some FF motor units).
The differences in the physical properties of motor units correlate
with their different biochemical and morphological characteristics.
Consider three major sources of energy used for muscle contraction.
The first is ATP contained in myofibrils; the importance of this source
may be assessed by the level of activity of the enzyme ATPase that
participates in metabolizing ATP. The second source is oxidative
metabolism occurring in mitochondria. Its rate may be assessed by
the activity of a couple of enzymes, succinic dehydrogenase and
NADH dehydrogenase. The third source is glycogen, whose
metabolism is anaerobic. Table 7.1 shows the three major types of
motor units and the relative representation of different physiological
and biochemical characteristics of their muscle fibers. The table
shows that S motor units contain mostly slow, oxidative fibers
characterized by a high level of mitochondrial oxidative processes
and a well-developed blood supply network. Fast motor units use
more energy from ATP and glycogen metabolism. FR motor units
have a rich capillary supply comparable to that of S motor units,
while FF motor units have a sparse capillary supply reflected in the
color of muscles with a high percentage of FF motor units: these are
pale muscles.

Table 7.1 Properties of Motor Units

Type FF FR S
Fiber diameter Large Medium Small
ATPase High High Low
Glycogen High High Low
Succinic dehydrogenase Low Medium High
NADH dehydrogenase Low Medium High

FF = fast fatigable; FR = fast fatigue resistant; S = slow.

Most muscles contain a mixture of motor units of different types,

although the percentage of slow and fast motor units may differ.
Slow muscles (i.e., those with a high percentage of S motor units)
are typically red (an example is the soleus), while fast muscles (i.e.,
those with a high percentage of FR and FF motor units) are typically
pale (an example is the lateral head of the gastrocnemius). Within a
muscle, the central nervous system has come up with a rule that
coordinates the order in which different motor units are recruited
during natural muscle contractions. This rule is extremely important
by itself but also as a unique example of a coordinative rule.
Figure 7.3 The Henneman principle (size principle). Small motor units recruit first
at low muscle forces. An increase in muscle force leads to recruitment of larger
motor units. Derecruitment (with relaxation) follows the opposite order.

7.3 The Henneman Principle

The Henneman principle (also know as the size principle) states
that the recruitment of motor units within a muscle goes from small
motor units to large ones. That is, if a person contracts a muscle at a
low force, nearly all the force is produced by the slowest motor units
(figure 7.3). If the contraction force is increased, larger motoneurons
start to generate action potentials, recruiting larger motor units. At
the highest level of muscle contraction (maximal voluntary
contraction force), the largest motor units are recruited. Note that
derecruitment of motor units during a decline of muscle force follows
the inverse order: The largest motor units are the first to be turned
off, while the smallest ones are the last to stop firing.
The contribution of a motor unit to total muscle force depends
upon two factors, the size of the motor unit and the frequency of
action potentials generated by its α-motoneuron. Larger motor units
have larger forces generated in response to single action potentials,
while all motor units generate more force (up to a limit, of course,
defined by the force during smooth tetanus) when action potentials
arrive at a higher frequency. The importance of the frequency of
firing for the force contribution of a motor unit gives the central
nervous system options for developing the same level of muscle
force with different combinations of motor units. Figure 7.4 illustrates
that the same level of muscle force may result from the recruitment
of fewer motor units at higher frequencies or from the recruitment of
a larger number of motor units at lower frequencies. Recruitment and
changes in firing frequency are the two major mechanisms of
regulating muscle force.
During sustained contractions, one can commonly see
derecruitment (turning off) of some motor units accompanied by
recruitment of new motor units or changes in the firing frequency of
already recruited motor units. Thus, the Henneman principle does
not by itself define which motor units are going to be recruited and at
what frequencies for a given level of muscle force. However, it limits
the area where the solutions can be searched for to particular
combinations of motor unit recruitment patterns. As such, it can be
compared to rules of grammar, which do not define exact word
combinations that have to be used to express particular meanings
but limit the freedom of choosing such combinations to
grammatically acceptable ones. The Henneman principle is a
coordinative rule rather than a prescribing rule.

Figure 7.4 Muscle force is kept constant. A change in the number of recruited
motor units correlates (negatively!) with their mean frequency of firing.

PROBLEM 7.3
 Formulate the size principle for the order of motor unit
involvement when the contraction is induced by progressively
increasing the strength of electrical stimulation of the muscle
nerve.

There are situations when the Henneman principle does not work
perfectly, although these are relatively rare. In particular, if a muscle
participates in a task where it is not the primary mover (for example,
it participates in a postural task component), the order of motor unit
recruitment within this muscle may change, leading to a violation of
the size principle for certain pairs of motor units: A larger motor unit
may be recruited before a smaller one. A reversal of the size
principle can also be seen in certain reflex responses, in particular in
responses to cutaneous stimulation.

7.4 Functional Roles of

Different Motor Units
The functional role of motor units is largely defined by their
properties. That is, tasks that require prolonged exertion of muscle
force are mostly carried out by slow, fatigue-resistant motor units,
while tasks that require a quick but short-lasting increase in muscle
force are mostly performed by fast motor units. In particular, many of
the postural muscles have a large proportion of S motor units. On
the other hand, muscles that participate in quick limb movements,
such as kicking, hitting, or catching, typically have a large proportion
of FR and FF motor units. Most muscles, however, have a relatively
wide range of motor units of different types reflecting their
participation in a variety of motor tasks. Some muscles, such as the
triceps surae, are composed on several heads with substantially
different compositions of motor units reflecting the roles of the
different heads in various motor tasks. In particular, the soleus has a
large proportion of slow muscle fibers and is best suited for postural
tasks such as prolonged standing. In contrast, the lateral
gastrocnemius has a large proportion of fast-twitch, fatigable fibers
and is used for quick actions, such as jumping.

PROBLEM 7.4
Which motor units would you expect to find in abundance in a
marathon runner, in a weightlifter, and in a swimmer?

The rates of sustained firing of motoneurons are commonly rather

high (from about 6 Hz to about 35 Hz) so twitch contractions of
individual motor units overlap, leading to a tetanus, although fully
fused (smooth) tetanus is observed rarely.
During natural voluntary movements, the central nervous system
uses both methods of force modulation: recruiting more motor units,
and increasing the firing frequency of already recruited motor units
(figure 7.5). The relative role of recruitment versus an increase in the
firing frequency differs across muscles and tasks. For example, hand
muscles are known to show full motor unit recruitment at relatively
moderate force levels (about 40% to 50% of the maximal voluntary
contraction force) such that further increase in the force can only be
accomplished by an increase in the average firing frequency. In
contrast, large leg and trunk muscles show recruitment of new motor
units up to very high forces.
Figure 7.5 To increase muscle force, the central nervous system may recruit
new motor units or increase the frequency of firing of already recruited motor units.
Here, F is force, f is a monotonically increasing function, and ƒ is frequency of
firing.

During most voluntary movements, individual motoneurons do not

demonstrate any substantial level of synchronization. At very high
levels of muscle force, however, during fatigue, and in some
neurological disorders (those accompanied by loss of voluntary
muscle force, such as following a spinal cord injury), synchronization
of motor unit firing becomes a way of achieving higher forces or
maintaining a required level of force for a considerable time. Motor
unit synchronization has both positive and negative features. The
gain is obvious: Synchronized discharges sum up to higher total
muscle force compared to asynchronous motor unit firing. However,
the smoothness of the contraction will suffer. There is also a
possibility of quicker fatigue.

PROBLEM 7.5
You have invented a way to induce abrupt synchronization of
motor units in human muscles. What groups of athletes would you
 recommend this method to, and what groups of athletes would
you suggest not even try it?

Synchronization of motor units can be measured directly, with the

method of cross-correlation of action potentials, or indirectly, by
performing spectral analysis of the “summed” (interferential)
electromyogram (see section 7.5). If one records the activity of a
couple of motor units for a long time, the cross-correlation function
shows a peak at about zero delay, if the motor units are well
synchronized.
So far, muscles have been discussed as separate units that are
controlled by the central nervous system to produce movements.
This is a major simplification. In particular, the notion of muscle
compartments has gained prominence (English 1984; Fleckenstein
et al. 1992; Serlin and Schieber 1993). Muscle compartments are
groups of muscle fibers that show similarities in their behavior in
physiological tests and motor tasks. In a way, muscle compartments
are muscles within a muscle. Compartments have been described in
both animal and human muscles. For example, human extrinsic
finger flexors have their muscle bellies in the forearm, while their
distal tendons, four tendons per muscle, attach to phalanges of
different fingers of the hand (see chapter 27 on prehension). Several
studies have suggested that motor units of these muscles form
groups that act preferentially (or even nearly exclusively) to produce
force in only one of the tendons. This allows us to achieve individual
control of finger motion such as that seen in professional musicians.

7.5 Electromyography
There are two basic methods of recording muscle activity:
Intramuscular or needle electromyography and surface or
interferential electromyography.
In the first method, a thin needle (with a diameter of less than 1
mm) is inserted into a muscle (figure 7.6). Inside the needle, there is
a very thin wire that is electrically isolated from the needle. The tip of
the wire is not isolated. An amplifier picks up the difference of
potentials between the tip of the wire and the needle. Since the
dimensions of each electrode and the distance between the two
electrodes are very small, the electrodes selectively pick up signals
(action potentials) in the closest proximity to the tip of the wire. Such
electrodes are designed to record the patterns of activity of individual
motor units. Note that each motor unit contains many muscle fibers,
but they all generate action potentials synchronously, so that the
electrode picks up the compound action potential of the whole motor
unit. Typically, a needle electrode can record the electrical activity of
a few motor units whose muscle fibers happen to be in close
proximity to the electrode. However, because each motor unit has a
somewhat different number of muscle fibers and also because the
location and orientation of these fibers with respect to the electrode
are different, each motor unit has a different, unique pattern of
voltage changes, a unique signature (figure 7.7). These differences
make it possible to record several motor units with one electrode and
identify their compound action potentials with a high degree of
certainty. Needle electromyography is frequently used in clinical
tests.
Figure 7.6 Intramuscular electromyography uses thin needle electrodes. Inside
the needle, there is a very thin wire that is electrically isolated from the needle. The
difference of potentials (Δφ) between the tip of the wire and the tip of the needle is
amplified and recorded.

Figure 7.7 A schematic of a typical recording with an intramuscular needle

electrode reveals a few motor units with different shapes of the compound
potentials MU1, MU2, and MU3.

The other method is interferential electromyography, which is

more frequently used in studies of voluntary movements of healthy
persons. The main goal of interferential electromyography is to
represent the activity of as many motor units as possible across a
muscle and to obtain a reflection in its overall involvement in the
task. Typically, two electrodes are taped on the skin over the muscle
belly, and the difference of potentials between the electrodes is
amplified (figure 7.8). On the one hand, the desire to sum up the
activity may suggest using very large electrodes and placing them as
far from each other as possible. On the other hand, in most studies,
a researcher would probably like to focus on just one muscle and to
avoid recording the activity of its neighbors. So there is a trade-off,
which is resolved differently by each experimenter and in each
particular case. For example, if one wants to record the activity of a
relatively small forearm or facial muscle, one cannot use very large
electrodes because they will pick up the activity of many other
muscles in the neighborhood. Alternatively, if one wants to record the
activity of a large postural muscle such as the latissimus dorsi or
biceps femoris, using large electrodes would probably be
appropriate. Typically, the size of electrodes used for surface
electromyography varies from 1 mm to 20 mm in diameter, while the
distance between the centers of the electrodes varies from 5 mm to
50 mm or even more. The choice of particular electrodes and their
placement is part of the art of electromyography.
Figure 7.8 Surface electromyography uses a pair of electrodes that are placed
on a muscle belly. A third electrode (ground) is used to reduce noise.

Absolute values of electromyographic signals recorded with

surface electrodes are typically on the order of tens to hundreds of
microvolts. There are numerous sources of electrical noise that can
obscure the biological potentials and make them indistinguishable
from the noise. The most frequently encountered sources of noise
are the 60 Hz voltage used as the power supply in every laboratory
and radio waves that are picked up by the subject’s body acting like
an antenna. Other possible sources include electric motors and
strong electrical magnets, even when these are located in an
adjacent room. In order to minimize the noise and ensure selective
recording of biopotentials, the body surface is usually grounded with
a large indifferent or reference electrode.
Developments in data acquisition and analysis during surface
electromyography have led to methods that combine the benefits of
surface and intramuscular electromyography (reviewed in Farina et
al. 2010, 2014; De Luca et al. 2015). The idea of these methods is to
record the same signals travelling within the muscle from different
locations on the skin over the muscle belly and then to extract from
them components (signatures) reflecting compound action potentials
of individual motor units. This is done by using arrays or clusters of
electrodes placed over the muscle belly. Such methods allow for the
exploration of recruitment patterns of individual motor units while
avoiding the unpleasant sensations associated with inserting
needles into one’s muscle. Data processing of such signals is
usually much more complex than typical data processing of surface
electromyographic signals.
Figure 7.9 illustrates the results of using this method to record a
set of motor units (the data for four motor units are illustrated) using
surface electromyography from an extrinsic finger flexor, the flexor
digitorum superficialis. The subjects performed cyclical force
production by the ring finger (shown in the middle panel). The
shapes of the motor unit action potentials are shown in the left panel,
the timing of their action potentials (raster plots) is shown in the
middle panel, and the changes in firing frequency with time are
shown in the right panel.
Figure 7.9 Arrays of surface electrodes can be used to identify individual motor
unit action potentials. (a) Examples of identified motor units. (b) Their firing
patterns during two cycles of force production. (c) The frequency profiles of the
motor units.
Reprinted by permission from S. Madarshahian, J. Letizi, and M. Latash, “Synergic Control
of a Single Muscle: The Example of Flexor Digitorum Superficialis,” The Journal of
Physiology 599 (2020).

7.6 Processing
Electromyographic Signals
It is impossible to recommend a universal method to record and
process electromyograms. There are several standard types of
procedures; however, each researcher selects his or her own
methods of data processing based on the actual goals of the study
and the researcher’s own imagination. Three operations are
frequently used in processing a surface electromyogram.
The first is filtering. Note that action potentials are very fast events
with typical times of potential changes of about a few milliseconds.
So a high-pass filter is frequently used, which cuts off all the
frequencies equal to or below 60 Hz. As a result, the 60 Hz noise is
reduced as well as possible reactions of the electrodes to purely
mechanical factors that are usually much slower than changes in
biological potentials. On the other hand, the upper limit of filtering
frequency is chosen based on the characteristic times of events that
are of interest for the experimenter. If the experimenter is not
interested in the microstructure of the electromyographic signals
such as the shapes of individual action potentials, a low-pass cut-off
frequency is commonly chosen on the order of a few hundred hertz.
The second operation is rectification. The purpose of this
procedure is to be able to get a quantitative estimate of an
electromyographic signal. If an action potential runs along a muscle
fiber under a pair of recording electrodes (figure 7.10), the difference
of potentials at the electrodes will change gradually, leading to a
reversal in its sign. Actually, many biopotentials show a nearly
symmetrical picture with respect to zero level. Integrating an
unrectified signal over a reasonably long time will yield a very small
number (close to zero) because the signal is composed of an
approximately equal number of positive and negative values.
Integrating a rectified electromyogram will result in a value reflecting
the average magnitude of the activity over the time of integration.

Figure 7.10 An action potential runs under a pair of electrodes. The difference of
potentials recorded by the electrodes will change its sign (the upper record).
Rectification means making all the values of the difference of potentials positive
(the lower record).
 There are two types of rectification. Full-wave rectification consists
of turning all the negative values of the difference of potentials into
positive values of equal magnitude (figure 7.10). Half-wave
rectification cuts off all the negative values and substitutes them with
zeros. Full-wave rectification is used more frequently. However,
recent analyses have shown that using this method may lead to
misleading outcomes, in particular in the identification of the timing of
bursts of muscle activation (Farina et al. 2004).

PROBLEM 7.6
Imagine that you have an EMG record. You can filter it and then
rectify it or, alternatively, you can rectify it and then filter. Which
method is better? Why?

The third procedure is integration. Actually, there are two types of

integration used for different purposes. If a researcher is interested
in the overall shape of the electromyogram rather than in its
microstructure, an EMG envelope is calculated. The EMG envelope
represents a time function each point of which is the result of
integration over small time periods, such as several tens of
milliseconds. The other integration procedure is used when an
overall measure of the amount of muscle activity over a certain
period of time is needed. Integration of a rectified EMG gives a value
reflecting total current between the electrodes as well as total
resistance. Skin resistance is very hard to control; it can vary in a
wide range, and it may even change during an experiment—for
example, if the subject sweats. So, in order to compare integral
electromyographic measures across subjects, one needs to
normalize these indices. Normalization commonly means dividing a
measured value by a number that is likely to reflect the differences in
the conditions of recording (for example, skin resistance) but not the
differences in the signal of interest:
where EN is normalized EMG, ∫EMG is a calculated integral of a
signal that one is interested in, and ∫EMGst is a calculated integral
over the same time interval during a standard task. This procedure is
another component of the art of EMG processing: It is very
subjective, and different experimenters use different methods of
normalization. Typically, integrated EMGs are normalized with
respect to the value recorded when the subject exerts maximal
voluntary contraction of the muscle or, alternatively, when the subject
exerts a standard level of force. Normalizing the EMG with the signal
recorded when the subject is asked to relax is typically not a good
idea; it is very sensitive to noise and borders on dividing the signal
by zero.
Figure 7.11 illustrates the effects of different filtering and
rectification procedures on an EMG signal recorded from the left
biceps brachii muscle of one of the authors. The upper trace is the
raw (unprocessed) EMG signal amplified and sampled at a high
frequency (1,000 Hz) by a computer. The next panel shows a full-
wave rectified signal without any additional filtering. The next two
panels show the effects of low-pass filtering (with a commonly used
second-order Butterworth filter; if you are interested, look into a
textbook on signal processing) at the cutoff frequencies of 100 Hz
and 20 Hz. The lowest panel shows the same signal rectified and
processed with a moving-average window of 100 ms to create an
EMG envelope signal. Note that filtering can affect characteristics of
the signal rather dramatically. It is always up to the experimenter to
select appropriate signal processing techniques based on the
purposes of the recording.
Figure 7.11 The effects of different filtering and rectification procedures on an
EMG signal recorded with surface electrodes from a human biceps muscle during
a series of brief voluntary contractions. The upper signal is the “raw”
(unprocessed) EMG signal sampled at a high frequency (1,000 Hz) by the
computer. Note the similarities (e.g., burst timing) and differences in the signal
under different filtering.

PROBLEM 7.7
Suggest methods to normalize EMG signals during very fast
movements and during very small changes in the level of muscle
activity.

CHAPTER 7 IN A NUTSHELL
A motor unit is a motoneuron and all
the muscle fibers innervated by its
axon. There are three major types of
motor units: slow, fatigue resistant;
fast, fatigue resistant; and fast,
fatigable. Slow motor units contain
neurons with a smaller body, thinner
axon, slower conduction velocity of
action potentials, and fewer
innervated muscle fibers. During
natural muscle contractions, motor
units are recruited in a fixed order,
from the smallest to the largest (the
size or Henneman principle).
Derecruitment follows the opposite
order, from the largest to the
smallest. Electromyography is a method
of studying muscle activation levels
and patterns that requires learning
many nuances to achieve accurate
results.
Problems for Part I
Self-Test Problems
1. You have an atypical neural cell in which the concentration of
K+ ions inside the cell is not 150 mmol/L, but only 50 mmol/L.
Everything else is exactly as in “regular” cells. Calculate the
equilibrium potential for K+. How will the equilibrium potential
on the membrane and the action potential in this cell differ
from those in “regular” neurons?
2. You observe two fibers under changes in temperature. A
stimulator is placed at one end of each fiber, and you record
the response at the other end. One fiber decreased the speed
of transmission of action potentials with a decrease in
temperature and eventually stopped transmitting them. The
other fiber did not transmit action potentials at higher
temperatures, started to transmit them at lower temperatures,
and stopped transmitting them at very low temperatures. What
can you conclude about these two fibers? Explain the
differences in their behavior.
3. You have a neural cell with one excitatory and one inhibitory
synapse. At a certain frequency of stimulation of both
presynaptic fibers, the neuron does not generate an action
potential. You increase the frequency of stimulation of the
inhibitory input without changing the frequency of stimulation
of the excitatory input. After some time, the neuron starts to
generate action potentials. Why? In another experiment, you
increase the frequency of stimulation of the excitatory input.
The neuron generates several action potentials and then
becomes silent. Why?
4. You study the response of a neural cell to a single excitatory
input. The cell generates action potentials at a certain
frequency. You add an excitatory neurotransmitter to the
extracellular space and the cell stops firing. What happened?
5. You induce a twitch contraction of a muscle by a direct
electrical stimulus. The external load is zero. Draw time
changes in the frequency of firing of a primary spindle ending,
 of a secondary spindle ending, and of a Golgi tendon organ.
Prior to the contraction, each receptor showed steady firing at
a constant frequency. Solve the same problem for isometric
conditions, that is, when the “muscle + tendon” complex
cannot change its length.
6. A person generates 5% of the maximal voluntary contraction
force of a muscle. Then muscle force increases slightly so that
only one new motor unit is recruited. What can you say about
the properties of this motor unit? The same person generates
95% of the maximal voluntary contraction force. Again, muscle
force increases slightly so that only one new motor unit is
recruited. What can you say about the properties of this motor
unit?

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. A quick shortening of a muscle
a. leads to an increase in the activity of secondary spindle
endings
b. leads to a decrease in the activity of Golgi tendon organs
c. leads to a burst of activity of primary muscle spindle
endings
d. all of the above
e. none of the above
Why?
2. A person generates a force of 5% of the maximal voluntary
contraction. Then a new motor unit is recruited. What can be
said about this motor unit?
a. It is large and will fatigue quickly.
b. It is small and will not fatigue.
c. It is small and will fatigue quickly.
 d. It is large and will not fatigue.
e. It consists of one γ-motoneuron and several muscle fibers.
Why?
3. During an isometric muscle contraction
a. the length of the muscle fibers does not change
b. Golgi tendon organs show a drop in their activity
c. tendon stiffness is higher than the muscle stiffness
d. secondary spindle endings show a steady increase in their
activity level
e. none of the above
Why?
4. An activated muscle is quickly stretched by an external force
to a new length. What will happen?
a. The activity of Golgi tendon organs will drop quickly and
then increase back to the original level.
b. Muscle force will drop.
c. The activity of Golgi tendon organs will increase quickly,
and then drop somewhat.
d. The activity of Golgi tendon organs will induce a
monosynaptic reflex in the muscle.
e. The muscle will show a period of silence in its electrical
activity (EMG).
Why?
5. In a neuromuscular synapse, the mediator (ACh) is quickly
destroyed by ACh-esterase. What will happen if ACh-esterase
is removed from the muscle?
a. Actively generated muscle force will depend on muscle
length.
b. The time profile of force during a single twitch contraction
will change.
c. The muscle will demonstrate a prolonged contraction in
response to a short sequence of action potentials in the
motor nerve.
d. All of the above
e. None of the above
Why?
 Part II

Neuroanatomical Foundations of
Motor Control
 Chapter 8

Cerebral Cortex

KEY TERMS AND TOPICS

primary motor cortex
premotor cortex
supplementary motor area
corticospinal tracts
brain–machine interfaces

The vast majority of visually guided voluntary movements that we

make on a daily basis involve many areas of the cerebral cortex. The
cortex receives visual information in the occipital cortex, and then
integrates this information with signals from other sensory
modalities, such as the vestibular and somatosensory systems in the
parietal cortex. It then relays this information to the frontal lobe for
movement planning and execution. The roles of these different
cortical areas have been understood through numerous studies that
have defined the motor challenges faced by individuals with different
neurological disorders affecting distinct regions in the cerebral
cortex. For example, a stroke in the frontal lobe causes motor and
cognitive impairments, whereas a stroke in the parietal cortex
primarily causes visuospatial neglect.
In this chapter, we focus primarily on the most extensively studied
motor areas in the frontal lobe of the brain. These areas include the
primary motor cortex (M1), the premotor cortex (PM), and the
supplementary motor area (SMA). M1 has been implicated in
direct activation of α-motoneuronal pools (Evarts 1968, Asanuma et
al. 1979) and also specifying higher-order movement parameters,
such as direction of movement (Georgopoulos et al. 1986). The PM
is primarily involved in coupling sensory cues to motor actions (e.g.,
stopping when traffic light is red and driving when it is green),
whereas the SMA appears to participate more in guidance and
planning of internally generated movements that do not rely on
external sensory cues.

8.1 Structure of the Cerebral

Cortex
The cerebral cortex consists of two hemispheres, four lobes in each
hemisphere, separated by the corpus callosum and anterior
commissure. These are the occipital, parietal, frontal, and temporal
lobes. The insular cortex is also part of the cerebral cortex. The
insula is an integration hub and shares strong connectivity to many
cortical and subcortical brain regions that serve sensory, emotional,
motivational, and cognitive functions. It also makes reciprocal
connections with the limbic system. Recent studies have also
implicated the insula in motor control and decision-making, but its
exact role is not clear (Gogolla 2017).
The corpus callosum consists of approximately 200 million heavily
myelinated white matter tracts that connect the left and right cerebral
hemispheres, and these tracts integrate and transfer information
from both the hemispheres to process sensory, motor, and cognitive
information. A view of the left cerebral hemisphere from the side is
shown in figure 8.1. The illustration shows the main gyri and sulci
that are used as landmarks to define the location of different cortical
motor areas.
Figure 8.1 A left sagittal view of the human brain showing the four lobes, the
central sulcus, the precentral gyrus, and the postcentral gyrus. The primary motor
cortex is in the precentral gyrus, and the primary somatosensory cortex is in the
postcentral gyrus.

To understand how the cortical motor areas control movement,

let’s take a simple example of a person reaching for a glass of water
that rests on a table along with other items, such as computers and
books. To make this simple movement, the nervous system performs
a series of operations. First, the visual system discerns the location
of the glass with respect to the body and other items on the table.
Then the visuospatial location of the glass is used to create a motor
plan to perform the movement of a hand. The proximal limb muscles
are activated first to initiate the transport of the arm toward the glass.
Once the movement is underway, the distal arm and digit muscles
are activated to open the hand so that the grip aperture is gradually
scaled to the diameter of the glass. Finally, as the hand approaches
the glass, the fingers are gradually closed to make contact with the
glass, lift it, and bring it to the lips.
 When the visual system detects an object (e.g., the glass), it
sends the sensory information along two parallel neural streams
(figure 8.2) to process the information related to the various
properties of the object. First, the dorsal visual stream that extends
from the occipital to the parietal cortex processes the spatial location
of the glass with respect to the body and other objects on the table.
Specifically, the dorsal visual stream can process different objects at
different locations on the table but may not be able to judge what
they are (e.g., differentiate a glass from a book). The ventral visual
stream will serve that function by identifying objects, but this stream
does not specialize in spatial localization.

Figure 8.2 The dorsal visual stream facilitates visuomotor function. This stream
is further divided into the dorsomedial and dorsolateral streams. The dorsomedial
stream facilitates reaching movements, and the dorsolateral stream facilitates
grasping movements. Both of these streams project to different parts of the
premotor cortex. The ventral visual stream subserves visual perceptual functions.
 The dorsal stream is further divided into the dorsomedial and
dorsolateral streams, where the former facilitates reaching
movements and the latter grasping movements. The dorsomedial
and dorsolateral streams project to the dorsal premotor (PMd) and
the ventral premotor (PMv) cortices, respectively. These premotor
areas then project to different areas in the primary motor cortex (M1).
Some motor neurons in M1 project directly to the spinal cord, and
others project indirectly through the brainstem to activate the
interneurons and α-motoneurons in the spinal cord that activate the
limb muscles to move the arm and fingers. As stated earlier, we will
focus only on the structure and functions of M1, PMd, PMv, and SMA
in this chapter. We will discuss the dorsal and ventral visual streams
in more detail in chapter 14.

PROBLEM 8.1
During a reach-to-grasp movement, the grip aperture gradually
scales as a function of the remaining distance between the hand
and the object. What does that tell us about how visual sensory
information is integrated during a limb movement?

We will refer to the different cortical regions using the mapping of

the cortex suggested by the German anatomist Korbinian Brodmann
(figure 8.3). Brodmann’s areas (BA) of the cortex refer to 52 regions
of the cerebral cortex identified based on cytoarchitectonic (cell
size, packing density, etc.) differences. The cortical motor areas in
the frontal lobe consist of the primary motor cortex (BA 4), premotor
cortex (BA 6), and supplementary motor area (BA 6c). BA 4 is part of
the precentral gyrus and lies anterior to the central sulcus and
posterior to the precentral gyrus. It borders the somatosensory
cortex (BA 3, 1, and 2) posteriorly. BA 6 lies anterior to the precentral
sulcus and its medial part is BA 6c. Both BA 6 and BA 6c are
involved in motor planning, movement initiation, and movement
inhibition, and are parts of two major loops involving subcortical
structures such as the basal ganglia and the cerebellum. It is
important to underscore that these abbreviations are often used in
scientific publications but are not used as frequently in the clinical
world. You will learn a little bit more about Brodmann areas 1, 2, and
3 in chapter 12.

Figure 8.3 The numbers shown in the map indicate Brodmann’s areas.
Brodmann assigned these numbers to the brain regions based on their
cytoarchitectonic structure.

8.2 Cells in the Cerebral

Cortex
There are two major types of neural cells in the cerebral cortex
(figure 8.4). These are pyramidal cells and stellate (or granule)
cells. The cortex has a characteristic layer structure that can be
seen in vertical sections. The uppermost layer is called the molecular
layer. It is composed mostly of axons and apical dendrites and
contains only a few cell bodies. The next is the external granular
layer containing a large number of small pyramidal and stellate cells.
It is followed by the external pyramidal layer containing mostly
pyramidal cells. The first three layers are the primary origin and
termination of intracortical connections. The next, internal granular
layer is composed of stellate and pyramidal cells and receives
projections from the thalamus. The fifth, internal pyramidal layer
contains large pyramidal cells. And the last, sixth layer, the multiform
layer, consists of different neurons. The last two layers connect the
cerebral cortex with other subcortical regions.
The stellate cells play the role of interneurons within the cerebral
cortex (i.e., their axons do not leave the cortex). In contrast, the
axons of pyramidal cells leave the cortex and form its most
conspicuous output. Some of the dendrites of pyramidal cells are
oriented toward the surface of the cortex and may reach the
molecular level. Other dendrites are oriented horizontally in layers 2,
3, and 4, and may be a few millimeters long.
Input signals (afferents) to cortical neurons come mainly from
thalamic nuclei and from other cortical neurons. Thalamic nuclei play
the role of a relay processing and transmitting information from
peripheral afferents, the cerebellum, and the basal ganglia. As stated
earlier, thalamic inputs make synaptic connections mostly in layer 4,
which contains many stellate cells with vertically oriented dendrites
that make synapses on pyramidal cells. As a result, the pyramidal
cells receive information that has been processed in both the
thalamus and the cortex. The vertical (column) input–output
organization is typical for the cortical structures. It is combined with
intercolumn connections with the help of horizontally oriented
dendrites.
Figure 8.4 Cortical layer 4 is rich in stellate neurons with local axons primarily
restricted to the primary sensory cortices (neuron on the left). These neurons
receive input from the thalamus. Layer 5 contains pyramidal neurons whose axons
leave the cortex (neuron on the right).

8.3 Premotor Cortex and

Supplementary Motor Areas
Before we look at the detailed structure and function of the primary
motor cortex (BA 4), we will discuss how premotor areas transform
sensory signals to facilitate the generation of descending signals by
the primary motor cortex. The premotor cortex (PM) serves the
important role of transforming visual sensory signals that it receives
from the dorsal visual stream in the parietal cortex to generate
descending motor signals for arm transport and hand grasping.
Compared to the motor cortex (see section 8.4), the premotor areas
of the brain have been identified more recently and are only isolated
from the motor cortex based on the cytoarchitectonic distinctions
made by Brodmann in his cortical atlas.
The premotor cortex is divided into the dorsal premotor cortex
(PMd) and the ventral premotor cortex (PMv) (figure 8.2). PMd
receives input from a parietal area called the medial intraparietal
area (also known as the parietal reach region), which is a node in the
dorsomedial pathway that encodes the direction of the reaching
movements and integrates visual information with proprioceptive
information from the somatosensory cortex to facilitate reaching
movements. PMv has reciprocal connections with cells in the
anterior intraparietal area in the parietal cortex. The anterior
intraparietal area is a part of the dorsolateral “grasping” circuit that
encodes an object’s shape. PMv then transforms the visual sensory
signals into neural signals to adapt the hand (wrist and fingers) to the
object’s shape. It is hypothesized that a special class of neurons
called canonical neurons in PMv that encode shapes of objects and
discharge during motor preparation are responsible for performing
these transformations. These neurons discharge selectively during
observation of graspable objects and when those objects are
grasped. Canonical neurons were first observed in monkeys by
Rizzolatti and colleagues in Italy (Rizzolatti and Gentilucci 1988), but
since then they have also been shown to exist in humans.
PMv contains another special class of cells called mirror neurons.
Mirror neurons modulate their activity both when an individual
executes a specific motor action and when they observe the same or
similar act performed by another person (reviewed in Rizzolatti and
Craighero 2004). When mirror neurons were first discovered,
researchers assumed that they had to play a critical role in action
understanding and observation-based motor learning. Mirror neurons
have now been reported to exist in both ventral and dorsal premotor
cortices as well as in the parietal cortex across species. These
neurons fire when one tries to repeat actions performed by others
and also during imitation learning. Note that performing “the same”
movement by different animals or by different individuals necessarily
involves different muscle activation patterns and different joint
rotations. Hence, mirror neurons likely participate in encoding more
general, topological properties of movements. Recently, it has also
been suggested that these neurons may facilitate learning of reach-
to-grasp movements in infants (Oztop et al. 2004). However, some
researchers believe that the function of mirror neurons in motor
learning is limited—that they play a major role in social cognition but
not in the learning of movements.
Both PMd and PMv are densely interconnected with the primary
motor cortex (M1). To facilitate visually guided reaching movements,
PMv encodes higher-level planning of goals, such as which target to
reach (a glass with water or a plate with snacks). PMv also plays a
major role in processing object properties relevant for grasping and
specification of hand shape and grip force. In contrast, PMd encodes
the relative positions of the target (the glass versus the plate), hand,
and gaze. PMd neurons are active both during the preparatory
phase of the movement and when the movement is underway,
correcting any deviations from the intended movement trajectory. For
example, while making a movement toward the glass, you realize
that there is an obstacle (e.g., a book); PMd will facilitate the
correction of the movement trajectory. Thus, the main roles of PMd
are integration of sensory information into motor commands and
specification of movement amplitude, direction, and speed.
The supplementary (SMA) and presupplementary motor areas
(pre-SMA) are located in the dorsomedial frontal cortex (figure 8.5).
The SMA is located on the medial surface of the cortex anterior to
the precentral sulcus, and the pre-SMA lies just anterior to the SMA.
Studies in primates suggest that pre-SMA and SMA are involved in
the sequencing and initiation of movements, with the pre-SMA
playing a more abstract, higher-level cognitive role and the SMA
playing a more motor-specific role. Indeed, the pre-SMA is strongly
connected with the prefrontal cortex, whereas the SMA is more
heavily connected with the motor cortex, suggesting a functional
division with the pre-SMA involved in higher-level motor planning
and the SMA with motor execution. More recently, the SMA has
been shown to facilitate response inhibition in reaction to sudden
changes in a task, such as those encountered when a traffic light
suddenly changes from green to yellow, prompting a swift change of
moving the foot from the accelerator to the brake (Nachev et al.
2008).
A few final points to note about the premotor cortex and
supplementary motor areas is that electrical stimulation of these
areas induces muscle activity, but the magnitude of stimulation
current required in premotor cortex areas to elicit muscle activity
tends to be much higher than the motor cortex. Such stimulation
induces more complex movements that frequently involve a number
of joints. Both the premotor and supplementary motor areas also
contain somatotopic maps of the body.

Figure 8.5 The frontal cortex areas involved in motor planning. The premotor
cortex (dorsal and ventral) and SMA receive extensive projections from the parietal
cortex and other brain areas. The pre-SMA receives strong projections from the
prefrontal cortex and is considered to be involved in cognitive motor control. The
motor planning areas are connected to the primary motor cortex (M1).

PROBLEM 8.2
While driving a car in the countryside, you decide to increase the
speed of the car by 5 mph. Which frontal premotor area will be
primarily responsible for facilitating the change in the car’s speed?
At another time during the drive, you see that cars ahead are
moving slowly, and many of them have their emergency flashers
 on. You then gradually slow down the car. Which premotor area
will be primarily responsible for this action?

PROBLEM 8.3
How can you interpret the fact that stimulation of the premotor
areas requires higher currents to induce visible muscle
contractions than does stimulation of the primary motor cortex?

8.4 Primary Motor Cortex

Brodmann area 4 (BA 4) or the precentral gyrus is also known as the
primary motor cortex (M1). Together with the premotor cortex (PMd
and PMv) and supplementary motor areas (SMA and pre-SMA), the
M1 makes up the cortical motor pathway. An important distinction
between the premotor and supplementary motor areas and the
motor cortex is that the motor cortex contains giant pyramidal cell
bodies known as Betz cells. The Betz cells are rarely present in the
premotor areas. Another important distinction between the premotor
areas and M1 is that the threshold of electrical stimulation for
initiating movements is much lower in M1. This suggests the
presence of relatively large and direct pathways from M1 to the
spinal cord and brainstem.
In the 19th century, the British neurologist John Hughlings
Jackson became interested in the cortex as a critical area for motor
control based on his observations of epileptic seizures. He saw that
seizures would often begin in the hand and then spread
systematically up the body toward the face. This led Jackson to
hypothesize that a motor map in the cortex might be responsible for
movement in different parts of the body. At around the same time,
the German physiologists G. Theodor Fritsch and Eduard Hitzig
used electrical stimulation on the canine cortex to show that the
motor cortex elicits contraction of muscles in the contralateral limbs.
 A little later in the 20th century, Sir Charles Sherrington mapped
the motor cortex using minimum currents to elicit the smallest
discernable limb movements in the great apes. One of Sherrington’s
students, Wilder Penfield, extended this work to humans and
showed that the motor cortex contains a somatotopic map of the
contralateral body. He showed that stimulation of small regions in the
motor cortex elicited activity in specific muscles, suggesting that
vertical columns of cells in the motor cortex may encode individual
muscles in specific body parts. Penfield and his collaborators called
this representation the motor homunculus (figure 8.6). A similar
map also exists in the somatosensory cortex, and it is called the
sensory homunculus (chapter 12).
However, according to many contemporary motor physiologists,
the idea that the motor cortex encodes a simple somatotopic map of
the body is not tenable anymore. Though in the motor cortex, the
head and the limbs have largely separate representations, the distal
and smaller body parts have widely distributed mosaic
representations within these major regions. More recent studies
have shown that even small currents capable of eliciting a limb motor
response initiate activity in multiple muscles, suggesting that the
motor map in the cortex might encode movements instead of
individual muscles (Kakei et al. 1999). Even within the separate
representations (e.g., face or upper limb) in M1, specific movements
can be elicited by stimulation of widely separated neural areas,
suggesting that intracortical connections interlink the cortex
throughout major body parts, and the convergent output of many M1
motor neurons reaches the same spinal motor neuron pools to
organize those movements. In addition, divergent output from many
M1 motor neurons reaches multiple spinal motor neuron pools.
Experiments on monkeys by Graziano and his colleagues have
shown that application of longer episodes of electrical stimulation to
the same cortical areas produces smooth, complex multijoint actions
that seem to resemble components of the everyday motor repertoire
of the animal (Graziano et al. 2002). These studies further suggest
that the motor cortex might encode movements rather than control
the individual contractions of muscles.
Figure 8.6 The motor homunculus is a topographic representation of the body in
the motor cortex. Note that the hands and face have a much larger representation
than the legs.

8.5 Efferent Output From the

Cortical Motor Areas
Efferent projections of the cortical motor areas have been
extensively studied with the help of direct electrical stimulation of the
cortex in animal studies. The major output from the cortical areas
extends to the basal ganglia, the cerebellum, the red nucleus, the
reticular formation, and the spinal cord. With the exception of the
projections to the basal ganglia and cerebellum, all other cortical
output projections contribute directly or via spinal interneurons to
muscle activation. The pyramidal tracts consist of pyramidal cells in
layer 5 of the cortex and are part of the upper motor neuron system
—a system of efferent nerve fibers that carry motor signals from the
cerebral cortex to either the brainstem (corticobulbar tract) or the
spinal cord (corticospinal tract).
The output neurons of the cerebral cortex reside in M1, the
premotor cortex (both PMd and PMv), and the SMA. In contrast, the
pre-SMA has a sparse projection in the corticospinal system. M1
also shares strong reciprocal connections with PMd, PMv, and SMA,
but not with the pre-SMA. The dense interconnections between
these areas mediate the planning and initiation of complex temporal
sequences of voluntary movements. The pyramidal cells in layer 5 of
M1 consist of Betz cells (~5% of all the cells) and non-Betz
pyramidal cells (~95% of all the cells). Though there are relatively
few Betz cells in M1, they do play a critical role in activating α-
motoneurons in the spinal cord that control muscles. The non-Betz
cells are also found in the premotor and supplementary motor areas.
The myelinated axons of the output cortical neurons descend in
the corticospinal tract and terminate in the spinal cord. The cell
bodies of these upper motor neurons lie in M1 (~40%), premotor
areas (30%), supplementary motor areas (15%), and the
somatosensory cortex. The corticospinal tract contains more than 1
million axons. There are two corticospinal tracts coming from the left
and right hemispheres. Approximately at the level of the medulla,
80% to 90% of the axons switch sides (decussate), forming the
lateral corticospinal tract (figure 8.7). Decussation is when nerve
fibers cross from one side of the body to another. Most of the axons
that form the lateral corticospinal tract from the right hemisphere
travel on the left side of the spinal cord and innervate muscles of the
left limbs, and vice versa. The lateral corticospinal tract forms a
direct pathway from the cortex to the spinal cord, and it projects to
the lateral portions of the central horn in the spinal cord. About 20%
of the axons from this tract directly project to α-motoneurons in the
spinal cord that innervate distal muscles in the forearm and hand.
These axons are critical for dexterous control of finger muscles and
allow us to make fine precision movements, such as lifting a glass
grasped with the fingertips, writing with a pen, and tying shoelaces.
The remaining 80% or so of the axons in the lateral corticospinal
tract terminate on spinal interneurons that coordinate activation of
the α-motoneurons that innervate different muscles. The axons that
do not decussate at the level of the medulla make up the ventral
corticospinal tract and terminate bilaterally. The ventral tract primarily
affects activation of postural muscles in the trunk and proximal limb
muscles.

Figure 8.7 About 90% of the axons of the corticospinal tract decussate and form
the lateral corticospinal tract. They innervate α-motoneurons of the distal muscles.
The remaining 10% of the axons form the ventral corticospinal tract and innervate
trunk and proximal limb muscles.

The activity of corticospinal tract neurons has been examined

during relatively simple movements, such as flexion or extension in a
joint, mostly in experiments on monkeys. If a monkey is trained to
make a simple movement in response to a sensory stimulus, the first
changes in the muscle activity (EMG) occur at a delay (latency) of
about 150 ms. Changes in the activity of pyramidal neurons can be
seen up to 100 ms prior to the EMG changes. These changes are
more tightly coupled with the EMG changes than with the sensory
stimulus, suggesting that they are related to movement production,
not to the perception of sensory stimuli. The magnitude of the
changes in the firing rate of pyramidal neurons has been reported to
be more closely related to the magnitude of force produced by the
animal (Evarts 1968).

PROBLEM 8.4
Does the last finding prove that pyramidal neurons control muscle
force? Why?

More careful, later studies by D. Humphrey (1982) have

suggested that there may be two subpopulations of cortical neurons.
One subpopulation shows reciprocal changes in the activity during
movements in the opposite directions (e.g., flexion and extension).
The other subpopulation changes its activity with a change in
cocontraction of agonist and antagonist muscles, which modifies
apparent joint stiffness without causing a major change in the net
joint torque or joint movement. There is, however, considerable
overlap between these groups, as there is among virtually all other
groups of neurons identified in brain structures. The equilibrium point
hypothesis (Feldman 1966) is based on two variables: one related to
joint equilibrium position and the other to joint apparent stiffness. The
presence of two subpopulations of cortical cells provides indirect
support for this view.

8.6 Afferent Input Into the

Cortical Motor Areas
The cerebellum and the basal ganglia are the primary sources of
noncortical input to the motor areas. The cerebellum and basal
ganglia project to the ventrolateral nuclei of the thalamus. These
projections in turn make synapses onto the primary motor cortex,
supplementary motor area, and premotor cortex (figure 8.8). Figure
8.8 oversimplifies the actual picture of thalamo-cortical projections,
which is characterized by considerable overlaps.

Figure 8.8 The projections from the cerebellum and the basal ganglia to the
frontal motor areas, premotor cortex (PM), and primary motor cortex (M1) go
through the thalamus.

The parietal cortex and cortical inputs are a major source of input
into the cortical motor areas. The parietal cortex integrates
multisensory information from the somatosensory system
(proprioception and tactile) and the visual and vestibular systems
and provides this information to the cortical motor areas for motor
planning and execution.
The premotor and supplementary areas also receive input from
the prefrontal cortex. The prefrontal cortex is involved in higher-order
executive functions, such as representation of abstract and complex
rules, short-term working memory, and decision-making. The
connections between the prefrontal and premotor areas indicate that
the output of the prefrontal cortex targets specific areas of the
premotor cortex involved in motor control (Lu et al. 1994). The
prefrontal cortex also shares strong connections with the pre-SMA,
which suggests that the pre-SMA plays an important role in cognitive
motor control.

8.7 Hemispheric Lateralization

in the Cortical Motor Areas
Most neural fibers within the corticospinal tract cross the midline of
the body, so the primary motor cortex in the right hemisphere
predominantly sends signals to the left limbs, and vice versa.
Besides the anatomical lateralization seen in many vertebrates, the
two cerebral hemispheres also specialize in distinct behavioral
functions. For example, speech and language comprehension is
controlled by Broca’s and Wernicke’s areas, respectively, and both
these areas are located in the left hemisphere in about 95% of right-
handers and about 70% of left-handers. In contrast, the right
hemisphere, particularly the parietal cortex, primarily guides visual
attention to both the left and right halves of the body. This separation
of neural circuitry to divide labor between the two sides of the brain
is considered an organizational attribute of vertebrate nervous
systems that may permit efficient behavior.
This form of organizational lateralization has also been proposed
for the cortical motor areas (Sainburg 2005). Though the primary
motor cortex contralateral to the arm has the greatest influence on
arm motor behavior, it has been proposed that both hemispheres
contribute to the control of unique aspects of behavior (Sainburg
2005). In right-handed people, the left hemisphere is specialized in
controlling movement dynamics and the right hemisphere contributes
to controlling the postural aspects of the movement. As an example,
when you reach out to lift a glass of water, the left hemisphere
specifies the spatiotemporal sequence of signals that lead to muscle
activations to control the trajectory of the arm movement, whereas
the right hemisphere specifies how the muscles are activated to
respond to any unexpected perturbations that the arm might
encounter (such as the arm bumping against a book).
 The premotor cortex is also considered to be functionally
lateralized. The left dorsal premotor cortex is activated during
unimanual movements of the left or the right hand and is particularly
involved in the selection of movements in response to visual stimuli.
In contrast, the right dorsal premotor cortex is active during control of
bimanual movements. However, it is unclear if functional
lateralization also exists in the ventral premotor cortex; the mirror
neuron system in the ventral premotor cortex does not seem to
exhibit functional lateralization.

8.8 Preparation for a Voluntary

Movement
Voluntary movements are typically initiated in one of two ways. The
first way is through spontaneous internally guided intentions that do
not originate in the sensory system. For example, you may feel
thirsty or you may remember that you should rehydrate regularly
before you reach out to grab a glass of water. This is typically
described in the literature as top-down motor control, or control that
originates in the prefrontal areas of the brain. Here “top” refers to
prefrontal regions that are critical for controlling behaviors in
accordance with intentions or rules. However, this is a gross
simplification, and intentions for many actions that humans perform
may originate not in the prefrontal areas but in the limbic system.
The second way actions might originate is in the sensory systems.
For example, you may just be glancing around from your computer
when your gaze lands on the glass of water, which may cause a
desire to take a sip. This is labeled as bottom-up, sensory feedback–
initiated motor control, where sensory signals in the peripheral
nervous system trigger the chain of sensorimotor events that
eventually lead to the initiation of movement.
In a series of experiments with monkeys, Miller and colleagues
(Buschman et al. 2007; Siegel et al. 2015) implanted electrodes and
recorded neural spikes (action potentials) and local field potentials
from multiple areas in the visual, parietal, frontal, and temporal
cortices to elucidate how neural signals associated with sensory
processing (bottom-up) and action selection (top-down) flowed in the
brain. Local field potential is the electric potential in the extracellular
space around neurons that can be recorded by electrode arrays.
In one experiment, monkeys were led to make saccades (rapid
eye movements between fixation of gaze on objects) to a visual
target. In a “bottom-up” kind of condition, the target was located
among three identical distractors that differed from the target in two
properties (color and orientation), making it incredibly easy to spot
the salient target. In another “top-down” condition, the distractors
matched the target on either color or orientation, forcing the
monkeys to search for the target while using their short-term working
memory (an executive function associated with the prefrontal cortex)
to memorize the target appearance. They showed that top-down and
bottom-up signals arise in the frontal and sensory cortices,
respectively, and contribute to the initiation of the saccades.
The supplementary motor areas mediate action selection based
on top-down, internally guided actions. For example, if action
sequences are initiated based on memory (e.g., without external
instructions or visual cues), the prefrontal and supplementary motor
areas are activated. The supplementary areas have also been
associated with the control of motor readiness during voluntary
initiation of movements. Electroencephalographic recordings from
humans have revealed a slowly increasing negative potential, known
as the Bereitschaftspotential, or the readiness potential (Deecke et
al. 1969), that is centered over the SMA and pre-SMA before
movement initiation (see figure 8.9). This would suggest that for
these movements, neural activity would start in the prefrontal cortex,
followed by the supplementary motor areas, and then the upper
motor neurons of the primary motor cortex. Changes in the resting
electroencephalogram (EEG) pattern can be seen as early as 1.5 s
prior to the first signs of changes in the background muscle activity.
The relatively long duration of the readiness potential is surprising.
Humans can make a decision to move and to start a movement in
much less time than 1.5 s. Actually, the shortest reaction time to a
visual or auditory stimulus is just over 100 ms.

PROBLEM 8.5
Suggest an explanation for the difference between the typically
short reaction times (commonly under 200 ms) and the rather long
readiness potential.

Figure 8.9 The electroencephalographic (EEG) signals before self-initiated

voluntary finger movements (right index finger flexion) show a negative potential
shift that begins approximately 1.5 s before the finger movement–related surface
electromyographic (EMG) muscle activity is detected. The EEG signal was termed
Bereitschaftspotential by Kornhuber and Deecke and provided the first insights into
the neural origins of self-initiated movements.
Distributed under the terms of the Creative Commons Atrtibution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/).

For many other real-world activities that originate in the sensory

system, the nervous system forms arbitrary associative maps
between bottom-up sensory signals and motor responses (e.g., go
when light is green, slow down when it’s yellow, and stop when it’s
red). Neurons in the dorsal premotor cortex (PMd) fire during
observation of the visual cues (e.g., go for green and stop for red)
and facilitate the selection of an appropriate action based on the
cues. Thus, the PMd plays an important role in action selection when
movements are guided by sensory signals.
 Movement preparation and action stopping, where a planned
movement is abruptly aborted, have been studied using transcranial
magnetic stimulation (TMS) over M1 (reviewed in Duque et al. 2017).
TMS is a noninvasive technique that can induce rapid and transient
(~250 μs) electrical currents in the human cortex. When the
stimulating coil is placed over M1, TMS elicits descending volleys in
the corticospinal tracts. TMS over M1 activates corticospinal neurons
not only directly but also indirectly via the stimulation of intracortical
circuits that project to corticospinal neurons. These tracts synapse
on spinal α-motoneurons that innervate peripheral muscles
contralateral to the hemisphere being stimulated. The evoked
response, called the motor evoked potential (MEP), is measured
using surface electromyography (EMG).
Over the course of three decades, TMS protocols have been
developed to probe M1 connectivity with other neural areas. In
paired-pulse protocols, a low-intensity subthreshold conditioning
pulse is applied first, and then a suprathreshold test pulse is
generated in the same coil. The two TMS pulses are applied over M1
at specific intensities and different time intervals. When the
interstimulus interval between the conditioning pulse and the test
pulse is 2 to 5 ms, it causes test MEP responses to be inhibited
(intracortical inhibition), and when they are between 8 to 15 ms, test
MEP responses are facilitated (intracortical facilitation). The
conditioning pulse is assumed to probe GABAergic inhibitory
neurotransmission in the cortical areas. As a reminder, the
GABAergic system is the main inhibitory neurotransmitter in the
central nervous system (see also chapter 4).

8.9 Neuronal Population Vectors

In the late 1960s, Ed Evarts used recordings in M1 of monkeys and
showed that single M1 neurons were active during movement of one
joint but less active or not active at all for other joints. In addition,
neurons fired when the preferred joint moved in one direction and
were less active when the joint moved a different direction. This
activity was observed 50 to 150 ms before the onset of agonist
muscle activity of the preferred joint.
This work was extended in an exciting series of studies performed
in the 1980s by Georgopoulos and his colleagues in investigations of
the behavior of large populations of neurons in the primary motor
cortex (and later, in other areas). In these experiments, monkeys
were trained to perform hand movements to visual targets that might
appear in different parts of the screen, which means they performed
movements in different directions. A large number of neurons were
recorded with implanted electrodes. The changes in the firing level of
each neuron demonstrated a peak during movements in a certain
preferred direction (figure 8.10). Movements in directions close to the
preferred one were accompanied by slightly lower increases in the
resting activity. Movement in the opposite direction could lead to
suppression of the resting activity.
Each neuron was associated with a unitary vector in the direction
for which the neuron demonstrated the largest increase in its resting
activity. Then the animal performed movements in different
directions, and the activity of all the neurons was recorded during
movements in all the directions. If a neuron demonstrated an action
potential in a time interval just prior to the movement, a unitary
vector in its preferred direction was drawn. If the neuron
demonstrated two or three action potentials, two or three vectors
were summed up, and so forth. Then all vectors were summed up
across all the neurons. Note that the population vector points in a
direction that is very close to the direction of the movement. A
change in movement direction in response to a change in the
position of the visual stimulus is accompanied by a rotation of the
population vector from the first to the second target.
The procedure itself is at least partly responsible for the result,
which can be obtained for any array of units (neurons or
nonneurons) that satisfy two conditions: (1) their activity is related to
movement direction by a cosine function, and (2) they cover the
whole range of movement directions. In particular, if one records
EMGs of all the limb muscles during the same task, the result will be
very similar. The same thing may happen if an experimenter records
and processes in a similar fashion the activity of muscle spindle
endings or Golgi tendon organs. So these results, by themselves, do
not prove that the population of neurons in the cortex controls
movement direction. This is an example of a very important
distinction between correlation and causation.

Figure 8.10 (a) The figure shows a raster plot of a single neuron during arm
movements in eight directions. Here 0 on the x-axis indicates movement onset.
Each vertical line on the raster plot indicates an action potential. The activity of this
particular neuron increased when a movement was made between 90º and 225º
and decreased below baseline levels when the movement was made between 45º
and 315º. (b) The firing frequency of a cortical neuron demonstrates a cosine-like
dependence on the direction of voluntary movement. For this particular neuron, the
preferred direction is 135º.
Adapted from Georgopoulous (1982).

Later, this experiment was performed with a modification when the

monkeys were trained to move not in the direction of the target but in
another direction shifted by a constant angular value from the target.
This procedure apparently requires a mental calculation (or a mental
rotation) in order to move in the required direction. In this task, the
monkeys demonstrated a considerably larger delay between the
stimulus and the initiation of a response; apparently this was related
to the task’s complexity. Recording the activity of a population of
cortical neurons revealed that the population vector rotated from the
direction of the stimulus to the movement direction during the
prolonged preparatory period. These very elegant experiments
strongly suggest that cortical neurons participate in processes that
encode the direction of voluntary movements.

8.10 Encoding Movement

Parameters in the M1
Since the pioneering work of Georgopoulos and colleagues, where
they showed that the activity of a population of cells in M1 encodes
higher-order movement parameters such as hand direction, many
researchers have questioned the tenability of this hypothesis. Using
physiological models of the musculoskeletal system, they have
shown that given the complex properties of the limb and
musculature, the relationship between neural activity and limb
movement can be shown to encode a variety of movement
parameters, such as target location in peripersonal space, hand or
joint motion, joint torques, and even muscle activation patterns.
Furthermore, there is now growing evidence that M1 cells discharge
both during action and during action observation, much like mirror
neurons in the ventral premotor cortex. Furthermore, somatosensory
responses in M1 neurons have been reported using tactile
stimulation, mechanical perturbation, and passive movements of the
limbs. These studies from the 1970s and 1980s showed that M1
cells integrated sensory information from muscles about the
particular joint that was perturbed and then activated synergistic
muscles to generate corrective motor responses.
 More recently, Pruszynski and colleagues have provided
compelling evidence to show that the motor cortex integrates
sensory feedback from not one but multiple muscles spanning
different joints (Pruszynski et al. 2011). The authors perturbed either
the shoulder or the elbow joint and observed different neural
responses in a population of shoulder-tuned M1 neurons
approximately 50 ms after the perturbations. The interesting aspect
of these perturbations was that both the shoulder and the elbow
perturbation produced the same shoulder motion. Despite the
unclear signals from the shoulder muscles, M1 cells integrated
peripheral sensory information from the muscles spanning the two
joints to generate an appropriate motor response. This suggests that
M1 cells may also serve intelligent sensory functions.
Though there is no doubt that the cortical motor areas serve
important functions for voluntary movements, the sensory responses
observed in the premotor and motor cortices suggest that we should
think of the cortical motor areas as nodes or layers in sensorimotor
neural networks. Almost 90 y ago, Bernstein (1935) wrote, “No area
of the cortex can currently be viewed as the origin or the final
destination of a neural process … Every area and every layer of the
cortex represent only transit points of the neural process (p. 326).”
Researchers who try to model the neural functions with the method
of artificial neural networks would express this idea somewhat
differently: The cortex is neither the input layer nor the output layer of
any meaningful neural process; it is always a hidden layer (figure
8.11). Note that at the hidden layer level, the information is typically
mixed up to such a degree that no clear reflections of the input or
output variables can be found.
Figure 8.11 Schematic of a neural network with three layers. The input layer
could process sensory information, and the output layer could output the motor
response processed by the hidden layer.

8.11 Brain–Machine Interfaces

In the last two decades, neuroscientists have collaborated with
engineers and clinicians to create devices based on interfacing
neural activity in the brain with control machines that can then be
used to restore functions lost to neurological insults and diseases.
The design of brain–machine interface (BMI) devices involves four
components:
a neural recorder for the acquisition of brain signals from a
population of neurons that encode motor goals,
a decoder to process and decode neural signals using artificial
neural networks and convert them into control signals for the
machine,
a device that takes the control signals from the decoder to
operate an end-effector (a robot, mechanical system, or virtual
control program), and
implementation of sensory feedback to facilitate adaptive
plasticity and control of the BMI device.
 A schematic representation of a BMI device is shown in figure
8.12. There are three important things to note here. First, even
though most intracortical microelectrode arrays are planted in M1 to
obtain neural recordings, signals have also been obtained from the
posterior parietal and premotor cortices to improve function. Second,
though there is no agreement among neuroscientists on what
aspects of movements are encoded in M1, a decoder does not need
to know that. It simply uses machine learning algorithms to extract
important features of movement to implement control signals for a
robot or end-effector. Finally, BMIs have been developed using
different forms of recordings, such as single- or multiunit recordings
of action potentials, electrocortography, electroencephalography, and
functional magnetic resonance imaging. This shows the versatility of
time scales over which neural information can be integrated to drive
brain–machine interfaces.

Figure 8.12 A brain-machine for upper limb motor function. The patient is asked
to imagine making movements of the paralyzed arm. The neural activity is
recorded with electrodes (often implanted in the brain). Then those signals are
processed and decoded. The decoded information is then used to generate control
signals to drive the motors of the prosthetic device.
Reprinted by permission from E. López-Larraz, A. Sarasola-Sanz, N. Irastorza-Landa, N.
Birbaumer, and A. Ramos-Murguialday, “Brain-Machine Interfaces for Rehabilitation in
Stroke: A Review,” NeuroRehabilitation 43, no. 1 (2018), with permission from IOS Press.
The publication is available at IOS Press through 10.3233/NRE-172394.
CHAPTER 8 IN A NUTSHELL
The motor areas of the brain are
involved in both the planning and
execution of actions. The primary
motor cortex (M1) provides the most
input to the corticospinal tract that
is involved in the control of fine
upper-limb arm and hand functions. M1
receives input from the premotor areas
of the brain. The premotor areas are
strongly interconnected with the
parietal and prefrontal cortices. The
premotor areas are involved in the
transformation of sensory input to
parameters of limb motor control.
These areas also project to the
corticospinal tract. The premotor
cortex is divided into the dorsal and
ventral areas, which are both involved
in movements initiated by sensory
signals. The main role of the dorsal
premotor cortex is integration of
sensory information into motor
commands and specification of reaching
movement amplitude, direction, and
speed. In contrast, the ventral
premotor cortex is involved in
preshaping of fingers for grasping
movements. It also contains mirror
neurons that respond to action
observation. The supplementary motor
area is also an important region that
abuts the premotor cortex. It receives
extensive input from the prefrontal
regions and is involved in intentional
movements.
 Chapter 9

Basal Ganglia

KEY TERMS AND TOPICS

caudate nucleus
putamen
globus pallidus
substantia nigra
subthalamic nucleus
dopamine
habit formation
reinforcement learning
Parkinson’s disease
Huntington’s disease

For many years, the basal ganglia have been viewed as a very
important set of brain nuclei for motor function. In particular, the
multilevel hierarchical scheme for the neural control of movement
introduced by Nikolai Bernstein (1947) involved two levels with
crucially important contributions from structures within the basal
ganglia. These were the level of synergies and the level of spatial
field. Bernstein postulated an important role for the basal ganglia in
forming task-specific groups of motor elements and ensuring the
dynamic stability of natural movement (done at the level of
synergies) and in the production of movements to spatial targets
(controlled at the level of spatial field).
The basal ganglia consist of several large subcortical nuclei: the
putamen and caudate nucleus—which are addressed together as
the striatum—the globus pallidus, the substantia nigra, and the
subthalamic nucleus. These nuclei neither receive direct inputs from
nor send direct outputs to the α-motoneurons in the spinal cord. The
importance of the basal ganglia for control of voluntary movements
has been assumed based mostly on clinical observations. Basal
ganglia disorders bring about quite different clinical pictures ranging
from excessive involuntary movements to slowness and a lack of
movement. Because of these clinical observations, it was supposed
that the basal ganglia were major components of an extrapyramidal
system that was thought to participate in the control of movements in
parallel with and largely independent of the pyramidal (corticospinal)
system. However, we now know that the pyramidal and
extrapyramidal systems are not independent but dynamically
cooperate for movement control. Basal ganglia also play an
important role in motor learning and habit formation. The function of
the basal ganglia is not limited to controlling movements; these
structures have a role in cognitive and emotional functions as well.

9.1 Anatomy of the Basal

Ganglia
Three of the nuclei of the basal ganglia lie deep in the cerebrum,
laterally to the thalamus (figure 9.1). Two nuclei, the caudate
nucleus and the putamen, form the striatum. The striatum is the
major input structure of the basal ganglia, and it receives input from
the cerebral cortex, thalamus, and brainstem. The two nuclei in the
striatum are separated from each other by the internal capsule. The
striatum is organized into modules called striosomes and matrix.
Striosomes are islands of relatively densely packed cells that are
parts of a larger, less densely packed compartment called matrix.
These areas of the striatum differ in projections they receive from the
cortex. The matrix receives inputs from many areas of the cortex,
while striosomes receive inputs mostly from the prefrontal cortical
areas.

Figure 9.1 A sagittal and frontal view of the brain showing the basal ganglia and
the two input nuclei of the basal ganglia, the caudate nucleus and the putamen.
These two structures constitute the striatum. The striatum is the major input
structure of the basal ganglia. One of the major output structures of the basal
ganglia is the globus pallidus.

The phylogenetically oldest nucleus in the basal ganglia is the

globus pallidus, which is also known as the paleostriatum. The
globus pallidus consists of two discrete nuclei, internal (GPi) and
external (GPe) segments. The internal segment is a major output
structure, and the external segment is part of the intrinsic basal
ganglia circuitry.
The other two nuclei of the basal ganglia, the substantia nigra
and the subthalamic nucleus, are located in the midbrain. The
substantia nigra is the largest nucleus of the human midbrain. It is
divided anatomically into two parts; its dorsal region is called the
pars compacta and its ventral region is called the pars reticulata. It
contains dopaminergic cells that project to the striatum and other
basal ganglia nuclei. The subthalamic nucleus is situated between
the thalamus and the substantia nigra; it receives projections from
the GPe, cerebral cortex, thalamus, and brainstem, and it sends
outputs to the globus pallidus and substantia nigra pars reticulata.
These nuclei have been historically united with the group of the
basal ganglia. There are, however, other nuclei that lie ventrally to
the striatum and the globus pallidus that may also be considered to
be within the basal ganglia group. The nuclei ventral to the striatum,
the nucleus accumbens and olfactory tubercle, are sometimes called
the ventral striatum. They receive inputs from the limbic and olfactory
area of the cortex and are similar to the striatum in receiving
dopaminergic inputs from the ventral tegmental area. The ventral
pallidum differs from the GP in that it receives direct inputs from the
amygdala and projects to the limbic area of the cortex. These nuclei
are assumed to play a role within the limbic structures of the brain,
possibly contributing to motivation and emotions (Pessoa et al.
2019).

9.2 Inputs and Outputs of the

Basal Ganglia
The cerebral cortex is a major source of input to the basal ganglia,
with numerous projections coming in from all of the cortical lobes in
the brain. Cortical input into the caudate nucleus and putamen
originates in different areas of the cerebral cortex. The caudate
nucleus receives input from the parietal and frontal lobes, which are
involved in integrating multimodal sensory information, as well as
from areas in the frontal lobe involved in eye movements (figure 9.2).
In contrast, the putamen receives input from primary and secondary
somatosensory and visual areas, and this input topographically
maps onto different regions of the putamen. In addition, the putamen
receives input from the premotor and motor cortices. These
pathways into the caudate and putamen remain parallel and
segregated even in the output structures of the basal ganglia.
Cortical neurons make projections onto the striatal neurons; these
projections are glutamatergic—that is, they are excitatory and use
glutamate as the neurotransmitter. Most cortical projections to the
striatum are to medium spiny neurons. The input from the cortex to
the spiny neurons is convergent as the axons from cortical neurons
terminate on fewer spiny neurons. The large dendritic trees of the
spiny neurons receive input from different cortical, brainstem, and
thalamic structures, allowing for the integration of information from
multiple neural regions. The projections from the motor and
somatosensory areas into the putamen follow a somatotopic
mapping (i.e., inputs from the arm, face, and leg areas are kept
separate throughout the basal ganglia). For example, neurons
responding to the leg areas are concentrated in the dorsolateral
putamen, and those from the arm are located in more ventromedial
areas.

Figure 9.2 (left) The main cortical structures that project to the basal ganglia and
the input nuclei of the basal ganglia, the caudate and the putamen. (right) The
internal projections within the basal ganglia.

Medium spiny neurons also receive information from local

interneurons in the striatum. The activity of these neurons is also
modulated by dopaminergic neurons in the substantia nigra pars
compacta. A schematic drawing of a spiny neuron is shown in figure
9.3. Note the typical spines on the dendrite of such a neuron and the
convergence of projections from the cerebral cortex, substantia
nigra, and other spiny neurons. These inputs use different
neurotransmitters: glutamate, dopamine, acetylcholine, and GABA.
 The medium spiny neurons from the caudate nucleus and
putamen make inhibitory projections onto both segments of the
globus pallidus and the substantia nigra pars reticulata using GABA
as the neurotransmitter; these projections are both inhibitory and
convergent. There are fewer neurons in the pallidum and substantia
nigra than in the striatum. As mentioned earlier, an interesting
feature of these projections is that the motor pathways remain
largely segregated through the basal ganglia networks and are
believed to serve different motor functions, such as motor planning,
motor execution, and motor sequencing. The efferent neurons from
the motor areas of the globus pallidus and substantia nigra pars
reticulata project to motor-related areas of the ventral lateral and
ventral anterior nucleus of the thalamus and the superior colliculus,
respectively. The thalamic nuclei project directly to the motor areas
of the cerebral cortex in the frontal lobe, completing vast parallel
motor circuits that both originate and terminate in the cerebral cortex
and go through the basal ganglia and the thalamus (figure 9.4). In
addition to these pathways that go through the thalamus, there are
direct projections from the substantia nigra pars reticulata that
directly innervate neurons in the superior colliculus that project to
neurons that innervate extraocular muscles and muscles controlling
head movements.
Figure 9.3 Medium spiny neurons of the striatum project to the globus pallidus,
ventral pallidum, and substantia nigra. They receive excitatory inputs from the
cortex and thalamus. In addition, dopamine afferents from the substantia nigra
modify the responsiveness of the medium spiny neurons to excitatory input. They
also receive inhibitory GABA inputs from other spiny neurons as well as excitatory
glutamatergic input.
Figure 9.4 The efferent projections from the basal ganglia (globus pallidus)
project to the motor areas in the frontal cortex through the thalamus. The output
from the substantia nigra pars reticulata projects to the superior colliculus. The
superior colliculus is a midbrain area that is involved in the control of eye
movements.

An important aspect of basal ganglia efferent output from the

globus pallidus and substantia nigra pars reticulata is that it is
inhibitory. The efferent neurons from these structures are
spontaneously active and inhibit neurons in the thalamus and the
superior colliculus. Since the medium spiny neurons are also
inhibitory (and GABAergic), when transient excitatory input from the
cortex reaches the striatum, it disinhibits the inhibitory neurons of the
globus pallidus and substantia nigra pars reticulata and creates a
burst of activity in the cortical neurons that then initiate movements.
The medium spiny neurons in the putamen and caudate are active
seconds before movement is initiated. This suggests that they are
involved in the decision to move.

PROBLEM 9.1
 Would putamen neurons fire as a person is reaching for an ice
cream cone, or in anticipation before the reaching movement is
initiated? Why?

9.3 Direct and Indirect

Pathways Within the Basal
Ganglia
There are two main pathways within the cortico-basal-thalamic-
cortical loop. The direct pathway consists of projections of the
medium spiny neurons of the striatum to the internal segment of the
globus pallidus (GPi) through the basal ganglia (figure 9.5). This
pathway facilitates the initiation of voluntary movements by releasing
cortical neurons from the tonic inhibition of the thalamic cells. The
indirect pathway (figure 9.5) involves projections of another
population of medium spiny neurons of the striatum to the external
segment of the globus pallidus (GPe). The GPe sends projections to
the GPi as well as to the subthalamic nucleus. The subthalamic
nucleus also receives direct projections from the cerebral cortex
through the hyperdirect pathway and makes excitatory
(glutamatergic) projections to the GPi and the substantia nigra pars
reticulata. Though the indirect pathway projects to the output nuclei
that project to the thalamus, it actually ends up opposing the activity
of the direct pathway. This is important for action selection as
explained below.
When cortical neurons activate the indirect pathway, the medium
spiny neurons fire and inhibit the tonically active GABAergic neurons
of the GPe. The subthalamic cells are glutamatergic and excitatory
and project to the GABAergic GPi and substantia nigra pars
reticulata neurons. When the subthalamic cells receive excitatory
input from the cerebral cortex, they end up increasing the inhibitory
output of the basal ganglia. This is obviously different from the output
of the direct pathway that disinhibits the cortical neurons from the
tonic inhibition of the thalamic cells and consequently facilitates the
initiation of voluntary movements. Thus, the direct pathway is
activated in anticipation of upcoming movements, and the indirect
pathway simultaneously suppresses competing motor plans from
being executed. Let’s try to understand this with an example.
Imagine you are driving a car and as you approach a traffic light,
you see the light turn from green to yellow. Now you have multiple
options; you could speed up to pass the light before it turns red, slow
down to bring the car to a stop, or not do anything because you are
certain that you have enough time to pass the light without hustling.
Let’s say you choose to slow down; the direct pathway would
facilitate moving the legs from the accelerator to the brake to stop
the car, while the indirect pathway would suppress the other two
motor alternatives. This has been studied in the laboratory using the
go/no-go paradigm (reviewed in Aron et al. 2007). On each trial,
participants are presented with one of two possible Go signals: press
left arrow if target moves to the left and right arrow if it moves to the
right. On a minority of the trials (typically between 25% and 33%), a
Stop signal is presented after the Go signals. Participants are
instructed to respond as fast as possible on the Go trials and to do
their best to stop their motor response when the Stop signals occur.
The timing of the Stop signal is varied systematically, and the Stop
signal reaction time is calculated. This paradigm has also been
combined with functional imaging to obtain the neural correlates. The
imaging studies have revealed that the nervous system stops the Go
movement via a frontal cortex basal-ganglia network.
Figure 9.5 The direct and indirect pathways of the basal ganglia. The direct
pathway consists of projections of the medium spiny neurons of the striatum to the
internal segment of the globus pallidus and facilitates the initiation of voluntary
movements by releasing the cortical neurons from the tonic inhibition of the
thalamic cells. The indirect pathway opposes the activity of the direct pathway.
There is also a hyperdirect pathway that conveys information from the motor,
associative, and limbic brain areas on to the subthalamic nucleus (STN),
bypassing the indirect inhibitor circuit and leading to excited globus pallidus/pars
reticulata of the substantia nigra (GPi/SNr) activity.

PROBLEM 9.2
If a baseball batter decides not to swing for a curveball, how
would the direct and indirect pathway contribute to that?

9.4 Dopamine Modulation of

Basal Ganglia Circuits
Within the basal ganglia, the caudate nucleus and the putamen
receive dopaminergic inputs, mostly from the substantia nigra pars
compacta. The medium spiny neurons of the striatum project to the
substantia nigra pars compacta and then receive dopaminergic input
back from the substantia nigra pars compacta. The projections from
the pars compacta provide excitatory input to the spiny neurons that
project to the GPi (direct pathway) and inhibitory input to the GPe
(indirect pathway). This is made possible by different types of
dopamine receptors, D1 (direct pathway) and D2 (indirect pathway),
on the medium spiny neurons (figure 9.6).
It has been hypothesized that dopaminergic neurons are involved
in “reward-related” changes in motor behavior, or reinforcement
learning. Here, “reward” is used in an abstract sense. Whereas
animals can be rewarded with food for performing a task
successfully, rewards almost certainly have a different meaning for
humans. For example, when you are learning to drive, you might
pick up good or bad habits based on outcomes of your decisions and
the praise received from your instructor. The praise itself might serve
as a reward.
Apparently, the firing rate of dopaminergic cells signals the
difference between the expected reward and the received reward
during a motor action. The release of dopamine strengthens the
synapses between neurons that fire whenever an action with a high
reward is performed. So, if you bring your car to a smooth stop and
instantly receive praise from your instructor, that behavior might get
reinforced and eventually become a habit. It is important to note that
dopaminergic neurons in the direct pathway only fire during the
learning process and stop firing when the action becomes a habit.
Similarly, if an action is performed that results in a punishment (e.g.,
you are rebuked by the instructor for making a hasty and unsafe
driving decision), that action might be inhibited, possibly with
contributions from the indirect pathway.
As stated earlier, movements become habitual or stereotyped
when they lose their dependence on explicit or implicit rewards. The
basal ganglia may facilitate habit formation by strengthening
stimulus–response associations. This relationship is beautifully
described by Ashby and colleagues (Ashby et al. 2010) in a
theoretical model of stimulus–response associations facilitated by
the basal ganglia (figure 9.7). Imagine that two corticostriatal inputs
code sensory events (S1, S2) and synapse on striatal output
neurons that lead to two distinct motor actions (R1, R2). Let’s
imagine that R1 leads to the delivery of a reward and causes a
transient or phasic increase in midbrain dopamine activity and a
release of dopamine in the striatum and strengthening of one
synapse, say S2 to R1. This synapse continues to strengthen with
repeated reinforcement until an asymptotic level of strength is
reached. In contrast, synapses that are activated in the absence of
the phasic dopamine-reward signal (S1 to R1) undergo long-term
depression, reducing the possibility that S1 will depolarize cells to
produce neural activation to produce R1. Thus, dopamine-mediated
long-term potentiation strengthens the synapses (S2 to R1) and
increases the chance that the motor action (R1) that is rewarded
occurs in response to the same sensory stimulus (S2) in the future.
Figure 9.6 The substantia nigra is the source of important dopaminergic input to
the basal ganglia; loss of neurons in this area is the cause of Parkinson’s disease.
In the direct pathway (shaded) of the basal ganglia, when the striatum is excited, it
inhibits the tonically active internal segment of the globus pallidus. That in turn
disinhibits the VA/VL complex of the thalamus and excites the motor cortex. In the
indirect pathway, the excited striatum inhibits the tonically active neurons of the
external segment of the globus pallidus (GPe). The GPe projects to the
subthalamic nucleus, which also receives excitatory input from the cortex. The
excitatory projections from the subthalamic nucleus to the internal segment of the
globus pallidus (GPi) counter the disinhibitory actions of the direct pathway.
Adapted from Purves et al. (2017).
Figure 9.7 Corticostriatal inputs S1 and S2 encode different sensory events that
eventually cause different behavioral responses, R1 and R2. If R1 produces a
reward, it causes an increase in dopamine activity in the basal ganglia. This will
strengthen the active synapses between S2 and R1 and increase the probability
that the reward-producing behavior R1 will occur in response to the sensory event
S2.

PROBLEM 9.3
A person is learning a skill that requires the initiation of rapid eye
movement (a saccade) at the onset of a visual stimulus. If the
basal ganglia played a role in the acquisition of this skill, would
you expect the reaction time to increase or decrease with
practice?

9.5 Motor Circuits Involving

the Basal Ganglia
Figure 9.8 illustrates major pathways originating from different areas
of the cerebral cortex, passing through the basal ganglia and the
thalamus, and terminating back in the cortex. Inputs to the basal
ganglia may originate in different cortical areas such as the motor
cortex, the premotor area, the supplementary motor area, the
somatosensory cortex, and the superior parietal cortex (which lies
dorsally to the somatosensory areas). The projections from the
somatosensory cortical areas to the putamen are organized
somatotopically (i.e., one more primitive drawing of a human figure
may be found on the surface of the putamen). This somatotopy is
relatively preserved in the projections from the putamen to both
internal and external segments of the globus pallidus. Both direct
and indirect loops participate in the motor circuit of the basal ganglia;
they are both mediated by thalamic neurons. Most of the thalamic
projections to the cortex are directed at the premotor cortex and
supplementary motor area. These areas have connections with each
other (mostly inhibitory) and with the motor cortex, and all of them
project directly to brainstem motor centers and to the spinal cord.

Figure 9.8 The motor and oculomotor loops of the basal ganglia.

Most somatosensory and motor areas of the cortex project

exclusively to matrix, while many striosomes receive inputs from
limbic structures. Remember that limbic structures include the
hypothalamus, the fornix, the hippocampus, the amygdaloid nucleus,
and the cingulate gyrus of the cerebral cortex. They are believed to
be involved in such elements of behavior as attention, motivation,
and emotion. Striosomes, which receive projections from limbic
structures, in turn, project on dopaminergic cells in the midbrain.
Midbrain neurons then project back upon the matrix and striosomes.
The latter projections may modulate the effectiveness of cortical
inputs to the same neurons. So this circle provides means for the
brain to change the effectiveness of transmission in the basal
ganglia motor loop based on attentional and emotional factors.

9.6 Activity of the Basal

Ganglia During Movements
The role of the motor network that involves the basal ganglia has
been examined in animal studies and by clinical observations of
movements in patients with basal ganglia disorders. These studies
have shown that this network is involved in action selection,
movement preparation, movement sequencing, and reinforcement
learning.
The medium spiny neurons increase their firing rate before an
impending movement. In particular, neurons in the putamen fire
before limb movements are initiated and caudate neurons before eye
movements are initiated. This firing can sometimes occur seconds
before the movement is initiated, suggesting that these neurons are
involved in action selection and movement preparation. The
involvement of the basal ganglia in action selection also implicates it
in reinforcement-based motor learning that is driven by rewards and
punishments. As stated earlier, this is accomplished through
simultaneous activation of the direct and indirect pathways.
Animal studies have shown that, at rest, there is substantial
activity of neurons in the globus pallidus and in the pars reticulata of
the substantia nigra (at frequencies of about 50 to 100 Hz), while
neurons in other structures of the basal ganglia are rather quiet. For
example, the baseline discharge of neurons in the striatum is
typically under 1 Hz. Neurons in the GPi fire at a relatively constant
high rate, while neurons in the GPe demonstrate high-frequency
burst-like activity (DeLong et al. 1983).
 PROBLEM 9.4
Suggest a functional reason that the neurons in the globus
pallidus have a high rate of activity at rest.

Many neurons in the basal ganglia show phasic modulation of

their firing frequency during voluntary movements of the contralateral
side of the body. Some of these neurons show correlation of their
firing frequency with such movement parameters as velocity, force,
and amplitude (Crutcher and DeLong 1984; Middleton and Strick
2000). Discharge patterns of other neurons, however, are both
movement and context dependent. In particular, movement is
associated with an increase in the frequency of action potentials
generated by neurons in different nuclei of the basal ganglia. This
increase is seen somatotopically—that is, it is seen in separate
groups of neurons during arm, leg, and facial movements (DeLong
and Georgopoulos 1979). A change in the discharge rate may be
seen about 20 ms prior to movement. Though most striatal neurons
are active before movement initiation, some also show changes in
their activity in association with movement termination rather than
with movement initiation.
Changes in the firing patterns of cells in other structures of the
basal ganglia typically occur rather late. For example, in reaction
time tasks, most of the cells show changes in their firing after
movement initiation, although there have been reports of relations
between changes in the basal ganglia activity and movement
initiation (Hauber 1998). Note that many neurons of the motor cortex
change their firing prior to the movement initiation. So it is possible to
conclude that neurons of the basal ganglia do not initiate movements
under these experimental conditions and that they are related more
to the control of movements that are already underway.
Another important characteristic of the basal ganglia neurons is
that their firing correlates more closely with movement direction than
with forces that are needed to perform a movement (Crutcher and
DeLong 1984).
9.7 Movement Disorders
Associated With the Basal
Ganglia
The two most common movement disorders associated with basal
ganglia dysfunction are Parkinson’s disease and Huntington’s
disease. Parkinson’s disease is a hypokinetic (decrease in
voluntary movement) disorder that was first described by James
Parkinson in 1817. The main symptoms of the disease are tremor
during rest that lessens during movement and sleep, slowness of
movement (bradykinesia), muscular stiffness and rigidity, speech
difficulties, and lack of facial expressions. When walking, patients
with Parkinson’s disease take shorter steps, have a stooped posture,
and have a reduced, often asymmetrical arm swing. The loss of
dopaminergic neurons in the substantia nigra pars compacta
contributes to this disease. Huntington’s disease was first reported
by George Huntington in 1872. This disease is a hyperkinetic
(increase in undesired movements) disorder and is characterized by
an increase in involuntary movements, abnormal gait, coordination
deficits, and rapid and jerky movements. This disease is caused by
gradual atrophy of the striatum (caudate and putamen). When the
medium spiny neurons do not provide sufficient inhibitory input to the
GPe, the cells in the GPe become very active and reduce the
excitatory input from the subthalamic nucleus to the GPi. This
causes a decrease in the inhibitory output of the basal ganglia and
makes the cortical motor neurons respond to small inputs, resulting
in undesirable movements. These disorders will be covered in more
detail in chapter 37.

9.8 Other Functions of the

Basal Ganglia
The function of the basal ganglia is not limited to motor control. Two
additional circuits play roles in modulating nonmotor aspects of
behavior. These circuits originate in different areas of the cortex,
pass through the basal ganglia and the thalamus, and terminate in
areas outside the motor and premotor cortices. The first circuit
serves cognitive functions and originates in the dorsolateral
prefrontal cortex and involves the caudate nucleus. A number of
recent studies have shown that this loop is related to such functions
and processes as short-term memory, attention, and cognition
(reviewed in Pessoa et al. 2019). The second circuit regulates
emotional and motivational behavior and is also involved in mood
regulation. This loop originates in the orbitomedial prefrontal cortex,
amygdala, and hippocampus, and passes through the ventral sector
of the striatum. Disorders in these circuits are associated with
psychiatric disorders, such as hallucinations, delusions, obsessive-
compulsive disorder, depression, and anxiety.

CHAPTER 9 IN A NUTSHELL
The basal ganglia are a group of
subcortical nuclei that are involved
in motor control, motor learning,
executive functions, and regulation of
emotions. The basal ganglia have input
nuclei, output nuclei, and intrinsic
nuclei. Input nuclei structures are
the caudate nucleus and the putamen,
and they receive incoming information
from cortical and other sources. The
output nuclei consist of the internal
segment of the globus pallidus (GPi)
and the substantia nigra pars
reticulata (SNr). These structures
send basal ganglia information to the
thalamus. The intrinsic nuclei consist
of the external segment of the globus
pallidus (GPe), the subthalamic
nucleus, and the substantia nigra pars
compacta (SNc), and they are located
between the input and output nuclei.
The basal ganglia need dopamine as an
input for proper functioning. Dopamine
is critical for reinforcement-based
motor learning, and dopamine
dysfunction is associated with several
movement disorders, such as
Parkinson’s disease and Huntington’s
disease.
 Chapter 10

Cerebellum

KEY TERMS AND TOPICS

cerebrocerebellum
spinocerebellum
vestibulocerebellum
cerebellar peduncles
cerebellar nuclei
Purkinje cells
inferior olive
vestibulo-ocular reflex
cerebellar ataxia
error-based motor learning

The cerebellum is a large structure in the caudal part of the brain,

which contains more than half the neurons in the brain. The
symptoms of cerebellar damage in humans as well as in
experimental animal models provide compelling evidence in support
of its involvement in motor control. The cerebellum receives afferent
input from the cerebral cortex, spinal cord, and vestibular nuclei. In
fact, the cerebellum receives many more input fibers (afferents) than
it has output fibers (efferents); the ratio is about 40:1. As with the
basal ganglia, the output neurons of the cerebellum do not directly
project to spinal α-motoneurons. A substantial number of efferent
projections from the cerebellar nuclei, which mediate all the output
projections of the cerebellum, are to cortical neurons, to the red
nucleus, and to vestibular nuclei. It is widely believed that one of the
main movement-related functions of the cerebellum is to provide
closed-loop feedback through error detection between the intended
movement and the actual movement. This can occur either while a
movement is underway or across different repetitions of a movement
as a form of motor learning. This form of motor learning is often
referred to as error-based learning. Damage to the cerebellum
causes major problems in postural control, movement coordination
across joints and muscles, and motor learning and adaptation. The
cerebellum has also been implicated in cognitive functions and has
been linked to developmental problems across a variety of
conditions including autism, Asperger’s syndrome, and Down
syndrome (see chapters 33 and 38).

10.1 Overall Structure of the

Cerebellum
The human cerebellum has two hemispheres separated by a midline
ridge. The cerebellum is broadly divided into three parts that receive
inputs from different areas of the nervous system. The largest and
the most lateral area is called the cerebrocerebellum, and it receives
extensive input from the cerebral cortex. Medial to the
cerebrocerebellum is the spinocerebellum, which receives afferent
input from the spinal cord, mostly from distal limb muscles (figure
10.1). The medial ridge that separates the two hemispheres is called
the vermis, evolutionarily the oldest part of the cerebellum, and it
primarily receives input from proximal limb muscles. The third and
final major area of the cerebellum is the vestibulocerebellum, and it
is in the most inferior part of the cerebellum. It includes two areas
called the nodulus and flocculus. The vestibulocerebellum is
primarily concerned with posture control and maintenance of
equilibrium.
The cerebellum connects to the rest of the central nervous system
through three pathways called the cerebellar peduncles (figure
10.1). These are the superior cerebellar peduncle, the middle
cerebellar peduncle and the inferior cerebellar peduncle. These
peduncles are the sources of all afferent input into and efferent
output from the cerebellum. The deep cerebellar nuclei (figure 10.2)
are the source of all efferent output of the cerebellum and form part
of the cerebellar system of neural circuits connected to the red
nucleus, vestibular nuclei, sensorimotor regions (primary motor and
premotor areas), associative cortices (parietal cortex), and limbic
system. The deep cerebellar nuclei send their output through the
superior cerebellar peduncle, which is primarily an efferent pathway.
The middle cerebellar peduncle is primarily an afferent pathway that
takes neural information into the cerebellum. The inferior cerebellar
peduncle contains both afferent and efferent pathways. Efferent
neurons in the inferior cerebellar peduncle project to the vestibular
nuclei and reticular formation (see chapter 11). Afferent projections in
this peduncle arise from the vestibular nuclei and the spinal cord.

PROBLEM 10.1
Which part of the cerebellum is involved in posture stabilization?
Figure 10.1 (a) The flattened cerebellar surface with the three major
subdivisions. (b) The cerebellar peduncles in the flattened view. (c) The sagittal
view of the cerebellum.
Figure 10.2 The frontal and parietal areas project to the cerebellum through the
ipsilateral pontine nuclei. Axons from the pontine nuclei decussate and enter the
contralateral cerebellum. These fibers are the largest source of input to the
cerebellum.

10.2 Inputs and Outputs of the

Cerebellum
Two major excitatory afferent systems act as inputs into the
cerebellum. These are mossy fibers and climbing fibers. The mossy
fibers originate from a variety of brainstem nuclei and from neurons
in the spinal cord whose axons form the spinocerebellar tracts. The
spinocerebellar tracts primarily convey somatosensory information.
Different portions of the cerebellum receive mossy fibers from
different sources. The medial zone (closest to the vermis) receives
information mostly of vestibular, somatosensory, visual, and auditory
modalities. The intermediate zone receives proprioceptive and
somatosensory information from the spinal cord, as well as
information from the motor cortex mediated by nuclei in the pons.
The lateral zone receives information mediated by pontine nuclei
from different areas of the cerebral cortex, including the motor
cortex.
Mossy fibers make excitatory synapses on the granule cells. The
axons of the granule cells ascend into the molecular layer, where
each axon splits into two and joins the system of parallel fibers. Each
granule cell receives inputs from many mossy fibers (this is an
example of convergence of neural information), while each mossy
fiber innervates a few hundred granule cells (this is an example of
divergence). Each Purkinje cell receives inputs from numerous
parallel fibers (up to 200,000).
The system of climbing fibers is organized very differently. These
fibers originate in the medulla, in the inferior olivary nucleus (also
known as inferior olives). Their axons enter the cerebellar cortex and
wrap around the soma and proximal portions of the dendrites of
Purkinje cells. Their synapses are excitatory and very strong. Each
Purkinje cell receives synaptic inputs from only one climbing fiber
that forms more than a hundred synapses on the soma and the
dendrites of the Purkinje cell it innervates. One climbing fiber may
innervate a few Purkinje cells. A single action potential in a climbing
fiber always induces a complex action potential in Purkinje cells it
innervates (i.e., its action is obligatory).
The largest source of mossy fiber input to the cerebellum is from
the frontal, prefrontal, and parietal areas of the cerebral cortex.
These cortical axons primarily project to the cerebrocerebellum
through the ipsilateral pontine nuclei (figure 10.2). The axons from
the pontine nuclei then decussate and enter the contralateral
cerebellum via the middle cerebellar peduncle. The decussation
implies that the input from one hemisphere of the cerebral cortex is
received and processed by the contralateral cerebellar hemisphere.
The output of the cerebrocerebellum is projected through the dentate
nucleus via the superior cerebellar peduncle. These output axons
exit the cerebellum, decussate, and enter the contralateral thalamus
before projecting to the motor, premotor, and prefrontal areas of the
cerebral cortex. These circuits through the cerebrocerebellum are
thought to be primarily involved in the planning and execution of
complex movements. The projections to the prefrontal areas may
also play a role in cognitive aspects of motor and nonmotor
behaviors.
The spinocerebellum consists of the vermis and intermediate
zones of the cerebellar cortex. The vermis receives somatosensory
(tactile and proprioceptive) input from the spinal cord and also
receives visual, vestibular, and auditory input. The somatosensory
input is topographically mapped onto the spinocerebellum and
provides representations of the body in the cerebellum. The spinal
and vestibular input into the spinocerebellum enters through the
inferior cerebellar peduncle and remains ipsilateral. This means that
the cerebellum contains topographic maps of the same side of the
body. Output projections from the vermis go through the fastigial
nuclei and the inferior cerebellar peduncle and reach cortical and
brainstem (reticular formation) areas involved in the control of
proximal muscles of the body and limbs, in particular for posture and
locomotion. The vermis is also involved in the control of eye
movements. In contrast to the vermis, the intermediate zones of the
spinocerebellum also receive somatosensory information from the
limbs and send output axons through the interposed nucleus and
superior cerebellar peduncle to the thalamus. Projections from the
thalamus to the frontal lobe provide input to neurons that form the
corticospinal tract (see chapter 8). This tract affects spinal circuits
participating in the control of distal muscles that are involved in fine
motor control.
The vestibulocerebellum is the phylogenetically oldest part of the
cerebellum and consists of the flocculonodular lobe. It receives
projections from the vestibular and visual systems through the
inferior cerebellar peduncle. Projections from the
vestibulocerebellum also go through the inferior cerebellar peduncle
to the vestibular complex in the brainstem. These nuclei control eye
movements as well as neck and head movements.

PROBLEM 10.2
 If there is focal damage to the central part of the cerebellar cortex,
which types of movements will be affected?

10.3 Pathways Within the

Cerebellum
There are three distinct layers in the cerebellar cortex with different
kinds of neurons that serve very different functions. The deepest
layer, the granule cell layer (also called granular layer), is the input
layer of the cerebellum (figure 10.3). The axons from the pontine
nuclei and from the brainstem and spinal cord that form the input to
the cerebellum are part of the mossy fiber system, and they
terminate in this layer. Within the granule cell layer, mossy fibers
make excitatory synapses on the granule cells that give rise to axons
called parallel fibers.
The granule cell layer contains structures called glomeruli, where
cells from the granular layer make synaptic contacts with the bulbous
expansions of afferent mossy fibers. A single glomerulum consists of
an incoming mossy fiber, clusters of small dendrites (called rosettes)
from a few dozen granule cells, and the axons of Golgi cells. A single
mossy fiber may innervate many glomeruli.
Climbing fibers from the contralateral inferior olives project on
Purkinje cells, which are the sole output neurons of the cerebellum.
The activity of Purkinje cells is modulated by two types of GABAergic
interneurons called basket cells and stellate cells. Purkinje cells have
elaborate dendritic trees (see figure 10.3), and these cells
themselves are GABAergic and make inhibitory projections to the
deep cerebellar nuclei and the vestibular nuclei in the brainstem.
Though the mossy and climbing fibers provide afferent input to the
cerebellum, their collaterals also provide excitatory input to the deep
cerebellar nuclei. Thus, the deep cerebellar nuclei end up receiving
both excitatory and inhibitory input.
 Parallel fibers—which receive input from the granule cells—
ascend to the outermost layer of the cerebellum, called the
molecular layer, and make excitatory synapses on the Purkinje cells.
Purkinje cells are GABAergic inhibitory projection neurons whose
cell bodies are in the Purkinje cell layer. The molecular layer is the
processing layer of the cerebellum. Parallel fibers also provide input
to the stellate cells. Thus, projections from the climbing fibers
eventually make synapses on the Purkinje cells through this
pathway. This layer also contains the dendrites of Golgi cells, which
are inhibitory neurons and have their cell bodies in the granular cell
layer.
Figure 10.3 The cerebellar cortex is organized into three layers: the molecular
layer, the Purkinje cell (PC) layer, and the granular layer (GL). Purkinje cells, Golgi
cells, stellate cells, and basket cells are inhibitory neurons. Granule cells are
excitatory. Cerebellar afferents are the mossy and climbing fibers, and both are
excitatory. The cerebellar cortex receives inputs from mossy fibers originating in
various brainstem and spinal cord nuclei and from climbing fibers originating from
the inferior olive. Climbing fibers contact Purkinje cells and the deep cerebellar
nuclei. The only output of the cerebellar cortex is provided by the Purkinje cells,
which project to the deep cerebellar nuclei.
© 2021 Consalez, Goldowitz, Casoni, and Hawkes. Redistributed under the terms of the
Creative Commons Attribution 4.0 International License (http://creativecommons.org
/licenses/by/4.0/).

The GABAergic Purkinje cells make inhibitory projections to the

deep cerebellar nuclei and are the only output cells of the cerebellar
cortex. This implies that the output of the cerebellum to its nuclei is
completely inhibitory.
 The cerebellum’s organization reflects a series of modules with
excitatory and inhibitory cells. In each module, mossy fibers provide
target cells in the deep cerebellar nuclei with inhibitory input
(indirectly through the cerebellum) and excitatory input (directly
through collaterals). This creates both excitatory and inhibitory local
circuits within the cerebellum (figure 10.4). With this form of input,
each of these modules subserves two important functions: (a)
continuous control of ongoing movement, and (b) long-term changes
in control to facilitate motor learning.

Figure 10.4 Mossy and climbing fibers provide strong excitatory input to the
Purkinje cells. Purkinje cells are GABAergic and project to the deep cerebellar
nuclei; they are the only output cells of the cerebellar cortex. The output of the
Purkinje cells is completely inhibitory.
 PROBLEM 10.3
GABAergic Purkinje cells provide the inhibitory output of the
cerebellar cortex. So all the output of the cerebellum is inhibitory.
But the deep cerebellar nuclei receive excitatory input from the
mossy and climbing fibers. What might be the role of such an
arrangement?

10.4 Distinct Cerebellar

Regions Control Discrete Motor
Functions
As stated earlier, the cerebellum is divided into three areas: the
cerebrocerebellum, spinocerebellum, and vestibulocerebellum. The
cerebrocerebellum primarily receives information from the cerebral
cortex. The cerebrocerebellum loop starts with the cortical
projections to the pontine nuclei and proceeds through the middle
cerebellar peduncle to the contralateral cerebrocerebellum, to the
dentate nucleus, and back to the cortex of the large hemispheres via
the thalamus.
The cerebrocerebellum has been implicated in motor planning of
hand movements. It has also been implicated in the temporal control
of movements, movement initiation, and the timing of different
components of movements. These findings have led to the
hypothesis that the cerebrocerebellum may be involved in the timing
of motor and serial events (Ivry and Keele, 1989; Diedrichsen et al.
2007). Damage to the cerebrocerebellum affects patients’ ability to
judge the passage of time and consequently impairs their ability to
judge if one time interval was longer or shorter than another.
The spinocerebellum consists of the vermis and intermediate
parts of the cerebellar hemisphere. The spinocerebellar tracts
provide extensive somatosensory input from the spinal cord about
limb position, touch, and pressure. The dorsal spinocerebellar tract
conveys proprioceptive information from muscle and joint receptors
to the spinocerebellum. This feedback provides the cerebellum with
a continuous and real-time estimate of the sensory consequences of
movement, irrespective of whether the movement is generated
actively or passively. In contrast, the ventral spinocerebellar tract
only provides sensory feedback during active movements. This has
led to the idea that the cerebellum processes active and passive
movements differently and that it may be involved in comparing the
sensory consequences of the movement with the actual movement
(Person 2019).
Purkinje neurons in the spinocerebellum project somatotopically to
different descending motor pathways. First, the vermis sends axons
to the fastigial nucleus, and the fastigial nucleus projects bilaterally
to the brainstem, the reticular formation, and the vestibular nuclei.
This way the vermis provides input to the medial descending motor
pathways (see chapter 11) that primarily control postural (neck and
trunk) muscles and proximal limb muscles. Purkinje neurons in the
intermediate parts of the cerebellar hemispheres project to the
interposed nucleus. These neurons exit through the superior
cerebellar peduncle and decussate to terminate on the red nucleus
(see section 11.4). Axons from the red nucleus decussate and
descend to the spinal cord, forming the rubrospinal tract. The
remaining axons from the interposed nucleus terminate in the
thalamus. Thalamocortical neurons then project to the primary motor
cortex (M1), where the corticospinal tracts originate. In this way, the
intermediate cerebellum contributes to the control of distal limb
muscles.
The vermis is also involved in regulation of two types of eye
movements, saccades and smooth pursuits. Eye movements will be
described in more detail in chapter 14. Briefly, saccades are rapid
eye movements that shift gaze between targets, and smooth pursuit
eye movements are used to track moving objects. Lesions in the
vermis impair the accuracy of both types of eye movements.
The vestibulocerebellum receives sensory input from the
vestibular system (i.e., the otolith organs and the semicircular canals
that relay information on the head’s movement with respect to
gravity). The otolith organs and semicircular canals project to the
vestibular nuclei in the brainstem that provide input to the
vestibulocerebellum. The output of the vestibulocerebellum
bypasses the deep cerebellar nuclei and directly reaches the
vestibular nuclei. Purkinje neurons in the medial parts of the
vestibulocerebellum project to the lateral vestibular nucleus to
modulate the lateral and medial vestibulospinal tracts. These tracts
control head and neck muscles and limb extensors and primarily
contribute to posture regulation. Purkinje neurons in the lateral parts
of the vestibulocerebellum project to the medial vestibular nucleus
that controls eye movements and coordination between head and
eye movements (i.e., the vestibulo-ocular reflex). The vestibulo-
ocular reflex (see chapter 14) is a gaze-stabilizing reflex: the
vestibular system transforms sensory signals related to head
movements into motor commands to generate compensatory eye
movements in the opposite direction of the head movement,
ensuring the stabilization of gaze direction. This reflex is critical to
stabilizing our gaze on someone’s face while we are nodding our
head in agreement.

PROBLEM 10.4
A person is having difficulty reading road signs while walking.
Which part of the cerebellum might be damaged?

10.5 Cerebellar Control of

Movement
Both Purkinje cells and deep cerebellar nucleus neurons discharge
during voluntary movement. These cells are tonically active at rest
but change their discharge rates as movements occur. Their
discharge rate varies as a function of movement speed and
direction, as well as whether the muscles are relaxed or contracted.
More importantly, the change in the firing frequency with respect to
movement initiation is similar to the initiation of activity in the primary
motor cortex. This suggests that the cerebellum and the primary
motor cortex work together for movement control.
The importance of cerebellar involvement in movement control
has been observed through studies of rapid arm movements.
Ballistic arm movements exhibit triphasic muscle activity where the
agonist muscle fires first to initiate the movement (see chapter 23).
This is followed by a burst in the antagonist muscles to slow down
the movement and bring the arm to rest (figure 10.5). The
contraction of the antagonist starts very soon after the agonist and
before any sensory feedback can be incorporated by the nervous
system. Finally, there is a second burst in the activity of the agonist
to prevent oscillations of the endpoint (e.g., the hand). In patients
with cerebellar deficits, the timing of the onset of the antagonist
activity is delayed until the limb has moved past the target (Flament
and Hore 1986). In other words, the patients have to rely on sensory
feedback to accomplish this movement, whereas healthy persons
seem to perform the movement using feedforward neural control.
Figure 10.5 In healthy individuals, muscle activity during rapid movements (e.g.,
elbow flexion) exhibits a triphasic pattern. The agonist fires first (AG1) at the
beginning of the movement. This is followed by the antagonist around the peak of
the movement velocity (ANT2) to slow down the limb and terminate the movement.
Finally, the agonist fires again (AG3) to prevent oscillations at the end of the
movement. In patients with cerebellar disorders, the antagonist’s activity is delayed
(dashed line).

10.6 Consequences of Cerebellar

Lesions on Movements
One of the defining features of a spinocerebellar lesion is an inability
to make smooth and coordinated limb movements. These lesions
cause movements to become fractured into jerky and inaccurate
components. This is called cerebellar ataxia. Lesions of the vermis in
the spinocerebellum impair the ability of the oculomotor system to
reduce motor errors. Specifically, these lesions can cause dysmetric
eye movements in the form of hypermetric saccades (overshooting
targets) or hypometric saccades (falling short of targets), and the
oculomotor system is unable to adequately correct for those errors
(Barash et al. 1999).
Damage to the vestibulocerebellum affects the ability to keep
postural balance during an upright stance. Since the
vestibulocerebellum is also involved in the control of eye
movements, lesions in this part of the cerebellum affect the ability to
maintain a steady gaze on a fixed location. Consequences of
cerebellar lesions will be covered in more detail in chapter 38.

10.7 Cerebellar Contribution to

Motor Learning
The idea that the cerebellum is involved in motor learning of new
skills was based on observations of animals with experimental
cerebellar injuries of differing severity, up to the complete removal of
the cerebellum. After the animals recovered, they demonstrated an
ability to use their movement repertoire that had been learned prior
to the surgery. However, they were unable to learn new motor skills.
These and other studies led to a prominent theory that inputs from
the climbing fibers provide a teaching signal that causes changes in
synaptic binding within the cerebellum. Changes in synaptic binding
occur through processing of sensory feedback. But how can sensory
processing in the cerebellum contribute to motor learning?
The cerebellum forms what are called closed-loop networks for
motor control. As stated earlier, the cerebellum receives input from
the inferior olive in the medulla oblongata, which in turn receives
information from the cerebral cortex. Collateral outputs from the
cerebellum are also sent to the red nucleus in the midbrain, which
projects back to the inferior olive, providing a mechanism for
cerebellar output to feed back into the cerebellar input, forming an
important closed-loop network (figure 10.6). This network provides a
mechanism for the cerebellum to modulate its own input and is
considered to be involved in learning through error reduction in
motor performance.

Figure 10.6 The red nucleus, inferior olive, and cerebellum form a closed-loop
feedback system that facilitates motor learning through error correction.

The main proposed mechanism is error reduction in motor

performance through long-term depression in the synapses
between parallel fibers and Purkinje cells (Hansel and Linden 2000;
Hirano 2018). Long-term depression reduces the functional strength
of excitatory synapses for hours or even longer. In contrast, long-
term potentiation involves strengthening of synapses that leads to
a long-lasting increase in synaptic transmission between neurons.
When a movement is inaccurately performed, the climbing fibers
respond to those errors and depress the strength of the synapses
between the parallel fibers and the Purkinje cells. If another
erroneous movement occurs, the parallel fiber inputs that carry the
flawed motor signal are further suppressed. Eventually, after enough
repetitions and suppression of the error signal in the climbing fibers,
a more appropriate pattern of activity emerges in the Purkinje cells.

PROBLEM 10.5
If a person sustains an injury to the inferior olive only, which
aspect of motor behavior might be affected the most?

10.8 Cerebellar Interactions

With the Basal Ganglia and
Cortex
The basal ganglia and the cerebellum are distinct systems that
perform different functions but project to the same cortical areas
through distinct thalamic nuclei. In the past, it was thought that the
cerebellum and the basal ganglia only interacted at the level of the
cortex through distributed modules (Houk and Wise 1995). These
modules are hypothetical structures that consist of recurrent loops
connecting the spiny neurons of the basal ganglia, the Purkinje cells
of the cerebellum, and the pyramidal neurons of the cortex. These
modules are hypothesized to function in a parallel cooperative
manner for motor planning and execution.
More recently, direct projections between the cerebellum and the
basal ganglia have been found (reviewed in Bostan and Strick 2018).
The subthalamic nucleus in the basal ganglia makes disynaptic
projections to the cerebellar cortex, and the dentate nucleus in the
cerebellum is the source of projections to the striatum. These
observations have provided anatomical support for the idea of
distributed modules by identifying the neural substrates that serve as
nodes to form an integrated network between the cerebellum, the
basal ganglia, and the cortex for motor control and cognitive
function.

CHAPTER 10 IN A NUTSHELL
The cerebellum is located at the back
of the brain, underlying the occipital
and temporal lobes of the cerebral
cortex. The cerebellum is considered a
motor region because most of its
output goes to the motor areas, and
cerebellar damage causes impairments
in motor control and postural balance.
The main roles of the cerebellum are
coordination of voluntary movements,
maintenance of balance, motor
learning, and cognitive functions. The
cerebellum can be divided into two
parts, the cerebellar cortex and the
cerebellar nuclei. The cerebellar
cortex consists of the
cerebrocerebellum, the
spinocerebellum, and the
vestibulocerebellum. The cerebellar
nuclei are the output structures of
the cerebellum. Purkinje cells make
inhibitory connections onto the
cerebellar nuclei and are the sole
source of output from the cerebellar
cortex. Three cerebellar peduncles
carry the input and output of the
cerebellum. Damage to the cerebellum
causes cerebellar ataxia and deficits
in motor learning.
 Chapter 11

Brainstem and Extrapyramidal

Tracts

KEY TERMS AND TOPICS

reticular formation
red nucleus
superior colliculus
vestibular nuclei
cranial nerves
descending tracts

The neural structures in the brainstem are primarily involved in

postural control, locomotion, and reflex-based gaze control. The
output neurons from four structures—the reticular formation, superior
colliculus, vestibular nuclei, and red nucleus—project to the spinal
cord to innervate interneurons that ultimately lead to changes in the
activation of axial and larger proximal limb muscles that serve the
postural and locomotor functions. Recall that cortical motor areas are
involved in the control of complex and fine motor skills that involve
distal muscles, such as those involved in reaching and grasping
movements, writing, and typing. The brainstem areas work together
with cortical motor areas to support these actions by facilitating
necessary postural control. Within the neural control hierarchy
developed by Bernstein (1947), descending projections from the
brainstem structures ensure background corrections for actions
controlled by cortical areas.

Figure 11.1 The brainstem is made up of the midbrain, pons, and medulla
oblongata. It connects the cerebrum of the brain to the spinal cord and cerebellum.
It is involved in both sensory and motor functions.

11.1 Brainstem Anatomy

The brainstem consists of the midbrain, pons, and medulla
oblongata (figure 11.1), and it serves as a conduit between the brain
and the spinal cord. It acts as a passageway for the ascending tracts
from the spinal cord, the sensory and motor tracts of the head and
neck, and the descending motor tracts from the cortex. There are
also many local neural networks in the brainstem that regulate eye
movements. It is also the source or the destination of most of the 12
cranial nerves that are involved in motor and sensory functions of the
head, neck, and eyes.
The irregular shape of the brainstem reflects the fact that it
houses many different nuclei and white matter tracts. The inferior
olives, which are a part of a closed-loop cerebellar network (see
chapter 10), are a prominent landmark on the lateral brainstem. The
corticospinal tracts go through the medullary pyramids, which form a
prominent bulge on the ventral brainstem. The pons is rostral to the
medulla, and the pontine nuclei of the pons are a source of major
input to the cerebellum. The cerebellum is attached to the pons
through the cerebellar peduncles (see chapter 10), which contain
both afferent and efferent tracts to and from the cerebellum.

11.2 Reticular Formation

The reticular formation is a collection of nuclei of different size
scattered over the brainstem in an ill-defined manner. The phrase
“reticular” itself is derived from the Latin word rete, meaning net,
which alludes to the diffuse structure of this neural region. The
reticular formation extends from the midbrain to the medulla, but in
contrast to other subcortical areas such as the red nucleus, the
reticular formation has no clearly defined cytoarchitectural boundary.
The neurons within the reticular formation facilitate limb motor,
oculomotor, autonomic, sensory, circadian, and mood-related
functions. The axons of neurons in the reticular formation form the
reticulospinal tract, a descending tract that terminates in the medial
spinal cord. From here, local interneurons project to α-motoneurons
that activate axial and proximal limb muscles.
Despite its diverse functions, we will focus only on how the
reticular formation contributes to control of posture and balance. The
neurons of the reticular formation course through the midbrain, pons,
and medulla. The reticular formation at the level of the midbrain is
called the midbrain reticular formation, at the level of the pons is
called the pontine reticular formation, and at the level of the medulla
is called the medullary reticular formation (see figure 11.2). Each of
these areas serves a different function. The mesencephalic
locomotor region is located in the midbrain reticular formation and is
considered to be involved in locomotor rhythm generation
(Roseberry et al. 2016) and the control of tonic activation of postural
muscles (Garcia-Rill and Skinner 1987; Takakusaki 2017). Neurons
in this region project to the medulla and then further to the spinal
cord. These projections are assumed to participate in the initiation
and regulation of locomotion, in particular based on classical studies
of locomotion in decerebrate cats induced by electrical stimulation of
this area (Shik et al. 1966; also see chapter 26).

Figure 11.2 Neurons in the reticular formation are scattered in different parts of
the brainstem. The reticular formation at the level of the midbrain is called the
mesencephalic reticular formation. At the level of the lower pons, it is called the
pontine reticular formation. At the level of the medulla, it is called the medullary
reticular formation. The superior colliculus is in the posterior segment of the
midbrain.

The paramedian pontine reticular formation is a part of the pontine

reticular formation and is a collection of cells in the pons that is
involved in the control of two types of eye movements, saccades
and smooth pursuits (Keller 1974). Saccades are rapid eye
movements that shift the gaze from one part of the visual field to
another. Smooth pursuit eye movements are slower tracking
movements of the eyes that keep a moving visual stimulus on the
fovea. This region contains neurons that are involved in the control
of saccades. Finally, the reticular formation neurons in the medulla
integrate sensory feedback signals and descending signals from the
cortical motor neurons and cerebellum and organize the output to
the spinal cord.
The motor centers in the reticular formation receive projections
from the motor cortex. These projections assist the reticular
formation in the control of posture in a feedforward manner to offset
the effects of impending limb movements on postural stability
(Schepens and Drew 2004; Cordo and Nashner 1982). For example,
if you reach to grab a cup from a high kitchen shelf, your center of
pressure (the point of application of the vertical resultant force acting
from the support) will move closer to the edge of the base of support,
and your posture may become unstable (also see chapter 24). The
reticular formation contributes to contraction of the lower limb
muscles in a feedforward manner, before any of the upper limb
muscles are activated (anticipatory postural adjustments) so that
posture remains stable, and you don’t fall or have to make a
protective step.
The reticular formation is also involved in the startle response, a
rapid motor response to a sudden and unexpected loud auditory
stimulus (Koch et al. 1992). The startle response has two
components. The first component is reflexive and occurs at short
latencies. It can be seen primarily in extensor muscles of the
extremities and has been interpreted as a “fight or flight” response.
The second component occurs at longer latencies that resemble
voluntary reaction times (figure 11.3). The reflex component of the
response originates in the auditory nerve fibers that project to the
cochlear nucleus, which is the first processing hub in the auditory
system. From there, neurons project to the pontine reticular
formation. From there, output neurons of the brainstem project
bilaterally to the spinal cord areas and lead to activation of
craniocervical muscles. The first burst of muscle activity in the
sternocleidomastoid muscles has been recorded within 55-80 ms
after stimulus presentation, supporting the reflexive nature of this
component. The second component involves an “orienting” response
toward the spatial location of the stimulus and occurs later (>200 ms)
(Brown et al. 1991; Dreissen and Tijssen 2012; Nieuwenhuijzen et al.
2000).

Figure 11.3 EMG responses in lower limb muscles to auditory stimulation during
locomotion. The EMG data were averaged and subtracted from both individual and
averaged stimulus trials. Figure shows two facilitatory startle responses at short
(~80-100 ms) and long latencies (~150-200 ms).
Adapted from Nieuwenhuijzen et al. (2000).

PROBLEM 11.1
The startle response shows a phenomenon called prepulse
inhibition. When a weak auditory stimulus (that does not elicit a
startle response) is followed by a strong stimulus within 0.5 s that
would normally elicit a response, the weak stimulus suppresses
the response of the strong stimulus. In Huntington’s disease
 patients, the prepulse inhibition is reduced. What would this
suggest about the interactions between the basal ganglia and the
reticular formation?

11.3 Superior Colliculus

The superior colliculus is another brain structure in the midbrain part
of the brainstem that transforms sensory input from multiple sensory
modalities into motor output (figure 11.2). Output neurons in the
superior colliculus project to neurons in the reticular formation that in
turn project to the spinal cord, leading to the activation of muscles,
primarily axial and proximal limb muscles (Werner 1993; Stuphorn et
al. 1999; Philipp and Hoffmann 2014). The firing patterns of cells in
these areas resemble those of neurons in other sensorimotor
regions, such as the motor cortex and premotor areas, and it
appears that these neurons are involved in the control of reaching
movements. However, the precise role that they play in conjunction
with neurons from the cortical areas is still unknown.
One of the main functions of the superior colliculus is to direct
sensory structures, such as the eyes, toward visual stimuli of
interest. This behavior is known as orienting and includes both head
movements and saccades that redirect the gaze. The superficial
layers of the superior colliculus receive direct projections from the
retina, providing sensory input, and the intermediate and deep layers
serve the oculomotor function by controlling saccadic eye
movements. The deep layer contains an organized motor map in
each hemisphere for controlling movements aimed at visual stimuli in
the contralateral visual field. Thus, visual sensory inputs originating
in the left hemifield reach the right superior colliculus, which in turn
generates leftward orienting movements, and vice versa. This
multilayer structure of the superior colliculus allows it to serve an
important function, rapid transformation of sensory signals into motor
output.
 The superior colliculus receives substantial projections directly
from retinal axons and cortical visual association areas (Boehnke
and Munoz 2008). They also receive input from the ascending
somatosensory and auditory pathways. In fact, many cells in the
superior colliculus are multimodal, showing responses to visual,
auditory, and somatosensory stimuli derived from cortical and
ascending inputs. This allows the superior colliculus to integrate
sensory information across multiple modalities. The multisensory
neurons in the superior colliculus also receive descending input from
multisensory neurons in the association cortices (parietal and frontal
cortex).
Multisensory integration allows for consolidation of sensory
information from specific sensory modalities (vision, audition, etc.) by
pooling of sensory information from different modalities into a single
multisensory stream to facilitate faster motor responses (Senkowski
et al. 2008). Imagine that you are sitting in a dark room and are
asked to make a pointing movement toward a stimulus that can
either suddenly flash or make an audibly loud sound. You will be able
to initiate the pointing movement sooner if the stimulus both flashes
and emits a sound at the same time, compared to a situation when it
only flashes or emits a sound.

11.4 Red Nucleus

The red nucleus is a relatively large group of cells located in the
rostral midbrain and is reddish pink in color, likely because of the
presence of iron-protein complexes (figure 11.2). The ascending and
descending projections from the red nucleus were first studied in
monkeys by Hans Kuyper and Don Lawrence in 1968. They found
that the descending tracts originating from the red nucleus terminate
in the lateral regions of the spinal cord, where α-motoneuronal pools
are located, innervating distal limb muscles. This suggests that the
red nucleus may be involved in the control of upper extremity
muscles. However, the rubrospinal tract, the descending tract that
originates in the red nucleus, originates from magnocellular neurons,
and these neurons constitute only a small fraction of the neurons in
the red nucleus in primates and humans, making it unlikely that the
red nucleus is a significant contributor to the neural control of
movements in higher mammals. A majority of the neurons in the red
nucleus are parvocellular neurons that do not project to the spinal
cord but instead facilitate communication between the red nucleus
and the cerebellum through the inferior olive. This implicates the red
nucleus in motor networks involving the cerebellum, especially in
functions that involve limb posture control (Herter et al. 2015).

PROBLEM 11.2
You are talking to a friend at an outdoor party when suddenly
through your peripheral vision you see a large object (an inflated
ball) rapidly approach your head. What kind of motor response
would you expect to make, and which brainstem structure would
make that possible?
Figure 11.4 The four components of the vestibular nuclei serve important
functions for controlling eye movements and postural equilibrium. The efferent
projections from the vestibular nuclei project to the nuclei of other cranial nerves
(III, IV, and VI). Lat = lateral; Sup = superior; Med = medial; Inf = inferior.

11.5 Vestibular Nuclei

Located in the medulla and pons of the hindbrain, the vestibular
nuclei are a highly interconnected complex consisting of four major
nuclei that integrate information from the vestibular afferents, the
somatosensory receptors, and the cerebellum (figure 11.4). They are
also the origins of the vestibulospinal tracts that play a major role in
maintaining equilibrium, posture control, and eye movements during
head rotations. Projections from these nuclei affect the activity of
extraocular muscles as well as of postural muscles of the back,
neck, and extremities. Thus, the vestibular nuclei are responsible for
feedback, or responding to a disturbance of body posture and
stability.
The lateral vestibular nucleus serves the vestibulospinal reflex.
This reflex contributes to the control of vertical posture by its effects
on the proximal extensor muscles of the limbs. The medial vestibular
nucleus mediates the vestibulo-ocular reflex (VOR). This reflex is a
gaze-stabilizing reflex where sensory signals encoding head
movements are transformed into signals to extraocular muscles,
which generate compensatory eye movements in the opposite
direction of the head movement. As a result, we can fixate gaze on
objects during head movements. To maintain balance, this nucleus
also controls head and neck movements via the medial
vestibulospinal tracts.
The third nucleus is the superior vestibular nucleus, and it
facilitates the coordination of posture and movements by contributing
to postural adjustments, as well as by providing neural signals for
eye movements. It is the most rostral of the vestibular nuclei. Finally,
the inferior vestibular nucleus is the most caudal nucleus. It receives
information on head tilt and gravity from the otolith organs.

PROBLEM 11.3
Which brainstem structure would play an important role in posture
stabilization if we were unexpectedly pushed from behind?
Figure 11.5 The 12 cranial nerves of the nervous system shown along a
transverse view of the brainstem and cortex.

11.6 Cranial Nerves

The brainstem is the source and target of a vast majority of the 12
cranial nerves in the nervous system (figure 11.5). The cranial
nerves are involved in sensory and motor functions of the head and
neck and create local circuits in the brainstem for integrating both
afferent and efferent information. Though the cranial nerves are
considered a part of the peripheral nervous system, not all cranial
nerves are arrayed along the brainstem. Two cranial nerves, the
olfactory nerve (I) and the optic nerve (II), enter the forebrain directly.
Except for the trochlear nerve (IV), the remaining cranial nerves
leave and enter on the ventral surface of the brainstem. The
trochlear nerve emerges from the dorsal side of the brainstem and
originates at the level of the inferior colliculus. The sensory and
motor functions of the cranial nerves and their points of origin on the
brainstem are listed in table 11.1.

11.7 Descending Tracts

Descending tracts convey motor information from the brain areas to
the spinal cord and cranial nuclei, and can be broadly divided into
two categories, pyramidal and extrapyramidal tracts. The pyramidal
tracts, the corticospinal and corticobulbar tracts, originate in the
cerebral cortex and carry axons to the spinal cord and brainstem.
The pyramidal tracts are described in greater detail in chapter 8.
Briefly, the corticospinal tract affects the neuronal apparatus of the
spinal cord involved in the control of the musculature of the limbs
and the trunk. The corticobulbar tract affects motor cranial nerves
that innervate the musculature of the head and the neck. The
corticospinal tract originates mainly from the primary motor cortex,
premotor cortex, supplementary motor areas, and somatosensory
cortex.
The extrapyramidal tracts originate in different areas of the
brainstem. These pathways are the reticulospinal, colliculospinal or
tectospinal, rubrospinal, and vestibulospinal tracts (figure 11.6). The
main roles of these pathways are in providing background
corrections such as regulating posture during voluntary movements.

Table 11.1 Cranial Nerves and Their Main Sensorimotor

Functions
Sensory or
Cranial nerve Nerve motor Function
I Olfactory Sensory Smell
II Optic Sensory Vision
III Oculomotor Motor Eye movements
IV Trochlear Motor Eye movements
V Trigeminal Sensory and motor Somatic sensation
from face,
mastication
muscles
 Sensory or
Cranial nerve Nerve motor Function
VI Abducens Motor Eye movements
VII Facial Sensory and motor Facial expression
muscles, taste
from tongue
VIII Vestibulocochlear Sensory Hearing, balance
sense
IX Glossopharyngeal Sensory and motor Sensations from
tongue,
chemoreceptors
X Vagus Sensory and motor Vocal cord
muscles,
swallowing
XI Spinal accessory Motor Shoulder and neck
nerve muscles
XII Hypoglossal nerve Motor Tongue movement

The reticulospinal tract originates in the pontomedullary reticular

formation and is an important descending tract for motor control and
a hub for sensorimotor integration that allows the nervous system to
make limb movements while stabilizing overall posture. The
pontomedullary reticular formation receives input from the motor
areas of the cerebral cortex. The tract that starts in the medulla
constitutes the lateral reticulospinal tract, and the fibers in this tract
descend bilaterally in the spinal cord near the corticospinal axons.
The fibers from the pontine region travel in the ventromedial region
of the spinal cord. Both these tracts descend ipsilaterally and project
to axial and proximal limb muscles and are important regulators of
posture.
The pontine and medullary nuclei that give rise to the
reticulospinal tracts receive cortical input from the premotor cortex
and to a lesser extent from the supplementary motor cortex. This
system is called the cortico-reticular system. Because reticulospinal
systems primarily influence extensor muscles, including the
paravertebral extensors as well as those of the limbs, the cortico-
reticulo-spinal system provides the cortex with the means to
influence antigravity extensor musculature in parallel with its
regulation of purposeful actions while standing.
The tectospinal tract (or the colliculospinal tract) is a bilateral
nerve tract that originates from cells in the superior colliculus and
projects to the neurons in the cervical spinal cord that coordinate
eye, head, and neck movements. The axons in this pathway
descend around the periaqueductal gray matter and then decussate
at the point called the dorsal tegmental decussation. Along with the
reticulospinal and vestibulospinal tracts, the fibers of this tract are a
part of the medial system of the descending tracts.

Figure 11.6 The approximate origins of the four main brainstem tracts involved in
motor control.

The tectospinal tract is considered to be mainly involved in

orienting eyes and head toward auditory and visual stimuli. It is
involved in responding to sudden and loud auditory stimuli, aka the
startle response. The afferent sound is detected by the brainstem,
which processes the sound and responds by activating cervical
muscles. The tectospinal tract connects with interneurons in the
cervical spinal cord, which then project to neck motor nuclei. In this
way, the auditory stimuli cause the instinctive movement of the head
and neck toward the perceived sound. One surprising discovery
made by Nudo and Masterton (1989) was that the number of fibers
in the tectospinal tract in mammals is surprisingly small. Raccoons
and cats have 628 and 909 fibers in the tectospinal tract,
respectively, but in seven species of primates that these authors
studied, they only found an average of 220 fibers. Thus, it is unclear
how important a role this pathway plays in human motor control.
The vestibulospinal tracts are one of the medial pathways in the
extrapyramidal system, and they arise from the vestibular nuclei of
the hindbrain. The vestibulospinal neurons receive neural input from
the vestibular division of the eighth cranial nerve labyrinthine
receptors in the inner ear, from the vestibulocerebellum mediated by
the fastigial nucleus, and from some proprioceptive afferent input
from the spinal cord. These tracts are involved in the initiation of
coordinated postural extensor activity in the limbs and trunk.
The first major projection of the vestibular complex to the spinal
cord is the lateral vestibulospinal tract, which arises from the lateral
and inferior vestibular nuclei at the level of the pons and medulla and
projects to all levels of the ipsilateral spinal cord, where it terminates
in the ventral gray horn and leads to changes in the activation of
proximal limb muscles. This tract activates the limb extensor
muscles when the vestibular system communicates deviation from
upright posture (vestibulospinal reflex). This reflex induces changes
in the activity of limb muscles at the level of the lumbar spinal cord
when unexpected head movements stimulate labyrinthine receptors
and destabilize posture.
The second projection of the vestibular complex is the medial
vestibulospinal tract, which arises from the medial and inferior
vestibular nuclei at the level of the pons and medulla and descends
bilaterally into the spinal cord. This tract innervates spinal segments
that control the muscles involved in supporting the head and
therefore only passes to the cervical region. This tract controls head
position through reflex activation of neck muscles (vestibulocervical
reflex) when the head suddenly rotates to indicate an unstable
posture. Note that the vestibulocervical reflex is also a
vestibulospinal reflex, but it acts at cervical spinal levels instead of
lumbar levels.
 The final descending tract in the extrapyramidal system is the
rubrospinal pathway. The fibers of this pathway originate in the red
nucleus. However, unlike the other three pathways described earlier,
this tract is located in the lateral white matter of the spinal cord, and
the axons of this pathway terminate in the lateral regions of the
spinal cord. Recall that the neural circuits involved in the control of
distal muscles of the upper extremity are located in the lateral white
matter. This would suggest that the rubrospinal pathway may be
involved in fine upper extremity motor control. However, as stated
earlier, this hypothesis is questionable because the axons of the
rubrospinal tract arise from large magnocellular neurons, and these
neurons are a relatively small percentage of the neurons in the red
nucleus of primates and humans. Therefore, it is unclear if this
pathway plays a major role in motor control in primates and humans.

CHAPTER 11 IN A NUTSHELL
The reticular formation, superior
colliculus, red nucleus, and
vestibular nuclei are four subcortical
regions involved in motor control. The
descending tracts originating from
these areas are called the
reticulospinal, tectospinal,
rubrospinal, and vestibulospinal,
respectively. The reticular formation
is located in the rostral midbrain to
the caudal medulla. It is involved in
the coordination of axial and proximal
limb muscles, in particular during
standing and walking. The superior
colliculus is a major hub for
processing sensory information and for
reflexive actions that involve
orienting movements toward or away
from salient visual and auditory
stimuli. The main role of the red
nucleus is in facilitating
communication between motor nuclei in
the cortex and the cerebellum. The
vestibular nuclei are the major
destination of the axons that form the
vestibular division of the eighth
cranial nerve. The projections from
the vestibular nuclei control axial
muscles and proximal limb muscles.
They play an important role in gaze
stabilization. There are a total of 12
cranial nerves that serve sensory and
motor functions, and most of them
arise from the brainstem.
Problems for Part II
Self-Test Problems
1. Under the action of a new drug, the outputs of all the
cerebellar nuclei are increased 10-fold. What changes in
movements can be expected during standing, walking, and
reaching?
2. The neuronal population vector in a brain structure points
consistently in the direction of hand movement when a
monkey performs a reaching task from a standard position to
targets distributed along a circle (center-out task). The
researcher draws a conclusion that the brain structure
encodes a hand velocity vector in space. Offer alternative
interpretations of this finding.
3. A subject is asked to imagine performing a motor task, such
as a fast arm movement. No actual movement or change in
the muscle activation levels occurred. In what structures of the
central nervous system would you expect to see changes in
the background activity of the neurons? In what structures
involved in the generation of actual movement would you
expect to see no changes in the neuronal activity?
4. The coils of two transcranial magnetic stimulators have been
placed over the right and left primary motor cortices. When
one of them generates a single stimulus, a response is seen
in the contralateral arm muscles at a latency of about 20 ms.
When the other stimulator produces a stimulus a few
milliseconds prior to the first one, the response decreases.
What neurophysiological mechanisms are likely to be involved
in the modulation of the response?
5. A drug has been discovered that increases the efficacy of the
projections from the putamen to the external part of the globus
pallidus by a factor of 10. A healthy person takes a single
dose of the drug. What changes would you expect to see in
their motor performance?
6. If a patient with Parkinson’s disease makes a saccade to a
target, where would you expect the saccade to land? On the
 target, short of the target, or beyond the target? Why?

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. The output neurons of the cerebellum
a. project on alpha-motoneurons in the spinal cord
b. excite their target neurons
c. project on the locomotor generator in the spinal cord
d. can generate action potentials of different shapes
e. are the smallest cells in the cerebellum
Why?
2. The two major pathways through the basal ganglia differ in the
following ways:
a. One of them is a positive feedback loop and the other is a
negative feedback loop.
b. One of them projects on the thalamus, while the other one
projects directly on the cortex.
c. One of them involves projections on the cerebellum, while
the other does not.
d. Only one of them is modulated by signals from the
substantia nigra.
e. All of the above
Why?
3. How do we know that basal ganglia neurons are involved with
a decision to move?
a. Neurons in the basal ganglia fire long before movement
initiation.
b. Neurons in the basal ganglia fire after motor cortex
neurons.
c. Neurons in the basal ganglia contain a topographic map of
the body.
 d. The basal ganglia receive extensive projections from the
cerebellum.
e. All of the above
Why?
4. How do the roles of the premotor cortex and motor cortex
differ for the control of limb movements? Select the correct
statements.
a. The primary motor cortex is responsible for movement
initiation.
b. The premotor cortex is involved in motor planning.
c. The corticospinal tract consists of axons originating from
both the premotor cortex and the motor cortex.
d. The primary motor cortex has more monosynaptic
connections to α-motoneurons than the premotor cortex.
e. All of the above
Why?
5. The cerebellum does not project to the spinal cord. Then how
does cerebellar activity influence how motor neurons in the
spinal cord affect movement?
a. The cerebellum projects through the thalamus to the
primary motor cortex and is indirectly involved in movement
planning and posture control.
b. The cerebellum is responsible for the initiation and
termination of movement.
c. The cerebellum is involved in movement coordination.
d. Both a and b are correct.
e. Answers a, b, and c are correct.
Why?
 Part III

Sensory Basis of Motor Control

 Chapter 12

Central Processing of
Somatosensory Information

KEY TERMS AND TOPICS

first-order neurons
second-order neurons
third-order neurons
proprioceptive system
primary and secondary somatosensory cortex
integration of somatosensory input with other
sensory modalities
injuries to somatosensory pathways

In this chapter, we continue the discussion of mechanisms involved

in processing sensory information. Chapter 6 reviewed peripheral
sensory endings sensitive to such mechanical variables as muscle
length, velocity and force, joint angle, and skin deformation. Here we
cover transmission of this information through the central nervous
system to cortical areas believed to play an important role in
perceiving relevant variables and guiding limb movements in the
environment. This topic is to be continued in chapters 28, 29, and
30, where we discuss in more detail both perceptual and motor
functions that rely on somatosensory information.

12.1 First-Order Neurons

The first-order neurons in the peripheral nervous system carry
information about somatosensory (tactile and proprioceptive) stimuli
to the central nervous system. The cell bodies of these first-order
neurons are located in the dorsal root ganglia of the spinal nerves.
These neurons are pseudounipolar, meaning that they have a cell
body with a T-shaped axon (but no dendrites). The distal end of the
axon innervates somatosensory receptors (sensory endings), and
the proximal end synapses with second-order neurons in the spinal
cord of the brainstem.
When these first-order axons enter the dorsal horn of the spinal
cord, they send collateral branches into different spinal segments
and synapse on motor neurons in the ventral horn or interneurons. A
major portion of the incoming axons travel rostrally and ipsilaterally
in a bundle called the dorsal column to the lower medulla. The dorsal
column is organized topographically with fibers from the lower limbs
traveling medially in a bundle called the gracile tract. The fibers from
the upper limbs and the trunk travel in a more lateral bundle called
the cuneate tract. These tracts terminate in different subdivisions of
the dorsal column nuclei, called the gracile nucleus and the cuneate
nucleus, respectively, where they synapse on second-order neurons
(figure 12.1).
Figure 12.1 The ascending pathways of the somatosensory system. The dorsal
column carries signals from the same side of the body to the medulla. The cuneate
nucleus in the medulla receives afferent input from the hand and forelimb. The
gracile nucleus is more medial to the cuneate nucleus and receives information
from the lower body. Second-order neurons in the medial lemniscus carry
information to the thalamus. Third-order neurons from the thalamus carry
information to the somatosensory cortex.

The axons entering the dorsal roots divide into ascending and
descending systems. It is important that some neurons entering the
dorsal roots synapse directly on motoneurons in the ventral horn.
This circuit forms the basis of the monosynaptic stretch reflex. The
stretch reflex or the myotatic reflex (an obsolete term) refers to a
muscle contraction in response to stretching of muscle fibers. The
proprioceptive afferent axons in the ascending branches travel along
with axons of tactile afferents in the dorsal column.

12.2 Second-Order Neurons

The dorsal column nuclei are the source of input to the second-order
neurons, which then project to the thalamus. These axons, called the
internal arcuate fibers, cross the midline of the body and form a tract
called the medial lemniscus. The crossing-over of tracts is called
decussation and seems to occur in vertebrates because of an axial
twist of the forebrain during evolution (de Lussanet and Osse 2012).
The axons of the medial lemniscus carrying tactile and
proprioceptive information from the lower limbs and upper limbs
travel separately before they enter the ventral posterior lateral (VPL)
nucleus of the thalamus. Surprisingly, and in contrast to the
organization of neurons in the dorsal column and later in the
somatosensory cortex, the axon bundle in the medial lemniscus
rotates 90º laterally and flips the axons from second-order neurons
carrying information from the upper body into the medial section of
the tract and the axons from second-order neurons carrying
information from the lower body into the lateral section. Therefore, if
an injury occurs in the lateral section of the medial lemniscus, it will
likely affect the lower body.

12.3 Third-Order Neurons

Third-order neurons from the thalamus send their axons via the
internal capsule to the ipsilateral postcentral gyrus, the primary
somatosensory cortex. These neurons will terminate in layer IV of
the somatosensory cortex. The internal capsule is a white matter
structure composed of myelinated fibers that carry information from
the brainstem to the cerebral cortex and vice versa.

12.4 Proprioceptive System

There is a key difference between the proprioceptive and tactile
systems. In particular, axons carrying proprioceptive information
reach both the primary somatosensory cortex and the cerebellum.
The pathway to the somatosensory cortex is considered critical for
sensing limb position, force, and stereognosis, whereas the
proprioceptive pathway to the cerebellum is involved in
“unconscious” proprioception and in the control of the timing of
muscle contractions. This “unconscious” proprioceptive information
from muscles and the Golgi tendon organs is carried from one side
of the body to the same side of the cerebellum (ipsilaterally) through
the spinocerebellar tract.
First-order proprioceptive neurons that enter the spinal cord via
the dorsal roots travel through the dorsal column and synapse on
Clarke’s nucleus. This nucleus is found approximately from levels T1
to L3 of the spinal cord. The second-order neurons from Clarke’s
nucleus project their axons via the ipsilateral posterior column of the
spinal cord in a tract called the dorsal spinocerebellar tract. Most
axons traveling in the dorsal spinocerebellar tract arise from the cells
of Clarke’s nucleus. The dorsal spinocerebellar tract ascends to the
level of the medulla and then enters the cerebellum. Interestingly,
collaterals of the axons from the dorsal spinocerebellar tract
decussate and join the medial lemniscus on the way to the thalamus.
This suggests that even the pathways that we consider to be
involved in unconscious proprioception indirectly communicate with
the primary somatosensory cortex. The functional role of these
pathways is not entirely clear, but it is believed that they may be
involved in long-term motor learning and formation of motor
memories.

PROBLEM 12.1
Why would the nervous system only send a copy of the
proprioceptive information, and not tactile information, to the
cerebellum?

12.5 Primary and Secondary

Somatosensory Cortex
The primary somatosensory cortex (S1) in the postcentral gyrus is a
primary receptor of sensory data from the skin, muscles, tendons,
and joints of the body. It comprises four distinct Brodmann’s regions
(1, 2, 3a and 3b) and mainly receives information from the ventral
posterior complex of the thalamus. Each of these four Brodmann
areas contains a distinct somatotopic map of the body; the lower
limbs and trunk are represented more medially, and the upper limbs
and face are represented more laterally in S1. Neurons in areas 1
and 3b respond primarily to tactile stimuli. Area 3a receives
proprioceptive information, originating mainly from the muscle
spindles. Area 2 neurons respond to both tactile and proprioceptive
stimulation (figure 12.2). Area 3b sends many projections to areas 1
and 2, and consequently plays a very important role in the
processing of tactile information and integration of tactile and
proprioceptive information. Lesions to area 3b will almost certainly
have significant consequences for the loss of tactile sensation.
Figure 12.2 Brodmann areas 1, 2, 3 of the primary somatosensory cortex. Area
3a is more ventral than area 3b in the postcentral gyrus. The output of the primary
somatosensory area reaches Brodmann areas 5 and 7 in the parietal cortex,
where it gets integrated with sensory information from other stimuli.

In the 1950s, Penfield and Rasmussen published a detailed

account of how different body parts are represented in S1 (Penfield
and Rasmussen 1950). They showed that different body parts do not
have a proportionate size-based representation in S1. The face and
fingers have much larger representations in S1. In contrast, the trunk
and proximal upper and lower limbs have much smaller
representations. This organization reflects the fact that humans need
high spatial acuity and dedicated neural circuits for complex motor
maneuvers such as object manipulation, speech production, and
generation of facial expressions. The density of sensory afferents is
much greater in the face and fingers than in the other parts of the
body, which have a smaller representation in S1. This organization of
S1 has also been called the sensory homunculus (see figure 12.3).
Note that the sensory homunculus is a topographic map of the
contralateral side of the body parts along the postcentral gyrus of the
cerebral cortex. The homunculus is where the sensations from the
body parts are processed and mapped.
 S1 neurons send out somatosensory information to other cortical
areas, such as the Brodmann areas 5 and 7 in the parietal cortex
(figure 12.2). For example, all S1 areas project to the secondary
somatosensory cortex (S2). Unlike S1, which primarily responds to
contralateral stimulation, tactile stimulation leads to bilateral
activation in S2. S2 is involved in the processing and integration of
somatosensory stimuli for further high-order processing and
functions. S2 is also known to project to limbic areas, the amygdala
and the hippocampus, and these projections are considered to be
involved in memory and learning of tactile information. More recent
studies suggest that S2 plays a role in the conversion of touch
sensation to a conscious perception (Rossi-Pool et al. 2021).

Figure 12.3 Frontal view of the sensory homunculus is a map of the body in the
somatosensory cortex. Adjacent regions of the body are represented next to each
on the map. Body parts with smaller receptive fields of tactile neurons have a
disproportionately larger representation in the somatosensory cortex.
12.6 Integration of
Somatosensory Input With Other
Sensory Modalities
Neurons in S1 also project to areas 5 and 7 in the parietal cortex.
The projections to the parietal cortex relay the current muscle state
(muscle length, velocity, and force) to the visual areas in the parietal
cortex. The parietal cortex is also the site for the integration of
somatosensory and visual information with information from the
vestibular system (chapter 13). By virtue of these projections,
parietal areas integrate and process multisensory information. These
parietal areas then project the processed and integrated signals from
multiple modalities to the premotor and motor cortex in the frontal
cortex for preparation and execution of goal-directed movements.

12.7 Injuries to Somatosensory

Pathways
Tabes dorsalis occurs because of an untreated syphilis infection, and
it causes slow degeneration of the nerve cells and fibers that carry
somatosensory information to the brain. These nerves are in the
dorsal columns and carry somatosensory information from the same
side of the body. The disease occurs more frequently in males than
in females and begins commonly during midlife. Persons with tabes
dorsalis are severely impaired in functionally important movements
such as standing and walking. They can perform these tasks only
with eyes open (i.e., using visual information to substitute for the
lacking or severely impaired somatosensory information). If not
treated properly, tabes dorsalis can lead to paralysis.
Stroke-induced lesions to the somatosensory cortex affect both
proprioceptive function (Kenzie et al. 2014; Findlater and Dukelow
2017) and tactile acuity (Tyson et al. 2008). The loss of these
functions has a direct impact on motor behavior. In any activity of
daily living, where people don’t overtly look at the arm during the
execution of a task, loss of proprioceptive function will have a
deleterious impact on the ability to dexterously control the arm during
skill execution (also see chapter 39 for stroke-related deficits).

PROBLEM 12.2
Tabes dorsalis is a disorder in which there is demyelination of the
axons in the dorsal column of the spinal cord. If an individual with
this disorder has suffered unilateral damage to axons, would
somatosensory function be altered on the ipsilateral side of the
body or the contralateral side?

There is a rare disorder, peripheral large-fiber neuropathy, that is

accompanied by loss of conduction of action potentials along first-
order myelinated afferents. Patients with this condition, sometimes
referred to imprecisely as “deafferented persons,” have no
segmental reflexes and do not feel their bodies (Cole and Paillard
1995). They have to learn how to perform functional movements with
the help of visual information, and even after many years of
adaptation, they still present slow and deliberate movements with
impaired interjoint coordination (Rothwell et al. 1982a; Sainburg et
al. 1993; Sarlegna et al. 2010).

CHAPTER 12 IN A NUTSHELL
Somatosensory information originates
in the peripheral receptors in the
skin, muscle spindles, and tendons and
makes its way to the central nervous
system through a series of neurons.
The first-order neurons carry
information from the ipsilateral side
of the body to the spinal cord and the
lower medulla. This bundle of axons is
called the dorsal column. The first-
order neurons in this column synapse
on the second-order neurons that
decussate and carry somatosensory
information to the contralateral
nuclei of the thalamus. The bundle of
axons that carry the second-order
neurons are called the medial
lemniscus. The third-order neurons
carry information from the thalamus
through the internal capsule to the
primary somatosensory cortex. The
primary somatosensory cortex (S1) is
topographically organized. The trunk
and the legs are represented more
medially, and the hand and the face
areas are represented more laterally.
Information from S1 is projected to
higher-order sensory areas such as S2
and areas 5 and 7 in the parietal
cortex for further processing and
integration with other sensory
modalities.
 Chapter 13

Vestibular and Auditory Systems

KEY TERMS AND TOPICS

otolith organs
semicircular canals
stereocilia, kinocilium
vestibular nuclei
vestibulocochlear nerve
vestibulo-ocular reflex
oscillopsia
vestibulocervical reflex
vestibulospinal reflex
cochlea
inferior colliculus
primary auditory cortex
Wernicke’s area

The vestibular system provides the sense of whole-body balance

and information about head position in space. It also allows the
nervous system to initiate fast compensatory movements to restore
balance in response to both externally generated and self-imposed
mechanical perturbations. The vestibular system also projects to
cortical areas to provide perception of gravity, movement, and
orientation. In addition, the vestibular nuclei also provide signals to
motoneurons innervating extraocular muscles and play a critical role
in the vestibulo-ocular reflex. These reflexes are important for
stabilizing gaze in space during head movement. The auditory
system allows us to localize the origin of sound in space, as well as
understand speech and language. There are neuroanatomical
pathways that reciprocally connect the auditory and motor cortices
and provide a link for motor-related neural signals to influence
auditory cortical activity, but the functional significance of these
pathways is not yet clear.

13.1 Transduction in the

Vestibular System
The peripheral vestibular system encodes translational and rotational
head motion in three dimensions using two different sensors: the
otolith organs, which detect linear motion, and the semicircular
canals, which detect rotational motion. Otolith organ afferents
respond to linear accelerations in all three dimensions, including
static head tilts relative to gravity. This makes the otoliths well suited
to convey to the central nervous system changes in head orientation
and to contribute to the sense of vertical balance. In contrast,
semicircular canal afferents encode angular velocity of the head
during yaw, pitch, and roll in three dimensions (see figure 13.1).
The peripheral component of the vestibular system is the
labyrinth, a set of continuous chambers that consists of two otolith
organs, the saccule and the utricle, and the three semicircular
canals. Between the membranous labyrinth and the bony walls is the
perilymph, whose ion composition is close to that of the
cerebrospinal fluid. Ion pumps provide the unusual ion composition
of the endolymph so that its potential is +80 mV with respect to the
surrounding perilymph. Hair cells within the inner ear have an
intracellular potential of −60 mV, which adds up to 140 mV of
electrical driving potential to drive K+ across open transduction
channels into the hair cell. This depolarizes the hair cell and opens
voltage-gated channels for K+ and Ca++ located in the hair cell.

Figure 13.1 (a) The vestibular organs in the inner ear consist of otolith organs
(i.e., utricle and saccule, which sense linear acceleration) and three almost
orthogonal semicircular canals, which sense rotational acceleration. The vestibular
nerve projects signals from otoliths and semicircular canals to the central nervous
system. (b) The vestibular system encodes linear movement in three-dimensional
space denoted as front, back, left, right, up, and down directions (by otolith organs)
and rotational movements (i.e., yaw by the horizontal canal and pitch and roll by
both anterior and posterior canals).
 Linear motion of the head is detected by the two otolith organs,
the saccule and the utricle. These two similar organs lie against the
walls of the inner ear. The receptors for these two organs, maculae,
are hair cells. Overlying the hair cells is a gelatinous layer called
otolithic membrane that consists of calcium carbonate crystals called
otoconia. The relatively heavier otoconia add inertia to the otolithic
membrane. So, when the head tilts, the gravitational pull causes the
membrane to move relative to the macula, and the resulting parallel
force displaces the hair bundles and produces a change in the
electric potential on the receptor membrane. Stereocilia are the key
mechanosensors of hair cells, and the orientation of the stereocilia
toward the kinocilium changes the receptor potential. Kinocilium is
the longest cilium located on the hair cell next to many stereocilia.
When the stereocilia lean toward the kinocilium, the cell is
depolarized, and there is an increase in the vestibular nerve activity.
In contrast, when the stereocilia tilt away from the kinocilium, the cell
is hyperpolarized, and there is a decrease in afferent nerve activity.
In the utricle, the kinocilia are oriented toward the middle line called
striola, and in the saccule they are oriented away from it. The
different orientations of the saccule and utricle allow the vestibular
system to detect head tilts in different directions.
Semicircular canals detect head rotations arising from either
voluntary movements or from angular accelerations caused by
external forces, such as on a roller coaster ride. The three
semicircular canals, almost at right angles to each other, encode
head rotations along three perpendicular axes. Each of the three
canals has at its base an expansion called an ampulla that houses
the crista. The crista contains hair cells that extend out into a
gelatinous mass called the cupula, which extends along the entire
length of the ampulla. In contrast to the saccular and macular hair
cells, all hair cells in the crista are organized with their kinocilia
pointing in the same direction (see figure 13.2).
Figure 13.2 The ampulla of the semicircular canal and a cross-sectional view of
the vestibular hair cells. Head rotation causes the membranous canal to distort the
cupula.

When head rotation occurs in the same plane as a semicircular

canal, the inertia produces a force across the cupula, moving it away
from the direction of the head rotation. When the cupula moves in
one direction, the entire population of hair cells is depolarized, and
vice versa; movement in the opposite direction causes hair cell
hyperpolarization. Orthogonal rotations have no effect on the cupula.
The semicircular canals work in pairs in the opposite ears. Head
rotations move the cupula in opposite directions for the two opposite
canals, causing one to depolarize and another to hyperpolarize. The
hair cells in the canals toward which the head is turning are
depolarized, and those on the opposite side are hyperpolarized.
Translational movements, in contrast, produce equal forces on the
two sides of the cupula, so no hair cells are displaced. This allows
the semicircular canals to encode only rotational movements arising
from either voluntary movements or external forces.

PROBLEM 13.1
If an animal has a damaged otoconia, which type of head
movement would it have difficulty detecting?
13.2 Vestibular Afferents
Respond to Head Motion
Axons of the otolith organs and semicircular canals exhibit
spontaneous levels of high activity. Thus, they can convey
information by both increasing and decreasing their firing rates. For
otolith organs, the firing rates increase or decrease based on the
direction of the head tilt. The response rate of the neurons remains
elevated (or depressed) as long as the head remains tilted—
meaning that the tonic activity of the neurons encodes the static
force on the head. In contrast, when there is a sudden translational
movement of the head, such as what the head experiences when a
vehicle suddenly accelerates, there is a transient increase or
decrease in the firing rate of the otolith organs.
The firing rate of the receptors in the semicircular canals
increases when the head rotation accelerates and the cupula is
deflected in one direction, and the firing activity decreases when the
head decelerates and the cupula is rotated in the opposite direction.
This receptor apparatus is very sensitive and can detect angular
acceleration as small as 0.1°/s2. Note that physical displacements of
the cupula are less than 10 nm, comparable to those produced by
low-amplitude sound in the auditory system. During rotation at
constant speed, the firing rate returns to baseline levels. Head
rotations at constant speed can be encountered during flights and
slow rides in amusement parks.

PROBLEM 13.2
A woman finds her shoelaces undone, and she bends down and
tilts her head downward to tie her shoes. She then straightens her
head back up again. How would the firing rate of the vestibular
nerve change during this activity?
13.3 Central Projections From
the Otolith Organs and
Semicircular Canals
The vestibular organs communicate via cranial nerve VIII
(vestibulocochlear nerve), which synapses in the brainstem and the
cerebellum. The cell bodies of the neurons of the vestibular section
of the vestibulocochlear nerve are in the vestibular nerve ganglion
(also called Scarpa’s ganglion). These neurons are bipolar. The
peripheral processes innervate the otolith organs and the
semicircular canals, and the central processes project to the
vestibular nuclei and the cerebellum. The vestibular nuclei are
located in the medulla and pons of the hindbrain. Afferent signals
from the otolith organs and semicircular canals converge in the
vestibular nuclei.
An important thing to note is that the receptor potentials
generated by the hair bundles in the otolith organs do not distinguish
between static head tilts and translational motions of the head.
Otolith organs are excited tonically by static head tilts and phasically
by translational movements of the head, whereas the semicircular
canals are only excited by head rotations that accompany tilts.
Otolith afferents respond to inertial motion (e.g., head motion of a
passenger when a bus suddenly accelerates) and a change in head
orientation relative to gravity. To disambiguate head motion arising
from different sources, the vestibular system also uses signals from
the semicircular canals and combines those signals more centrally in
the vestibular nuclei and the cerebellum. In fact, the mixing of the
signals from otolith organs and the semicircular canals in the
vestibular nuclei makes it possible for the nervous system to
disambiguate translational and rotational head movements. Thus,
when there is a head tilt, integration of information from both the
otolith organs and the semicircular canals in the vestibular nuclei and
the cerebellum can be used to disambiguate head tilts from
translational movements of the head.
The cerebellum, specifically the vestibulocerebellum and
spinocerebellum (see chapter 10), is a major destination of vestibular
afferent information, and the cerebellum in turn projects back to the
vestibular nuclei. The major nodes in the cerebellum include the
flocculus, nodulus, uvula, and rostral fastigial nuclei. The nodulus
and uvula integrate information from otolith organs and semicircular
canals to disambiguate head tilts from translation. Another important
role played by the cerebellum is that it differentiates vestibular
signals that arise from self-motion and external forces. For example,
if a ballet dancer slips and their head rotates backward, the
vestibular system will trigger protective reflexes, but if the same
posture is acquired by the dancer during a performance, then the
cerebellum will suppress any protective reflexes. It has been
hypothesized that predictive signals in the cerebellum cancel out the
ascending vestibular information in the rostral fastigial nucleus
(reviewed in Cullen 2019) during self-motion (see figure 13.3).
An alternative interpretation is that voluntary movements are
produced by changes in referent body configurations (see chapter
21), which are used as the origins of the coordinate systems to
measure and interpret information from relevant sensory receptors
(see chapter 28). When signals from vestibular receptors change
without a change in the referent body configuration, they are
interpreted as deviations from that configuration caused by external
forces (perturbations) and lead to quick corrective changes in muscle
activation. If the referent configuration changes to produce a
voluntary movement, similar changes from the vestibular system do
not produce such corrections.
In addition to the cerebellum, the information from the vestibular
nuclei also reaches the thalamus on its way to the cerebral cortex.
The regions in the cerebral cortex include Brodmann’s area 3a, 2v,
and the parietoinsular vestibular cortex. Activation in these regions is
important for perceptions arising from vestibular sensations. The
parietoinsular vestibular cortex is also a multisensory area that
responds to proprioceptive and visual information. This suggests that
most cortical processing of vestibular information is multisensory in
nature.

PROBLEM 13.3
A figure skater has sustained a concussion that has temporarily
disrupted communication in the ascending vestibular pathways to
the cerebellum. How could the skater’s performance be affected
on skills such as the layback spin?

Figure 13.3 The vestibular projections to the cortex and the cerebellum. The
inputs from the vestibular labyrinth project to the vestibular nuclei. From there the
projections extend to both the cerebellum and the cortex.

13.4 Central Pathways That

Stabilize Gaze, Posture, and
Head Movements
The vestibular system stabilizes eye movements and posture by
initiating a variety of reflexes. In that way, the vestibular system by
definition is multisensory because it integrates information from the
visual as well as the somatosensory systems. Neurons in the
vestibular nuclei, the first processing node for afferent vestibular
information in the brainstem, themselves receive converging visual
input. In addition, vestibular nuclei also receive other sensory inputs,
including those from the proprioceptive and auditory systems. This
makes the system for postural control inherently multisensory in
nature and underscores the importance of multisensory input for
maintaining upright stance and for performing many activities of daily
living that require precise spatiotemporal coordination between
multiple sensory systems.
One such activity that requires precise spatiotemporal
coordination is the ability to read street names while walking. When
we walk, our head undergoes passive oscillations in upward-
downward and left-right directions. During these oscillations, the
vestibulo-ocular reflex acts to stabilize gaze on a point in space.
Eye muscles contract in a coordinated manner to stabilize the gaze
with respect to the external world during both rotational and
translational head movements. This gaze stabilization allows us to
read texts on street signs and billboards without having to focus too
hard.
The neural circuitry for this reflex is shown in figure 13.4. Briefly,
when the head moves, the otolith organs and semicircular canals
communicate those movements via the vestibular nerve to the
vestibular nuclei in the brainstem. The vestibular nuclei project to the
contralateral abducens nuclei. From the abducens nuclei, one
pathway directly makes excitatory projections to the ipsilateral lateral
rectus muscles. The second pathway makes excitatory projections to
the contralateral oculomotor nucleus and the medial rectus muscle.
At the same time, the vestibular nuclei also inhibit the ipsilateral
abducens nuclei. This architecture causes the left (right) lateral
rectus and the right (left) medial rectus muscles to be activated
together to compensate for rightward (leftward) head movement.
Damage to the vestibulo-ocular reflex can cause a condition called
oscillopsia. Patients suffering from oscillopsia would encounter
difficulties fixating their gaze on visual targets while their head is
moving.
The vestibulocervical and vestibulospinal reflexes are changes in
muscle activation induced by movements of the head. The goal of
these fast reflexes is to maintain the head in an upright posture. The
vestibulocervical reflex acts on the neck muscles and the
vestibulospinal reflex acts on the limb muscles to stabilize the
position of the head in space. The vestibulocervical reflex is
mediated by the medial vestibular nucleus and descending axons in
the medial longitudinal fasciculus that reach the cervical spinal cord.
When posture is perturbed, the semicircular canals are activated,
and the head muscles act reflexively to pull the head upright. Head
movement produced by the vestibulocervical reflex is also
accompanied by forelimb (arm) extension and hindlimb (leg) flexion
to protect the body.
Figure 13.4 The neural network involved in the control of the vestibulo-ocular
reflex. This is a gaze-stabilizing reflex: the sensory signals encoding head
movements (sensed through the semicircular canals) are transformed into motor
commands to generate compensatory eye movements in the opposite direction of
the head movement. This ensures stable vision during head movement. The
neural network involves vestibular afferents, central neurons, oculomotor neurons,
and extraocular eye muscles and can operate without cortical input and control.
Dashed and solid lines indicate inhibitory and excitatory actions.

The vestibulospinal reflex is mediated by the lateral and medial

vestibulospinal tracts as well as the reticulospinal tracts. The otolith
organs project to the lateral vestibular nucleus. The axons of the
projections from this nucleus in the lateral vestibulospinal tract reach
the ipsilateral spinal cord and excite the extensor motoneuronal
pools and inhibit the flexor motoneuronal pools. Thus, the strong
activation of the extensor muscles in the trunk and limbs supports
balance and maintenance of upright posture.

PROBLEM 13.4
 A person reports feeling dizzy while nodding their head in
agreement while looking at someone’s face. What would be the
most likely cause of this disorder?

13.5 Peripheral Auditory System

Sound waves are pressure waves that move in three-dimensional
space caused by the vibration of air molecules. Sound waves carry
energy, and the energy increases with the amplitude and frequency
of the waves. In particular, the higher the frequency, the higher the
energy. The audible frequency range for healthy humans is between
20 and 20,000 Hz.
The ear consists of three parts: the external, middle, and inner
ear. The external ear gathers the sound energy, focuses it on the
eardrum, and amplifies its pressure. The main role of the middle ear
is to prepare the sound energy that is traveling in a low-resistance air
medium for a high-resistance aqueous medium of the inner ear. If
the middle ear did not perform this function, most of the sound
energy would reflect off the ear. The cochlea of the inner ear is
where the original signal is transduced by the sensory hair cells and
transformed into electrical nerve impulses in the auditory nerve
fibers. The cochlea is a fluid-filled coiled structure that contains the
organ for hearing.
The cochlea consists of sensory hair cells that are displaced by
the traveling sound waves. The hair bundles consist of stereocilia,
and these inner hair cells project to the auditory nerve fiber. Unlike
the hair cells of the crista or the maculae of the saccule and utricle of
the vestibular system, hair cells of the cochlear duct in humans (and
in most mammals) do not have kinocilia. The kinocilia disappear
shortly after birth in most mammals. Thus, displacement of the hair
bundle toward the tallest stereocilia opens mechano-electrical
transduction channels that allow K+ ions to flow into the hair cells
(down the electrochemical gradient) and depolarize them. That
opens voltage-gated Ca++ channels, allowing calcium entry and
release of neurotransmitters. Movement in the opposite direction
hyperpolarizes the hair cells. As the stereocilia move, the graded
potential follows the movement of the stereocilia. The receptor
potential leads to transmitter release from the hair cell. That in turn
triggers action potentials in the cranial nerve VIII fibers. Hair cells
only release neurotransmitters when they are depolarized.

13.6 Central Auditory

Projections From the Cochlea
The auditory nerve and the vestibular nerve together constitute the
cranial nerve VIII. The auditory nerve enters the cochlear nuclei in
the brainstem and from there projects to both the ipsilateral and
contralateral superior olives (see figure 13.5). The superior olive
plays a crucial role in localizing the spatial location of a sound. The
pathways from both the superior olives and cochlear nuclei project to
the midbrain auditory center, the inferior colliculus. The inferior
colliculus neurons exhibit spatial selectivity for the origin of the sound
source and also process the temporal patterns of sound.
Figure 13.5 The anatomy of the ear and the major auditory pathway through the
brainstem, thalamus, and cerebral cortex.

13.7 Auditory Integration

Humans use two different strategies for spatial localization of a
sound source. For low frequencies (<3 KHz), the interaural time
differences of the neural information reaching the nervous system
from the two ears provides cues for auditory source localization and
space perception. Above these frequencies, interaural intensity
differences are used as cues. The neural circuitry that is responsible
for calculating the time differences of the auditory signals from the
two ears is the medial superior olives. The intensity differences are
processed by the lateral superior olive. These two pathways receive
information from the cochlear nuclei and eventually converge in the
midbrain auditory centers.
13.8 Auditory Thalamus and
Cortex
The medial geniculate nucleus of the thalamus is the hub and the
relay for all incoming auditory sensory information. Most of the
afferent information reaching the thalamus originates from the
inferior colliculus, and the destination of the efferent information is
the auditory cortex. The auditory cortex can be divided into the
primary and peripheral (secondary) areas. The primary auditory
cortex (A1) is located in the superior temporal gyrus of the temporal
lobe. Similar to the primary visual cortex (V1) and primary
somatosensory cortex (S1), A1 has a topographical map of its
sensory epithelium, the cochlea. A1 comprises a tonotopic map
where sounds of similar frequencies are processed in adjacent
regions and sounds with dissimilar frequencies are processed farther
apart.
The secondary areas of the auditory cortex receive diffuse input
from the thalamus and are less precise in their tonotopic
organization. Besides providing a tonotopic map for sound
processing, the auditory cortex is also involved in speech
processing, and this function is primarily performed by the superior
temporal gyrus (STG) of the secondary auditory cortex. Recordings
in the STG have shown that neural activity correlates with the onset
and offset of syllables. STG neurons also exhibit context-specific
activity. For example, when we are trying to pay attention to what
one person is saying at a large and loud party, neural activity in STG
neurons would exhibit context-specific activity to speech produced
by that person and not by other people in the room.
Wernicke’s area, which is critical for spoken and written language
comprehension, is involved in speech processing. This area is also
located in the posterior section of the STG, in the secondary auditory
area. The area is named after Carl Wernicke, a physician who
identified this area in the 1870s while observing patients with severe
language comprehension deficits. This area also receives significant
input from the visual cortex and is involved in transforming the visual
impression of letters in an alphabet into representations of speech
sounds.

PROBLEM 13.5
A patient reports intact hearing in both ears, but difficulties with
spatial localization of sound sources. This is called the central
auditory processing disorder (CAPD). Which are the likely neural
substrates involved in CAPD?

13.9 Auditory Cortex and Limb

Motor Control
Listening to music is an aesthetically pleasing experience that not
only engages many areas of the auditory cortex, but also engages
the motor cortex (Grahn and Brett 2007; Martín-Fernández et al.
2021). This is not surprising because often listening to music triggers
a desire to move, dance, or tap fingers, or even just hum along with
the song. Indeed, the premotor and supplementary motor areas
exhibit increased activity while a subject listens to music, and they
are considered to be involved in rhythm detection (Merchant et al.
2015). In contrast to these results, a recent study has shown that
during the preparation phase of a limb movement, there is an
increase in neural activity in the auditory cortex (Gale et al. 2021a).
Though the exact reasons for limb movement–related increases in
neural activity in the auditory cortex are not clear, the increase in
activation is not unique to the auditory cortex. It is also observed in
the primary somatosensory cortex (Gale et al. 2021b; Ostry and
Gribble 2016) and the primary visual cortex (Gutteling et al. 2016).
The increase in neural activity in the early sensory areas could be to
tune these sensory areas for sensory prediction.
 The startle response (see chapter 11) is a rapid brainstem-
mediated motor response that is generated in response to a loud
auditory stimulus. Triggering the startle response produces muscle
activity in the orbicularis oculi (eyelids) and sternocleidomastoid
(neck muscles) and strong flexor activity in limb muscles.

CHAPTER 13 IN A NUTSHELL
The vestibular system is involved with
the maintenance of equilibrium and the
orientation of the body in space. The
auditory system is concerned with the
perception of sound, its spatial
location, and the perception of
language. The vestibular system
encodes both translational and
rotational motion of the head. Axons
of the otolith organs and semicircular
canals exhibit spontaneous levels of
high activity and convey information
both by increasing and decreasing
their firing rates. The vestibular
system plays an important role in
stabilizing gaze during head
movements. The cerebellum and
parietoinsular vestibular cortex are
important areas in the brain for
processing vestibular signals. The
cochlea, inferior colliculus, and
primary auditory cortex play an
important role in processing auditory
signals.
 Chapter 14

Visual System

KEY TERMS AND TOPICS

retina
rods
cones
optic nerve
optic tract
optic radiation
visual cortex
extrastriate areas
MST
V4
magnocellular neurons
parvocellular neurons
cerebral akinetosopia
cerebral achromatopsia
saccades
smooth-pursuit eye movements
vergence
The visual system is one of the most important sensory systems. It
provides information on where objects of interest are in space and
where they are with respect to each other as well as on salient
properties of those objects, such as color, shape, and brightness.
The system can measure spatial distances and sizes of objects with
sub–millimeter level accuracy. The visual system is instrumental in
helping guide a vast majority of goal-directed movements. The
visible spectrum of light for humans is between the wavelengths of
ultraviolet light and infrared light. It is believed that, within this range,
humans can distinguish up to 10 million colors (Kuehni 2016). These
abilities bestow humans with extraordinary capabilities to sense,
perceive, and act in the external world.

14.1 Structure of the Eye

Each eye is capable of focusing on objects far and near, constantly
adjusting the amount of light it lets in and producing visual images
that are then transmitted to the brain. The eye can be considered as
a fluid-filled sphere with three discrete layers (see figure 14.1). Most
of the outer layer is composed of a tough white fibrous opaque
tissue, the sclera, which transforms into the transparent cornea in
the front of the eye. The cornea permits light rays to enter the eye.
The iris, the ciliary body, and the choroid are three distinct structures
that constitute the middle layer of the eye tissue. The size of the
pupil is under neural control and is controlled by two muscles in the
iris with opposing actions. The second of the middle structures, the
ciliary body, is a ring of tissue that adjusts the refractive power of the
lens. The final structure in the middle layer, the choroid, supplies
blood for the photoreceptors in the retina. The final and innermost
layer is the retina. The retina is where all the photoreceptors reside.
The photoreceptors are nerve cells that are sensitive to light and
transmit visual signals to the brain.
Figure 14.1 Anatomy of the eye. The retina contains the neurons that are
sensitive to light and is the site of phototransduction.

14.2 Structure of the Retina

The retina’s neural circuitry converts the electrical activity of
photoreceptors into action potentials that are communicated to the
brain via axons in the optic nerve. Despite its apparent peripheral
location, the retina is actually a part of the central nervous system.
There are five different types of neurons in the retina:
photoreceptors, bipolar cells, horizontal cells, amacrine cells, and
retinal ganglion cells (see figure 14.2). The cell bodies and
processes of these neurons are arranged in five layers. The cell
bodies are located in the inner nuclear, outer nuclear, and ganglion
cell layers, whereas the processes are located in the inner plexiform
and outer plexiform layers.
Figure 14.2 The structure of the retina and the photoreceptors in the retina.
Rods and cones are the photoreceptors that form a direct pathway with bipolar
cells and retinal ganglion cells to relay visual information to the brain. The
horizontal and amacrine cells facilitate lateral interactions between adjacent
photoreceptors.

There are two light-sensitive photoreceptors in the brain called the

rods and cones. Rods are specialized for operating at low light
levels, and cones are specialized for high visual acuity and the
perception of color under high light levels. A major route of
information flow from the retina to the optic nerve consists of a direct
pathway from the retinal photoreceptors to the bipolar cells to the
retinal ganglion cells. The two other types of neurons in the retina,
horizontal cells and amacrine cells, facilitate lateral interactions
between different photoreceptors. They primarily integrate
information at the bipolar and retinal ganglion cell layers. The lateral
interactions are mainly responsible for the visual system’s sensitivity
to changes in luminance contrast over different light intensities.
The macula is a small (5 mm diameter) structure located near the
center of the retina, and it processes clear color vision. Within the
macula, the fovea is a depression at the center of the macula that
provides the greatest visual acuity. The central portion of the fovea is
called the foveola; it provides the highest possible resolution to light.

14.3 Rods and Cones

Under well lighted (photopic) conditions, almost all our vision is
mediated by cones, but cones comprise only about 5% of the retinal
photoreceptors (~4.5 million). In contrast, rods contribute to human
vision at very low light levels (scotopic conditions), but they comprise
the vast majority of the photoreceptors in the retina (~90 million).
Despite this huge difference in numbers, the cones are crowded at
extremely high density into the fovea, and these cones provide high
visual acuity in the central area of vision.
Besides the differences in their distribution across the retina, rods
and cones have different shapes (see figure 14.3), photopigments,
and patterns of synaptic connections. The rod system has low spatial
resolution, but it is extremely sensitive to light. Cones, on the other
hand, have very high spatial resolution, but they are relatively
insensitive to light. The transduction mechanisms—through which
light energy is transformed into receptor potentials—also differ
between the rods and cones. Even a single photon produces a
reliable response in a rod, whereas more than 100 photons are
required to produce a comparable response in a cone.
Figure 14.3 Rods and cones are generally similar in structure, but they differ in
size and shape. They also differ in terms of how membranous disks are arranged.

The projections of rods and cones to bipolar cells and retinal

ganglion cells also contribute to the differences in their responses to
different light conditions. Rods and cones converge on the same
retinal ganglion cells in most parts of the retina, and the ganglion
cells respond differently based on the level of ambient light. There is
one fundamental difference between the pathways emanating from
the two photoreceptors, though: Cone bipolar cells directly contact
retinal ganglion cells, whereas rod bipolar cells contact amacrine
cells. The processes of these cells make synaptic contacts on the
dendrites of retinal ganglion cells. Each rod bipolar cell is also
contacted by many rods, and many rod bipolar cells contact a given
amacrine cell. The cone system is far less convergent. Thus, each
retinal ganglion cell receives information from only one cone but
many rods.
Because of this high degree of convergence, rods amplify signals
and are much better at detecting even dimly illuminated sources of
light. But at the same time, the convergence reduces the spatial
resolution of rods because the source of a signal in a rod bipolar cell
or in the retinal ganglion cell could have emanated from a relatively
large retinal surface area. Retinal ganglion cells transmit these
signals to the brain via their axons that make up the optic nerve.
The high density of cones in the fovea, along with their one-to-one
relationship with bipolar and retinal ganglion cells, provides the
cones with high spatial acuity along the line of sight. Foveal vision
corresponds to the central 3° of the field of vision and provides the
greatest spatial acuity for reading, driving, and perceiving the colors
and shapes of objects. Cone density declines rapidly away from the
foveal vision and drops by about 75% just 6° away (Purves et al.
2017). In contrast, the density of rods increases away from the
fovea. Thus, the threshold for detecting light stimuli away from the
central field of vision is lower; it is easier to detect an indistinct
source of dim light in the dark by looking away from the source of
light.
Cones also make it possible to perceive the colors of objects
based on the wavelengths of light that they reflect to the eye. There
are three types of cones with different photopigments (in contrast,
rods have only one kind of photopigment) that are sensitive to light of
different wavelengths. These cones are referred to as the blue
(short), green (medium), and red (long) wavelength cones. The blue
cones comprise about 5% to 10% of all the cones, and the ratio of
the other two cones varies among individuals. It can be anywhere
from 1:1 to 4:1 in favor of the green cones (Purves et al. 2017).
Because of these three different types of cones, human color vision
is trichromatic. However, many people suffer from color blindness, a
deficiency in perceiving colors that most people with trichromatic
vision can easily perceive. These people have dichromatic vision
that arises because of a loss of either green cones or red cones.
Loss of either of these cones makes it difficult to distinguish red and
green colors.

PROBLEM 14.1
A person with dichromatic vision will have difficulties performing
which common activities of daily living (and why)?
14.4 Optic Nerve, Tracts, and
Radiations
Retinal ganglion cells exit the retina through a region called the optic
disk in a bundle called the optic nerve. The optic nerve runs to the
optic chiasm, where more than half (~60%) of the fibers of the optic
nerve decussate and enter the contralateral cortex. The remaining
axons continue toward the lateral geniculate nucleus (LGN) of the
thalamus and other structures in the midbrain, such as the superior
colliculus. Once the axons go past the optic chiasm, they form the
optic tract. Because of the decussation, the optic tract contains nerve
fibers from both eyes. The thalamic neurons send axons to the
primary visual cortex through the optic radiations in the internal
capsule (see figure 14.4).
Figure 14.4 The visual pathway from the retina to the primary visual cortex.

If the retina or one of the optic nerves is damaged, a loss of vision

occurs that is limited to the eye of origin. Damage to the optic chiasm
results in deficits that involve the visual fields of both the eyes. The
visual field is the entire field of vision that can be seen without
moving the head when the eyes are fixated on a single point. Finally,
damage to the optic tract, LGN, optic radiation, and visual cortex
results in deficits that are limited to the contralateral visual hemifield.

PROBLEM 14.2
 If a person sustains an injury that leaves a complete lesion of the
left optic tract, how would that affect their vision? How could this
person compensate for the lesion of the left optic tract?

14.5 Striate Cortex

Visual information leaves the LGN and travels to the primary visual
cortex (V1, Brodmann area 17), the first area of the visual cortex,
that is in the occipital lobe of the cerebral cortex. V1 is also known as
the striate cortex. V1 is the cortical sensory region of the brain that
receives, integrates, and processes this information. V1 is divided up
into six distinct cortical layers. Layer 4 in V1 is the location that
receives neural information from the LGN. V1 neurons are
orientation selective—that is, they respond strongly to lines, bars, or
edges of a particular orientation (e.g., horizontal) but not to the
orthogonal orientation (e.g., vertical). They are also sensitive to the
direction of motion but not to its velocity.
Area V2 (secondary visual cortex) in the visual cortex receives
and integrates information from V1 (see figure 14.5) and encodes
more complex features of objects, such as color and orientation. It
sends feedback projections back to V1 and has feedforward
connections to visual areas V3 to V5. Areas V3 to V5 comprise the
extrastriate cortex (Brodmann’s areas 18 and 19) and are involved in
higher-order visual functions, such as motion and color processing.
Most importantly, the visual information leaving area V2 splits into
two separate streams of information processing.
Figure 14.5 Area V2 splits visual information processing into two streams. The
first stream (dorsal stream) that projects toward the parietal cortex is shown by the
solid black lines, and the second stream that projects toward the temporal cortex
(ventral stream) is shown as the dashed black line. MT = middle temporal area;
MST = medial superior temporal area; LGN = lateral geniculate nucleus.

Most neurons in these early stages of the visual system are

monocular—that is, they only carry information from one eye. Even
in the LGN, inputs from the ipsilateral and contralateral eye terminate
on separate layers. This causes the thalamic neurons to also be
largely monocular. Even in the earliest stages of V1 in layer 4 of the
cerebral cortex, the input from both the eyes remains segregated.
Beyond layer 4, in layers 5 and 6, neurons receive input from both
the eyes and are consequently binocular (Purves et al. 2017). These
binocular neurons in V1 provide the sensation of depth in three
dimensions, called stereopsis (see figure 14.6). Stereopsis arises
because each eye sees an object from a slightly different angle. The
differences between those images are integrated in V1 and cause
the visual system to perceive depth. This is the basis of binocular
depth perception. There are also monocular mechanisms with which
humans perceive depth, but those are beyond the scope of this
textbook.
Figure 14.6 The relative movement of the binocular images on the retina is the
basis of stereopsis. When the object is closer, the image of the object is displaced
more laterally, and vice versa. This provides the basis for depth perception.

14.6 Retinotopic Organization

of V1
The visual space is mapped on V1 in terms of a retinotopic map. The
central (foveal) part of the visual field is represented in the very
posterior part of V1, and more peripheral regions of the visual field
are represented relatively anteriorly. The retinotopic map is
lateralized so that the left hemisphere V1 represents the right half of
the visual field and vice versa. In addition, central vision (~10º of the
central visual field) occupies about half of V1. This suggests that the
visual cortex dedicates a large amount of its neural machinery to
process the fine details of the space where the gaze is currently
fixed. The remaining approximately 170º of visual space is
processed by the other half of V1, and consequently we have poor
spatial resolution in the visual periphery.

PROBLEM 14.3
A person was born with congenital damage to the right retina.
Which normal visual function would this person have a difficulty
performing?

14.7 Extrastriate Cortex

Anatomical and physiological studies in monkeys have revealed
areas in the occipital, parietal, and temporal lobes that also contain
maps of the visual field and receive inputs from V1. In contrast to
cells in V1, cells in the extrastriate cortex specialize in specific
aspects of visual processing. For example, cells in the middle
temporal area (MT) respond selectively to the direction and velocity
of motion of a moving edge but are insensitive to color. In humans,
area MT and the medial superior temporal (MST) areas encode
direction and velocity of motion (Dukelow et al. 2001). In contrast to
motion-sensitive cells in MT and MST, neurons in area V4 of the
monkey respond only to the color of a visual stimulus. In humans,
area V8 is considered to be the homolog of area V4 in monkeys that
is responsible for color detection (Hadjikhani et al. 1998).
Extrastriate cortical areas are organized into two almost
independent systems that send neural information into association
areas in the temporal and parietal lobes (see figure 14.7). The first
stream is called the dorsal stream, which includes the middle
temporal area; it begins in the striate cortex and extends into the
parietal lobe. This stream is considered to be responsible for spatial
aspects of vision, such as the analysis of motion, positional
relationships between objects in the visual scene, and visuomotor
functions. The second system is called the ventral stream, and it
includes area V4. This stream begins in the striate cortex and
extends into the inferior part of the temporal lobe. The ventral stream
is considered to be responsible for high-resolution form vision and
object recognition.

Figure 14.7 The spatial and perceptual aspects of vision are processed by
different but interdependent streams. The figure shows the interconnections
between the different nodes of the dorsal stream (solid black) and the ventral
stream (dashed black). SPL = superior parietal lobule; IPL = inferior parietal lobule;
MT = middle temporal area; MST = medial superior temporal area; LGN = lateral
geniculate nucleus; FFA = fusiform face area; IT = inferotemporal cortex.

PROBLEM 14.4
A healthy adult slipped on ice and suffered a concussion. The
patient later tells the family doctor about experiencing difficulties
at four-way intersections with two-way stop signs. They cannot
gauge how fast a car is approaching to determine if it is safe to
pull through the intersection. Which area of the extrastriate cortex
most likely was affected by the concussion?
14.8 Neurons of the Two Visual
Streams
Layers in the LGN receive input from individual retinas from each
eye. The layers also differ based on the cell size of the neurons. In
each LGN, the two ventral layers are primarily composed of large
magnocellular neurons (M cells), and the four dorsal layers are
composed of the relatively smaller parvocellular neurons (P cells).
These neurons receive inputs from different retinal ganglion cells
that differ in morphology. The axons of M ganglion cells have larger
diameters, and the cells have more extensive dendritic trees and
larger cell bodies. M ganglion cells also have larger receptive fields
and faster conduction velocities. Besides these morphological
differences, M cells generate phasic responses to presentations of
visual stimuli, whereas P cells produce tonic sustained responses. P
ganglion cells are sensitive to differences in wavelengths of light in
their receptive fields, and consequently they respond to color. In
contrast, M cells do not respond to the color of the stimulus.
The M and P cells give rise to the magnocellular and parvocellular
pathways, respectively. The magnocellular pathway forms the dorsal
ventral stream, and the parvocellular pathway forms the ventral
visual stream (see figure 14.8). The magnocellular pathway carries
information about large and fast-moving objects. It has low spatial
resolution but high temporal resolution, and it is agnostic to the color
of the stimulus. The parvocellular pathway carries information about
small objects and also encodes their color and has high spatial
resolution and low temporal resolution.
Figure 14.8 The neuronal connections between the retina and the parvocellular
and magnocellular layers of the lateral geniculate nucleus (LGN). The
parvocellular and magnocellular neurons are the bases of the ventral and dorsal
visual streams, respectively.
Adapted by permission from C. Meissirel, K.C. Wikler, L.M. Chalupa, and P. Rakic, “Early
Divergence of Magnocellular and Parvocellular Functional Subsystems in the Embryonic
Primate Visual System,” Proceedings of the National Academy of Sciences of the United
States of America 94, no. 11 (1997): 5900-5. Copyright (1997) National Academy of
Sciences, U.S.A.

PROBLEM 14.5
A monkey sustained an injury to layers 4 through 6 of the
thalamus. Which visual function or functions are most likely to be
affected by this injury?
14.9 Visual Deficits Due to
Area-Specific Visual System
Damage
Visual deficits can occur due to damage to the retina, optic nerve,
tracts, radiations, or the magnocellular or parvocellular pathways.
Damage to the retina causes macular degeneration, which results in
loss of central vision. Age-related macular degeneration is the
leading cause of severe, permanent vision loss in people over age
60. Glaucoma is caused by high pressure in the eye, and it results in
damage to the optic nerve. It is also one of the leading causes of
blindness for people over age 60.
Damage to the primary visual pathway can cause specific visual
field deficits, called anopsias. Damage to one of the optic tracts of
one eye causes a complete loss of sight in the contralateral visual
field. This is referred to as homonymous hemianopsia. In contrast,
damage to the optic chiasm causes loss of bilateral vision in the
lateral visual fields. This is because about half of the optic nerve
fibers from each eye that map the lateral visual field decussate at the
optic chiasm. This bilateral visual field deficit is called heteronymous
hemianopsia. Similarly, damage to the optic radiations causes visual
deficits in one half of the contralateral visual hemifield (see figure
14.4). For example, damage to optic radiations will result in loss of
vision in the inferior contralateral visual field. This is called
homonymous quadrantanopsia.
Cerebral akinetopsia is a rare disorder that occurs because of
damage to the motion-processing area MT/MST along the
magnocellular pathway. Patients suffering from this disease have
difficulty crossing the street because they are unable to perceive the
motion of an oncoming vehicle. Cerebral achromatopsia occurs
because of damage to areas V4 and V8 along the parvocellular
pathway. Patients suffering from this disease are unable to see the
world in color, though most other aspects of their vision remain
intact.

14.10 Ocular Movements

Eye movements were extensively studied by the great German
physicist and physiologist von Helmholtz in the 19th century. In
particular, von Helmholtz noticed that voluntary eye movements were
associated with a feeling of a stationary visual field, while eye shifts
created “artificially” (for example, when you press on your eyeball
with a finger) led to a sensation of a moving environment. These
observations led him to a hypothesis on an important role of
oculomotor commands for visual perception.
Each eye is controlled by six muscles (see figure 14.9). These
muscles involve four recti muscles (superior, inferior, lateral, and
medial) and two oblique muscles (superior and inferior) that can
produce eye rotations around three major axes: adduction/abduction
(horizontal movements), elevation/depression (movements up and
down), and intorsion/extorsion (rotation without changing the
direction of gaze). The levator palpebrae superioris is a small muscle
that moves the upper eyelid but is not involved in controlling eye
movements.
Figure 14.9 The six extraocular muscles controlling the human eye.

Eye movements can be broadly divided into three categories:

conjugate, disconjugate, and vestibulo-ocular. Conjugate eye
movements imply that both eyes move in the same direction
(left/right or up/down) and at the same time. Saccades and smooth-
pursuit eye movements belong in this category. Disconjugate eye
movements are when the eyes move in opposite directions
(converge or diverge) along the horizontal axis to stabilize gaze on a
moving target toward or away from the body. Vergence eye
movements belong in this category. These eye movements are
critical for depth perception. Finally, vestibulo-ocular movements
stabilize the eyes on an object fixed with respect to the external
world, while the head moves. Thus, these eye movements are
reflexive and respond to head movements to prevent visual images
from “slipping” on the surface of the retina.
 Saccades are the most frequent voluntary eye movement that
humans make during everyday life. These eye movements rapidly
redirect gaze to objects in the environment by bringing the image of
an object onto the fovea (the central part of the retina). We make
saccades during almost every activity of daily living, and it is widely
believed that on average, humans make about three saccades every
second. Saccades can be suppressed when we are trying to focus
on something very small, such as when we are trying to thread a
needle or playing darts. Saccades are elicited both voluntarily, such
as when we are actively searching for an item, or reflexively, such as
when a bright flashy source of light comes on in our peripheral
vision. Saccadic eye movements are ballistic, and the fastest eye
movements can be up to 800°/s to 1000°/s. Actual peak velocity
depends upon the angular distance between the objects and is
higher for larger distances. Humans can voluntarily control the
direction and amplitude of saccade movements but not their velocity.
Saccades are separated by fixations on objects. These fixations
can last anywhere from 50 ms to 500 ms (or more), depending on
how many visual features need to be extracted. For example, if we
are detecting the shape of an object (circle versus rectangle), we
might only fixate on it for less than 100 ms, but if we are reading a
word, we might fixate on it for approximately 300 to 500 ms.
Smooth-pursuit eye movements are voluntary but much slower
eye movements than saccades that go up to 80°/s to 100°/s (Meyer
et al. 1985) to keep the images of moving stimuli on the retina.
These eye movements are engaged when we track a tennis ball in
play or when we follow the trajectory of a basketball in flight.
Previously, moving stimuli were considered essential for smooth-
pursuit eye movements to occur. However, there is evidence that
smooth-pursuit eye movements can also be made over short
distances in the absence of moving stimuli (Becker & Fuchs 1985).
These eye movements are considered important for forming visual
predictions about moving objects and for guiding interceptive arm
movements toward those objects, such as while playing ball sports
(Spering et al. 2011). Oftentimes when pursuing a fast-moving
object, smooth-pursuit eye movements are accompanied by short
catch-up saccades to land the gaze on the object and then pursue it.
The neural circuitry responsible for smooth-pursuit and saccadic
eye movements is fairly well characterized. These eye movements
are controlled by similar networks of cortical and subcortical regions
and, in some cases, by the same set of neurons (Krauzlis 2005). The
overlap in the functional architecture of these eye movements is
shown in figure 14.10. However, there are some differences. Area
MT/MST are involved in the control of pursuit eye movements, but
not in the control of saccadic eye movements.
Vergence eye movements are disconjugate eye movements that
align the fovea for each eye on targets that move toward or away
from the person. These eye movements involve rotation of the eyes
in opposite directions and can be made as either ballistic (saccades
at different depths) or as slow eye movements (pursuit of targets
moving toward or away from the body).
Figure 14.10 Sensory and motor neural substrates involved in saccadic (black
arrows) and smooth-pursuit (gray arrows) eye movements. SPL = superior parietal
lobule; FEF = frontal eye fields; MT = middle temporal area; MST = medial
superior temporal area; CN = caudate nucleus in the basal ganglia; SNr =
substantia nigra pars reticulata; PMN = brainstem premotor nuclei; PON = pontine
nuclei; SC = superior colliculus.

The main purpose of eye movements facilitated by the vestibular

system is to retain the image of an object on the retina while the
head is moved, either voluntarily or involuntarily. These eye
movements are engaged when our gaze is fixated in space while our
head moves up/down (e.g., when we nod in agreement while looking
at someone) or left/right (e.g., when we shake our head in
disagreement). The first type of eye movement initiated by the
vestibular system is the vestibulo-ocular reflex (VOR). This reflex
acts at a relatively short latency (about 14 ms) during head
movements. It changes the activity of ocular muscles and induces
motion of the eyes during head rotations so that the retinal image of
the visual field is stabilized. If head rotation continues after the eyes
reach the edge of the orbit, the eyes do not stay at the extreme
position but rapidly reverse the direction of their movement and jump
quickly back. The entire phenomenon is called vestibular nystagmus;
the rapid reversal is called the quick phase. The VOR is coordinated
at the brainstem level. Vestibular receptors send signals about the
velocity of head rotation to vestibular nuclei that project on the
oculomotor nuclei. The cerebellum participates in modulation of the
gain of the VOR.

PROBLEM 14.6
Damage to the dorsal visual stream in the parietal lobe will affect
which type of eye movement?

The second vestibular mechanism is the optokinetic system,

which uses visual information to complement the VOR. It also acts to
keep stable features of the environment (faces, buildings, etc.) on
the fovea during head movements. Its latency is longer than the
VOR, and it takes time to build up. The loop of this reflex, called the
optokinetic reflex, involves both cortical and subcortical structures.
Both reflex systems demonstrate habituation and change their gain
adaptively—for example, if a person is wearing glasses.

CHAPTER 14 IN A NUTSHELL
Photoreceptors transform light falling
onto them into action potentials.
There are two types of photoreceptors,
rods and cones. Rods are responsible
for vision at low light levels. Cones
are densely packed in the fovea, and
they are active at higher light
levels. There are three types of
cones: short (blue), medium (green),
and long wavelength (red) cones. These
cones mediate trichromatic color
vision. A loss of green or red cones
contributes to color blindness. The
retinal ganglion cells carry action
potentials to the lateral geniculate
nucleus in the thalamus. From the
thalamus, visual information reaches
the primary visual cortex (V1).
Extrastriate cortical areas receive
information from V1 and are organized
into two independent systems, the
dorsal and ventral visual streams,
that specialize in visuomotor and
visuoperceptual processes,
respectively. The eyes make different
types of movements to gather
information from the visual scene and
to keep the image of an object on the
retina. Saccades are ballistic eye
movements used to gather information.
Smooth-pursuit eye movements are
slower eye movements that track moving
objects to keep their image on the
retina. Vergence eye movements involve
rotation of the two eyes in opposite
directions to view objects in depth.
Problems for Part III
Self-Test Problems
1. High-frequency muscle vibration typically leads to illusions of
motion in a joint spanned by the muscle. Most frequently, the
illusory movement corresponds to muscle elongation. Under
some conditions, however, the direction of the illusory
movement reverses. Offer an interpretation for these
observations.
2. If a person sustains injury to the left medial lemniscus, they
will lose tactile sensations on which side of the body?
3. A person reports “unstable bouncing vision” and an inability to
read road signs during walking. What is this person suffering
from?
4. A person demonstrates normal ocular reflexes and no
changes in the peripheral eye structures. What pathology can
you suspect if the person reports no visual experiences
(blindness)?
5. A person has suffered a lesion in the left middle temporal
areas of the brain. What ability have they most likely lost?

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. Someone’s spouse has bought them a piece of clothing as a
surprise birthday gift. They ask their spouse to guess the
fabric. The spouse gently grasps the clothing with their
fingertips and moves them over the clothes. Which
mechanoreceptor helps them identify the fabric?
a. Merkel disks
b. Meissner corpuscles
c. Ruffini endings
d. Pacinian corpuscles
e. Free nerve endings
 Why?
2. Which cortical areas either send or receive direct projections
to or from the primary somatosensory cortex (S1)?
a. secondary somatosensory cortex (S2)
b. parietal cortex
c. thalamus
d. both a and b are correct
e. a, b, and c are correct
Why?
3. Damage to the vestibular system does NOT produce which of
the following symptoms?
a. poor balance
b. reduced postural stability
c. inability to perceive motion
d. difficulty visually fixating on objects and road signs during
walking
e. rotation of eyes during head movement
Why?
4. If a person suffers damage to the left optic nerve, which
normal visual function are they most likely to lose?
a. color vision
b. stereopsis
c. motion perception
d. face perception
e. ability to read fine text
Why?
5. If a person suffers damage to the right optic nerve, which part
of the visual field won't they be able to see?
a. extreme left
b. central left
c. central right
d. extreme right
e. extreme left and right
Why?
 Part IV

Reflexes and Reflex-Like

Movements
 Chapter 15

Reflexes

KEY TERMS AND TOPICS

reflex
voluntary movements
general scheme of a reflex
reflex arc
latency
gain
conditioned reflexes

In the beginning of the century, muscle reflexes were considered

building blocks for voluntary movements. This view was based on a
very impressive series of experiments on spinalized animals
performed by Sir Charles Sherrington and his colleagues (for a
review see Stuart et al. 2001). These studies suggested, in
particular, that locomotor movements in dogs could represent
sequences of reflexes to paw stimulation during contact with the
ground—a theory that later was shown to be wrong (see chapter 26).
The notion of reflexes as building blocks for animal and human
behavior was also developed by I.P. Pavlov (reviewed in Bernstein
2003; see also Meijer 2002), who viewed behavior as a set of
complex input–output relations between sensory signals and motor
acts. His experiments and theoretical views are described in section
34.4. Later, the growing awareness of the complexity and variability
of voluntary movements led N.A. Bernstein (1967; see also
Bongaardt 2001) to conclude that movements could not be viewed
as combinations of reflexes. Ultimately, this led to another extreme
when muscle reflexes started to be considered relatively
insignificant, nearly atavistic mechanisms that play an important role
only when a movement gets off its planned track. Recently, muscle
reflexes have once again been brought to the center of attention by
hypotheses of motor control, which consider voluntary movements to
be consequences of central modulation of parameters of certain
reflexes (reviewed in Feldman 2015).

15.1 Definition of a Reflex

The noun reflex means a phenomenon that reflects or is a
consequence of another phenomenon. Correspondingly, a muscle
reflex is a muscle contraction induced by an external stimulus.
However, many actions are induced by external stimuli that are not
considered reflexes. For example, if a driver sees a red light, he or
she presses the brake pedal. Is this a reflex? Most people would
probably argue that if the driver presses the pedal automatically,
“without thinking,” this is a reflex (remember, there are drivers with
“good reflexes” and with “bad reflexes,” which commonly means
quicker or slower reaction to the red light). If the same driver saw the
light from far away, and then decided to brake, this is probably not a
reflex but a voluntary action. This example shows that it may be
rather difficult to formally distinguish a reflex from a voluntary
movement. The driver example suggests that there may be a
continuum of reactions to sensory stimuli ranging from reflexes to
voluntary movements.
Although the term reflex has been used in studies of movements
and movement disorders since the time when it was introduced by
French physiologist Jean Astruc (1684–1766), it is still discussed
actively. See, for example, a discussion written by several prominent
researchers in the field of movement neuroscience (Prochazka et al.
2000), which presents a range of opinions from “reflex is a
misnomer, and this word should be avoided” to “reflex is a useful
concept well-grounded in neurophysiology.” Before introducing a
definition for reflex, let us consider a number of characteristics of
responses to sensory stimuli that distinguish between actions that
would almost unanimously be classified as reflexes and those that
would almost unanimously be classified as voluntary movements
(figure 15.1).
The first characteristic, which is obvious in the aforementioned
examples of the driver pressing the brakes, is the time between the
stimulus and the action. Reflexes are responses that take little time,
whereas voluntary movements require more time to be initiated. The
second, related characteristic is the neural chain involved in the
generation of an action. Reflexes are commonly viewed as more
simple in nature, involving “short circuits” through the central nervous
system, which naturally leads to shorter time delays between the
stimulus and the action. In contrast, voluntary movements may use
longer, more complex chains of neurons.

Figure 15.1 Reflexes and voluntary actions differ in a number of their

characteristics. These differences, however, are usually quantitative, not
qualitative.

The next couple of characteristics are less universally applicable.

Reflexes are usually viewed as local (i.e., involving an action close to
the location of the stimulus). For example, a tap on the patellar
tendon induces a knee jerk of the same leg—a proverbial reflex. In
contrast, voluntary movements can involve muscles, joints, and
extremities that are not directly affected by the stimulus. Along
similar lines, reflexes are typically viewed as less variable or more
stereotypical: Applying the same stimulus is expected to induce very
similar motor responses. In contrast, variability is one of the most
universal features of voluntary movements, even in conditions when
movements are produced in response to a sequence of similar
sensory stimuli applied when the initial states of the body are also
similar.

PROBLEM 15.1
Suggest examples of actions, commonly addressed as reflexes,
that violate the suggested criteria.

Finally, reflexes, by definition, are always induced by a sensory

stimulus. In contrast, voluntary movements can happen with a
sensory stimulus or without a clear external stimulus. Figure 15.1 is
purposely simplified, and it is easy to find counter-examples to all the
aforementioned contrasts between reflexes and voluntary
movements. Some of these contrasts point at reactions to sensory
stimuli that are located somewhere in between the two extremes. We
discuss them in a future chapter. Other examples suggest that there
is no strict border between these two classes of movements.
Nevertheless, we are going to assume that the concept of reflex is
useful and will try to use it as cautiously as possible.
For the purposes of this textbook, let us accept the following
definition of muscle reflex and stick with it, even when it apparently
fails: A reflex is a muscle contraction induced by an external stimulus
that cannot be changed by “pure thinking” (i.e., by a volitional act or
a change in the instruction to the subject that is not accompanied by
another muscle contraction). This definition is far from perfect, but at
least it provides a starting point.
 Many reflexes have been studied in animal preparations (i.e., in
animals whose central nervous system had been damaged by an
experimental procedure). Such procedures frequently induce
surgical separation of the spinal cord from the brain, in which the
animal is then identified as spinalized or as a spinal preparation. If
one assumes that every volitional act comes from the brain, all
muscle reactions to external stimuli in a spinalized animal are
obviously reflexes because signals from the brain cannot reach the
spinal cord.
Note that figure 15.1 does not include a “Spinal cord–Brain” axis.
This is done on purpose. Indeed, there are animals whose nervous
systems represent groups of neurons (neuronal ganglia) that are not
yet organized into structures typical of those in higher animals, such
as mammals. These animals have neither spinal cord nor brain. But
they can show typical reflexes to sensory stimuli and actions not
associated with any obvious sensory stimuli, which, for the lack of a
better term, can be classified as voluntary (spontaneous may be a
better term). For example, a slug would cringe when pricked by a
sharp object—a typical reflex. But in the absence of clear external
stimuli, it would not stay motionless forever but would crawl
somewhere—for example, in search of food—a spontaneous action
that can be addressed as voluntary.

PROBLEM 15.2
What about plants? Their parts (e.g., roots, branches, and fruits)
grow and change location in space (move!). Can these
movements be classified into reflex and voluntary?

15.2 Reflex Arc, Gain, and

Latency
A central notion common to all muscle reflexes is that of the reflex
arc (figure 15.2). A reflex arc consists of an afferent neuron (or, more
commonly, a number of neurons) with a sensory ending (receptor)
that senses an external stimulus, central processing neural
structures, and an efferent neuron that induces a muscle contraction.
The central processing neural structures may be very simple,
consisting of just one synapse on a neuron that induces muscle
contraction, or very complex, involving numerous synapses and
integration of information from different sources.

Figure 15.2 A reflex arc consists of a sensory ending (receptor), an afferent

(sensory) nerve, central processing structures, an efferent (command) nerve, and
an effector (e.g., a muscle).
Figure 15.3 The delay between a stimulus and a reflex reaction is called reflex
latency. It consists of an afferent conduction time (ΔTAF), a central delay (ΔTC),
and an efferent conduction time (ΔTEF).

Each reflex involves a time delay between a stimulus and a

reaction. This time delay is called reflex latency (figure 15.3). It
consists of three components: time of afferent conduction, central
delay, and time of efferent conduction. Conduction time apparently
depends on the speed of action potential propagation along the
involved neural fibers and the length of the fibers. For example, a
myelinated, fast-conducting neural fiber from a sensory ending in the
foot to the spinal cord (about 1 m at about 70 m/s) will take about 15
ms. An efferent fiber from an α-motoneuron to innervated muscle
fibers will take about the same time. Conduction delays from more
proximal sensors and to more proximal muscles are proportionally
shorter. These delays obviously depend on anthropometric
characteristics of the person (e.g., on the length of one’s arms and
legs). Central delay depends mostly on the number of synapses
involved in processing the afferent volley and generating an efferent
command. The simplest reflexes have a single central synapse with
the central delay of about 0.5 to 1 ms. An increase in the number of
synapses leads to a proportional increase in the central delay.

PROBLEM 15.3
What is missing in the suggested definition of reflex latency as the
sum of three components?

Another important characteristic of a reflex response is its gain (G)

defined as the ratio of the response magnitude (MR) to stimulus
magnitude (MS): G = MR/MS. While delays are relatively easy to
measure and compare across muscles, types of reflexes, persons,
and species (they are always measured in time units), gains are
more complex. Their values depend on physical units that are used
to measure the magnitude of the stimulus and the magnitude of the
reflex response. For example, the well-known tendon tap reflex is
induced by a stimulus causing changes in muscle force, length, and
velocity in addition to deformation of the skin and subcutaneous
tissues. Each of those can potentially be used to define MS. It
produces changes in muscle activation (units of electrical potential),
force (newtons) and its derivative (newtons/second), muscle
shortening (meters) and its derivatives (meters/second), and joint
angle changes (radians and radians/second). Any of these variables
can be used to define MR. To compare reflex gains across conditions
or populations (for example, healthy persons and patients with an
injury to the central nervous system), one has to agree on units of MS
and MR to make observations comparable.
There are a few additional characteristics used to describe
reflexes. In particular, most reflex responses become smaller during
successive applications of the same stimulus. This phenomenon,
known as response attenuation or habituation, can be seen across
a spectrum of reflexes, from the simplest to relatively complex ones
such as, for example, the startle reflex, which is seen in response to
an unexpected very loud sound. Response habituation is defined, in
particular, by the fact that the presynaptic fiber in a neuro-neural
synapse releases smaller amounts of neurotransmitter in response
to successive presynaptic action potentials because the original
amount of the neurotransmitter in the vesicles takes time to restore.
Note also that neuro-neural synapses are weak and, as a result,
even a small decrement in the amount of released neurotransmitter
can make a difference between reaching the activation threshold on
the postsynaptic membrane and failing to reach the threshold. That
is why, during neurophysiological studies of reflexes, there are
obligatory rest intervals between successive stimuli. Rate of
habituation can be quantified as the decrement of the responses to
successive stimuli presented at fixed time intervals. As illustrated in
figure 15.4, the drop in the response amplitude is typically
exponential and can be described with a parameter related to the
quickness of the drop (τ in figure 15.4).

Figure 15.4 Application of a sequence of stimuli commonly leads to a drop in the

amplitude of the induced reflex response, which is described as habituation or
attenuation of the response. It commonly follows an exponential drop with a
parameter τ.

15.3 Reflex Classifications

The simplest reflexes involve only one central synapse (in addition to
the neuromuscular synapse, which is obviously involved in any
muscular contraction). These reflexes are called monosynaptic
(figure 15.5a). In this book, we will consider only one type of
monosynaptic reflexes (see chapter 17). Note that monosynaptic
projections from afferent fibers to α-motoneurons contribute to
muscle activation during other reflexes and voluntary movements. In
most situations, however, these projections are not strong enough to
produce muscle contractions by themselves. The term monosynaptic
reflex will be used for muscle contractions induced primarily or
exclusively by projections of sensory afferents to α-motoneurons.
Reflexes that involve two to three central synapses are called
oligosynaptic (figure 15.5b). A number of oligosynaptic reflexes have
been described in mammals, including humans. Reflexes that
involve many synapses are called polysynaptic. Unlike monosynaptic
and oligosynaptic reflexes, the neuronal loops of polysynaptic
reflexes are typically unknown. Hence, the central part of these
reflexes may be viewed as a black box with unknown connections
(as in figure 15.2).
Another classification is based on typical time profiles of the motor
response to an adequate stimulus. If the response is short-lasting
and only to a change of the stimulus magnitude, not to its steady-
state magnitude, such a reflex is known as phasic. If the response
continues for a long time following a change in the magnitude of a
stimulus, such a reflex is addressed as tonic. For example, imagine
that a quick stimulus is applied to a muscle—for example, it is
stretched quickly and held at the new length for some time (figure
15.6). Its response—for example, a change in its level of activation
or force—can be seen in response to the stretch process as well as
some time after the stretch, at the new length. This illustrates both
phasic and tonic components of the reflex to stretch (or simply
stretch reflex).
Figure 15.5 (a) Monosynaptic reflexes have only one synapse in their reflex arc.
This synapse is between an afferent fiber and an α-motoneuron. (b) Oligosynaptic
reflexes have one or two interneurons (IN) in the reflex arc. They can be inhibitory
or excitatory.

Figure 15.6 A quick increase in a stimulus can produce both phasic and tonic
reflexes. Phasic reflexes are transient. Tonic reflexes are steady state.

Finally, reflexes are classified into excitatory and inhibitory.

Assuming a nonzero level of activation of a muscle, a stimulus can
lead to an increase or a drop in its activation level. Note that
projections of afferent fibers to neurons in the spinal cord are always
excitatory. This means that monosynaptic reflexes can only be
excitatory, but oligosynaptic and polysynaptic reflexes can involve
inhibitory interneurons and can produce both excitation and
suppression of muscle activation level.

PROBLEM 15.4
What would you expect to see in a person lacking inhibitory
neuromediators in the spinal cord?

15.4 Conditioned Reflexes

Different animals are born with different sets of reflexes and motor
abilities. As described in more detail in chapter 33, some animals are
capable of performing basic functional movements very soon after
birth. Others, including human babies, are quite helpless. However,
they demonstrate several basic reflexes that are crucial for survival,
including the sucking and swallowing reflexes as well as the
grasping reflex. Newborns are also able to demonstrate stepping if
they are held in the air and their feet touch a moving treadmill
(Dominici et al. 2011; Ivanenko et al. 2013; Sylos-Labini et al. 2020).
Some of the reflexes in newborns show atypical patterns, partly due
to the continuing process of myelinization over the first months
following birth.
A very impressive series of experiments were performed by Ivan
Pavlov (1849–1936, Nobel Prize winner in 1904) and his colleagues
in Russia in the late 19th and early 20th centuries (reviewed in
Asratyan 1953). In those studies, Pavlov explored the salivation
reflex in dogs. This rather complex reflex induces salivation when a
hungry animal sees or smells food. In Pavlov’s studies, the dogs
salivated naturally when they saw an experimenter entering the room
with their feeding bowl in the morning. For several days, a bell rang a
few minutes before the feeding time, and the dogs started to salivate
when they heard the bell, even before they could see or smell food.
Such new reflexes have been termed conditioned reflexes in
contrast to the inborn reflexes.
If we assume that the original salivation reflex involved a
polysynaptic reflex loop originating from light- and smell-sensitive
receptors in the retina and in the nose, respectively (figure 15.7),
Pavlov’s experiments show that the loop could be modified and
could involve new receptors, those sensitive to the sound produced
by the bell. Pavlov and his students explored properties of such
conditioned reflexes and came up with a theory that any behavior, no
matter how complex it is, could represent a combination of inborn
and conditioned reflexes.

Figure 15.7 Conditioned reflexes emerge when a natural stimulus for a particular
response is replaced with another stimulus as a result of multiple presentations of
the second stimulus. This can happen for the salivation reflex, which is seen
naturally in response to seeing or smelling food, if food presentation is preceded
by a stimulus of a different modality (e.g., a bell ring).

PROBLEM 15.5
Imagine that you developed a typical Pavlovian conditioned reflex
and then rang the bell a few times without bringing food to the
dog. And then, once again, ring the bell and bring food at about
the same time. What will happen with salivation?
 Pavlov’s views dominated the field for many years and were
considered ultimate truth in the Soviet Union. The only person who
dared to criticize Pavlov openly was Nikolai Bernstein (2003).
According to Bernstein, reflexes played the very important role of a
foundation for functional movements, but behaviors were initiated
primarily within the central nervous system, not in response to
sensory stimuli. Based on this “principle of activity,” Bernstein
created his multilevel control scheme for biological movements
described in chapter 1. During the later years of his life, Bernstein
(1966) started to sketch a new field in physiology that he termed
physiology of activity. Unfortunately, this field remains barely
developed.

CHAPTER 15 IN A NUTSHELL
Reflex is a loosely defined notion
meaning an involuntary motor reaction
to an external stimulus. The simplest
reflex consists of a receptor (sensory
ending), an afferent nerve, at least
one synapse on a neuron within the
central nervous system, an efferent
nerve, and an effector. Three main
characteristics of reflexes are time
delay (latency), gain, and
habituation. Monosynaptic reflexes
contain only one central synapse; they
always induce excitatory effects on α-
motoneurons. Oligosynaptic reflexes
contain two or three synapses within
the central nervous system. The neural
loop of polysynaptic reflexes is
typically unknown. Reflexes can induce
short-lasting reactions (phasic
reflexes) or steady-state changes in
muscle activation (tonic reflexes).
They can induce excitation or
inhibition of baseline muscle
activity. Some reflexes are seen at
birth. Conditioned reflexes represent
responses to an unusual stimulus that
replaces the natural one as a result
of repetitive presentation of both
stimuli. According to Pavlov’s theory,
conditioned reflexes form the basis of
all animal behaviors. This theory was
criticized by Bernstein, who suggested
that voluntary actions could be
initiated without any external
stimulus by the central nervous system
(the principle of activity).
 Chapter 16

Excitation and Inhibition

Within the Spinal Cord

KEY TERMS AND TOPICS

anatomy of the spinal cord
excitation
postsynaptic and presynaptic inhibition
recurrent inhibition
Renshaw cells
reciprocal inhibition
Ia interneurons
persistent inward currents

Before moving to analysis of various reflexes, we have to consider

how different neurons interact with each other within the spinal cord
and with signals supplied by afferent fibers from peripheral
receptors. The wiring of the spinal cord is rather complex, and only a
tiny fraction of all the connections have been deciphered with any
degree of certainty.
Figure 16.1 The spinal cord has a laminar structure. It “flows” along the body,
preserving the general picture of its cross-sections. At each level, the gray matter
forms a characteristic butterfly picture consisting of 10 Rexed’s laminae (shown in
the rightmost drawing).

16.1 The Spinal Cord

The spinal cord has a laminar structure (figure 16.1). This word
means that it more or less “flows down” from the brain without abrupt
changes in its cross-sectional structure, like crossings-over or other
discontinuities. The meaning of “lamina” is very close to that of
“layer.” If the spinal cord is transected at an intermediate level, a
characteristic butterfly-like picture can be seen. The “butterfly”
consists of gray matter (mostly bodies of the spinal neurons), while
the rest of the section is white matter (mostly neural tracts
transmitting information to and from the brain). At each level, the
same areas (laminae) can be identified in the gray matter. These
areas are named after a Swedish neuroscientist, Bror Rexed (1914–
2002), and are numbered with Roman numerals from laminae I to
laminae X.

PROBLEM 16.1
What is the antonym of “laminar”? What could be the origin of the
laminar structure of the spinal cord?

It is necessary to introduce a few “geographical” terms that will

help in future chapters as well (figure 16.2): Dorsal means oriented
toward the back of the body, ventral means oriented toward the
stomach, rostral means closer to the head, caudal means closer to
the tail (or what is left of the tail in humans), medial means closer to
the center, lateral means closer to a side, proximal means close to
an origin of coordinates (commonly, the trunk), distal means far from
the origin of coordinates. Three orthogonal planes are commonly
identified with respect to the human body. These are the frontal
plane, which means a plane parallel to the mirror into which a person
looks when combing his or her hair; sagittal plane, which means a
plane parallel to the plane of arm and leg movements during normal
walking; and coronal plane, which means a plane perpendicular to
the spine and parallel to the ground if one is standing upright (figure
16.2).
Figure 16.2 These terms will be helpful in future descriptions. The figure of a
person is drawn in a sagittal plane.

The spinal cord is a very important part of the central nervous

system; an injury to the spinal cord may lead to complete and
irreversible paralysis. It is protected from possible damaging
influences by the spine. The spine consists of vertebrae (bone
structures) separated from each other by elastic spinal disks
(cartilage structures). This construction provides for flexibility of the
spine, resistance to possible compression forces, and protection of
the spinal cord. Each vertebra has two pairs of horns, which are
neural structures, the dorsal horns (closer to the back) and the
ventral horns (closer to the stomach or chest) (figure 16.3). The
dorsal horns serve as input paths for information from peripheral
receptors. Remember that bodies of the receptor cells are located in
spinal ganglia, just outside the spinal cord. These cells have T-
shaped axons whose distal branches travel to the sensory endings,
somewhere in the periphery, while the proximal branches enter the
spinal cord through the dorsal horns. Axons from numerous
peripheral receptors form a dorsal root and come together through
the same dorsal horn. The ventral horns serve as output paths of
signals to peripheral structures, in particular to muscles (the axons of
α-motoneurons) and to muscle spindles (the axons of γ-
motoneurons). The axons of these neurons form ventral roots.

Figure 16.3 Each vertebra has a body and a spinous process. Peripheral
information gets into the spinal cord through the dorsal roots, while efferent signals
are sent from the spinal cord through the ventral roots.

Spinal vertebrae form four major groups: the cervical spine (7

vertebrae), the thoracic spine (12 vertebrae), the lumbar spine (5
vertebrae), and the sacrum (figure 16.4). The vertebrae are
numbered through each group, starting from the most rostral
vertebrae. So, starting from the connection between the spine and
the skull, the vertebrae are C1 to C7, Th1 to Th12, L1 to L5, and the
sacrum. The spinal cord is commonly described as consisting of
spinal segments; this classification refers to the spinal roots, which
enter the spine and leave it at the level of each vertebra. Each spinal
segment receives peripheral information through one pair of dorsal
roots and sends command signals through one pair of ventral roots.
The segments are also classified into cervical, thoracic, lumbar, and
sacral, but this classification does not exactly correspond to the
classification of vertebrae. Spinal segments, starting from the brain,
go in the following order: C1 to C8, Th1 to Th12, L1 to L5, and S1 to
S5. The anatomical length of the spinal segments along the spinal
cord does not correspond exactly to the size of the corresponding
vertebrae, so there is a mismatch between the numbering of the
segments and that of the corresponding vertebrae. The mismatch
increases in the caudal direction, leading to an accumulation of
ventral and dorsal roots that have not had a chance to exit or enter
the spinal cord. As a result, the spinal cord ends at the level of L1
vertebrae. The caudal part of the spine does not contain the spinal
cord at all, only the accumulated axons forming the dorsal and
ventral roots of the lower spinal segments. This part is called the
cauda equina, or the horse tail, because of its appearance.
Figure 16.4 Vertebrae are numbered starting from the skull: C1 to L5, and then
the sacrum. Spinal segments are numbered from C1 to S5, but this classification
does not exactly correspond to the vertebrae classification. The spinal cord ends
at the L1 vertebra; lower, the roots of the lower segments form the cauda equina.

Each segment of the spinal cord receives peripheral information

from a well-defined area of the body and sends command signals to
muscles within approximately the same area. Some muscles have
very large bodies and extend over a number of segments. These
areas go like zebra stripes across the body and down the limbs. One
can say, for example, that the anterior thigh area is innervated by
segments L2 and L3, while the foot is innervated by segments L5
and S1. The topography of these areas is well known to neurologists
who use it, in particular, to define the level of a spinal cord injury.
Let us get back to the “butterfly,” which can be seen on a cross-
section of the spinal cord at any level. The back parts of the “wings”
are the places where dorsal roots enter the spinal cord. The axons
forming the dorsal roots spread out and make synapses in different
areas of the cross-section. The ventral (front) parts of the “wings” are
the places where the motoneurons (both α- and γ-motoneurons) are
located and where ventral roots leave the spinal cord. Since our
main interest is in the production of movements, we are going to be
particularly interested in what happens in that area. In this chapter,
the discussion is going to be limited to neurons and events that
occur in the ventral areas of the spinal cord. Note, however, that
most of the neurons within the spinal cord are not motoneurons but
interneurons. These neurons receive information from both afferent
fibers and axons of other neurons within the central nervous system,
including descending signals from various brain structures; they
generate action potentials that are transduced to other interneurons
or to motoneurons.

16.2 Excitation Within the

Central Nervous System
There are methods that can be used to trace the path of a given
axon, or, more frequently, of a group of axons from cells of one type
located in one area. These methods involve the injection of certain
chemicals into the cell bodies or into neural fibers. These chemicals
are transported, mostly by simple diffusion, along all the neural cell
branches, including the axon and the dendrites. Then, a
histochemical analysis can be performed showing the areas where
the chemical is present. Such experiments are performed on
anesthetized animals or even on slices of tissue. A common
substance used for tracing neural pathways is horseradish
peroxidase. It allows researchers to trace the patterns of termination
of different afferent fibers as well as of the axons of neurons within
the central nervous system.
Tracing excitatory neural pathways has shown that, within the
central nervous system, virtually every neuron is connected to every
other neuron through a certain number of synapses. So if an
excitation emerges somewhere, theoretically it has the chance to
spread to each and every neural cell and, consequently, to induce
contractions of each and every muscle of the body. It is important to
realize that all the synapses made by afferent fibers on neurons
within the central nervous system are excitatory. This means that the
central nervous system is always under the influence of an inflow of
excitatory stimuli that may lead to undesirable motor effects, unless
balanced by inhibitory stimuli. In some pathological conditions,
something like this can actually happen, when pinching an arm or
moving a leg joint may lead to a spasm in virtually all the body
muscles (see chapter 36, which discusses clinical signs and
mechanisms of spasticity). It is obvious that the central nervous
system needs strong and reliable mechanisms to prevent
uncontrolled spread of excitation. One may even go a step further
and state that the essence of information transmission within the
central nervous system is to cut off undesirable routes. Let me
suggest the following analogy: In order to control the flow of a river,
you need to build dams to prevent the water from flowing astray.
Thus, the central nervous system needs a way to make excitatory
synapses ineffective in task-specific ways. There are two basic
mechanisms of inhibition within the central nervous system; these
are termed postsynaptic inhibition and presynaptic inhibition.
The former makes a neuron less sensitive (or insensitive) to any
excitatory signal that may arrive. The latter is more subtle and
selective. It makes certain inputs (certain synapses) into the neuron
less effective without affecting other inputs.

16.3 Postsynaptic Inhibition

First, recollect (chapter 4) that a synapse consists of a presynaptic
membrane, a synaptic cleft, and a postsynaptic membrane (figure
16.5a). The term postsynaptic inhibition means that the inhibition
occurs on the postsynaptic membrane. The mechanism of
postsynaptic inhibition is rather straightforward. Excitatory synapses
lead to depolarization of the postsynaptic membrane—that is, to a
decrease in the absolute value of its negative resting potential (an
excitatory postsynaptic potential [EPSP]; figure 16.5a). If the
combined action of a number of excitatory synapses reaches the
threshold anywhere on the postsynaptic membrane, the postsynaptic
neuron generates an action potential.

Figure 16.5 A synapse consists of a presynaptic membrane, a synaptic cleft, and

a postsynaptic membrane. An excitatory synapse (A) leads to a depolarization of
the postsynaptic membrane (EPSP)—that is, bringing it closer to the threshold for
activation, while an inhibitory synapse (B) leads to a hyperpolarization (IPSP) of
the postsynaptic membrane.

Synapses between neurons within the central nervous system

may be either excitatory or inhibitory. Inhibitory synapses use
neurotransmitters (neuromediators) that lead to an increase in the
absolute value of the membrane resting potential (i.e., to membrane
hyperpolarization) (figure 16.5b). Thus, inhibitory postsynaptic
potentials (IPSPs) bring the membrane potential further away from
its threshold and make the membrane less likely to generate an
action potential in response to excitatory stimuli. Note that IPSPs
have properties similar to those of EPSPs, such as temporal and
spatial summation. So, a large number of inhibitory synapses may
cancel out the potential excitatory effects of a comparably large
number of comparably effective excitatory synapses (figure 16.6).

Figure 16.6 The response of a neuron is defined by the combined action of

excitatory (open circles, ES) and inhibitory (filled circles, IS) synapses. The relative
strength of synapses varies, in particular with distance from the axon hillock. In the
figure, synapses ES-1 and IS-1 are expected to be stronger than ES-2 and IS-2.

PROBLEM 16.2
 Can an inhibitory neuron be used to increase the excitability of
another neuron?

Consider two examples of postsynaptic inhibition within the spinal

cord that are particularly important for the control of voluntary muscle
contractions.

16.4 Recurrent Inhibition and

Renshaw Cells
Motoneurons (and some interneurons) are organized into pools—
that is, into groups with a similar function. In particular, all the α-
motoneurons innervating a single muscle are called a motoneuronal
pool. Figure 16.7 shows schematically a few motoneurons (α and γ)
innervating the same muscle. The axons of α-motoneurons travel to
the muscle and induce contraction of its fibers, while the axons of γ-
motoneurons innervate intrafusal muscular fibers and change the
sensitivity of spindle sensory endings to muscle length and velocity.
The axons of α-motoneurons branch very closely to the cell body,
within the ventral horns, and make excitatory synapses on special
interneurons that are called Renshaw cells. The axons of the
Renshaw cells go back to the bodies of α-motoneurons as well as to
the bodies of γ-motoneurons and form inhibitory synapses. Note that
inhibition by Renshaw cells spreads over all the motoneurons of the
pool and may also involve motoneurons innervating synergistic
muscles. This scheme is known as recurrent inhibition. It has been
viewed as an important mechanism stabilizing the output of
motoneuronal pools (van Heijst et al. 1998; Uchiyama et al. 2003).
Figure 16.7 Axons of α-motoneurons branch very close to the cell body and
make excitatory synapses on Renshaw cells, which in turn make inhibitory
synapses on α-motoneurons of the same pool as well as on γ-motoneurons
sending their axons to spindles in the same muscle.

At first glance, this scheme may look strange: The output of α-

motoneurons excites cells that inhibit the same α-motoneurons!
However, it makes a lot of sense as a mechanism limiting the level of
activity of a pool of motoneurons. Keep in mind that human muscles
are quite strong and are capable of tearing tendons off their points of
attachment (this actually happens sometimes in athletic
competitions). Limiting the peak force that can be produced by a
muscle is a very important protective mechanism for the integrity of
the peripheral structures involved in movements.
Schemes such as recurrent inhibition are called negative
feedback. Later, we will discuss quite a few examples of negative
feedback that take part in the control of voluntary movements.
Negative feedbacks are widespread, suggesting that the central
nervous system does not like any changes and that, if a change
occurs, there is always a mechanism to scale it down quickly and
efficiently. Negative feedback mechanisms allow the central nervous
system to minimize the effects of external perturbations, as well as to
keep down reactions to such perturbations. Such mechanisms can
be seen as being directed toward preserving homeostasis—a
combination of physiological states best suited for optimal
functioning of the body.

PROBLEM 16.3
A muscle is actively contracting against a constant load. What
effects do you expect from the action of Renshaw cells upon γ-
motoneurons? Is this a negative or a positive feedback?

Renshaw cells also receive descending inputs—that is, signals

from the brain. These signals may be considered as means of
controlling the effectiveness (gain) of the negative feedback. For
example, if the brain sends signals to achieve a high level of muscle
contraction force within the shortest possible time, it makes sense to
scale down the excitability of Renshaw cells. On the other hand, if
one wants to have precise control over the output to a muscle,
Renshaw cells must be activated to counteract any accidental
changes in the level of activity of α-motoneurons.

16.5 Reciprocal Inhibition

Another important group of inhibitory interneurons receive signals
from Ia spindle afferents (figure 16.8). Note that these signals are
always excitatory. These interneurons (Ia interneurons) send their
axons to α-motoneurons controlling an antagonist muscle, a muscle
whose contraction opposes the action of a muscle from where the Ia
afferents originate. Such neuronal projections are called heterogenic,
while projections from afferents within a muscle to motoneurons
controlling the same muscle are called autogenic. Ia interneurons
make inhibitory synapses on the membrane of the α-motoneurons
that belong to the antagonist pool.

Figure 16.8 Ia interneurons receive excitatory inputs from Ia afferents and make
inhibitory synapses on α-motoneurons innervating the antagonist muscle. Ia INs
are inhibited by Renshaw cells and also receive descending inputs.

This arrangement is another example of a negative feedback

(figure 16.9). Imagine that a muscle is being stretched. Typically, this
occurs because the antagonist muscle is shortening as a result of its
active contraction. The spindle endings of the stretched muscle will
become more active and will excite Ia interneurons. These
interneurons will inhibit the antagonist α-motoneurons, thus bringing
down the level of contraction of the antagonist muscle. Thus, the
action of Ia interneurons will counteract the original factor leading to
muscle stretch—an increase in the level of activity of the antagonist
α-motoneurons.
Ia interneurons are inhibited by Renshaw cells as shown in figure
16.9. This is another mechanism of negative feedback. Imagine that
α-motoneurons increased their level of activity. This would bring
about an active contraction of the muscle controlled by the α-
motoneurons and a corresponding joint motion. The Renshaw cells
inhibit the Ia interneurons and thus decrease their inhibitory effects
on antagonist α-motoneurons. Such an action is called disinhibition
and is equivalent to additional excitation of antagonist α-
motoneurons. An increase in the activity of the antagonist muscle will
apparently counteract the joint motion that would otherwise occur.

PROBLEM 16.4
You stimulate a pool of Ia interneurons. How will muscle force
change for an agonist and for an antagonist muscle? What kind of
change do you expect to see in the firing rate of Renshaw cells?

Figure 16.9 Renshaw cells inhibit Ia interneurons in addition to their action

illustrated in figure 16.7.

16.6 Presynaptic Inhibition

Both Renshaw cells and Ia interneurons act on the membrane of
their target motoneurons and hyperpolarize it. By this action, these
interneurons decrease the potential response of their target neurons
to all possible excitatory stimuli. The second major type of inhibition
within the central nervous system is more selective. Its purpose is to
decrease the effectiveness of just one type (or a few types) of inputs
to a neuron without affecting other inputs. Apparently, any action on
the postsynaptic membrane will affect the effectiveness of other
synapses because of the passive spread of membrane potential
changes (see figure 16.6). So, in order to shut only one synapse
down, the system has to act at a presynaptic level.

PROBLEM 16.5
Is postsynaptic inhibition equally effective in decreasing the
response of the cell to all the presynaptic excitatory synapses
(e.g., those on remote dendrites and those close to the axonal
hillock)?

This is achieved in a paradoxical way. An excitatory synapse acts

at the presynaptic axonal membrane near the synaptic cleft (figure
16.10). It uses gamma aminobutyric acid (GABA) as a mediator.
GABA induces a long-lasting depolarization of the presynaptic
membrane. This certainly does not prevent the action potential from
getting to the synapse. However, the amount of the excitatory
mediator released in response to an action potential very strongly
depends upon the peak-to-peak amplitude of the action potential.
This is illustrated in figure 16.11a: Note that a modest drop in the
amplitude of the action potential (ΔA) can lead to a substantial drop
in the amount of released neurotransmitter (ΔNT). Depolarization of
the presynaptic membrane leads to a decrease in the peak-to-peak
amplitude of the action potential (figure 16.11b) and, as a result, to a
substantial decrease in the amount of mediator released into the
synaptic cleft. Neuro-neural synapses are typically weak, and
therefore even a modest drop in the efficacy of a synapse effectively
turns it off completely. Consequently, the overall level of
depolarization of the postsynaptic membrane decreases and may
become unable to bring the membrane potential to its threshold.
Thus, an additional excitatory synapse becomes an effective tool of
inducing selective inhibitory effects!

Figure 16.10 Presynaptic inhibition acts selectively on certain synapses. It

involves an excitatory synapse acting on the presynaptic membrane, inducing its
steady subthreshold depolarization and thus decreasing the amount of mediator
released in response to a single presynaptic action potential.
Figure 16.11 (a) A relatively small change in the peak-to-peak amplitude of the
presynaptic action potential (ΔA) leads to a major change in the amount of
released neurotransmitter (ΔNT). (b) A small, steady depolarization leads to a drop
in the peak-to-peak amplitude of the action potential and a decrease in the
synapse efficacy.

PROBLEM 16.6
What will happen if an inhibitory synapse is acting on the
presynaptic membrane close to the synaptic cleft?

Figure 16.12 illustrates an example of the action of presynaptic

inhibition. In this figure, an excitatory synapse is acting at the
presynaptic afferent (sensory) fiber close to its synapse on an α-
motoneuron. The activity of the presynaptic synapse can lead to a
decrease in the reflex response to the afferent fiber activity, while the
activity of the motoneuron can be unchanged or even increased.
Such examples will be considered in later chapters.

Figure 16.12 An example of the action of presynaptic inhibition. In this figure, an

excitatory synapse is acting at a presynaptic afferent (sensory) fiber close to its
synapse at the target α-motoneuron.

16.7 Persistent Inward Currents

As mentioned in chapter 4, dendrites are not simply passive
conductors of local currents. They have the capacity to generate
strong, persistent inward currents that can modulate the excitability
of the dendritic membrane and even turn a dendrite into a self-
sustained generator of action potentials. These depolarizing currents
are produced via voltage-sensitive Ca++ channels that do not show
inactivation (i.e., a drop in the permeability to ions following a brief
episode of the channel being open—as is typical, for example, for
Na+ channels). This lack of inactivation justifies the term persistent.
Persistent inward currents are believed to be produced by
descending neural pathways, in particular those originating from the
brainstem, using such neurotransmitters as serotonin and
norepinephrine. These neurotransmitters have been implicated in
nonmotor functions, including emotional states of the body. Hence,
the mechanism of persistent inward currents provides a
neurophysiological link between the emotional and motor functions,
which is known to most people, especially to athletes.
Figure 4.8 illustrated how persistent inward currents could change
the relation between the overall transmembrane current and voltage.
This figure showed the potential effects of very strong persistent
inward currents that could create a new equilibrium potential on the
membrane over its threshold for the generation of action potentials.
However, even weaker persistent inward currents can be highly
efficient in changing the membrane threshold for action potential
generation, thus helping other inputs into the neuron to produce its
output.
Studies have suggested that persistent inward currents are rather
strong in human α-motoneurons during everyday motor tasks
(Heckman et al. 2005, 2008). They are strong enough to play a
major role in defining the patterns of recruitment and derecruitment
of motor units. These currents are, however, very sensitive to
postsynaptic inhibition and therefore, can be modulated by other
inputs into the α-motoneuron membrane.

CHAPTER 16 IN A NUTSHELL
The spinal cord is protected by the
spine; it has a laminar structure and
is divided into segments that
innervate certain areas of the body.
The spinal cord contains numerous
motoneurons, interneurons, and
conduction pathways that carry both
descending and ascending information.
Afferent information enters the spinal
cord through the dorsal roots, while
efferent nerves that carry motor
signals exit through the ventral
roots. Inhibition within the central
nervous system is vital for its proper
functioning. Postsynaptic inhibition
hyperpolarizes the postsynaptic
membrane and makes the neuron less
sensitive to all excitatory inputs.
Presynaptic inhibition works through
depolarization of presynaptic fibers
and selectively decreases the
effectiveness of only some of the
inputs. Renshaw cells are interneurons
that are excited by axons of α-
motoneurons and inhibit motoneurons of
the same pool (recurrent inhibition).
Ia interneurons are excited by Ia
afferents from primary spindle sensory
endings; they inhibit α-motoneurons of
the antagonist pool (reciprocal
inhibition). Persistent inward
currents through dendritic membranes
play a major role in modulating the
excitability of spinal motoneurons and
defining their recruitment patterns.
 Chapter 17

Monosynaptic Reflexes

KEY TERMS AND TOPICS

general scheme of a reflex
reflex arc
latency
H-reflex
M-response
T-reflex
effects of voluntary muscle activation
F-wave

Monosynaptic reflexes are the only ones with the afferent source and
reflex pathway relatively well defined. They originate from primary
spindle endings and make only one intraspinal excitatory connection
(synapse) with α-motoneurons of the muscle housing the spindle or,
sometimes, its agonist (i.e., a muscle causing joint movement in the
same direction). Not all human muscles show monosynaptic
reflexes. This depends on the density of synapses of Ia afferents on
the α-motoneurons and a number of other factors, such as the
strength of presynaptic inhibition of the Ia afferents.

17.1 H-Reflex and M-Response

Let us imagine that a pair of stimulating electrodes is placed close to
a muscle nerve (figure 17.1). Note that afferent fibers (axons of
muscle receptors) and efferent fibers (axons of motoneurons) travel
between the spinal cord and the muscle together (i.e., within one
nerve). Nerves contain numerous efferent fibers (axons of α-
motoneurons) as well as numerous afferent fibers (peripheral
branches of the axons of sensory neurons). So, when one stimulates
the nerve with a short pulse of electrical current, the same stimulus
acts on both afferent and efferent fibers. Since the fibers in both
directions from the site of stimulation are not refractory, an action
potential induced in a fiber travels in both directions, away from the
body of the neuron—called orthodromic conduction (in the “correct
direction”)—and toward the body of the neuron—called antidromic
conduction (in the “wrong direction”). If we replace schematically all
the neural fibers with one efferent fiber and one afferent fiber, a
single strong electrical pulse applied to the nerve is expected to give
rise to four action potentials traveling along the two fibers
orthodromically (along the efferent fiber toward the innervated
extrafusal muscle fibers and along the afferent fiber into the
intrafusal muscle fibers) and antidromically (along the efferent fiber
toward the α-motoneuron and along the afferent fiber toward the
spinal ganglion with the proprioceptor neuron). These are four
groups, or volleys, of action potentials traveling nearly synchronously
along many afferent and efferent fibers, which are shown in figure
17.1 as AP1, AP2, AP3, and AP4.

PROBLEM 17.1
Why does the last phrase contain the expression “nearly
synchronously”? What factors can lead to violations of
synchronicity?
Figure 17.1 A scheme of an experiment with electrical stimulation of a muscle
nerve. Note that the stimulus is applied to both afferent and efferent fibers,
resulting potentially in four volleys of action potentials (AP1, AP2, AP3, and AP4)
traveling orthodromically and antidromically along the afferent and efferent fibers.

Imagine that a short rectangular pulse of stimulation is applied (its

duration is typically about 0.5 to 1 ms), and then its amplitude is
slowly increased. The stimulus will induce a depolarization of the
membrane of all the axons within the nerve. The magnitude of the
depolarization in individual fibers will depend on two factors, the first
being the location of the fiber with respect to the stimulating
electrodes. In general, the fibers that are closer to the stimulation
site will show higher depolarization. Second, the magnitude of
depolarization will depend on the properties of individual fibers. In
particular, thick fibers show stronger depolarization than thin fibers. If
individual axons within the nerve are mixed randomly, the first factor
does not play a major role. The second factor is very important in
defining the pattern of muscle reaction to the stimulation.
The thickest fibers within a muscle nerve are Ia afferents that
originate from muscle spindles. The axons of the α-motoneurons are
just a little bit smaller. So when one starts to increase the strength of
the stimulation, the first fibers to react are typically Ia afferents. They
will conduct action potentials (AP1 and AP2 in figure 17.1) toward
the spinal cord (antidromically, AP1) and toward the intrafusal fibers
in muscle spindles (orthodromically, AP2). AP1 travel along Ia
afferents from muscle spindles toward the ganglion where the
neuron body is located. There the action potentials switch to the
central branch of the T-shaped axon (without a synapse), enter the
spinal cord, and make a number of synaptic connections, including
monosynaptic connection with α-motoneurons innervating the
muscle that houses the spindles. Thus, a burst of activity in Ia
afferents may be expected to induce a monosynaptic reflex.
Synapses made by Ia afferents on α-motoneurons are typically
weak, as most neuro-neuronal synapses are. Therefore, to observe
a muscle response, a large number of Ia afferent fibers have to be
stimulated, leading to spatial summation of their effects on the target
motoneurons.
The monosynaptic reflex induced by electrical stimulation of the
muscle nerve is called the H-reflex, named after a famous German
scientist, Paul Hoffman (1884-1962). H-reflexes can be studied in a
number of human muscles including, for example, the calf muscles
(triceps surae) in response to electrical stimulation of the tibial nerve.
The latency of the H-reflex in the triceps surae is about 30 to 35 ms,
depending mostly on the length of the subject’s leg.
The other group of action potentials induced in afferent fibers
(AP2) is not expected to produce any measurable effect. They travel
into the spindle sensory ending, where there are no targets. Since
neural fibers become refractory just after generating an action
potential, AP2 does not backfire.
Figure 17.2 Afferent fibers are typically the first to react to a slowly increasing
electrical stimulus. They induce a reflex muscle contraction (H-reflex). At higher
magnitudes of the stimulus, efferent fibers become excited and induce a direct
muscle contraction (M-response). Further increase in the strength of the
stimulation leads to an increase in the M-response and suppression of the H-
reflex.

If the amplitude of the stimulation is increased gradually (figure

17.2), the amplitude of the H-reflex increases because more and
more Ia afferents are excited by the stimulus and, consequently,
more and more α-motoneurons are activated. At some magnitude of
stimulation, the stimulus will also be able to induce action potentials
in the axons of α-motoneurons. It will induce two volleys of action
potentials, AP3 and AP4 in figure 17.1, traveling orthodromically
toward the muscle and antidromically toward the body of the neuron.
Since the place of stimulation is typically close to the muscle, the
action potential AP3 traveling along the axons of α-motoneurons will
take a short time to get to the muscle and to induce its contraction.
For example, during stimulation of n. tibialis in the popliteal fossa,
contraction in the triceps surae can be seen at a delay of about 8
ms. This contraction is addressed as the M-response or direct
response of the muscle in contrast to the reflex contraction, which is
mediated by Ia afferents.
Further increase of the strength of the stimulation would lead to an
increase in the amplitude of the M-response because the
membranes of more and more axons of α-motoneurons will be
depolarized to the threshold. The amplitude of the H-reflex will
continue to increase with an increase in the stimulation strength up
to a certain magnitude (HMAX in figure 17.3) and then it will start to
decrease. This nonmonotonic behavior of the H-reflex with an
increase in the strength of electrical stimulation is due to the
physiological mechanism of the H-reflex and the effects of the fourth
volley of action potentials, AP4. This group of action potentials will
travel to the spinal cord, then to the motoneuron body, and two
things may happen. AP4 may either disappear without an apparent
effect or it will induce an orthodromic action potential (i.e., backfire
leading to a muscle contraction not mediated by synaptic
connections, which will be discussed later in this chapter).
Figure 17.3 Typical dependences of the peak-to-peak amplitude of the H-reflex
and M-response (AH,M) on the strength of the stimulation applied to a muscle
nerve (STIM). Note the nonmonotonic H-curve with a peak magnitude (HMAX) for
intermediate STIM values and the monotonic increase in the M-response.

PROBLEM 17.2
What can happen with a motoneuron if an antidromic action
potential arrives at the axonal hillock? Consider as many
scenarios as possible.

PROBLEM 17.3
What will happen if two action potentials (one orthodromic and
one antidromic) are moving toward each other and meet at a
certain point on the axon?

Consider now what happens when a number of action potentials

comes nearly simultaneously along numerous Ia-afferent fibers to an
α-motoneuron. All these action potentials have excitatory effects on
the α-motoneuron membrane—that is, they all induce EPSPs. If the
EPSPs are strong enough, they may show spatial summation and
depolarize the membrane to the threshold, thus leading to the
generation of an action potential by the α-motoneuron. This action
potential will travel down the axon and induce a twitch contraction of
the corresponding motor unit, which will contribute to the H-reflex.
Now imagine that an action potential is already traveling
antidromically along the same axons induced by the same electrical
stimulus. Note that the velocity of an action potential in Ia afferents is
a little bit higher than that in the axons of α-motoneurons. However,
the path for an action potential in a Ia-afferent fiber is slightly longer,
and it also involves a 0.5 ms synaptic delay. So, one may expect
these two signals to arrive to the target α-motoneuron almost
simultaneously. In this case, they will extinguish each other because
of the refractory period of the membrane. For an analogy, imagine
that there is fire in a dry and open field. The best way to stop the fire
is to start another fire that would spread toward the first one. When
they meet, there is no fuel for the fire to move in either direction, and
it will extinguish itself. You may view a neural fiber as a “fire
conductor” that takes time to be able to conduct another “fire” (figure
17.4).

Figure 17.4 When an afferent fiber delivers a presynaptic action potential to an

α-motoneuron whose axonal hillock has just responded to an antidromic efferent
action potential, the motoneuron is unable to generate another efferent action
potential because of the refractory period.
 Thus, further increase in the strength of the stimulation leads to
more and more axons of α-motoneurons directly responding to the
stimulation, conducting action potentials both ortho- and
antidromically, and consequently, making more and more
motoneurons unable to respond to action potentials that are
conducted by Ia afferents. Eventually, all the efferent axons will be
stimulated directly, and the H-reflex will disappear while the M-
response will achieve its peak (MMAX in figure 17.3).
Figure 17.3 illustrates typical dependences between the strength
of the nerve stimulation and the peak-to-peak amplitudes of the H-
reflex and M-response. Since both responses are recorded by the
same pair of electrodes placed over the muscle, their amplitudes can
be compared directly. In particular, the M-response to a very large
stimulus, MMAX, typically represents the largest response one can
obtain from the muscle since it corresponds to the compound action
potential when all motor units are excited nearly simultaneously.
The peak value of the H-reflex is seen at an intermediate strength
of the stimulation. The ratio of this value to the peak M-response
magnitude reflects a number of factors, such as the relative
excitability of Ia afferents and α-motoneuron axons as well as the
processes on the α-motoneuron membrane. This ratio is used in
both clinical studies and studies of reflexes in healthy persons.
Figure 17.5 Successive stimuli at a high frequency induce similar M-responses
(top) but progressively smaller H-reflexes (bottom).

If two or more electrical stimuli are applied to a muscle nerve at

relatively short delays, the direct muscle response and the reflex
response will demonstrate dramatically different behaviors. M-
responses in response to successive stimuli will be very similar to
each other (unless you get to very high frequencies). The H-reflex in
response to the second stimulus will be smaller; if there is a third
stimulus, the reflex may even be smaller, and it may eventually
disappear after a few stimuli. The difference in the behaviors of the
M-response and H-reflex illustrated in figure 17.5 is due to the
properties of the central synapse that are part of the H-reflex arc but
not of the M-response. Recall that the absolute refractory period for
a typical membrane is due to the inactivation of sodium channels
and is rather short, on the order of 1 ms, so that a relatively short
time is required to restore the membrane’s sensitivity and ability to
conduct action potentials.

PROBLEM 17.4
 What is the frequency of stimulation at which you may expect to
see the M-response become smaller?

Synapses, on the other hand, need more time to restore the

original amount of the mediator in synaptic vesicles. Neuromuscular
and neuro-neuronal synapses differ dramatically in their response to
modest changes in the amount of stored neurotransmitter.
Neuromuscular synapses are strong and obligatory. The amount of
neurotransmitter stored in the presynaptic membrane is large, and
large amounts of the neurotransmitter are released in response to
each incoming action potential. Neuro-neuronal synapses are weak
and non-obligatory. This is reflected in the relatively small amount of
stored neurotransmitter and its amount released each time in
response to a presynaptic action potential. As a result, neuro-
neuronal synaptic transmission shows signs of attenuation even at
time delays of a few seconds. So, during experimental and clinical
studies of the H-reflex, there are usually at least 10 s intervals
between successive stimulus applications.

17.2 Tendon Tap Reflex (T-

Reflex)
A monosynaptic reflex, similar to the H-reflex, may be induced by a
more physiological stimulus—that is, by a quick muscle stretch
(figure 17.6). Primary endings of muscle spindles are very sensitive
to muscle velocity, so a quick muscle stretch leads to their
synchronized firing. This volley of action potentials travels along the
Ia afferents to the spinal cord and induces a reflex response of α-
motoneurons, leading to a twitch contraction of the muscle. This
reflex is called the T-reflex (or tendon tap reflex). Muscle stretch
may be induced by a tendon tap, leading to the well-known knee jerk
or ankle jerk. Note that quick muscle stretch does not induce direct
excitation of the axons of α-motoneurons, so this stimulus induces
only one volley of action potentials (AP1 in figure 17.2). As a result,
there is no direct muscle response (no M-response), and there is a
monotonic increase in the amplitude of the T-reflex with an increase
in the amplitude or velocity of the stretch until the reflex reaches its
peak value.
The tendon tap reflex is very easy to elicit, which is why it is very
commonly used in clinical practice. On the other hand, it is difficult to
standardize the stimulus magnitude while tapping a tendon, which
highlights the importance of H-reflex studies if one is interested in
possible quantitative changes in structures and connections involved
in the monosynaptic reflex loop.

Figure 17.6 A tendon tap excites spindle endings and may induce a
monosynaptic reflex contraction (T-reflex). Its reflex pathway is the same as for the
H-reflex.
17.3 Effects of Voluntary
Muscle Activation on
Monosynaptic Reflexes
Monosynaptic reflexes are rather poorly controlled by the subject’s
will, although prolonged training of monkeys has been shown to lead
to their ability to modulate the amplitude of a monosynaptic reflex
without changing the activation level of any muscle (Wolpaw 1987;
Wolpaw and Carp 1993). These studies used the operant
conditioning technique, which rewards desired responses by the
animal. Later studies have shown that humans can also show such
plastic changes in monosynaptic reflexes after a relatively short
training session, which can be useful in clinical practice (Thompson
and Wolpaw 2015).
Human subjects can modulate the amplitude of monosynaptic
reflexes indirectly, for example, by activating certain muscle groups.
Voluntary activation of a muscle leads to depolarization of the
membranes of many α-motoneurons innervating the muscle. Some
of the motoneurons generate action potentials, while others are
depolarized under the threshold for action potential generation.
These motoneurons are more excitable and can be more easily
recruited by a standard afferent discharge induced by an electrical
stimulus. As a result, more motoneurons respond to a standard
stimulus, leading to an increase in the overall reflex muscle
response. For example, voluntary activation of the calf muscle group
increases the H-reflex in this muscle induced by an electrical
stimulation of the tibial nerve (figure 17.7).
Figure 17.7 Voluntary muscle activation increases the amplitude of the H-reflex
in the activated muscle through an overall excitation of the motoneuronal pool.
Figure 17.8 Voluntary activation of an antagonist muscle leads to suppression of
the H-reflex amplitude via activation on Ia interneurons (Ia-IN).

On the other hand, voluntary activation of an antagonist muscle

decreases the amplitude of a monosynaptic reflex. For example, an
H-reflex in the calf muscle group decreases if the subject activates
the tibialis anterior muscle (figure 17.8). This effect is assumed to
result from the action of inhibitory Ia interneurons excited by the
descending voluntary command to the antagonist muscle. These
interneurons produce postsynaptic inhibition of the motoneuronal
pool (see chapter 16), which is activated by Ia afferents that respond
to the electrical stimulus.

PROBLEM 17.5
What will happen with the H-reflex amplitude if you strongly
coactivate the agonist–antagonist pair of muscles?
 Besides that, activation of distant large muscle groups can lead to
a modulation of the H-reflex in the calf muscles. This can be
achieved, for example, with the Jendrassik maneuver, which consists
of a steady, strong contraction of remote muscle groups. For
example, the Jendrassik maneuver may involve clenching the hands
in front of the chest and trying to separate them by a strong
contraction of the shoulder and back muscles. Alternatively,
clenching teeth strongly can lead to similar effects. After several
seconds, the Jendrassik maneuver may lead to a change (usually,
an increase) in the amplitude of the H-reflexes in the calf muscles.
The mechanism of this effect is not known, but it likely involves the
conduction of action potentials along propriospinal pathways, which
connect spinal segments at different levels of the spinal cord.

17.4 F-Wave
As mentioned earlier, the H-reflex can be readily seen only in some
human muscles. This may be due to a number of factors, in
particular to differences in the density and effectiveness of
monosynaptic connections of Ia afferents on α-motoneurons, and to
different relations between the diameters of Ia afferents and efferent
axons.

PROBLEM 17.6
How can a synapse on an α-motoneuron be more or less effective
than another synapse?

Some muscles, in particular a number of muscles in the upper

extremities, demonstrate a response to an electrical stimulation of
the muscle nerve at a latency similar to that expected of the H-reflex.
However, this response does not increase monotonically with an
increase in the stimulation amplitude, unlike the nonmonotonic
behavior of the H-reflex amplitude (as in figure 17.3). In addition, it
does not suffer from an increase in the stimulation frequency like the
H-reflex does. This response is termed the F-wave. The described
properties of the F-wave suggest that no synaptic transmission is
involved. The F-wave results from antidromic conduction along the
axons of α-motoneurons induced by action potentials arriving
orthodromically along those axons (AP4 in figure 17.2). As
mentioned earlier, commonly α-motoneurons do not backfire
because of the refractory period on the axonal membrane. However,
in some neurons, because of the peculiarities of the time course of
membrane depolarization at the neuron body, the membrane at the
axon hillock can get out of its refractory period in time to respond to
its continuing depolarization. In a sense, this is a unique example of
an antidromic action potential coming to a neuron and giving rise to
another, orthodromic action potential (figure 17.9). So, some neurons
can backfire!
Accurate measurement can show that the latency of the F-wave is
just a little smaller than the H-reflex latency. To be exact, the
difference is about 0.5 ms, corresponding to the typical synaptic
delay because the H-reflex arc involves a synapse, while the F-wave
arc does not. F-waves are used in both experiments and clinical
studies to explore the excitability of α-motoneuronal pools. Similar to
the H-reflex, the F-wave shows an increase in its magnitude in
response to voluntary muscle activation due to an overall increase in
the α-motoneuronal excitability.
Figure 17.9 An antidromic action potential in an efferent fiber, induced by an
electrical stimulus, can induce an orthodromic action potential, leading to muscle
contraction called the F-wave.

PROBLEM 17.7
For a given set of Ia afferents, is their stimulation more likely or
less likely to induce the H-reflex in a muscle innervated by very
thin and very thick axons of α-motoneurons?

CHAPTER 17 IN A NUTSHELL
Reflex is a loosely defined notion
meaning an involuntary motor reaction
to an external stimulus. The simplest
reflex consists of a receptor, an
afferent nerve, at least one synapse
on a central neuron, an efferent
nerve, and an effector. Monosynaptic
reflexes contain only one central
synapse; they always induce excitatory
effects on α-motoneurons. They can be
induced by an abrupt change in the
activity of primary muscle spindle
afferents, as during tendon tap (T-
reflex), or in response to an
electrical stimulation of a muscle
nerve (H-reflex). In the latter case,
motor axons are also stimulated,
leading to a direct muscle response
(M-response). Voluntary activation of
a muscle leads to an increase in its
monosynaptic reflexes. In some
muscles, an F-wave can be observed,
which is a direct response of α-
motoneurons to an antidromic volley
traveling up their axons, induced by
an electrical stimulus.
 Chapter 18

Oligosynaptic and Polysynaptic

Reflexes

KEY TERMS AND TOPICS

oligosynaptic reflexes
polysynaptic reflexes
flexor reflex and crossed extensor reflex
tonic stretch reflex
tonic vibration reflex
force-sensitive reflexes
interjoint and interlimb reflexes

Monosynaptic reflexes are the simplest among the variety of muscle

reflexes acting in the human body. Their functional importance,
however, is questionable. Because they induce brisk, short-lasting
contractions and are poorly controlled voluntarily, they are unlikely
candidates as important parts of the mechanism of voluntary muscle
control. Of course, this statement does not mean that monosynaptic
projections of Ia afferents on α-motoneurons are not functional. They
contribute to muscle activation patterns in collaboration with other
inputs, but rarely lead to bursts of activation typical of monosynaptic
reflexes.
Figure 18.1 Ia-interneurons receive excitatory inputs from Ia afferents and make
inhibitory synapses on motoneurons innervating the antagonist muscle. Thus, they
have an oligosynaptic (disynaptic) inhibitory reflex effect.

18.1 Oligosynaptic Reflexes

Oligosynaptic reflexes are assumed to involve, by definition, a small
number of neuro-neuronal synapses, but more than one. Typically,
oligosynaptic reflexes are assumed to involve two or three such
synapses. One of the best known inhibitory oligosynaptic
connections is from the primary endings of muscle spindles to α-
motoneurons of the antagonist muscle, mediated by one interneuron,
a Ia interneuron (figure 18.1). The latency of inhibitory muscle
reactions mediated by these connections is very close to the latency
of monosynaptic reflexes induced by the same stimulus in a muscle
housing the spindles. The difference is due to the additional
synapse, which adds about 1 ms to total transmission time.
Another important group of oligosynaptic reflexes originate from
Golgi tendon organs whose axons belong to the group of Ib afferents
(figure 18.2). The action of Ib afferents looks like the exact opposite
of that of Ia afferents. That is, Ib afferents from a group of Golgi
tendon organs induce a disynaptic inhibition of the homonymous α-
motoneurons (i.e., those controlling the muscle whose tendon
contains the Golgi tendon organs) and a disynaptic or trisynaptic
excitation of the antagonist α-motoneurons. Note that both excitatory
and inhibitory pathways from Ib afferents involve at least one
interneuron that belongs to the group of Ib interneurons. The action
of Ib afferents is another example of negative feedback: If a muscle
develops active force, its Golgi tendon organs become active and
inhibit the motoneurons of this muscle while exciting the
motoneurons of its antagonist. Both effects lead to a decrease in the
moment of force in the joint crossed by the agonist–antagonist pair.

Figure 18.2 Golgi tendon organs send their axons (Ib afferents) to Ib
interneurons, which exert an inhibitory action on the agonist α-motoneurons and
disinhibit (excite) antagonist α-motoneurons via another interneuron.

PROBLEM 18.1
What kinds of reflex effects can be expected from the action of Ib
afferents during a fast flexion movement against a constant
external load?

PROBLEM 18.2
 What can be expected from the action of Ib afferents during a fast
development of extension force against a stop?

Many of the properties of oligosynaptic reflexes are similar to

those of monosynaptic reflexes, with the only exception being the
slightly longer conduction time. However, the presence of at least
one additional synapse changes properties of the reflex arc, in
particular its reaction to high-frequency stimulation. Usually, when
more synapses are involved in the transmission of a stimulus to α-
motoneurons, more pronounced suppression of reflex effects is
observed with an increase in the stimulation rate. This is due to the
fact that each neuro-neuronal synapse is a potential bottleneck along
the reflex loop due to its weakness and sensitivity to the amount of
available neurotransmitter in the presynaptic terminal.
Since in everyday movements there are not too many twitchlike
contractions (and humans would probably like to avoid them!),
monosynaptic and oligosynaptic reflexes are likely to represent
phenomena that, by themselves, do not produce functionally
important reactions to peripheral stimuli. This statement certainly
does not mean that the mechanisms underlying these reflexes are
not functioning or that they can be ignored in an analysis of everyday
motor control. The point is that in everyday life, humans make
smooth movements, and most of the peripheral stimuli induce
muscle stretches that do not lead to strong, synchronized afferent
discharges that would give rise to monosynaptic and oligosynaptic
reflexes resembling those studied in the laboratories. Studies of
monosynaptic and polysynaptic reflexes have also been helpful in
motor control research, including studies of disordered movements.

18.2 Polysynaptic Reflexes

Most of the muscle reflexes are polysynaptic; that is, they involve
many (more than three; the exact number is typically unknown)
synapses in the reflex arc. These reflexes are generally
characterized by noticeably longer latencies and more complex
behavior than monosynaptic or oligosynaptic reflexes. Since they
involve more interneurons in their reflex arc, they can potentially
provide more information about other levels of the neural hierarchy
that are beyond the reach of the simpler reflexes.
As mentioned earlier, neurophysiological phenomena, including
the reflexes, are commonly classified into two groups, phasic and
tonic. This classification is far from clear and straightforward;
depending on the time scale (for example, a typical time of
observation), the same phenomenon can be qualified as tonic or as
phasic. Complex processes can have both tonic and phasic
components.
Phasic reflexes emerge in response to a change in the level of a
specific receptor stimulus. They usually represent a burst (or a
short-lasting depression) of muscle activity leading to a twitchy
movement or a series of twitchy movements. Note that all the
monosynaptic reflexes are phasic.
Tonic reflexes emerge in response to the level of a stimulus
itself. They lead to sustained muscle contractions and relatively
smooth movements. These reflexes are always polysynaptic.

PROBLEM 18.3
Reflex eye movements make it easy to maintain gaze direction
during voluntary head movements. Is this a tonic or a phasic
reflex?

For example, the activity of length-sensitive spindle endings can

lead to both types of reflexes (figure 18.3). If a muscle is quickly
stretched, monosynaptic and polysynaptic reflexes can be observed
in response to the process of stretch. If the muscle stays at a new,
stretched state, the phasic reflexes quickly disappear. However, if
the muscle was active before the stretch, it is possible to record tonic
changes in the muscle activity at any time after the stretch is
completed, and the muscle is at a new steady state. This mechanism
is frequently termed the tonic stretch reflex in contrast to the
phasic stretch reflex, which is another word for the T-reflex. If the
imposed stretch is slow (thin dotted traces in figure 18.3), the tonic
stretch reflex can be observed in the absence of the phasic reflex. To
avoid the tonic–phasic confusion, sometimes researchers refer to the
overall muscle reaction to stretch as simply the stretch reflex.

Figure 18.3 Tonic and phasic components of the muscle reflex to fast stretch
(thick black lines). Stretching the muscle over the same range slowly leads to the
tonic reflex only (gray dotted lines).

The phasic–tonic dichotomy is based on quantitative rather than

qualitative differences in the mechanisms. For example,
monosynaptic reflexes are mediated by monosynaptic connections
of Ia afferents with α-motoneurons. Spindle Ia afferents are sensitive
to both muscle length and the rate of its change. Their monosynaptic
projections certainly function during slow stretches or maintenance
of a constant muscle length, but they do not evoke monosynaptic
reflexes. Therefore, monosynaptic connections of Ia afferents can
and are likely to play a role in tonic reflexes. When the rate of muscle
stretch reaches a certain threshold, monosynaptic reflexes emerge;
when it is under the threshold, only tonic reflexes are observed.
 There is a big difference between monosynaptic and polysynaptic
reflexes. For the former, the afferent source and the reflex arc are
known with a degree of certainty. For the latter, next to nothing is
known about the exact neural loops leading to reflex muscle
contractions. So, polysynaptic reflexes are usually considered as
functional notions. Figure 18.4 illustrates a typical scheme for a
polysynaptic reflex. A stimulus is applied, leading to the activation of
a number of sensory endings. These are commonly of different
modalities. For example, a quick muscle stretch, such as that
induced by an external device, leads to changes in muscle length
and velocity. However, since the muscle and tendon have springlike
properties, muscle force also changes instantaneously. The stretch
violates the balance of forces between the tapped muscle and its
antagonist, which is expected to lead to changes in joint angle,
antagonist muscle length, and antagonist muscle force. In addition,
many cutaneous and subcutaneous receptors react to skin
deformation produced by the tap.
Figure 18.4 A functional scheme of a polysynaptic reflex. Commonly, these
reflexes receive contributions from receptors of different modalities, and their
central pathways are unknown. They are assumed to involve more than two
interneurons.

In polysynaptic reflexes, a stimulus leads to a change in the level

of activation of a muscle (commonly more than one muscle). The
muscle may be located in close proximity to the stimulus (such
reflexes are termed autogenic) or may be rather far away, even on a
different limb. As a result, polysynaptic reflexes are typically
described as input–output relations between a stimulus and its
mechanical or electromyographical consequence.

18.3 Flexor Reflex

The flexor reflex, as its name suggests, represents a reflex
contraction of flexor muscles in response to a stimulus. There are
many sensory endings contributing to the flexor reflex; they are
frequently united into one group and are called the flexor reflex
afferents. This group involves, among others, secondary endings of
muscle spindles (group II afferents), free endings scattered all over
the muscles and innervated by thin, slowly conducting axons of
groups III and IV, some cutaneous receptors, and nociceptors
(receptors of painful stimuli). A flexor reflex may be induced by a
painful stimulus to the skin or by an electrical stimulation of a
cutaneous nerve, such as the sural nerve (figure 18.5). It has a
relatively long latency (typically about 70 ms), which is due to both
slow conduction of most of the afferent fibers and long central
delays. An appropriate stimulus induces a quick but sustained
contraction of all the major flexor muscles of the limb to which the
stimulus is applied. If the limb is free to move, a withdrawal motion
from the stimulus occurs; that is why this reflex is sometimes called a
withdrawal response. Such reflexes look somewhat different in
babies and may be changed in certain pathologies, such as spinal
cord injury, brain trauma, and multiple sclerosis.
Figure 18.5 A flexor reflex is induced by a group of afferents called flexor reflex
afferents. These include secondary endings of muscle spindles, free nerve
endings, some cutaneous receptors, and some others. A flexor reflex leads to the
activation of flexor muscles within the limb. Shown here is a typical reflex reaction
in a flexor muscle of a hindlimb (for example, the tibialis anterior) to an electrical
stimulation of n. suralis.

The flexor reflex is commonly accompanied by a reflex contraction

of extensor muscles of the contralateral extremity. This crossed
extensor reflex will be discussed in more detail a bit later.

18.4 Tonic Stretch Reflex

We have already identified the tonic stretch reflex as a sustained
muscle contraction in response to the slow stretch or maintenance of
the muscle at a new, longer length. The tonic stretch reflex is a
mechanism contributing to a very important feature of human
muscles: the dependence of active force on length and velocity,
sometimes called viscoelastic behavior. This reflex was first
described by a great British physiologist, Sir Charles Sherrington,
and his colleague Eric Liddell (Liddell and Sherrington 1924) and
studied by many since that time (reviewed in Gottlieb and Agarwal
1972; Feldman and Levin 1995; Feldman 2015).

Figure 18.6 A muscle is slowly stretched by an external force. First, it resists the
stretching only due to its passive elasticity (passive characteristic, dashed). At a
certain threshold, recruitment of α-motoneurons begins, leading to active force
development (tonic stretch reflex). The force–length dependence is called the tonic
stretch reflex characteristic (thick, solid line). Muscle activation increases with
further stretch (illustrated with different sizes of the EMG label). The slope of the
characteristic (its apparent stiffness; see the dotted line) increases with length.

Let us consider what will happen if an intact muscle (i.e., a muscle

with all its neural connections in place) is slowly stretched by an
external force (figure 18.6). First, the muscle will resist stretching due
to its passive elastic properties (remember that the muscle has both
parallel and series elastic elements). Then, at a certain length, the
increased activity of muscle spindles will bring about an autogenic
recruitment of a few α-motoneurons whose activation will lead to
active force development by the muscle, also opposing the stretch.
The muscle length at which the recruitment of motoneurons begins is
termed the threshold of the tonic stretch reflex. If one continues to
stretch the muscle, the already recruited motoneurons will increase
their firing frequency, and more motoneurons will be recruited,
contributing to an increase in the muscle force. Eventually, such an
experiment will yield a tonic stretch reflex characteristic (a force–
length characteristic) of the muscle.
The slope of the characteristic is sometimes addressed as muscle
stiffness (note that it varies with changes in muscle length); actually,
it is better to use another term, apparent stiffness, because an intact
muscle with its reflexes cannot be considered a single, ideal spring
(Latash and Zatsiorsky 1993). Note that this apparent stiffness has
two components. The first component is purely peripheral and is
independent of any reflex effects; it can be observed in a
deafferented muscle (i.e., in a muscle without reflexes). The second
component has a reflex nature. Note also that the slope of the curve
changes with muscle length.
A study by Nichols and Houk (1976) quantified the force–length
characteristics of intact cat muscles and then of the same muscles
without reflexes (this can be done, for example, by cutting the
corresponding dorsal roots that carry sensory information into the
spinal cord). This study showed that reflexes contribute significantly
to the force–length characteristic of intact muscles. The slope of the
characteristic increases, and the dependence force on length
becomes more linear, more “springlike.”
The term tonic stretch reflex is somewhat misleading because this
reflex has a velocity-dependent component due to the velocity
sensitivity of the primary endings in muscle spindles. This means
that if a muscle is stretched at a certain velocity and moved through
a certain value of muscle length, its force of contraction will be higher
than if the same length is kept constant. This is an important feature
of the stretch reflex (discussed in more detail in chapter 22) that
contributes to damping in the muscle behavior and helps stop
movements efficiently without major terminal oscillations of
trajectories.
 PROBLEM 18.4
What will happen with muscle force when the muscle passes
through a certain length during its shortening as compared to the
force developed at the same length in static conditions?

18.5 Tonic Vibration Reflex

High-frequency, low-amplitude muscle vibration is a very potent
stimulus for spindle sensory endings. For example, in humans,
vibration of a muscle at a frequency of about 100 Hz and an
amplitude of about 1 mm may be strong enough to “drive” virtually all
the primary endings of all the spindles within the muscle (Brown et
al. 1967; Matthews and Stein 1969). Driving means inducing an
action potential (or, sometimes, several action potentials) in
response to every cycle of vibration. Secondary muscle endings are
also rather sensitive to vibration, as are numerous skin and
subcutaneous receptors. That is why skin vibration is sometimes
used to test individual skin sensitivity. Golgi tendon organs show
strong responses to vibration only in an activated muscle.

PROBLEM 18.5
If all the Ia afferents are driven by vibration, why don’t they induce
monosynaptic reflexes in response to each vibration cycle?
Figure 18.7 High-frequency muscle vibration leads to a slow reflex increase in
muscle force (tonic vibration reflex). It starts at a substantial delay (commonly, a
few seconds) and lasts for some time after the stimulus is turned off.

Vibration typically induces a tonic contraction of the muscle to

which it is applied. This contraction is termed the tonic vibration
reflex (Eklund and Hagbarth 1966). (Think about what a strange
word combination this is: Tonic means steady-state, while vibration
implies high-frequency oscillations. These two words become
compatible, however, when a high-frequency stimulus leads to a
relatively slowly changing muscle contraction.) The contraction starts
a few seconds after the beginning of the vibration (figure 18.7),
increases gradually, and stays at a relatively constant level until the
vibrator is turned off. Then the contraction gradually subsides over a
few seconds. An electromyogram (EMG) of a muscle during the tonic
vibration reflex looks very similar to the EMG during voluntary
muscle contractions. However, deeper analysis shows a large
number of motor units firing synchronously with vibration cycles
(which is not surprising because Ia afferents are being driven by the
vibration), while voluntary EMG is rather asynchronous.
An analysis of the excitatory postsynaptic potentials in α-
motoneurons during muscle vibration in animal experiments revealed
two types of EPSPs: Those synchronized with the vibration, and a
slow depolarization that is not synchronized with the vibration. The
slow depolarization was attributed to the action of a polysynaptic
pathway from the Ia afferents (with a possible contribution of other
muscle afferents), while the synchronized EPSPs were attributed to
the action of monosynaptic or polysynaptic connections between Ia
afferents and α-motoneurons.
Muscle vibration is accompanied by a number of rather unusual
phenomena. The first phenomenon is the ability of subjects to
suppress the tonic vibration reflex voluntarily (i.e., by “pure thinking”).
Sometimes the desire to suppress the reflex is so strong that
subjects need to be taught how not to suppress the tonic vibration
reflex or, in other words, to allow reflex muscle contraction to
develop. This fact makes the tonic vibration reflex a “nonreflex”
according to our earlier definition. However, recall that the definition
was going to fail, and here is just one example of its failure.
The second is suppression of monosynaptic reflexes on the
background of vibration (figure 18.8), sometimes leading to their total
elimination. These effects look rather paradoxical since a tonic
voluntary contraction of a muscle is typically associated with
postsynaptic facilitation of the motoneuronal pool, leading to an
increase in H-reflexes observed in the same muscle (discussed in
chapter 16). Thus, a voluntary muscle contraction mimicking a tonic
vibration reflex leads to an increase in the amplitude of H-reflexes.

PROBLEM 18.6
Suggest at least two explanations for the suppression of H-
reflexes during the tonic vibration reflex.

The suppression of monosynaptic reflexes by vibration has been

demonstrated to be of a presynaptic nature. That is, the complex
afferent inflow induced by vibration leads to selective presynaptic
effects on the terminals of Ia afferents at their synapses with α-
motoneurons. Other synapses, including those from descending
neural pathways involved in voluntary muscle contraction, do not
seem to suffer. As a result, the polysynaptic pathways induce a tonic
muscle contraction (tonic vibration reflex), while the monosynaptic
pathways become relatively ineffective.

Figure 18.8 A tonic vibration reflex is accompanied by suppression of

monosynaptic reflexes in the same muscle (for example, of the H-reflex). This
suppression has a presynaptic origin (upper drawing).

The third unusual effect of muscle vibration is its ability to induce

reflex contractions of muscles that are not subjected to vibration. In
particular, if vibration is applied to a muscle, reflex contractions may
be observed in the muscle or its agonists, or in the antagonist
muscles, or even in muscles acting at different joints of the limb.
Which muscle is going to be activated depends on a number of
factors, including, in particular, limb configuration (joint angles), its
orientation with respect to the field of gravity, the presence of support
under the foot, and a few other factors (figure 18.9). These patterns
suggest that muscle vibration involves rather complex,
intersegmental mechanisms that may also be used during such
common activities as standing and walking.
Recent studies have shown that vibration applied to leg muscles
can induce cyclical, locomotion-like leg movements if the subject’s
legs are suspended in the air (Gurfinkel et al. 1998; Solopova et al.
2016). These observations support the hypothesis that responses to
muscle vibration involve neural structures in the spinal cord that
participate in the production of locomotor movements—the spinal
central pattern generators (see chapter 26).
Figure 18.9 Muscle vibration can induce reflex contractions of different muscles
of the same limb (shown by black). Note that these muscles are the same as those
that would be active during locomotion when the postures shown are passed
through. The vibration is applied to the Achilles and the patellar tendons.

The fourth effect is kinesthetic illusions, and these are considered

in chapter 28.

18.6 Interaction Among Reflex

Pathways
In the previous sections, the reflex effects of certain afferents, such
as Ia and Ib afferents, were considered as separate phenomena, out
of their context. But what happens if all the afferents are acting
simultaneously, just like it happens in everyday life? The exact
answer is unknown. However, a series of experiments by an
outstanding Swedish research group headed by Lundberg and
Jankowska (Lundberg 1975, 1979; Jankowska 1979; Jankowska et
al. 1983) have demonstrated that most interneurons, including the Ia
and Ib interneurons, receive information that is very mixed (figure
18.10). That is, if the level of firing of Ib afferents is kept constant,
the output of Ib interneurons may change, depending on the level of
firing of Ia afferents. The neural wiring of the spinal cord creates an
impression of a total mess rather than of a well-designed set of
feedback circuits, each having its own function.
These findings should not be a cause for desperation, however.
They demonstrate the limits of the classical neurophysiological
approach, which is good for testing certain neural circuits, but which
is probably too simplistic to analyze the functioning of the central
nervous system as a whole. This is an illustration of the limitations of
the reductionist approach in studies of the functions of the human
body.

Figure 18.10 Ia and Ib interneurons receive mixed information from afferents

originating from different receptors.
18.7 Interjoint and Interlimb
Reflexes
In the previous section we encountered examples of interjoint
reflexes, when vibration of a muscle could induce a tonic reflex
contraction of another muscle within the same limb, but acting at a
different joint. The flexor reflex also induces muscle contractions of
virtually all the flexor muscles of a limb and may be considered an
interjoint or a whole-limb reflex.
A series of experimental studies by the group of T. Richard
Nichols (1989, 2002, 2018) focused on reflex effects from force-
sensitive receptors. Muscle force is sensed by Golgi tendon organs;
however, other receptors may also react to tissue deformation that
accompanies muscle force development. Studies in cats have shown
that force-sensitive feedback loops act on large muscle groups of the
limb involving muscles that cross different joints. These projections
are muscle-specific, depending on the joints crossed by the muscle
and its exact line of action. They have been interpreted as providing
for force-dependent across-joints coordination that can be used as a
building block to construct typical multijoint movements.
Interlimb reflexes are commonly observed in animal preparations
(i.e., in animals whose neural axis has been cut, separating the more
rostral brain structures from the rest of the central nervous system).
In such animals, many spinal reflexes become released from the
suppressing descending influence. In particular, an electrical
stimulation of a skin nerve or a pin prick will induce a typical flexor
response in the limb to which the stimulus is applied and an extensor
reflex in the contralateral limb. This reaction is termed the crossed
extensor reflex (figure 18.11). The observation of the flexor reflex in
one of the limbs and the crossed extensor reflex in the contralateral
limb led C.S. Sherrington to a theory that locomotion represented a
sequence of flexion and extension reflexes (Sherrington 1910). We
will consider this theory critically in chapter 26 on locomotion. Also,
some of the other more complex polysynaptic receptors involving the
vestibular, ocular, and postural systems will be discussed in later
chapters.

Figure 18.11 Stimulation of the flexor reflex afferents in one limb induces a reflex
response in flexor muscles of this limb and a crossed extensor reflex in extensor
muscles of the contralateral limb.

CHAPTER 18 IN A NUTSHELL
Oligosynaptic reflexes contain a few
central synapses, while polysynaptic
reflexes contain many central
synapses. Reflexes of these two groups
can be excitatory or inhibitory.
Phasic stimuli or responses are those
that change in time quickly; those
that are steady-state are called
tonic. Muscle spindles and Golgi
tendon organs are sources of
oligosynaptic reflexes that can be
viewed as negative feedback loops.
Polysynaptic reflexes can involve
muscles of a whole limb, as during the
flexor reflex, which is induced by the
activity of a variety of receptors
with relatively thin axons. The tonic
stretch reflex is an increase in the
activity of motoneurons innervating a
muscle that is being stretched by an
external force. The tonic vibration
reflex is a steady increase in the
level of activation of a muscle during
vibration at a high frequency and low
amplitude. This reflex can be seen in
other muscles, suggesting that its
neural loop involves a spinal central
pattern generator for locomotion. The
tonic vibration reflex is accompanied
by suppression of monosynaptic
reflexes induced by an increase in
presynaptic inhibition.
 Chapter 19

Long-Loop Reflexes and Reflex-

Like Reactions

KEY TERMS AND TOPICS

corrective reactions to perturbations
nature of preprogrammed reactions
postural corrections
corrective stumbling reaction

Human voluntary movements are typically performed not in a motor

control laboratory but in the real world full of unexpected changes in
sensory information, force fields, targets, and so on. As a result,
people are never able to predict ideally the external force field
(among other unpredictable things) and its possible changes. When
a command is issued for the purpose of performing a certain
movement or a certain posture, at the time the command is being
realized by muscles, external conditions may have changed, and the
planned movement or posture will be perturbed by these changes.
Posture and movement must be stable with respect to such
everyday perturbations, and human bodies are equipped with a
variety of mechanisms designed to ensure such stability. Within the
concept of stability here we understand an ability to return to a state
or trajectory after a small, transient perturbation or a change in
intrinsic body states. This notion will be discussed in more detail in
chapter 22.
Some of the mechanisms contributing to the stability of actions
have already been described briefly in previous chapters (for
example, chapter 5 on muscle and tendon elasticity and chapters 15,
17, and 18 on muscle reflexes that generate changes in muscle
force against the perturbing force). In this chapter, we are going to
look at another, very important group of muscle reactions to
perturbations that come at a relatively short delay (although the
delay is longer than that of monosynaptic reflexes) and provide
context-specific corrections of movement or posture in cases of
unexpected external perturbations.

19.1 Preprogrammed Reactions

There is a group of semiautomatic reactions to muscle length
changes (or, sometimes, to other stimuli) that may be tentatively
termed reflexes. In fact, there is a whole group of terms that have
been used to address these reactions: preprogrammed reactions,
long-latency reflexes, functional stretch reflexes, M2-M3, and
triggered reactions (Phillips 1969; Tatton et al. 1978; Nashner and
Cordo 1981; Chan and Kearney 1982; McKinley et al. 1983). The
variety of terms reflects different understandings of the nature and
functional significance of these reactions. For the sake of
convenience, we will address them as preprogrammed reactions.
One of the most common procedures for eliciting a
preprogrammed reaction is as follows: A subject maintains a
constant position in a joint with a steady muscle contraction against
an external load—for example, one provided by a motor. The subject
is given the instruction “Return to the starting position as fast as
possible in cases of external perturbations.” Unexpected rapid load
changes by the motor give rise to a sequence of EMG events in the
muscle (figure 19.1). The first one corresponds, according to its
latency, to monosynaptic transmission and probably represents the
phasic stretch reflex, similar to the one observed during a tendon tap
(T-reflex). After that, two peaks (sometimes poorly differentiated)
appear at an intermediate latency; the first peak is M2 while the
second one is M3. These peaks are followed by a voluntary reaction.
The latency of the intermediate reactions is typically in the range of
40 to 90 ms. The bottom panel of figure 19.1 shows an actual EMG
record of the activity of a human biceps brachii muscle in response
to its loading (stretch) with clear M2-M3 peaks.
Figure 19.1 An unexpected perturbation of a joint gives rise to a sequence of
EMG events in the stretched muscles. The first one comes at a short latency
(under 40 ms; M1); then, there are two peaks (M2 and M3) that come at a latency
of between 40 and 90 ms. M2 and M3 are considered preprogrammed reactions.
Later, a voluntary correction comes. The bottom panel shows an actual recording
of the reaction of the human biceps muscle to unexpected loading (perturbation).

The latencies of the M2 and M3 responses do not allow viewing

them as components of a very quickly generated voluntary action.
Indeed, the shortest reaction time from a signal (e.g., a beep or a
flash of light) to the initiation of an instructed action is at least 100
ms. In fact, the reaction time becomes much longer, typically at least
200 ms, if the subject does not know what action is needed and has
to choose it from a set of possible actions.
Preprogrammed reactions have been hypothesized to
represented a transcortical reflex (i.e., a reflex whose loop involves
neurons in the brain cortex; see chapter 8), and this idea is still very
much alive (Allum 1975; Marsden et al. 1976a; Chan et al. 1979;
Cheney and Fetz 1984; Day et al. 1991). However, these reactions
(or some other reactions that are very similar to M2-M3) were
reported in decerebrated and even spinalized animals—that is,
animals without neural pathways connecting the cortex with the
spinal cord (Ghez and Shinoda 1978; Miller and Brooks 1981). On
the other hand, studies of the effects of muscle fatigue have
suggested qualitatively different behaviors of M2 and M3, supporting
the idea of different neural loops bringing about these responses. In
particular (see also chapter 31), fatigue is accompanied by strong
suppression of M1, relatively mild suppression of M2, and an increase
in the magnitude of M3 (Duchateau et al. 2002).

PROBLEM 19.1
The difference between the latency of the monosynaptic response
and that of M2 is the same in arm muscles and in leg muscles.
What does this finding tell us about the possible transcortical
nature of M2?
Figure 19.2 Preprogrammed reactions (M2 and M3) demonstrate strong
dependence on the instruction. If the subject is instructed to resist perturbations,
the preprogrammed reactions are large (solid lines); if the subject is asked to let
the limb move (let go), the preprogrammed reactions are much smaller (dashed
lines). Note that the M1 response is the same.

Preprogrammed reactions differ significantly from other reflexes in

that they strongly depend on the instruction to the subject. For
example, if you repeat the same experiment as described above, but
give the subject the instruction “Do not resist external perturbations;
let the motor move your arm,” the amplitude of the preprogrammed
reactions will decrease significantly, and the reactions may even
disappear (figure 19.2). More recent studies have shown instruction
effects even on relatively early components of the responses to
external perturbations (Pruszynski et al. 2016; Reschechtko and
Pruszynski 2020), suggesting that instruction can modify responses
induced by neural loops that do not go beyond the spinal cord.

19.2 Preprogrammed Reactions

Versus Stretch Reflexes
Several experimental findings prevent one from considering
preprogrammed reactions as a kind of long-latency stretch reflex
because preprogrammed reactions do not demonstrate a consistent
dependence between the magnitude of the response and changes in
muscle length. Depending on the instruction to the subject,
preprogrammed reactions can be observed not only in stretched
muscles but also in a muscle shortened by the perturbation or even
in a muscle whose length is not changed by the perturbation at all
(Marsden et al. 1979; Nashner et al. 1979; Nashner and Cordo
1981). In addition, the amplitude of the preprogrammed reactions
does not correlate with the amplitude of the applied perturbation if
the latter cannot be predicted by the subject. Thus, the
compensation of the perturbation by preprogrammed reactions can
vary in different trials from 0 to 100% or even to overcompensation
(Houk 1979).
The fact that these responses are independent of the perturbation
magnitude suggests that the perturbation represents a nongraded
signal for the response generation, a trigger, and the response
magnitude is defined prior to the stimulus on the basis of other
factors. Then, certainly, these reactions can be called triggered or
preprogrammed.
Consider the following scheme for the generation of
preprogrammed reactions (figure 19.3). The instruction to keep a
joint position against a load requires the subject to generate a
voluntary command to muscles controlling this joint. If the subject
knows that a perturbation can occur, a corrective command can be
prepared in advance and is ready to be triggered by an appropriate
peripheral signal. Note that this scheme implies preparation of a
preprogrammed reaction by some “higher” center, such as the
cortex, while the loop of the reaction may be limited to the spinal
cord or brainstem.
Figure 19.3 A subject is holding a position in a joint against a load with a central
command to the muscles. If the subject knows that a perturbation can occur, he or
she can prepare a modification to the central command that would compensate for
the predicted perturbation. The preprogrammed command (ΔC) is triggered by
peripheral signals generated by the perturbation and attenuates the mechanical
effects of the perturbation.

19.3 Afferent Sources of

Preprogrammed Reactions
Viewing preprogrammed reactions as triggered responses to a signal
provided by the perturbation suggests that the search for a unique
afferent source of these responses is likely to provide unreliable
results. Indeed, if the perturbation provides only a triggering signal,
the source of this signal is not really significant (figure 19.4). It must
be sufficient to provide necessary information about the occurrence
of the perturbation. In this context, signals for preprogrammed
reactions can be provided by virtually any group of peripheral
receptors that deliver information on changes in load, position,
pressure on the skin, and, in certain experimental situations, also by
visual, auditory, and vestibular receptors.
That is why experiments with selective blocking of the
transmission along certain afferent systems have not provided
conclusive information on the role of these afferents in the
generation of the preprogrammed responses. The preprogrammed
responses disappear after total deafferentation of the limb (i.e., when
the central nervous system does not receive any signals about the
perturbation) (Marsden et al. 1979; Bawa and McKenzie 1981).

Figure 19.4 Preprogrammed reaction can be induced by triggering signals of

different modalities as long as the signals carry sufficient information. It can be
provided by proprioceptors, by a flash of light, or by a loud tone, for example.

An elegant example of the preprogrammed reactions can be

observed in experiments with the grasp reflex (Traub et al. 1980). A
subject is given an instruction to position his or her thumb and index
finger just near a glass, placed on a table in front of the subject, “as if
going to grasp it.” Although no command to grasp occurs, the
instruction clearly implies preprogramming of a grasping movement
by the subject. This movement is actually observed, at a
characteristic for the preprogrammed response latency, when the
subject’s arm is unexpectedly lifted so that the glass remains below
the hand. Note that in this case the perturbation has nothing to do
with the length of the muscles directly involved in the grasping task.

PROBLEM 19.2
What would you expect to happen if, in the previous example, the
glass suddenly started to fall (watched by the subject at all times)?

Imagine now that these experiments are performed under

anesthesia of different arm segments. If the hand is anesthetized,
the reaction is still there. Then the experimenter anesthetizes the
forearm; the reaction does not disappear. Only after the whole arm is
anesthetized up to the shoulder are preprogrammed reactions
eliminated. These observations may lead to the conclusion that the
afferent source of the observed reaction is in the proximal segments
of the limb. In fact, information from these segments has proven to
be sufficient as a signal for the preprogrammed reaction, but it has
not proven to be necessary. It is quite probable that in a naive
subject the same response could have been observed to an
unexpected loud sound.

19.4 Preprogrammed Reactions

During Movement Perturbations
Up to now, we have considered preprogrammed reactions that occur
when a limb posture is perturbed. Similar reactions, at similar
latencies, can be seen when a perturbation occurs unexpectedly in
the course of a purposeful movement.
 For example, imagine that a subject is performing a fast joint
flexion against a constant external load from a certain initial position
to a certain final position. An unperturbed movement is accompanied
by a typical triphasic EMG pattern, as described in chapter 23.
Unexpected changes in the load can be introduced in different ways,
depending on the equipment; for example, an electrical motor can be
programmed to generate force simulating an increase or a decrease
in the inertial load. If the load unexpectedly increases, an increase in
the activity of the joint flexors (agonists) and a decrease in the
activity of the joint extensors (antagonists) is seen at a latency of
about 70 ms (Gottlieb and Agarwal 1980; Rothwell et al. 1982b).
Alternatively, if the load unexpectedly decreases, a decrease in the
activity of the elbow flexors and an increase in the activity of the
elbow extensors will be seen at a similar latency. As typically
happens with preprogrammed reactions, these adjustments are
unable to fully compensate for the effects of the perturbation and are
followed by later, voluntary corrections.

PROBLEM 19.3
You are carrying a bundle of firewood with extended arms.
Suddenly, all the firewood drops. What kinds of reactions can you
expect in your biceps and triceps?

Similar reactions are seen during a variety of tasks—for example,

in response to unexpected perturbations applied to a handheld
object (Johansson and Westling 1988). If an external force tries to
yank the object away, the grip force increases at a typical delay for
preprogrammed reactions (50-70 ms). These responses are induced
primarily by activation of sensory endings in the glabrous skin of the
fingertips. They produce changes in the activation of digit flexor
muscles that are not directly perturbed by the external force. The
analysis of preprogrammed responses to perturbations applied
during various motor tasks allows us to formulate the following
hypothesis: The execution of any motor task is associated with
preprogramming of fast compensatory reactions to conceivable
external perturbations.
This suggestion is mainly based on common sense since the
execution of familiar movements in everyday life is inevitably
connected with certain unpredictable but frequently encountered
changes in the external conditions of movement execution. These
changes include rapid changes in the external load or obstacles on
the movement’s trajectory. Therefore, one may assume that any
motor task is always associated with preprogrammed compensations
triggered in response to unexpected external perturbations. The
magnitude of a preprogrammed reaction can be generated based on
previous experience of the subject.

19.5 Basic Features of

Preprogrammed Reactions
It has already been noted that these reactions depend significantly
on the instruction to the subject (on what is perceived as major task
components). Namely, the responses take place when the instruction
is, “To compensate as fast as possible…,” and they are small or
absent in the case of, “Do not pay attention.…” The origin of this
phenomenon is obvious since the subject is free to prepare or not to
prepare preprogrammed reactions.
The emergence of preprogrammed reactions in muscles
shortened by the perturbation or in muscles not directly affected by
the perturbation is also quite understandable since the subject can
preprogram any combination of neural commands to any muscle or
muscle group (depending on the given instruction or intention)
independently of the influence of a future perturbation on the muscle
length.
Since the amplitude of the preprogrammed movements should be
defined prior to the perturbation, random changes of the perturbation
amplitude cannot lead to any correlation of the preprogrammed
response amplitude with that of the perturbation. Reproduction of the
same perturbation amplitude in a series of trials should lead to an
improvement of the compensation due to the preprogrammed
responses. Such effects were observed experimentally (Nashner
1976).
As discussed earlier, high-frequency muscle vibration leads to
pronounced suppression of the muscle monosynaptic reflexes.
However, the preprogrammed responses remain unchanged on the
background of the vibration (Lee and Hendrie 1977; Agarwal and
Gottlieb 1980) (figure 19.5).

Figure 19.5 Muscle vibration has different effects on different components of the
responses to an external perturbation. The early response (M1) is suppressed, just
as the H-reflex is, while the preprogrammed response (M2-M3) is unchanged.

PROBLEM 19.4
You apply a sequence of perturbations in the same direction but of
different magnitudes and measure an integral of a
preprogrammed reaction in consecutive trials. Do you expect this
integral to correlate with any characteristic of the perturbations?
 Vibration-induced changes in muscle afferent activity may not be
directly related to preprogramming, and thus, they influence only the
amplitude of a signal for the preprogrammed response playback. On
the other hand, the amplitude of the preprogrammed response does
not depend upon the magnitude of the triggering signal giving rise to
the response, which explains the lack of vibration influence upon
preprogrammed reactions. Note that muscle vibration suppresses
monosynaptic reactions through the mechanism of presynaptic
inhibition, which is a selective inhibition mechanism. Apparently, it
acts on the terminals of Ia afferents on α-motoneurons but not on the
terminals of interneurons that participate in the hypothetical loop
bringing about preprogrammed reactions.

19.6 Preprogrammed Corrections

of Vertical Posture
Unexpected perturbations of the vertical posture (see also chapter
25) bring about compensatory reactions at latencies resembling
those of preprogrammed reactions in limb muscles (i.e., intermediate
between the latencies of the phasic stretch reflex and voluntary
reactions) (Nashner 1976; Nashner et al. 1979; Allum 1983). Such
reactions have been observed during maintenance of the vertical
posture and during walking (Nashner 1980; Dietz et al. 1984).
Perturbations of vertical posture come not only from the environment
but are frequently generated by a person’s own movements. As the
reader will see in the chapter on postural control, fast arm
movements create large reactive torques acting on the trunk that
effectively perturb vertical posture.
Maintenance of the vertical posture is probably the most common
motor task that is a component of many voluntary movements.
Therefore, it is reasonable to assume that the mechanism of vertical
posture control is very well defended against possible unexpected
changes in external conditions. This means, in particular, that
different preprogrammed corrections of the vertical posture are ready
to be initiated in response to certain triggering signals without any
special instructions. The complicated problem of maintaining the
vertical posture in the field of gravity requires relatively complex
corrective reactions involving the activation of different muscle
groups of the legs and trunk (Lisin et al. 1973; Nashner 1980).
Changes in the activity of postural muscles have been seen in
response to perturbations applied to different parts of the body of a
standing person. These corrective postural reactions are specific
to the mechanical effects of the perturbation on body equilibrium but
not to the exact point of application of the perturbation. They can be
seen, in particular, in muscles whose length is not directly affected
by the perturbation. According to our general view, these reactions
represent preprogrammed motor commands realized when
peripheral signals inform about a perturbation independently of the
afferent source.
One of the most common methods of studying postural reactions
to unexpected perturbations has been to introduce quick rotations or
translations of the platform on which the subject is standing. Such
perturbations lead to reactions in leg and trunk muscles at the
latencies typical of preprogrammed reactions (Nashner 1976;
Nashner and McCollum 1985). Patterns of these reactions have
been shown to be sensitive to the postural task; for example, they
change when a person stands on a narrow supporting surface
(Horak and Nashner 1986). These reactions will be discussed in
more detail in the chapter on postural control.
Sometimes, humans use arm muscles for additional support
during standing. Then corrective postural reactions can be seen in
these muscles in addition to reactions in postural muscles of the legs
and trunk (Nashner 1982). For example, imagine a subject who
stands grasping an object for additional support. When the object is
fixed to the floor (i.e., it provides reliable support), a certain pattern of
preprogrammed reactions can be seen in the shoulder muscles.
However, if the handheld object has low inertia (e.g., a cane), the
responses can invert and emerge in antagonistic muscle groups, and
the overall pattern of the movement resembles that observed in a
person holding a cup of tea in his or her hand in response to a
postural perturbation.
Consider another example. A person is standing on a platform
holding in front of him or her a cup loaded with playdough (figure
19.6). The platform starts to move unexpectedly. There will be
changes in the person’s posture, including the posture of the arm
holding the cup. If an experimenter records the activity of muscles
that maintain the vertical posture of the body and the posture of the
arm, there will be changes in the levels of muscle activation. Some
of these changes will happen at latencies typical for preprogrammed
reactions. Now, imagine that the same experiment is repeated, but
the cup is loaded not with playdough but with hot tea. The platform
starts to move in exactly the same fashion. Imagine that all the
forces, all the initial postures, and all the other conditions are the
same (by the way, this is impossible to assure). Preprogrammed
changes in the muscles involved in the person’s posture and his or
her arm movements will be quite different. This difference is due to
the difference in the person’s intentions. In the first case, the major
task was not to fall down, while there were no major restrictions on
arm movements. In the second case, not spilling the tea becomes
comparably important to not falling down. So human intentions
contribute to the patterns of preprogrammed reactions even when
the mechanics of the postural task remain seemingly the same.
Figure 19.6 Preprogrammed postural corrections to a perturbation created by
platform movement are context-dependent. They will be different when the cup is
filled with playdough compared to a situation when it is filled with hot tea.
Reprinted by permission from M.L. Latash, “The Bernstein Problem: How Does the Central
Nervous System Make its Choices?,” in Dexterity and its Development, edited by M.L.
Latash and M.T. Turvey (Mahwah, NJ: Erlbaum, 1996), 279. Reproduced with permission of
The Licensor through PLSclear.

PROBLEM 19.5
You are picking up a small object that turns out to be very heavy
(unexpectedly for you). What kinds of reactions do you expect to
see in the arm muscles?
19.7 Corrective Stumbling
Reactions
Locomotion is another very commonly used movement in everyday
animal and human activity (chapter 26). Therefore, it is reasonable to
expect the neural mechanism of locomotor movement generation to
always be associated with preprogramming of corrective actions to
perturbations. This is particularly true for bipeds, such as humans,
who have to combine the stability of locomotion with the stability of
the vertical posture.
A particular pattern of long-latency reactive responses has been
observed during cat (and human) locomotion associated with
overcoming an unexpected obstacle: the corrective stumbling
reaction (Forssberg et al. 1975, 1977; Duysens and Pearson, 1976;
Duysens et al. 1990; Prochazka et al. 2002). This pattern can be
observed in cats in response to weak mechanical stimulation of skin
areas of the paw or of the leg (even with an air puff) or during short
electrical stimulation of skin nerves.
The application of any of these stimuli during the swing phase
gave rise to a flexor reaction with the hindlimb transferring over a
hypothetical obstacle (figure 19.7). The same stimulation applied
during the stance phase could give rise to an extensor reaction
accelerating the step. The coordinated, functionally appropriate
pattern of this reaction and the relative independence of the stimulus
suggest that it is a preprogrammed response of a mechanism
responsible for the compensatory reactions needed during everyday
life. For example, imagine that one hits a stone or a branch with the
tip of the foot during the swing phase. The corrective stumbling
reaction will contribute to lifting the foot and stepping over the
obstacle. In another situation, imagine stepping on a sharp object
(e.g., a tack). Given that the other foot is in the swing phase, lifting
the foot off the sharp object is impossible because the body would
collapse. The only feasible strategy is to accelerate the step to
minimize the time the object is in contact with the foot.
Figure 19.7 A mechanical or electrical stimulation of the paw during locomotion
induces different reactions in the swing and the stance phases. In the swing phase
(A), there is a flexor reaction, so that the leg “steps over” a fictitious obstacle. In
the stance phase (B), there is an extensor reaction, leading to shortening of the
stance phase for this limb.

PROBLEM 19.6
Can you define the afferent source of the corrective stumbling
reaction with the method of successive limb denervation—that is,
eliminating afferent inputs from areas of the leg?

One can present an example of a very complex but obviously

preprogrammed reaction: Imagine an officer giving commands to
soldiers: “Lie down,” “stand up,” “lie down,” “stand up,” …, “lie down,”
“stand up,” “stand up!” It is obvious that some of the participants in
this experiment would lie down in response to the last command.
After numerous repetitions of the presented sequence, the motor
reactions become preprogrammed, and the voice of the officer plays
the role of a triggering signal.
CHAPTER 19 IN A NUTSHELL
A group of reactions to external
stimuli (commonly, to mechanical
perturbations) that come at a latency
longer than typical spinal reflex
latencies and shorter than voluntary
reaction time are called preprogrammed
reactions, triggered reactions, or
long-latency responses. These
responses can be modulated by prior
instruction. They produce quick, crude
corrective actions counteracting the
mechanical effects of the original
perturbation. They can be seen in
muscles whose length is increased,
decreased, or not changed by the
perturbation. The exact sensory source
of preprogrammed reactions is not
important as long as it is sufficient
to provide information about a
perturbation. Preprogrammed responses
are important components of all
everyday activities, particularly of
prehensile tasks, standing, and
locomotion.
Problems for Part IV
Self-Test Problems
1. A pair of very thin electrodes are placed close to a group of
neural fibers in a dorsal root of the spinal cord. Brief
stimulation of these fibers leads, after a delay of about 25 ms,
to an increase in the activity of a flexor muscle and a drop in
the activity of an extensor muscle at a joint innervated by the
corresponding spinal segment. When the joint is moved
passively into flexion, the fibers show an increase in their
activity. A twitch contraction of the extensor muscle produced
by an electrical stimulus applied to the muscle is accompanied
by an increase in the activity of these fibers. Where do these
fibers originate? Explain your answer.
2. A subject is performing a series of very fast elbow flexions
over 40° against a constant external load. Movement time is
200 ms. Unexpectedly, in one trial, the movement is
completely blocked. Draw time patterns of the biceps and
triceps EMGs.
3. Two identical vibrators are placed over the quadriceps femoris
muscle group and over the hamstrings of a standing person.
They vibrate at 100 Hz with the amplitude of 1 mm. What
effects on muscle activation levels and monosynaptic reflexes
can be expected?
4. A drug is taken orally by a person. As a result, the amplitude
of the H-reflex in a muscle is decreased by 80%, while the
amplitude of the tonic vibration reflex stays unchanged.
Suggest possible mechanisms of action of the drug. What
changes may be expected in the peak muscle force during
maximal voluntary contractions?
5. A new drug reverses the action of Renshaw cells, turning their
projections into excitatory synapses. What behavioral changes
would you expect under the action of this drug? What will
happen with the dependence of the H-reflex amplitude on the
strength of the electrical stimulus?
 6. A subject is holding a position in a joint against a constant
external load. The instruction is “Try your best to keep this
position in cases of possible perturbations.” Unexpectedly, the
external load increases and moves the joint. Describe all the
mechanisms that help the subject to follow the instruction.

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. A person unexpectedly steps on a sharp tack while walking.
What is the main mechanism leading to an adjustment of the
walking pattern?
a. tonic stretch reflex
b. monosynaptic reflexes from Golgi tendon organs
c. monosynaptic reflexes from muscle spindles
d. preprogrammed reactions (M2-M3)
e. all of the above
Why?
2. Ib afferents from Golgi tendon organs
a. mediate reciprocal inhibition
b. mediate recurrent inhibition
c. cause both inhibitory and disinhibitory reflex effects on α-
motoneurons
d. inhibit the activity of antagonist α-motoneurons
e. project on interneurons that cause excitation of their targets
Why?
3. A short electrical pulse applied to a muscle nerve in a human
leg
a. can induce a reflex muscle response at a delay of about 5
ms
b. excites both afferent and efferent fibers
 c. induces action potentials, which travel only toward the
spinal cord
d. can induce both an H-reflex and an F-wave in the same
muscle
e. all of the above
Why?
4. A person performs a series of fast elbow extensions against a
stop. In one trial, the joint is unexpectedly released. What will
happen during the first 5 ms as compared to regular trials?
a. Both biceps and triceps EMG bursts will increase.
b. Both biceps and triceps EMG bursts will decrease.
c. Both EMGs will stay the same, but muscle forces will
change.
d. Biceps EMG bursts will increase, triceps EMG bursts will
decrease.
e. Biceps EMG bursts will decrease, triceps EMG bursts will
increase.
Why?
5. Presynaptic inhibition
a. suppresses the excitability of a neuron to all inputs
b. involves depolarization of the presynaptic membrane
c. involves hyperpolarization of the postsynaptic membrane
d. blocks the action potential from moving toward the synapse
e. all of the above
Why?
 Part V

Control and Coordination of

Goal-Oriented Movements
 Chapter 20

Voluntary Control of a Single

Muscle

KEY TERMS AND TOPICS

elements of the control theory
feedforward and feedback control
servo hypothesis
α-γ coactivation
equilibrium-point hypothesis
the role of reflexes in muscle control

Functional movements are performed by groups of muscles recruited

in task-specific coordination fashion. So one can ask whether
studying the control of a single muscle is meaningful. There are good
reasons to explore the control of individual muscles, however,
including the fact that it is easier to introduce concepts using, as
examples, relatively simple systems such as intact muscles. Indeed,
muscles are the simplest systems that allow us to follow
Sherrington’s traditions on the role of reflexes in voluntary
movements and to discuss how the neurophysiological subsystems
described in part IV can be involved in the control of voluntary
movements. Even for a single muscle, there is evidence for
numerous reflex mechanisms originating from different types of
peripheral receptors and mixed at the level of interneurons. All these
mechanisms are likely to be mediated by thousands of neurons
during voluntary movements. It is certainly unrealistic to try to trace
each and every reflex pathway and to describe its functioning during
voluntary muscle activation. However, we can consider a different
question: Is it possible to describe the action of the whole variety of
muscle reflexes with just a few parameters? In order to do so, we
need to move from consideration of individual reflex pathways to a
new level of analysis that will be characterized by a new set of
functionally significant variables and parameters—just as we did
when we proceeded from the physiology of membranes and ions to
the physiology of cell interactions (reflexes). But first, we have to ask
a very basic question: What kinds of movements deserve to be
classified as voluntary? This question was partly addressed earlier
(see chapter 15) when we tried to define reflexes.

20.1 What Is Voluntary

Movement?
While there are a number of characteristics that distinguish voluntary
movements from the other end of the motor spectrum—reflexes—it
is hard to define voluntary movement at this point in the textbook.
The reason is that we have not yet introduced a scheme with
explicitly defined neurophysiological variables involved in the control
of movements that could be called voluntary. As it follows from figure
15.1 in chapter 15, compared to reflexes, typically, voluntary
movements are expected to:
take more time to be initiated or corrected,
involve longer neural loops with more synaptic connections,
show larger indices of variability,
involve large groups of muscles distributed across limbs and the
body, and
not be necessarily tied to an external sensory stimulus.
 Indeed, sometimes the time of movement initiation is taken as a
defining sign of movement being voluntary. For example, even in
very well trained persons, the first sign of changes in muscle
activation takes at least 100 ms. This is seen, for example, under
simple reaction time conditions when the subject is asked to initiate a
simple action (e.g., press the button) as quickly as possible following
a signal, typically auditory or visual. On the other hand, there are
preprogrammed reactions (chapter 19) that are not viewed as
voluntary movements, which can come at only slightly shorter
delays. There are also reflexes involving supraspinal structures,
such as, for example, the startle reflex (chapter 13) that come at
delays comparable with those of the quickest voluntary movements.
Neural loops involved in voluntary movements are typically
unknown. So the second criterion remains not easily applicable.
There are variable reflexes and there are those involving large
muscle groups, such as flexion and crossed extension reflexes
(chapter 18). So the only solid criterion is whether a person can
initiate a movement without any sensory stimulus. In such cases, the
movement can definitely be classified as voluntary. We will return to
this issue after introducing a hypothesis on the nature of
neurophysiological variables used to produce voluntary movements.

20.2 Feedforward and Feedback

Control
Moving into the area of control requires the introduction of a few
basic notions from the field of control theory. Consider figure 20.1.
We assume that the controller wants to induce changes in state
variables, X(t), that describe the object of control (e.g., length or
force of a muscle, angle, or moment of force in a joint, or spatial
coordinates of the limb or the whole body) from their current values
to some target values. A control signal, C(t), is coming from a
hypothetical central controller and, after some processing, produces
certain changes in state variables with an important role played by
the environment. Variables that are used by the controller to
formulate command signals are called control variables. Within the
central nervous system, these are neurophysiological variables that
can be expressed in units of voltage and time. It is commonly
assumed that these variables may be supplied by the controller,
ignoring possible changes in state variables or in other external
factors. Certainly, this does not mean that the controller cannot
change a control variable based on peripheral information. What is
important is that it has a choice of reacting or not reacting to this
information.

Figure 20.1 A scheme of feedforward control. The task is formulated as desired

changes in salient variables, X(t). The controller converts the task into control
variables, C(t). They serve as an input into the effector, which produces actual
magnitudes of the salient variables.

If the controller supplies the signal (a variable or a few variables)

independently of the values of state variables, this type of control is
termed feedforward control. A typical example of feedforward
control is kicking a soccer ball. The brain sends commands to the
muscles during a kick that are generated before the kick and
certainly before the outcome of the kick becomes known.
 If a controller changes command signals on the basis of their
effects on state variables, this type of control is termed feedback
control (figure 20.2). An important component of a feedback control
system is a comparator (i.e., a unit that compares the current values
of state variables with their desired values and changes command
signals based on the discrepancy between the actual and desired
effects). For example, when you drive a car at a certain constant
speed, you use visual information from the speedometer or from the
moving environment to change the force with which you are pressing
on the gas or brake pedal. This type of control allows you to maintain
a preferred speed when the road goes uphill or downhill, when the
wind changes, or when you see a police car.
A negative feedback loop acts to modify a control variable (x) by
an amount (Δx) related to a deviation in a peripheral state variable
(Δy) from its desired value, typically leading to a decrease in the
original deviation (figure 20.2a). Such a system tends to maintain a
certain value (or a certain time function) of the output variable y.
Positive feedback acts to modify a control variable by an amount
related to a deviation in a peripheral state variable typically leading
to an increase in the original deviation (figure 20.2b). Positive
feedback systems tend to amplify any deviations of peripheral
variables and commonly get out of control, leading to a qualitative
change in the behavior of the system (cf. action potential generation
described in section 3.3, chapter 3). Both negative and positive
feedback can use proportional control when Δx ~ Δy or any other
relation between the magnitudes of Δx and Δy.
Figure 20.2 Feedback control modifies control variables, C(t), based on the
actual output of the effector. The comparator (a) compares the desired and actual
X(t) and generates a correction, ΔC(t), to the ongoing C(t). Negative feedback (b)
changes the output of the comparator to bring down any possible deviations
(“error”) of the output, while positive feedback (c) amplifies any deviations of the
output.

Two important parameters characterize feedback loops, gain and

delay. Gain can be defined as the ratio of a change in a control
variable to a change in a peripheral state variable (Δx/Δy). Delay can
be measured in time units—for example, in seconds or milliseconds,
or in relative timing units (e.g., in percentage of a time interval typical
of the process). Both positive and negative feedback loops achieve
their functional purpose of amplifying or decreasing error only if their
gain is high enough and their delay is within certain limits. Large
delays can, however, lead to unexpected circumstances. For
example, figure 20.3 shows what can happen if a steady output
signal is perturbed by a function sin(t), and a negative feedback loop
is acting with the gain of 0.9 and time delay (Δt) of either 0 or half the
period of the perturbing function (π). In the first case, there is a
nearly perfect compensation of the perturbing function. In the second
case, the effects of the perturbation are actually amplified, although
a negative feedback loop is acting. Thus, while talking about
negative or positive feedback loops, one needs to distinguish
between actual structural components of the system and their overall
functional effects.

Figure 20.3 Large delays in a feedback loop may lead to unexpected

consequences. For example, in this figure a perturbing signal (a sine function with
the peak value of 1) is acting on the output of a system that includes a negative
feedback with the gain of 0.9. If the time delay in the feedback loop is zero, the
error is almost ideally compensated (thin line); if the time delay is π, the error is
amplified (bold line).

Time delay is an important drawback of feedback control.

Therefore, if speed is vital, feedforward control may become
preferred, while, if accuracy is more important, feedback control has
an advantage.

PROBLEM 20.1
Suggest examples of negative and positive feedback from the
material we have already considered in this book and from
everyday life.

20.3 Servo Control

Feedback and feedforward types of control are frequently combined
into schemes of different complexity that have an ability to generate
command signals in a feedforward manner and to correct the control
variables—if their effect is different from some desired outcome—
using feedback. When a cat chases a mouse, it uses a combination
of feedforward and feedback control. On the one hand, it tries to
predict what the mouse will do and to intercept it (feedforward
control), while on the other hand, it uses visual information and
corrects its movements based on actual movements of the mouse
(feedback control). Schemes combining feedforward and feedback
control frequently consist of a few control circuits. Among different
types of control circuits, let us single out the servo mechanism,
which is sometimes simply called the servo.

Figure 20.4 This circuit combines feedforward and feedback mechanisms of

control. The feedback loop (the servo) keeps a value of a variable (x0) specified by
the controller in a feed-forward way constant despite possible changes in the
external conditions that may potentially cause changes in this variable.
 In figure 20.4, a signal is sent by the controller in a feedforward
manner to the servo loop. The signal encodes a desired value of the
output state variable that needs to be kept constant. The servo loop
is taking care of keeping this value constant with the help of a
feedback mechanism. A sensor is used to measure current values of
the state variable and to supply these measurements to the
comparator. The comparator compares the measured value to the
specified one and changes its output (Δx) based on the error (i.e.,
the difference between the prescribed and actual values). Note that
the presence of errors is a necessary component of the functioning
of a servo. Good servos allow only very small errors to emerge and
correct them promptly; in other words, they have high gains and
small time delays, while bad servos may have considerable delays in
their corrective actions, so errors may be rather large.
A thermostat is a typical example of a servo controller (figure
20.5). It keeps room temperature constant by using a comparator
that compares actual room temperature to its preset value. If the
temperature differs from the preset value by a large enough margin,
a heater or an air conditioner is turned on. Note that you specify the
input into this servo by setting the dial of the thermostat.
Figure 20.5 A scheme of a simple thermostat that keeps room temperature
constant.

PROBLEM 20.2
Suggest an area of sports in which having a servo mechanism
would be vital for good performance. Suggest another area in
which such a mechanism would be disastrous.

We have already mentioned an important characteristic of any

servo (or any feedback system), namely its time delay. Apparently,
the bigger the delay, the more errors are likely to accumulate before
a corrective action by the servo takes place. In electrical systems,
the errors may be very small. However, in the human body, the
speed of information transmission is limited by the speed of action
potential propagation. Thus, delays of several tens and up to a
hundred milliseconds are common. These delays are comparable to
the movement time of the fastest voluntary movements; thus, even
the best servos within our bodies are not likely to function very
efficiently.
 The servo is an autonomic element of a control system. This
means that setting a desired value of an output parameter makes a
servo mechanism do its job independently of other factors as long as
the specified value remains constant. Using servos apparently
simplifies control within a complex system because part of the
responsibility may be delegated to the “lower” servos, and the
controller may ignore the details and concentrate on specifying more
general and important variables.

20.4 Servo Hypothesis

In the early 1950s, R.A. Merton suggested a hypothesis that was the
first control hypothesis in the area of human movement studies. This
hypothesis, called the servo hypothesis, was the first to formulate
the problems of generation of voluntary movements in terms of
control using the information about the neurophysiological
mechanisms of muscle reflexes. Merton suggested that the control of
muscle spindle endings with the γ-system was part of a servo
mechanism controlling muscle length (Merton 1953; also see
Matthews 1972).
The main idea of the servo hypothesis is illustrated in figure 20.6
and involves the following steps:
1. A descending signal comes to γ-motoneurons and thus
changes the sensitivity of the sensory endings in muscle
spindles to muscle length. The effects of an increased γ-
activity are similar to the effects of an increase in muscle
length (both lead to an increase in the spindle afferent activity
level), while a decrease in γ-activity is similar to the effects of
a decrease in muscle length. So, one may say that the
descending signal in figure 20.6 simulates a new value of
muscle length.
2. The activity of the spindle endings changes and, via the
mechanism of the tonic stretch reflex, leads to a change in
the level of activity of α-motoneurons innervating the muscle.
The original version of the Merton’s hypothesis was based on
 the monosynaptic action of Ia muscle afferents on
homonymous α-motoneurons. However, later, the mechanism
underlying the servo action was reconsidered and based on
the polysynaptic tonic stretch reflex.
3. The level of muscle contraction changes, leading to
movement (i.e., to a change in muscle length). Assume that
there are no changes in the external load (i.e., isotonic
conditions). Note that an increase in spindle activity will lead
to an additional muscle contraction leading to muscle
shortening, which will lead to a decrease in the spindle
activity. So, the tonic stretch reflex mechanism acts as a
negative feedback system.
4. The movement will continue until muscle length comes to a
new value at which the activity of muscle spindles leads to a
muscle contraction exactly balancing the external load (i.e., to
a new equilibrium state).
Figure 20.6 The servo hypothesis of Merton considers the feedback loop muscle
length => muscle spindle endings => tonic stretch reflex (TSR) loop => activity of
α-motoneurons => changes in muscle force => movement => changes in muscle
length to be a perfect servo. Descending signals specify a level of activity of γ-
motoneurons, thus setting a desired value of muscle length implemented by the
servo.

If the descending command remains constant, the mechanism of

the tonic stretch reflex is assumed to ensure constant muscle length
despite possible changes in external load (i.e., to work as a perfect
servo). For example, if the load increases, it leads to an increase in
the muscle length, which in turn leads to an increase in the activity of
α-motoneurons and an increase in muscle contraction force.
According to the Merton’s servo hypothesis, the increase in muscle
force will exactly balance the external load change so that muscle
length will not change.
Figure 20.7 The servo hypothesis of Merton may be illustrated with vertical
force–length curves whose position along the abscissa axis is defined by the
descending signal (γ1, γ2, and γ3). Note that even a large change in the external
force is assumed to be perfectly compensated by the tonic stretch reflex
mechanism (i.e., this mechanism has a very high [infinite] gain).

Figure 20.7 illustrates how the servo hypothesis functions with the
help of force–length muscle characteristics. The central command
specifies the location of a characteristic corresponding to a certain
value of muscle length. In order for the servo mechanism to ensure
perfect compensation of possible changes in external load, the
characteristic must be vertical; then muscle length will not depend on
muscle (and external) force. Voluntary movements are performed by
shifting the characteristics along the x-axis so that the independently
controlled variable may be associated with the signal to γ-
motoneurons (γ1, γ2, γ3 in figure 20.7).

PROBLEM 20.3
What will happen, according to the servo hypothesis, if you try to
activate a muscle to produce a movement but the movement is
unexpectedly blocked (isometric conditions)?
20.5 α-γ Coactivation
The servo model made a number of predictions that could be tested
in experiments. Unfortunately for the servo model, the findings in
those experiments did not follow the predictions. In particular, the
servo model predicts that voluntary movements are initiated by a
change in the activity of γ-motoneurons, while changes in the activity
of α-motoneurons follow at a delay characteristic of the tonic stretch
reflex arc. The development of new methods, in particular direct
recordings from human peripheral nerves pioneered by Swedish
scientist Vallbo (1971, 1981), allowed direct assessment of the
relative timing of changes in the activity of α- and γ-motoneurons
during voluntary muscle contractions. These observations showed
that during virtually all voluntary movements, changes in the activity
of α- and γ-motoneurons happen simultaneously. This phenomenon
is termed α-γ coactivation, and it is illustrated in figure 20.8.

Figure 20.8 The modified version of the servo hypothesis assumes signals are
sent by the brain simultaneously to α-motoneurons and γ-motoneurons (α-γ
coactivation).
 In the mid-1960s, researchers were very reluctant to abandon the
servo hypothesis, and they suggested that the servo mechanism
worked as postulated by Merton while voluntary movements were
initiated by a combination of a feedforward command signal (to α-
motoneurons) and a signal to the length-controlling servo (to γ-
motoneurons) (Matthews 1970, 1972).

PROBLEM 20.4
What will happen with the activity of spindle afferents in a flexor
muscle during a voluntary increase in flexion force against a stop?

PROBLEM 20.5
What will happen with the activity of spindle afferents in a flexor
muscle during a fast flexion movement against a constant external
load?

This is a very elegant model, indeed! Note, however, that it implies

a very high (actually an infinite) gain in the tonic stretch reflex loop,
so that any change in the external load (which may be rather large)
is readily balanced by a change in muscle force, while the change in
muscle length is assumed to be extremely small (in fact, zero).
Unfortunately for the servo hypothesis, later measurements of the
gain in the tonic stretch reflex arc demonstrated relatively low values,
so the mechanism of this reflex cannot be considered a perfect
servo. Thus eventually the servo hypothesis was replaced by new
models.

20.6 Voluntary Activation of

Muscles
At present, two views exist on the nature of voluntary muscle
activation. According to the first view, central neural commands
directly specify the desired levels of activity of α-motoneuronal pools,
and therefore the levels of muscle activation (Gottlieb 1996; Gottlieb
et al. 1989; Schiedt and Ghez 2007). Reflex mechanisms are
assumed to play a minor role, mostly in cases of unexpected
changes in the external forces (perturbations). This view, however, is
incompatible with some observations. For example, imagine that a
person activates a muscle to a high level against a large load and is
then instructed not to change the level of muscle activation (figure
20.9). Now imagine that the load is unexpectedly and quickly
removed. A fast movement will occur. Recording the electromyogram
of the muscle will show that, immediately after the unloading, there is
a period of virtually total silence in the muscle activity. This effect is
called the unloading reflex. Note that the unloading leads to a quick
shortening of the muscle (a decrease in its length at a high negative
velocity) so that the sensory endings in the muscle spindles become
silent, and their reflex effects on the homonymous α-motoneurons
disappear. So the unloading reflex may be considered an inverse of
the stretch reflex. The disappearance of the muscle activity during
the unloading reflex shows that the reflex effects may be strong
enough to eliminate 100% of voluntary muscle activation and
certainly cannot be considered “minor additions.” Thus, the
alternative view looks much more attractive. We will return to
arguments between the competing motor control hypotheses later in
chapter 21.
Figure 20.9 The subject is holding a position in the elbow by activating the
biceps. The load is suddenly removed. Biceps EMG shows a period of complete
silence (the unloading reflex) even if the subject tries to keep the biceps activity
constant. This means that muscle EMG is not an independently controlled
variable.

20.7 Equilibrium-Point
Hypothesis
According to the second view, central commands use the
mechanisms of muscle reflexes to induce changes in the levels of
muscle activity and to specify parameters of these reflexes. This
view is compatible with all the observations on the reflex effects on
voluntary muscle activation. On the other hand, it does not go to an
extreme of implying an infinite gain in any of the reflex arcs, and thus
it avoids the problems of the servo hypothesis. This view emerged
as a formal language for describing a body of experimental data on
observations of single-muscle force–length characteristic curves in
animal preparations and single-joint torque-angle characteristic
curves in human subjects (Matthews 1959; Asatryan and Feldman
1965; Feldman 1966; Feldman and Orlovsky 1972; reviewed in
Latash 1993). Animal experiments were typically performed on cats
with a lesion of the central nervous system, so that the cats lacked
the ability to make voluntary movements. Then, an electrical
stimulator was placed on the residual part of the brain. This
stimulator simulated different descending commands.

Figure 20.10 According to the equilibrium-point hypothesis, muscle reflexes

specify a relation between muscle force and muscle length—an invariant
characteristic (IC), which is defined by setting a magnitude of the tonic stretch
reflex (TSR) threshold. The system “muscle + load” is in an equilibrium when
muscle force is equal to the external force (load). This point is termed the
equilibrium point (EP). If the external load changes, both muscle force and length
will change, corresponding to a new EP (EP2). Note that muscle activity (EMG)
changes along the IC.

At a fixed level of stimulation and a constant external load, the

muscle-load system will be in an equilibrium at a certain length. The
combination of muscle length and force at the equilibrium is called
the equilibrium point (figure 20.10). This is a central notion within this
hypothesis, which is known as the equilibrium-point hypothesis
(Feldman 1966, 1986; reviewed in Latash 1993; Feldman and Levin
1995). A change in the external load leads to a change in muscle
length that induces changes in the level of muscle activation via the
tonic stretch reflex arc. Thus, a constant descending command does
not mean a constant level of muscle activation. Changed muscle
activation can be accompanied by changes in muscle length, force,
or both until a new equilibrium point is achieved. For a fixed
descending command, all the equilibrium points form a curve on the
force–length plane that is termed the invariant characteristic (IC).
The term invariant has been criticized because the shape of this
characteristic can change; in particular, it is not identical during
muscle loading and unloading (Gottlieb and Agarwal 1988).
Therefore, it is better to address it simply as a force–length
characteristic.
If the descending signals in an animal experiment were changed
by using a different level of the electrical stimulation, a new invariant
characteristic emerged, shifted with respect to the first one (figure
20.11). One can introduce a variable encoding the location of a
force–length characteristic—for example, the point at which
activation of α-motoneurons occurs (the threshold of the tonic stretch
reflex, shown as TSR threshold in figure 20.10). This variable may
be viewed as a control variable as defined earlier because changes
in the external load are able only to move the equilibrium point along
the force–length characteristic.
Figure 20.11 Central command specifies the location of an IC for the muscle.
This can be described as a shift in the threshold of the tonic stretch reflex (λ).
Depending on the external load, a shift in λ can lead to a change in muscle length
(isotonic conditions, EP1), or in muscle force (isometric conditions, EP2), or in both
(elastic load, EP3). Note that EMG changes will depend on both central λ changes
and the external load.

Within this scheme, movements may occur because of changes in

the external load, as in figure 20.10, or induced by shifts of the
force–length characteristic, as in figure 20.11. Note that a shift of the
characteristic may have different peripheral effects, depending on
the external load. In figure 20.11, a standard shift of the force–length
characteristic may lead to a change in muscle length (cf. EP0 and
EP1, isotonic conditions), or in muscle force (cf. EP0 and EP2,
isometric conditions), or in both (cf. EP0 and EP3, if the load is
elastic).

PROBLEM 20.6
How can one change the velocity of a voluntary movement
according to the equilibrium-point hypothesis?

As far as the neurophysiological mechanisms are concerned, the

equilibrium-point hypothesis assumes that the tonic stretch reflex,
providing for the muscle force–length characteristics, incorporates all
the reflex loops that can be influenced by levels of activity or
excitability of γ-motoneurons, α-motoneurons, and interneurons. It
does not single out one anatomical structure as the most important.
Moreover, the feedback loops from receptors other than muscle
receptors (e.g., skin and subcutaneous receptors) can also play an
important role in defining the shape of the muscle force–length
characteristics. Thus, it is assumed that central commands for
voluntary movements involve a balanced combination of signals to
all types of spinal neurons, including α-motoneurons, γ-
motoneurons, and interneurons. Actual levels of muscle activation
(electromyograms), muscle forces, and movements occur as
consequences of these central commands.

CHAPTER 20 IN A NUTSHELL
Control signals can be generated in a
feedforward fashion or based on
feedback signals. Positive feedback
tends to amplify the error, while
negative feedback tends to eliminate
errors; however, the exact effects
depend on gains and time delays of
feedback loops. A servo is a
particular case of feedback loop that
keeps output at a preset value.
Merton’s theory considers muscle
spindles to be the sensor in the
perfect servo controlling muscle
length; the theory has been proven to
be wrong. Voluntary activation of a
muscle is accompanied by simultaneous
activation of the α- and γ-motoneurons.
The equilibrium-point hypothesis
considers control of voluntary
movements as a process of central
modulation of the threshold of the
tonic stretch reflex for the
participating muscles. The tonic
stretch reflex and peripheral muscle
and tendon elasticity provide for the
springlike behavior of muscle. Reflex
effects play an important role in
defining levels of muscle activation.
 Chapter 21

General Issues of Motor Control

KEY TERMS AND TOPICS

force control
generalized motor program
internal models
equilibrium-point control
equifinality
dynamic systems

Motor control has two obvious components, “motor” and “control.” In

chapter 20, we reviewed a few basic notions from the control theory.
It is important to realize, however, that the notions and the
computational apparatus of the control theory have been developed
to describe control of inanimate objects built by humans. There are
many important differences between the typical designs of artificial
controllable objects, such as robots, and the human body. To name a
few:
1. Human (and animal) movements are produced by force-
generating structures that have features that may look
suboptimal to an engineer. In particular, muscles produce
forces that are length- and velocity-dependent. In addition,
they act on bones via another elastic structure, the tendon
(chapter 5). As such, muscle forces always depend on the
 actual external load because changes in the external load
obviously lead to changes in movement kinematics, thus
affecting muscle length and velocity. In artificial systems, most
commonly, movements are produced by nearly perfect force-
or torque-generating actuators (i.e., systems that can
generate a predefined time pattern of force or torque
independently of the external load).
2. Muscles are relatively slow (i.e., they take time to produce
force after getting a signal from the central nervous system)
(chapter 5). In robotics, actuators can produce forces that rise
much faster with little delay. Muscles change their activity in
response to a change in external conditions via reflex
pathways that take time to conduct action potentials (chapters
17 and 18). Control circuits in most artificial systems are
nearly instantaneous.
3. Engineers prefer to deal with linear systems (i.e., those that
react to a small change in an input signal by a proportionally
small change in the output). Muscles and their interactions
with the central nervous system are essentially nonlinear
because of the threshold properties of neurons.
From the point of view of a 21st century engineer, these
differences may look like major flaws or at least major complicating
factors in the design of the human body. To control movements of
this apparently poorly designed structure, the central nervous system
needs to be computationally powerful and equipped with a lot of
sophisticated tools to be able to predict the consequences of the
aforementioned properties of muscles and neural pathways on motor
performance and to introduce adequate corrections.

PROBLEM 21.1
Suggest one or two more examples of the apparently suboptimal
design of the body with respect to movement production.
 An alternative philosophical attitude to the design of the human
body is that this design is perfect given the whole variety of tasks
humans face in everyday life. This design allows motor performance
to display both stability and flexibility when external conditions
require a switch in motor patterns. According to this view, we should
not be ashamed of our bodies but proud of their design. And the
central nervous system does not need computational “crutches”;
instead, motor control researchers need better theories.
The differences that have been mentioned between the designs of
the human body and those of artificial systems engineered by
humans should also make one very cautious when trying to apply
schemes developed for the control of artificial systems to the neural
control of movements in humans and animals. Nevertheless,
knowledge of the fundamentals of the control theory is essential for
the first steps in the area of motor control. The exact formulations
and rigorous mathematical apparatus of the control theory are
stimulating examples for motor control researchers who should strive
to develop a comparably rigorous approach for the control of
biological movements.

21.1 Force Control

One of the most intuitive schemes of control comes from a rather
straightforward analysis of the human motor system as a system that
needs to produce a certain mechanical output. Most everyday tasks
require getting an effector or the whole body into a particular point in
space at a particular time. Examples include locomotion, pointing,
and reaching. Let us consider the task of touching an object with the
tip of the index finger of the human arm (figure 21.1). The desired
point of contact with the object can be described with three target
coordinates or a three-dimensional vector, XT, while the starting
position of the tip of the index finger can be described with another
three-dimensional vector XS. So, the control problem may be
formulated as producing movement of the arm that moves the tip of
the finger from XS into XT.
 However, the anatomical design of the arm allows only joint
rotations that ultimately lead to translation of the endpoint. Hence,
one needs to find out what combinations of joint rotations will be
adequate to move the endpoint to XT. Note that the number of
independent axes of joint rotations in the human arm is higher than
three; even if one considers only the shoulder, elbow, and wrist
joints, it is seven. The shoulder can flex or extend, abduct or adduct,
and rotate internally or externally, the elbow can flex or extend, the
wrist can flex or extend and abduct or adduct, and in addition the
forearm can pronate or supinate. So the space of joint angles is
seven-dimensional, and it is not a trivial task to find a target joint
configuration (which can be described by a seven-dimensional
vector ΦT) corresponding to the three-dimensional XT. This task is
similar to a problem of solving three equations with seven unknowns.
The problem of finding a joint configuration corresponding to a
particular location of the endpoint is known as a particular case of
the problem of inverse kinematics (Mussa-Ivaldi et al. 1989). It
belongs to the class of ill-posed problems. There is no general
solution for the problem of inverse kinematics and other similar
problems of finding a large number of unknowns based on a set of
fewer known variables. To solve it, one needs to impose additional
constraints (add more equations) or address it as a problem of
optimization of a particular cost function (Seif-Naraghi and Winters
1990; Latash 1993). We are going to consider this problem in more
detail in chapter 22.
Figure 21.1 To produce a reaching movement from an initial location in space
(XS) to a target location (XT), the controller apparently has to deal with several
computational problems described in the text. Multiple endpoint trajectories are
possible.

For now, let us assume that this problem has somehow been
solved, and a target vector ΦT has been found. The starting vector of
joint angles, ΦS, is supposed to be known. The next step of the
problem is to find patterns of joint torques that would move the arm
from ΦS to ΦT. This problem is known as the problem of inverse
dynamics (Hollerbach and Atkeson 1987; An et al. 1988; Atkeson
1989). Potentially, this is another ill-posed problem, partly because
the original and final arm configurations do not define arm trajectory:
There is an infinite number of time functions Φ(t) that can move the
fingertip to the target. Once again, let us, for now, assume that this
problem has also been solved, and appropriate patterns of joint
torques, T(t), have been computed.
However, the controller (the central nervous system) can only
produce active changes in joint torque by sending signals to the
physiological motors, the muscles. So the next step is to compute
patterns of muscle forces, FM(t), that are adequate to produce the
required function T(t). Now, we need to take into consideration that
typically joints are spanned by many muscles. For example, the
seemingly simple elbow joint has three flexors (biceps, brachialis,
and brachioradialis) and three extensors (the heads of the triceps
muscle). Some of these muscles span more than one joint, while
activation of a muscle frequently induces torque production along
several possible axes of joint rotation (Hogan 1985, 1990; Zajac and
Gordon 1989). Besides, torques in individual joints are defined not
only by the forces produced by muscles crossing the joint but by
other sources such as external forces (in particular, gravity) and
torques related to the motion of other joints of the limb. The latter are
commonly addressed as interaction torques (Zatsiorsky 2002).
Once again, assume that this problem is somehow solved by the
very powerful computing device somewhere in the central nervous
system.
The central nervous system produces active changes in muscle
forces by changing the levels of muscle activation A(t). These
functions need to be computed and realized to make sure that the
vector of muscle forces is FM(t). Note that muscle force depends on
both neural signals to the muscle and actual movement kinematics,
which is defined by many factors, including the external forces and
their time changes. That is why this stage requires predicting the
exact changes in muscle length and velocity over the time course of
the movement.
The last stage of the problem is to compute physiological control
signals, U(t), that are going to be sent by the brain to neurons in the
spinal cord to produce required levels of activation of α-motoneurons
innervating the involved muscles. These patterns should take into
account the time-varying excitation received by motoneurons from
peripheral receptors via reflex pathways. So the whole process of
computing descending signals from the brain to the spinal cord can
be represented (in a very much simplified way!) as:

After the computation is done, the brain issues command signals

U(t). These signals are delivered to spinal neurons that induce
changes in the levels of activation of appropriate muscles such that
the muscles produce required force patterns. The force patterns will
induce joint rotations from the initial to the final configuration, and the
task will be accomplished:
 There are several nontrivial computational steps involved in the
process. However, researchers commonly expect the brain to
possess immense computational power and to have little problem
performing all these computations in no time. This assumption is far
from being trivial because the brain spends considerable time even
when it faces a relatively trivial task to initiate a simple, quick finger
movement to one of the two predefined targets. Under such
conditions, the reaction time increases by 100 ms as compared to
similar movements under the simple reaction time instruction (i.e.,
when the target is known in advance).
This approach to control of human movement may be called a
force-control approach since it is based on the assumption that the
brain generates control signals based on precomputed patterns of
muscle forces that are adequate to perform the task. This
assumption looks very intuitive, but it has been challenged by
another view on motor control that starts the analysis of movements
from a very different set of facts. However, let us first consider two
influential developments stemming from the force-control approach.

21.2 Engrams and the

Generalized Motor Program
In the mid-1930s, Bernstein published one of his most influential
papers (Bernstein 1935). There, he reviewed some of the basic
features of the design of the human body and some of the known
features of motor behavior. In particular, he paid attention to such
facts as the ability of humans to perform similar semiautomatic
actions, such as signing one’s own name, with different effectors.
Figure 21.2 reproduces the famous original illustration by Bernstein,
who himself wrote the Russian word “coordination” with the pencil
grasped in the dominant and nondominant hand, attached to the
right elbow, attached to the foot, and gripped between the teeth.
Based on the apparent similarity of the writing samples, Bernstein
concluded that motor practice led to the elaboration of control time
functions expressed in some hypothetical variables that could be
scaled in time and magnitude and then projected onto different
effector systems. He called these functions engrams.
Bernstein appreciated the importance of the apparently
complicating factors, reviewed in the previous section, that affect
peripheral motor patterns in combination with neural commands
produced by the brain. Therefore, he did not associate engrams with
muscle forces, joint torques, or displacements, but rather left them
undefined, hypothetical variables. He emphasized that engrams
encoded only topological features of movement (i.e., their general
patterns expressed in salient variables, such as trajectory of the tip
of the pencil during handwriting), not their metrical characteristics
such as magnitudes of actual displacements, velocities, and forces.
Figure 21.2 Samples of Bernstein’s writing of the Russian word “coordination”
with the pencil held in the right and left hand, gripped by the teeth, and attached to
the right and left elbow. Note the apparently preserved features of the handwriting.
Reprinted from N.A. Bernstein, On the Construction of Movements (Moscow: Medgiz,
1947).

In the 1970s, the idea of engrams was developed by Richard

Schmidt in the form of a generalized motor program theory or
schema theory (Schmidt 1975, 1980). This theory is very close in
spirit to the force-control approach in assuming that the brain stores
“movement formulas” expressed in variables directly related to
mechanical patterns associated with the production of particular
actions. Experimental support for the generalized motor program
theory came, in particular, from studies that demonstrated invariant
timing patterns within a sequence of actions that could be produced
more slowly or more quickly. In particular, when professional typists
typed a standard phrase many times, they could demonstrate
substantial changes in the average speed of typing, while the relative
timing of key strikes was shown to be rather invariant (Viviani and
Terzuolo 1980).

PROBLEM 21.2
Speeding up a motor action has been shown to lead in some
cases to abrupt changes in the relative timing of the
subcomponents of the action. Do these findings disprove the
generalized motor program theory?

21.3 Internal Models

The line of thinking within the force-control and generalized motor
program approach has led to the development of an influential trend
in movement studies that is commonly known as internal models
(reviewed in Wolpert et al. 1998; Kawato 1999; Shadmehr and Wise
2005). The main idea of this approach is that the central nervous
system computes control signals that lead to the production of
adequate muscle force patterns. This idea sounds so obviously
correct that it requires an effort to realize that it may be wrong.
According to this view, to achieve a desired peripheral outcome of a
planned action, the brain needs to take into account two groups of
factors: (1) all the steps involved in transforming neural signals into
mechanical variables, and (2) time delays in information
transmission from the brain to muscles and from peripheral receptors
to the brain.
Correspondingly, the brain is assumed to contain two types of
models. Inverse models compute descending neural commands
based on a desired mechanical effect; they model transformations
presented in 21.1. The models are termed “inverse” because the
process illustrated in 21.1 moves in the opposite direction as
compared to the natural processes during movement generation.
The computations performed by inverse models are made based on
sensory signals that deliver somewhat outdated information on the
current state of the periphery. Moreover, computed signals from the
brain will also reach the muscles after a substantial time delay such
that the state of the muscles will likely change even more. To take
into account possible changes in the state of the muscles and limbs
due to the time delays in the neural pathways, another group of
models is used called direct models or predictors. Direct models
compute (predict) the effects of current neural commands on the
state of the periphery; they model transformations (21.2). The
inverse and direct processes are illustrated in figure 21.3.
Two sources of experimental material have been used to support
the idea that a combination of direct and inverse internal models is
involved in the production of all purposeful movements. First, there
are observations of anticipatory actions by the muscles, which are
apparently based on predictions of mechanical effects expected from
the planned actions (Belen’kii et al. 1967; Cordo and Nashner 1982;
Johansson and Westling 1988; Koshland et al. 1991; Massion 1992;
Lacquaniti and Maioli 1989). We will consider a few such examples
in the chapters on postural control and prehension. Second, there
are observations of the effects of motor learning in unusual external
force fields, commonly simulated by a combination of powerful
torque motors (Shadmehr and Mussa-Ivaldi 1994), or in conditions of
unusual sensory feedback, commonly achieved by wearing prismatic
glasses (Imamizu et al. 1995).
Figure 21.3 The direct and inverse processes involved in assumed computations
associated with the production of a simple reaching movement.

There have been attempts to link the idea of internal models

directly to particular neurophysiological structures such as the
cerebellum (Kawato and Gomi 1992; Wolpert et al. 1998). In those
studies, plastic changes in the cerebellum have been demonstrated
following learning movements in novel conditions. These plastic
changes have been interpreted as reflecting the formation of new
internal models.
There is no argument about the ability of humans and animals to
behave in a predictive fashion. For example, a predator chasing a
prey always tries to intercept it at a point in space where the prey is
expected to be in the future. However, does any anticipatory
behavior mean that the neural controller contains an internal model
precomputing peripheral variables expected during the planned
actions, such as muscle forces?
 Consider the following example. When you drive a car, you control
its motion by pressing the gas and brake pedals and by turning the
steering wheel. Does your brain contain a computational model that
predicts changes in the amount of gasoline injected into the cylinders
when you press the gas pedal, the mechanical effects of gasoline
explosion on the translational motion of the cylinders, the effects this
motion has on the rotation of the wheels, and the interactions
between the wheels and the pavement? Probably you would agree
that the answer is no. Rather, you have in your brain a set of rules
that map your actions (pressing the pedals and turning the wheel) on
car motion without any explicit modeling of all the intermediate
transformation processes.

PROBLEM 21.3
Is it possible to consider the driving example within the
generalized motor program theory? Is there a “driving program”?

An alternative to the ideas of force control, generalized motor

program, and internal models is offered by the equilibrium-point
hypothesis of motor control. We considered the basic ideas of this
hypothesis for the neural control of single-joint movements in chapter
20. Now the time has come for deeper analysis of this hypothesis.
Figure 21.4 If two muscles, an agonist and an antagonist, produce equal
moments of force (M) at a certain joint angle, muscle spring properties will lead to
a disbalance of moments of force if the joint moves to a new angle due to the
associated changes in muscle length (L). In the illustration, movement from
position A to position B is associated with shortening of the flexor (F) and
lengthening of the extensor (E). As a result, there will be a nonzero net change in
the moment of force (ΔM ≠ 0) acting to move the joint back toward position A.

21.4 Equilibrium-Point
Hypothesis: Main Ideas
The ideas of the equilibrium-point hypothesis can be traced back to
studies performed in the beginning of the 20th century. A famous
German physiologist, Kurt Wachholder (reviewed in Sternad 2002)
asked a question that may seem simplistic: How can humans relax
their muscles at different joint positions? Indeed, muscles are known
to have springlike properties (chapter 5). This means that if an
agonist-antagonist pair of muscles acting at a joint are relaxed at a
certain joint position (position A in figure 21.4), one of them will be
stretched and the other shortened when the joint moves to another
position (position B in figure 21.4). The balance of moments of force
will be violated (ΔM ≠ 0), and active muscle force production will be
required to prevent the joint from moving back to position A.
Wachholder measured muscle activity at different joint positions and
found out that humans could indeed relax completely at different
positions. This observation allowed him to conclude that the
muscle’s springlike properties had to be modified by neural
structures.
Later, in the 1950s, this insight was generalized in the form of a
posture–movement paradox (Von Holst and Mittelstaedt 1950). By
that time, it was known that joint postures were stabilized by
neurophysiological mechanisms, some of which have been
discussed in chapters 18 and 19. If a joint is at an equilibrium at a
combination of external load and angle (figure 21.5a, filled circle,
EP0), any deviation from this point (open circle, EP1) leads to the
generation of muscle forces trying to bring the joint back to its
equilibrium position. The question is: How can an active movement
occur without triggering the resistance of these posture-stabilizing
mechanisms?

PROBLEM 21.4
What physiological mechanisms contribute to posture stabilization
of a joint if it is moved away from an equilibrium?
Figure 21.5 (a) An equilibrium position of a joint (EP0) is defined by the joint
moment-angle characteristic (dashed line) and external load. If the joint is moved
to another position (P1) and released, it will tend to return to EP0. (b) If the
commands to joint muscles are changed to correspond to a new equilibrium joint
position, the moment-angle characteristic shifts (from the dotted to the dashed
line), and posture-stabilizing mechanisms are re-addressed to the new position,
the original EP0 becomes a non-EP, and movement occurs to the new equilibrium
point (EP1).

An answer was suggested by von Holst and Mittelstaedt (1950) in

the form of the principle of reafference. In their scheme, a copy of
the signals from α-motoneurons (termed efference copy; see
chapter 28) was used to predict changes in sensory signals from
proprioceptors expected from the planned action (reafference). If the
prediction was accurate, no correction was issued. If, however, the
prediction happened to be wrong, an error signal was sent to the α-
motoneurons leading to correction of the action. If movement was
produced by a change in the external force acting on the effector,
changes in signals from proprioceptors were addressed as
exafference. These changes always led to reflex-based corrections.
This scheme has been criticized recently (Feldman 2009, 2016)
because of its inability to explain how animals (including humans)
can relax after moving an effector to a different position. Indeed, if
the muscles in both positions are relaxed, the output of all α-
motoneurons is zero, and a copy of this output cannot predict
changes in the activity of position-sensitive sensory endings,
including those in muscle spindles. As a result, this change in
afferent signals has to lead to reflex-mediated changes in muscle
activation and produce motion of the effector in clear contradiction to
the everyday experience that humans can relax muscles at various
positions.
Nevertheless, the original idea that neural commands for any
active movement re-address afferent signals from proprioceptors to
a new posture has been accepted and developed within the
equilibrium-point hypothesis (chapter 20). The equilibrium-point
hypothesis assumes that the central control structures shift
physiological variables associated with muscle activation thresholds
(λ, threshold of tonic stretch reflex) and re-address posture-
stabilizing mechanisms to a new posture. These mechanisms turn
from posture-stabilizing into movement-producing. Figure 21.5b
shows that if postural stabilization is re-addressed to a new joint
position (corresponding to EP1), the original joint position becomes a
deviation from the new one, and posture-stabilizing mechanisms
produce a movement of the joint to the new position. The ability to
resolve the posture–movement paradox is unique to the equilibrium-
point hypothesis; other theories of motor control have so far failed
this litmus test.
The main physiological control variable within the equilibrium-point
hypothesis has been associated with changes in the average level of
subthreshold depolarization of α-motoneurons within a pool leading
to changes in the threshold of the tonic stretch reflex (λ, see chapter
20) in the muscle innervated by that pool. For more complex
systems, such as a joint crossed by several muscles, other variables
have been introduced, such as r-command and c-command
described in chapter 23. The commands (r, c, λ, etc.) represent tools
that allow the system to reach its goals, just like the gas pedal and
the steering wheel allow the driver to drive a car to a desired place.
21.5 Equilibrium-Point
Hypothesis: Subtle Details
At times, the equilibrium-point hypothesis has been described in a
simplified way as a mass–spring model, and changes in λ have been
associated with changes in the resting length of the spring (Feldman
1966; Latash 1993). Such a representation was useful in many
aspects, but it also was the root of much confusion. Starting from
Weber (cited in Latash 1993), other scientists, including Bernstein
(1935), applied the spring analogy to muscles and literally modeled
muscles as springs. However, already in 1938, Sir A.V. Hill
demonstrated that the mass-spring model misrepresented basic
dynamic properties of muscles, including its energy output. To
emphasize that the equilibrium-point hypothesis cannot be
adequately viewed as a mass-spring model of muscles, the muscle-
load system is essentially nonlinear because of its threshold
properties—its behavior at muscle length values shorter than λ—is
qualitatively different from the behavior at length values longer than λ
(see figure 20.10). Changes in the threshold value during voluntary
movements make the system nonlinear even locally.
Another controversial issue related to equilibrium-point control is
that of movement equifinality. A number of studies have shown
that, when a brief transient perturbation is applied unexpectedly
during a fast limb movement to a target, the trajectory shows major
deviations as a result of the perturbation, but the final position is
achieved very accurately (Schmidt and McGown 1980; Rothwell et
al. 1982b; Jaric et al. 1999). The equilibrium-point hypothesis
naturally predicts equifinality, but only if two conditions are met. First,
the patterns of λs are not changed as a result of the perturbation.
Second, the peripheral muscle force-generating properties are not
modified by the perturbation. Under these conditions, the final
equilibrium state of a system is defined only by the final values of the
commands to the muscle and external load. No effects of changes in
the trajectory induced by transient changes in the external load are
expected to have any effect.
Violations of equifinality have been observed under specific
experimental conditions. These experiments involved, in particular,
unexpected application of motion-dependent forces (for example, the
Coriolis forces or other destabilizing forces proportional to movement
velocity [Lackner and DiZio 1994; DiZio and Lackner 1995; Hinder
and Milner 2003]) that act only in the course of movement but not at
steady states. These results have been interpreted within the
framework of the equilibrium-point hypothesis as consequences of
changes in patterns of control signals that may occur despite the
explicit instruction to the subject “not to intervene” and the lack of the
subject’s awareness of such changes. More recent studies have
predicted and demonstrated violations of equifinality under relatively
natural conditions, in the absence of unusual force fields, based on
the equilibrium-point hypothesis (Zhou et al. 2015).
The most simplistic accounts of the equilibrium-point hypothesis
(including the one in chapter 20) consider the threshold of the tonic
stretch reflex (λ) as a variable under exclusive control of the brain.
This is not true. A number of experiments have shown that muscle
reactions to length changes depend on other factors, such as the
velocity of muscle length changes, the history of activation of the
motoneuronal pool, and the reflex effects from other muscles,
including those crossing other joints (reviewed in Feldman and Levin
1995; Feldman and Latash 2005; Feldman 2015). Figure 21.6
illustrates how the summed effects of all the aforementioned
contributors to λ affect the relation between muscle length and its
active force.
Figure 21.6 A dependence of active muscle force on muscle length depends on
the threshold of the tonic stretch reflex λ*. For the muscle to be activated, its length
should be higher than γ*. The value of λ* is defined by several components,
including central command (γ), velocity-dependent feedback (with the gain μ),
effects of intermuscular reflexes (ρ), and history-dependent changes (ƒ(t)).
Data from Feldman and Latash (2005).

To use the earlier example, when the driver turns the steering
wheel, the driver’s effort is only one of the factors that define how the
steering wheel turns. Another factor is, for example, the availability of
a power-assisted steering system whose gain can be different in
different makes and models.
The main control variable within the equilibrium-point hypothesis,
the threshold of the tonic stretch reflex, is assumed to be expressed
in units of muscle length—that is, in positional units. This assumption
fits well with the general notion that most of our motor tasks are
expressed in units of location in space.
Now that the reader is hopefully accustomed to the idea that
muscle elastic properties are an essential feature of the design of the
human body, return to the seemingly obvious earlier statement that
the central nervous system has to make sure that adequate force
patterns are produced for a required motion. Since force and
displacements are coupled in any elastic system, this statement
implies that movement kinematics is known perfectly in advance.
This is obviously unrealistic, as emphasized by Bernstein many
years ago (Bernstein 1947), for any system equipped with position-
dependent force generators.

PROBLEM 21.5
Some motor control hypotheses suggest that the controller
precomputes patterns of muscle activation or their direct
precursors, such as total presynaptic input to a pool of α-
motoneurons. Do these hypotheses avoid typical problems that
face force-control models?

Until now, we have discussed the equilibrium-point hypothesis as

applied to single muscles and single joints. While this is a natural
step in the development of any hypothesis, one would like to know
how this hypothesis handles the control of natural multijoint
movements. This issue has been developed as the theory of control
with spatial referent coordinates, described in more detail in chapter
24.

21.6 Dynamic Systems Approach

The dynamic systems approach to movement studies originated
from a particular area of mathematics, the theory of nonlinear
differential equations, and its applications to physics developed in
the 1960s (Haken 1977). This approach is applicable to any system
that changes in time; hence, the word “dynamic” here refers not to
an area of physics dealing with forces but to time changes. Dynamic
systems are typically assumed to be characterized by two types of
behavior. They behave in a predictable, deterministic (or close to
deterministic) way over short time intervals, while over large time
intervals their behavior may be rather complex and poorly
predictable. As a result, a small change in the initial state of a
dynamic system may lead to relatively small changes in its behavior
at the first, short time scale, but to qualitative, dramatic changes at
the second, large time scale. The possibility of analyzing such
qualitative changes in behavior makes the dynamic systems
approach very attractive to movement science, particularly if one is
interested in such characteristics of biological motion as its variability
and stability.
In general, any material object is a dynamic system. A brick and
the human body are both dynamic systems. However, the brick is a
boring dynamic system for motor control scientists, while the body
(and its subsystems) is an exciting one. This is due, in particular, to
the natural variability of processes within the human body reflected,
in particular, in the variability of human movements, which somehow
assures stability of motor outcomes for most everyday actions (see
chapter 22).
The dynamic systems approach has been very successful in
addressing such issues of motor performance as relative timing of
the involvement of different effectors, stability of actions in the
changing environment, and coupling between perceptual and motor
variables (reviewed in Kugler and Turvey 1987; Kelso 1995). One
classical example (Kelso et al. 1979; Kelso 1984) is that when a
person tries to move the two index fingers rhythmically, there are two
regimes of movement that feel easy and are characterized by
relatively low variability, in-phase (both fingers flex or extend
simultaneously) and out-of-phase (when one finger flexes, the other
extends). An increase in the frequency of motion does not change
this pattern for the in-phase regime. However, the out-of-phase
motion at some frequency shows destabilization of the movement
pattern (it becomes more variable) and switches to the in-phase
regime. A particular dynamic systems model has been able to
describe these results very well (Haken et al. 1985), while no other
approach to motor control has been able to handle them.

PROBLEM 21.6
Imagine that finger movements produce an opposite pattern of
movement on a computer screen—that is, when the fingers move
 in-phase, the cursors on the screen move out-of-phase, and vice
versa. Which of the two patterns of finger movement would you
expect to switch to the other pattern with an increase in the
movement frequency?

The equilibrium-point hypothesis may be viewed as a particular,

neurophysiology-based, dynamic model for the production of
voluntary movements. It assumes a certain coupling between
sensory and motor variables (the tonic stretch reflex), and it operates
with the notion of stability (the equilibrium point). As a result, analysis
of movements within this hypothesis is complicated, and any
simplification may lead to qualitative changes in the outcome.
Within the dynamic systems approach, it is necessary to
distinguish between state variables and parameters of systems.
Variables such as forces and displacements reflect the laws of
nature and are called state variables. Any variable that is a direct
precursor of a state variable or results from computations using state
variables is also a state variable. For example, such computed
variables as velocity, acceleration, rate of force change, stiffness,
and damping are all state variables. The same is true for the
magnitude of muscle activation. Equations of physics that link state
variables to each other include parameters. Some of them reflect
properties of the moving objects (such as, for example, mass). Other
parameters can be modified by the central nervous system. Their
changes modify the relations among state variables according to the
laws of physics.
Within the equilibrium-point hypothesis, λ is a parameter that links
two salient state variables for the muscle, its force and length.
Muscles move under changes in external forces: such movements
are commonly termed involuntary or passive. Voluntary movements
are produced by changes of λ (i.e., by changing the parameter of the
law of nature that links force, length, and λ). This type of control is
known as parametric control.
Consider a simple pendulum, a mass on a weightless rope (figure
21.7). When it swings in the field of gravity, its position and forces
acting on the mass all change according to the laws of physics. Its
behavior will depend on such parameters as the length of the rope,
the coordinates of the suspension point, and the direction of gravity.
These parameters can be changed, causing a change in such
features of the pendulum as its period and the point about which it
oscillates. If you observe changes in the period and location of the
pendulum, it makes little sense to assume that some hypothetical
controller models and implements the forces acting on the mass to
assure these changes. A controller can act naturally by changing
parameters of the pendulum (for example, you can move it in space
or change its length), not its state variables, while changes in the
state variables emerge based on the laws of physics.
Figure 21.7 Motion of a perfect pendulum, a point mass on a rigid cord, is
produced by forces of gravity and by the cord (FG and FC). If one wishes to control
a pendulum, he or she should change such parameters as the length of the rope
(L), the coordinate of suspension, {X0, Y0, Z0}, and the direction of gravity (g).

This example illustrates a general rule: Although in an equilibrium

state all forces are balanced, not the forces but a system’s
parameters predetermine where, in spatial coordinates, this state is
reached (Glansdorf and Prigogine 1971). Therefore, in order to bring
a system from one equilibrium state to another, desired equilibrium
state, the central nervous system must change parameters that are
independent of state variables. The equilibrium-point hypothesis
suggests a set of parameters (muscle activation thresholds, λ) to
implement such control, while the alternative motor control theories
fail to do so.
CHAPTER 21 IN A NUTSHELL
There are two basic approaches to
motor control. The first views it as a
neural process directed at achieving
desired time profiles of performance
variables such as forces and torques.
Within this approach, the generalized
motor program theory assumes that the
brain stores “movement formulas”
expressed in variables directly
related to mechanical patterns
associated with the production of
particular actions. The idea of
internal models is based on an
assumption that the neural structures
combine direct and inverse models to
be able to predict changes in the
peripheral system based on its current
state and neural commands and
precompute neural signals that are
adequate to achieve a desired set of
muscle activations and forces. An
alternative view is the equilibrium-
point hypothesis, which views neural
control processes as parametrization
of the neuromotor system. The
equilibrium-point hypothesis solves
the movement–posture paradox and
provides a powerful instrument to
analyze a variety of motor actions.
 Chapter 22

Motor Synergies

KEY TERMS AND TOPICS

motor redundancy
optimization
principle of abundance
synergies
stability of movement
uncontrolled manifold hypothesis

Motor synergies are hypothetical neural structures that play a central

role in the coordination of multiple elements (joints, muscles, motor
units, etc.) that take part in all natural movements. This concept was
developed by Nikolai Bernstein (see chapter 1), who introduced a
multilevel scheme for the construction of movements with one of the
levels termed “The Level of Synergies and Patterns” (Bernstein
1947). This concept has been developed over the years, and a
number of recent reviews address it (Turvey 2007; Tresch and Jarc
2009; Bruton and O’Dwyer 2018; Latash 2021c). It is intimately tied
to such important characteristics of movement as motor redundancy
and stability of performance.
22.1 The Problem of Motor
Redundancy
In the previous chapter, we encountered a few examples of the
problem of motor redundancy. These included, in particular, the
problem of inverse kinematics (Mussa-Ivaldi et al. 1989): how to
define angular rotations in a set of joints within a multijoint limb
corresponding to a prescribed final location of the limb’s endpoint.
Since the number of independent axes of joint rotation is higher than
three (the dimensionality of the external space), this problem does
not have a unique solution. Another example is the problem of
inverse dynamics (Hollerbach and Atkeson 1987): how to define
patterns of muscle forces (or joint torques) that would lead the
endpoint to the required location.
Similar problems emerge at other levels of analysis of the system
for the production of voluntary movements. For example, the
relatively simple elbow joint is spanned by six major muscles, three
flexors and three extensors. Imagine that a person produces a
constant flexion moment of force (or flexion joint torque, TFl) against
an external resistance. The problem is to define muscle forces that
produce TFl. This problem is similar to trying to solve the following
equation with six unknowns:

where FFL,i are forces of individual flexor muscles, FEX,j are forces of
individual extensor muscles, and Li,j are corresponding lever arms (i
and j stand for the individual flexor and extensor muscles,
respectively). Obviously, this equation cannot be solved. That is why
such problems are sometimes called ill-posed.
A similar problem can be formulated at the level of a single
muscle: Define what motor units should be recruited and at what
frequencies to produce a required level of muscle activation. The
size principle (chapter 7) imposes constraints on possible solutions,
but it does not prescribe a unique solution to the problem.

PROBLEM 22.1
If motor units are always recruited in a fixed order, how can the
same muscle force be produced by different combinations of
motor units?

Nikolai Bernstein viewed the problem of motor redundancy as a

central problem of motor control. He formulated it as the problem of
elimination of redundant degrees of freedom. This problem is
sometimes referred to as the Bernstein problem (Turvey 1990;
Latash 1996). There have been two qualitatively different
approaches to addressing the Bernstein problem. The first approach
assumes that the controller finds a unique solution each time such a
problem emerges. The second approach assumes that the controller
facilitates groups of equally acceptable solutions rather than tries to
find unique solutions.

PROBLEM 22.2
Suggest a couple of examples of the Bernstein problem from
everyday life. Does the controller find unique solutions in your
examples?

22.2 Optimization Approaches

One of the common approaches to the problem of motor redundancy
has been to apply optimization criteria that enable the selection of a
unique solution for such problems. A central notion common across
all such approaches is that of cost function (or objective function).
Cost function is a particular function of a system’s performance that
the controller tries to keep at an optimal value, commonly at a
minimal or at a maximal possible value. Movement studies have
used a variety of cost functions related to mechanical performance
by the system, hypothetical control processes within the system, and
psychological factors such as effort and comfort (reviewed in Nelson
1983; Seif-Naraghi and Winters 1990; Prilutsky and Zatsiorsky
2002).
Imagine that a movement in a single joint from a certain initial to a
certain final position is to be performed as quickly as possible—that
is, within the minimal time. In this case, movement time can be used
as a cost function that has to be minimized. The optimal solution
would involve the “bang-bang control” (figure 22.1a; Nelson 1983)—
that is, using the maximal levels of acceleration and deceleration.
This is also known as the teenager driving principle: pressing the gas
pedal as strongly as possible when one sees the green light, and
pressing the brake pedal as strongly as possible when the light turns
red. The velocity profile for this mode of control does not look like a
typical bell-shaped pattern observed in experiments (see chapter
24). For a purely inertial system, the optimal switching time between
the accelerating and decelerating “bangs” is exactly in the middle of
the movement.
Figure 22.1 (a) The optimal solution for movement of a purely inertial system
from an initial to a final position within a minimal movement time is “bang-bang”
control. (b) If a similar movement is performed minimizing the impulse, the control
is “bang-zero-bang.” In both cases, movement velocity profiles are rather different
from the typical bell-shaped velocity seen during natural movements.

Another cost function is the impulse of the system defined as the

time integral of force. For systems with low damping, minimizing the
system’s impulse is equivalent to minimizing the peak velocity. For a
given movement time, the optimal strategy would be to quickly
accelerate the system to an optimal speed, keep the speed constant,
and quickly decelerate the system. This type of control is known as
“bang-zero-bang” (figure 22.1b).
Arguably, the most influential cost function in movement studies
has been an integral measure of squared jerk shown as CJERK in
equation 22.2. Jerk is the time derivative of acceleration (or the third
time derivative of displacement). Application of this cost function has
been termed the minimum jerk criterion (Hogan 1984; Flash and
Hogan 1985):

where a stands for acceleration and MT stands for movement time.

The minimum jerk criterion leads to smooth trajectories with a bell-
shaped velocity profile and a symmetrical double-peaked
acceleration.

PROBLEM 22.3
Another optimization approach has been proposed based on
minimization of a torque-change function similar to jerk
minimization. For what kinds of mechanical systems do the two
criteria lead to identical results?

A particular optimization approach has been suggested based on

an assumed relation between muscle force generation and fatigue
(Crowninshield and Brand 1981; Prilutsky 2000). In particular, this
minimal fatigue criterion has been successfully applied to the
problem of defining patterns of individual muscle force production
during tasks that involve multiple muscles.
Another group of optimization criteria include less strictly defined
movement characteristics such as comfort and effort (Cruse and
Bruwer 1987; Bruwer and Cruse, 1990; Hasan 1986). They have
sometimes been combined with more mechanistic criteria to form
complex cost functions. In general, adding a criterion to a cost
function leads to an improvement of the fit between predictions
based on this cost function and data.
A rather sophisticated optimization approach has been developed
based on an idea that movement planning is performed in a posture
space based on a set of memorized postures (Rosenbaum et al.
1993). This approach includes selecting a target posture and a
trajectory moving the effector from the initial posture to the target
posture based on a number of criteria associated with moving the
endpoint, individual joints, and other features.
A common feature of most optimization approaches is estimation
of the cost function value over the movement time, which is done
before the movement is initiated. This is not a trivial assumption
because it assumes that the central nervous system somehow can
predict very accurately both movement time and all the performance
variables that are used in the computation of the cost function. This
assumption goes against the conclusion by Bernstein (1947) that the
brain cannot prescribe details of peripheral mechanics during natural
movements performed in the typical, not perfectly predictable,
environment.
Another feature of optimization approaches that does not look
attractive is the relatively arbitrary choice of cost functions, which is
typically done based on common sense and personal views of
researchers. This problem has been addressed in several studies
that tried to introduce methods of inverse optimization (Bottasso et
al. 2006; Terekhov et al. 2010). These methods try to compute cost
functions based on observations of the system’s behavior, assuming
that such functions exist. So far, application of these methods has
been limited due to the associated computational challenges.
Before moving further, consider the following example of the
apparent problem of redundancy: How does the brain select a
particular subset of Na+ ions that should cross the membrane during
the generation of an action potential from zillions of available ions
close to the membrane? Probably, most readers would agree that
this problem makes no sense. The central nervous system does not
micromanage at such a level. It simply does not care which of the
ions cross the membrane, as long as the action potential is
generated. But how do we know that it cares about specific patterns
and magnitudes of motor units, muscle forces, joint torques, and joint
rotations in the previous examples? It seems more realistic that the
brain does not select perfectly optimal solutions (no matter what the
criterion is) but uses solutions that are “good enough” (Loeb 2012;
Akulin et al. 2019). This conclusion is supported by the classic
experiment by Bernstein (1930) on professional blacksmiths showing
that consistent trajectories of the hammer are associated with much
more variable joint trajectories (described in section 22.5).
Now consider the following situation: You are a conductor of an
orchestra of 100 drummers, and you need to produce a noise level
of 100 dB. One solution is to tell each drummer how hard to beat
based on some optimization criterion—in other words, to eliminate
the redundant degrees of freedom. An alternative solution is to tell
each drummer: “Listen to the level of noise. If it is above 100 dB,
beat lighter; if it is below 100 dB, beat harder.” The second solution
does not eliminate the apparently redundant degrees of freedom, it
uses them and links them to the task with the help of sensory
feedback loops from auditory receptors. Note that in the second
case, the solution will lead to an acceptable outcome even if one of
the drummers falls ill and decides to go home. In other words, it will
possess the very important feature of stability discussed in more
detail in section 22.6.

22.3 Bernstein’s Level of

Synergies
In his main book (Bernstein 1947), Nikolai Bernstein described a
multilevel hierarchical scheme for the neural control of movements.
The second from the bottom level was called “The Level of
Synergies and Patterns or the Thalamo-Pallidar Level” or, briefly,
Level of Synergies (also see chapter 1). The name suggests that
Bernstein associated the organization of synergies with the
functioning of the basal ganglia and thalamus. As discussed in more
detail in other chapters (see chapters 9 and 37), recent studies,
including observations of motor patterns in patients with disorders of
the basal ganglia, have corroborated this insight. Bernstein
suggested that the Level of Synergies had two main functions:
dealing with the problem of motor redundancy at the level of muscle
groups and ensuring dynamical stability of movements.
Under dynamical stability we imply the ability of a system to come
back close to its original state or its originally planned trajectory
following brief, transient force perturbations. For example, when a
ball rolls down a concave half-pipe, small variations of its initial point
of release and small bumps on its way down would still result in its
trajectories converging toward the center of the half-pipe (assuming
some friction, which is always present). In contrast, if a ball is
pushed to roll along a flat surface or along a convex surface, its
trajectories would diverge with variations in the initial state or small
force perturbations. Figure 22.2 illustrates trajectories of an object
along dynamically stable and dynamically unstable directions,
assuming that the starting states vary. In the former case, the
trajectories converge; in the latter case, they diverge.
Bernstein emphasized on many occasions (Bernstein 1935, 1947,
1967, 1996) that the central nervous system (CNS) was in principle
unable to predict all the forces acting on segments of moving
effectors. Indeed, external forces are never 100% predictable. This is
also true for internal states of the segmental apparatus of the spinal
cord, which mediates the effects of descending neural signals and
reflex loops in muscle activation and force. Given the conduction
time delays in the body and the dependence of muscle force on its
length and velocity, the brain is in principle unable to predict perfectly
the future mechanical effects of its signals at the time these signals
are generated.

Figure 22.2 A person producing a sequence of movements will start from slightly
varying initial states (black dots). Trajectories in unstable directions are expected
to diverge (top). Trajectories in stable directions are expected to converge
(bottom).
 In addition, body segments are linked mechanically. So, any
unexpected deviation in the motion of a joint produces deviations of
motion in other joints, leading to even more unpredictable
mechanics. Based on this analysis (much more detailed in the
original works!), Bernstein concluded that dynamical stability was
essential for any functional movement and, as a result of evolution
and practice, neural control patterns converged on those that
resulted in dynamically stable trajectories of salient performance
variables such as the trajectory of the fingertip during pointing and
the body’s trajectory during walking. Finding such neural control
patterns and ensuring the stability of movements with respect to
important performance variables with the help of feedback loops was
the second major function of the Level of Synergies.

22.4 Uniting Muscles Into

Groups
Let us consider the two functions of the Level of Synergies in more
detail. The first one assumes that muscles participating in typical
actions form stable groups, and then involvement of each group is
modified with a single time-varying coefficient. This method of
control, illustrated in figure 22.3, reduces the number of variables
manipulated by the brain as compared to the number of involved
muscles and, hence, contributes to solving the problem of motor
redundancy. Such muscle groups have been referred to as modules,
modes, factors, primitives, or synergies (reviewed in Ivanenko et al.
2006; Giszter 2015; Overduin et al. 2015; Latash 2020). Since
forming such groups is only one function of the Level of Synergies,
we are going to call them modes or muscle modes to avoid
confusion.
Figure 22.3 The concept of synergy implies that a neural network unites
effectors into groups (modes) controlled by one variable per mode in a task-
specific way.

Muscle modes have been described for a variety of movements,

including whole-body postural tasks, walking, and reaching (Saltiel et
al. 2001; d’Avella et al. 2003; Krishnamoorthy et al. 2003; Ivanenko
et al. 2004; Ting and Macpherson 2005; Tresch and Jarc 2009). This
has been done with methods that analyze correlations between
muscle activation levels (Mi) over repetitive trials or over time and
then look for a few variables representing linear combinations of
muscle activations that can describe the behavior of all the muscles
with acceptable accuracy:

where k stands for constants, n is the number of muscles, and j

stands for individual muscle modes (j < n). Such methods are known
as matrix factorization, and typical examples include principal
component analysis (PCA), factor analysis, independent component
analysis (ICA), and non-negative matrix factorization (NNMF).
Different methods are selected based on the goals of researchers.
For example, PCA produces a set of muscle modes that are
orthogonal to each other, which is important if this is the first step in
analysis of muscle synergies that also addresses their second
function, ensuring dynamical stability (see later). On the other hand,
NNMF takes into consideration that muscle activation levels cannot
be negative. However, it produces a set of modes that are
independent of each other but not orthogonal.
Figure 22.4 illustrates a set of three muscle modes that were
defined using the PCA-based method. Note that the modes unite
muscles that cross different joints. Muscles within a mode are
commonly united by their combined action at the body level. For
example, the first two modes illustrated in figure 22.4 can be called
“push-back” and “push-forward” modes based on their effect on the
trunk orientation.
The muscle mode composition is typically robust across subjects
and variations of task parameters. It can change, however, under
significant changes in the task, in particular those that affect stability
of action (Krishnamoorthy et al. 2004) and also in the processes of
motor development and motor learning (Asaka et al. 2008; Dominici
et al. 2011; Ivanenko et al. 2013).
The methods of identification of groups of elemental variables
(those produced by elements at a selected level of analysis) have
been applied not only to muscle activation levels but also to
mechanical variables. For example, if a standing person performs a
voluntary body sway task, angular rotations of the three major leg
joints (the ankle, the knee, and the hip joints) in the sagittal plane of
the body co-vary in such a way that they can be related to each
other by a single linear equation with high accuracy (Alexandrov et
al. 1998). This relation may be viewed as a reflection of a kinematic
mode involved in stabilization of the vertical posture. Studies of
finger and hand kinematics during a variety of actions, in particular
during grasping and fingerspelling (Santello et al. 2002; Jerde et al.
2003), have shown that a few modes could account for patterns of
motion in numerous joints.
Figure 22.4 An illustration of muscle modes observed during actions performed
by a standing person. Note that muscles of the leg and trunk are united into groups
(shown with different colors) that include muscles spanning different joints. SOL =
soleus, GL = lateral gastrocnemius, BF = biceps femoris, ES = erector spinae, TA
= tibialis anterior, RF = rectus femoris, VL = vastus lateralis, RA = rectus
abdominis, TF = tensor fascia latae, GMED = medial gluteus.

22.5 Principle of Abundance

Optimization approaches assume that the central nervous system
cares about making choices for all the described problems of motor
redundancy and somehow computes a unique solution for each of
them. If this assumption is wrong, the whole problem of motor
redundancy becomes wrongly formulated. There are reasons to
believe that the controller does not precompute specific motor
patterns for each and every movement. One of the primary reasons
is motor variability.
 In the 1920s, Bernstein (1930) performed a study of the
kinematics of hitting movements when professional blacksmiths
stroked the chisel with the hammer (figure 22.5). His subjects were
perfectly trained: They had performed the same movement hundreds
of times a day for years. However, they performed each strike
differently from other strikes. Bernstein noticed that the variability of
the trajectory of the tip of the hammer across a series of strikes by a
blacksmith was smaller than the variability of the trajectories of
individual joints of the subject’s arm holding the hammer. Since
apparently the brain could not send signals directly to the hammer,
Bernstein concluded that the joints were not acting independently but
correcting each other’s errors. This observation suggested that the
brain did not try to find a unique solution for the problem of kinematic
redundancy by eliminating redundant degrees of freedom but rather
used all the apparently redundant degrees of freedom at the level of
joint rotations to ensure more accurate performance of the task (less
variable hammer trajectory).
Figure 22.5 An illustration of the experiment of Bernstein on professional
blacksmiths. The trajectory of the tip of the hammer was reproduced more closely
across repetitive strikes than the trajectories of individual joints

So, an alternative view of the problem of coordinating a large set

of elements has emerged that views large sets of elements not as
redundant but rather as abundant. The numerous degrees of
freedom existing at each and every level of description of the system
for movement production may be viewed not as posing
computational problems for the brain but as providing the brain with
a powerful apparatus that allows it to combine flexibility of movement
performance with stability of movement outcome. The problem of
motor redundancy is substituted with a solution of motor abundance:
The controller does not look for unique solutions but unites the
degrees of freedom to facilitate families of solutions that are equally
able to solve the task and ensure its stability in the changing
environment. This attitude toward the design of the human body is
known as the principle of abundance (Gelfand and Latash 2002;
Latash 2012).
The principle of abundance has already been illustrated with the
example of a drum orchestra. It can also be illustrated at the level of
a single muscle viewed as a set of a large number of elements:
motor units. The tonic stretch reflex (chapter 18) provides for a
relation between muscle force and length. Imagine that a muscle is
at a static state at an equilibrium point, where muscle force exactly
balances the external load, corresponding to a certain force and
length combination (EP0 in figure 22.6). Imagine now that one of the
active motor units within the muscle stops producing action
potentials (shown with the dashed image). This motor unit’s
contribution to the muscle force disappears. The drop in the muscle
force leads to a disbalance of the external and muscle forces, and
joint motion occurs corresponding to muscle stretch. The muscle
stretch leads to an increase in the activity of sensory endings in the
muscle spindles leading, through the stretch reflex mechanism, to an
additional excitation of the α-motoneurons innervating the muscle.
This leads to an increase in the active muscle force, partly
compensating for its original drop.
This example shows that the organization of motor units with the
help of stretch-sensitive feedback loops leaves room for various
solutions to the problem of finding a combination of motor units and
their firing frequencies to solve the apparent Bernstein problem. The
motor units are united with the help of the tonic stretch reflex in such
a way that they compensate for each other’s errors and contribute to
stability of muscle performance. Note also that, similar to the tonic
stretch reflex, all the reflexes discussed in earlier chapters are
negative feedback loops, which act against possible spontaneous
changes in the body and environment and stabilize states of the
peripheral motor apparatus.

22.6 Ensuring Stability of

Movements
The principle of motor abundance is a natural consequence of a view
on the coordination of multi-element systems that was started by
Israel Gelfand, a brilliant mathematician, and Michael Tsetlin, a
brilliant physicist, who offered an analysis of biological systems
based on a set of definitions and axioms (Gelfand and Tsetlin 1966).
They assumed that the central nervous system operated with
structural units. A structural unit is a task-specific neural
organization of elements, while each element within a structural unit
is itself a structural unit at a different level of analysis. Synergies
were defined as purposes of structural units, while behaviors were
viewed as extrinsic patterns reflecting a synergy in particular external
conditions.
For example, a proverbial synergy, locomotion, was assumed to
be based on a structural unit, a neural network uniting different
extremities. Motion of a multijoint extremity could be viewed as a
structural unit based on individual joint rotations. Each joint rotation
is a structural unit based on muscle actions as elements. Each
muscle is a structural unit of its motor units. The term structural unit
is not used frequently in our times. So further, for simplicity, we will
use the word synergy to address both neural organizations of
elements and their functional purpose.
Figure 22.6 (a) An illustration of the principle of abundance using the
equilibrium-point control of a muscle. (b) Possible spontaneous changes in the
number of recruited α-motoneurons are partly compensated by the action of the
tonic stretch reflex. (c) Imagine that the neuron whose axon is shown with the
dashed line stops firing. The force of the muscle will drop, and (d) it will extend
under the action of the external load. This will lead to higher activity in the tonic
stretch reflex loop, leading to higher activity of the other motoneurons of the pool.

PROBLEM 22.4
Suggest the smallest and the largest structural unit you can
imagine.

Earlier, we mentioned two main reasons why dynamical stability of

functional movements is crucial: the time-varying external forces that
are commonly unpredictable and the spontaneous changes in the
states of neurons, which are amplified by their threshold properties.
There seem to be two main strategies to ensure accurate motor
performance given the aforementioned factors. First, the controller
has to predict all the variations in the external and internal sources of
motor variability and to introduce corrections to signals to all the
effectors on-line based on these predictions (see the discussion of
the concept of internal models in chapter 21). Alternatively, the
controller may opt to organize the elements into a synergy that
naturally generates dynamically stable behaviors in the ever-
changing environment. Then, the controller only has to set a task
and delegate the responsibility for its stable performance to the lower
structure, the synergy (figure 22.7).

Figure 22.7 Performance of a multi-effector system is defined by an interaction

between the poorly predictable environment and the output of the synergy network
to the effectors.

One of the expected features of synergies is compensation of

errors introduced by individual elements into salient, task-specific
performance variables. The blacksmith study by Bernstein was
arguably the first to document compensation of errors. Indeed, the
small inter-trial variance of the hammer trajectory (compared to the
variances of individual joint trajectories) suggests that, when a joint
in a specific trial got off its expected trajectory (estimated as its
average trajectory over many trials), other joints showed deviations
from their respective expected trajectories in directions that
compensated (partly) for the error in the hammer trajectory
introduced by the first joint.
A number of later studies documented error compensation
phenomena across sets of effectors and tasks, including gripping an
object between the index finger and the thumb, producing constant
force while pressing with several fingers, and speech (Abbs and
Gracco 1984; Cole and Abbs 1987; Latash et al. 1998). These
studies used both unexpected external perturbations applied to one
of the involved effectors and self-generated actions by one of the
effectors. In both cases, the apparently uninvolved effectors showed
deviations in their performance that compensated (partly) for the
effects of the original error in salient task-related performance
variables. In cases of external perturbations, such error corrections
could be seen at very short time delays, much shorter than the
simple reaction time (confirmed in recent studies; see Pruszynski et
al. 2016; Reschechtko and Pruszynski 2020). In cases of self-
generated actions, there were no delays in the corrective deviations
of the apparently uninvolved effectors, suggesting that the
corrections were introduced based on feedforward mechanisms.

22.7 Uncontrolled Manifold

Hypothesis
The ideas of abundance and dynamical stability of task-related
performance variables have been merged within the uncontrolled
manifold (UCM) hypothesis (Schöner 1995; Scholz and Schöner
1999; Latash et al. 2002b, 2007). The UCM hypothesis suggests that
the central nervous system can ensure desired stability properties of
a salient performance variable by acting in a space of elemental
variables and allowing relatively large amounts of inter-trial variance
within a space where this performance variable does not change (the
UCM for that variable) as compared to the amount of variance within
the orthogonal to the UCM space (ORT space). The term
uncontrolled manifold implies that elemental variables do not have to
be controlled as long as they are within that subspace because it
corresponds to perfect performance. The signature inequality, VUCM >
VORT (both variance indices have to be quantified properly!), has
been used as a sign of performance-stabilizing synergy in many
studies (reviewed in Latash et al. 2007; Latash 2008, 2019).
Such an organization can use both short-latency feedback loops
within the central nervous system and feedback loops from
peripheral sensory endings (Latash et al. 2005; Martin et al. 2009).
We have already considered two examples of physiological loops
that can be viewed as a performance-stabilizing organization: the
system of recurrent inhibition within the central nervous system (see
chapter 16) and the tonic stretch reflex involving sensory endings
(see chapter 18 and earlier in this chapter).
In more intuitive terms, variability of elemental variables may be
viewed as the sum of two components, “good” variability and “bad”
variability. Good variability (within the UCM) does not affect important
performance variables and therefore may be relatively large. Bad
variability (orthogonal to the UCM) changes important performance
variables and has to be kept low. For the earlier example of two
fingers producing a constant force, deviations of finger forces along
a line corresponding to the equation F1 + F2 = FTASK are “good”
variability because they keep the total force at the required level
(VGOOD in figure 22.8). Deviations of finger forces from that line are
“bad” variability (VBAD in figure 22.8).
Figure 22.8 An illustration of good and bad variabilities for the task of constant
force production with two fingers pressing in parallel. The ellipse shows a
hypothetical distribution of data points collected over many trials. UCM =
uncontrolled manifold; ORT = direction orthogonal to the UCM.

PROBLEM 22.5
Do you like the term “good variability”? It does not hurt
performance, but why is it good, not irrelevant?

Note that variability can be “good” or “bad” only with respect to a

particular performance variable. For example, imagine that the two
fingers produce forces while pressing on two sides of a pivot (figure
22.9). To preserve balance, any increase in the force produced by
one finger should be matched by a similar increase in the second
finger’s force. Then, good variability is along a line F1 - F2 = 0, while
bad variability is orthogonal to that line, in contrast to the illustration
in figure 22.8.
 To analyze the error compensation feature of synergies within the
framework of the UCM hypothesis, one has to know the relations
between small deviations in elemental variables and changes in a
performance variable that the system may try to stabilize. Such
relations are known as the Jacobian matrix of the system. These
relations can be found based on the mechanical properties of the
system—for example, on its geometry (Zatsiorsky 1998).
Sometimes, however, these relations are hidden and have to be
discovered in experiments. In particular, mappings between small
changes in muscle activation variables and task-relevant salient
mechanical variables are generally unknown. So, to study multi-
muscle synergies, one has to discover the Jacobian matrix based on
special experimental manipulations and computations
(Krishnamoorthy et al. 2003, 2007).
Analysis within the uncontrolled manifold hypothesis allows one to
ask a multi-element system questions like: Are you a synergy
keeping down the variability (stabilizing) of such-and-such
performance variable in such-and-such task? Moreover, this analysis
allows one to introduce quantitative indices of alleged synergies and
track their changes with practice, injury, rehabilitation, aging, and
development (reviewed in Latash and Huang 2015; Latash 2019).
Figure 22.9 An illustration of good and bad variabilities for the task of pressing
with two fingers on two sides of a pivot. The ellipse shows a hypothetical
distribution of data points collected over many trials. Note that what used to be
good variability in the task of force production (figure 22.8) has become bad
variability in the task of balancing on the pivot.

PROBLEM 22.6
Can the two methods, PCA and UCM, be combined within one
study? What kinds of questions can be answered by such a
combined approach?

CHAPTER 22 IN A NUTSHELL
The human neuromotor system is
characterized by redundancy at many
levels of analysis. The traditional
approach to the problem of motor
redundancy has been based on the idea
of elimination of redundant degrees of
freedom. Optimization approaches have
commonly been used to solve the
problem. An alternative approach views
the system as abundant, not redundant.
Within that approach, the controller
does not look for a unique solution to
the problem but facilitates families
of equivalent, good enough solutions.
Motor synergies may be characterized
by two features, organizing numerous
variables produced by elements at the
selected level of analysis (elemental
variables) into groups and ensuring
the stability of important performance
variables. The first feature is
commonly explored and quantified using
principal component analysis and
related methods. A quantitative method
to study the stability of performance
has been developed within the
uncontrolled manifold (UCM)
hypothesis. According to this
hypothesis, the controller acts in a
space of elemental variables, allows
large deviations in a subspace (UCM)
corresponding to a desired value of a
performance variable of the whole
system, and limits deviations
orthogonal to that subspace. The
method is based on partitioning motor
variability into “good” (which does
not affect important performance
variables) and “bad” (which does).
 Chapter 23

Patterns of Single-Joint
Movements

KEY TERMS AND TOPICS

multijoint and single-joint movements
isotonic movements
isometric contractions
triphasic EMG pattern
kinematic profiles
dual-strategy hypothesis

In this chapter, we discuss the patterns and control of movements

performed in a single joint, which rarely happen in everyday life. This
is, however, a natural step from the neural control of a single muscle
discussed in the previous chapter to the control of natural
movements discussed in the next few chapters.
Figure 23.1 If a person is trying to perform “the same” movement several times,
for example, to pick up a cup of tea, the trajectories of all the joints and of the hand
will be different in different attempts.

23.1 Isotonic Movements and

Isometric Contractions
Our everyday movements usually involve many joints and are
performed in conditions of changing external forces. Even if a person
tries to do “the same thing” a few times—for example, to pick up a
cup of tea from the table—the movements will be slightly different
during different attempts (figure 23.1). This phenomenon of motor
variability is always present, even in the best trained subjects. We
have already mentioned Bernstein’s classic study of professional
blacksmiths hitting the chisel held in one hand with the hammer
moved by the other hand (Bernstein 1930; reviewed in Bernstein
1967, Jansons 1992). These were the best trained subjects one can
imagine since they had performed such actions hundreds of times
every day for years. However, even these persons showed
substantial variability in the trajectories of both individual joints and
the hammer. An important message is that humans always show
variable motor patterns, even in tasks that may be viewed as
perfectly practiced and automatized. Besides that, all persons are
unique and are characterized by different dimensions of body
segments, different experiences, and different abilities to learn new
motor tasks. A scientific study, however, implies a possibility of
reproducing an experiment in another laboratory and with another
group of subjects and getting similar results.
In order to solve at least part of this problem, experimenters
frequently try to reduce the number of parameters describing motor
tasks and also the variables available to the subject for performing
the task. This is commonly done by confining the movements to a
single joint, one axis of rotation, and, frequently, constant external
load. Movements against an apparently constant external load are
termed isotonic. Isotonic conditions are achieved, for example, by
restricting movements to a horizontal plane when there are no
changes in the force of gravity.

PROBLEM 23.1
What will happen with the activity of Golgi tendon organs in the
agonist muscle and tendon during a fast movement in isotonic
conditions?

Another favorite among experimenters’ regime of muscle work is

termed isometric contraction. The word “isometric” implies a lack of
changes in muscle length, which can never be secured, even in
animal experiments. Any active contraction of muscle fibers leads to
a decrease in their length. Even if this decrease is relatively small, it
can be realized at high velocities and can lead to changes in the
activity of length- and velocity-sensitive receptors (spindle endings).
In human studies, tendons and other passive tissues surrounding the
joints and the muscles deform under changes in muscle force. This
leads to inevitable changes in muscle fiber length when the muscle
activation level changes. In particular, in typical isometric
conditions, the tendon stretches and muscle fibers shorten during
muscle activation so that the length of the “muscle plus tendon”
complex does not change.

PROBLEM 23.2
What will happen with the level of activity of muscle spindles
(primary and secondary endings) during a muscle contraction in
isometric conditions?

Unfortunately for the experimenters, most of the commonly

studied joints of human limbs (e.g., shoulder, elbow, wrist, and ankle)
have more than one axis of rotation (kinematic degree of freedom)
and are controlled by more than two muscles. Some of these
muscles are biarticular (i.e., their contractions directly induce
changes in joint torque or movements in two adjacent joints).
Besides that, a movement in a joint may lead to a change in the
relation between muscle force and joint torque due to changes in the
level arm; thus, different active muscle forces may be required to
balance a constant external load. So, we are coming to the
conclusion that single-joint movements against a constant external
load (isotonic) and without changes in muscle length (isometric) do
not exist in vivo.
Even if single-joint movements and contractions existed and were
obtainable in experimental conditions, one might question their
relation to movements performed during everyday activities and to
problems of motor control of voluntary movements in general.
However, experimenters stubbornly continue to study these fictitious
phenomena. They have good reasons to do so.
First, the progress of science goes from simple to complex, and
studies at a certain advanced level usually require understanding of
the lower functional levels.
Second, although results of single-joint studies cannot be directly
generalized to multijoint movements, they provide theoretical
frameworks and experimental approaches that can be helpful for
understanding general principles of motor control irrespective of the
number of joints and muscles involved.
Third, investigations of single-joint movements have proven to be
useful in clinical studies. In some patients, voluntary motor control is
restricted to single muscles or single joints; therefore, understanding
basic principles of control of such movements is likely to be useful.
And fourth, experimental studies require reproducible conditions
of performing experiments and taking into account all the important
factors that are likely to influence a subject’s performance. Even for
human single-joint movements, it is, strictly speaking, virtually
impossible to take into account all such factors. For multijoint
movements, the situation is even more complicated since joint
interaction forces come into play (reviewed in Zatsiorsky 2002),
biarticular muscles are likely to start playing more important roles,
the force of gravity changes its direction with respect to individual
joints for most of the movements, and the number of involved
variables increases dramatically. Therefore, the chances of
performing poorly controlled experiments increase for multijoint
movement studies.
So, let us consider experimental situations when the subjects are
asked to perform single-joint isotonic movements or isometric
contractions. Assume that these regimes of muscle work exist,
keeping in mind that this assumption is likely to be false. This is
commonly done in research for the purpose of discovering general
regularities in behavior unaffected by simplifying assumptions. For
example, geometry starts with accepting the definition of a point as
an object without length, width, and height. Such objects do not
exist, but the assumption has proven to be crucial for the
development of this field of mathematics.

23.2 Task Parameters and

Performance Variables
In most motor control experiments, relations between task
parameters (what is required from the subject) and performance
variables (what the subject is doing) are studied. For single-joint
movement studies, task parameters may include movement
amplitude, movement time (or speed), external load, accuracy
requirements, and other specific instructions to the subjects (e.g.,
“Make a smooth movement,” “Avoid oscillation at the end of the
movement,” “Do not correct the final position if you miss the target,”
etc.). Performance variables include kinematic variables (joint
position, speed, and acceleration), kinetic variables (joint torque and
its derivatives), electromyograms (EMGs), accuracy indices (e.g.,
variability of the final position or the percentage of trials hitting the
target), and other recorded or calculated indices. Note that any
indices computed from performance variables, such as their
derivatives, integrals, and ratios, are also performance variables.
Studying relations between performance variables and task
parameters is an example of a typical “black box” approach when
input–output relations in a complex system are explored in order to
check hypotheses on the internal structure or principles of
functioning of the “black box.” An absolute prerequisite for any new
hypothesis is to be experimentally disprovable. This means that a
hypothetical experiment studying relations between task parameters
and performance variables should be able to generate results
incompatible with the hypothesis, so that it can be rejected and
replaced with an alternative hypothesis. Therefore, studies of
relations among these two groups of parameters and variables are
crucial for the development of the field of motor control.
Note that performance variables are characterized by natural
variability (i.e., variations across repeated attempts at the same
task), even when performed by very well trained subjects in
reproducible conditions. Because of the variability, one cannot
expect absolutely identical signals even when the subject is
seemingly performing the same movement in the same conditions.
So any experiment is this area requires collecting several trials and
using appropriate statistical approaches.
 PROBLEM 23.3
You want to show that there are no changes in a performance
variable with a change in a task parameter. How can you do this?

Figure 23.2 A typical kinematic pattern of a single-joint movement (six elbow

flexions over 54° were averaged). The beginning of the movement can be defined
differently based on the first visible change in the acceleration (Tacc), velocity
(Tvel), or angle (Tang). It is even harder to define the end of the movement. One
can use an instant when the acceleration reverses its direction (T1), or when it
comes to zero (T2), or when the joint signal first hits the target (T3), or some
velocity-based criterion. Note that the figure shows a relatively smooth movement
without major oscillations at the end, which can only complicate the task.
Adapted by permission from M.L. Latash, Control of Human Movement (Champaign, IL:
Human Kinetics, 1993), 109.
23.3 Kinematic Patterns During
Single-Joint Isotonic Movements
In general, an infinite number of joint–time profiles can lead from an
initial joint position to a required final position. This is an example of
motor redundancy at the level of trajectories. Despite the broad
choice of possible trajectories for typical tasks, kinematic profiles of
natural single-joint movements show certain regular features
common across a range of amplitudes and velocities. Trajectories of
natural movements are commonly smooth; they show unimodal, bell-
shaped velocity profiles and double-peaked acceleration profiles
similar to those illustrated in figure 23.2. Such kinematic profiles are
typical of movements that are not very slow and not very fast. Slow
movements show a less smooth velocity profile, commonly with a
time interval of a relatively constant velocity. On the other hand, very
fast movements commonly show substantial terminal oscillations
about the final position, which are viewed as signs that human joints
are underdamped (i.e., they cannot dissipate kinetic energy quickly
enough) (see chapter 5; see also Cannon and Zahalak 1982; Milner
and Cloutier 1998).
Figure 23.2 also illustrates typical problems with defining such a
basic movement characteristic as movement time. Indeed, it is hard
to define when movement starts. The result will depend on the
selected kinematic variable: The moments when trajectory, velocity,
and acceleration show first visible changes obviously differ in figure
23.2 (compare TANG, TVEL, and TACC). Defining the moment of
movement termination is also far from trivial because of typical
oscillations at the end of even very smooth and fast movement:
compare T1, T2, and T3 in figure 23.2.
A particular computational criterion has been suggested to
account for the smoothness of natural single-joint movements.
According to this criterion, natural movements minimize an integral
function of jerk (CJERK) computed over the movement time. As a
reminder, jerk is the time derivative of acceleration (or the third time
derivative of joint trajectory) (Hogan 1984; Zatsiorsky 2002):

where CJERK stands for the integral jerk index, MT is movement time,
and α is joint angle trajectory (see also chapter 22). This criterion
has been shown to generate trajectories that show all the typical
features of the trajectories of natural movements.

23.4 EMG Patterns During

Single-Joint Isotonic Movements
Before proceeding any further, two basic notions of agonist and
antagonist muscles need to be introduced. The word agonist (or
agonist muscle) will be used for a muscle whose activation
accelerates the limb or increases joint torque in a required direction.
A muscle whose activation brakes the movement or resists joint
torque developed by the agonist is going to be termed an antagonist
(or antagonist muscle). These notions are as good or as bad as
those of isotonic movements and isometric contractions, but they
help to describe certain major findings in this area.

PROBLEM 23.4
Can one and the same muscle be both an agonist and an
antagonist for the same movement (or contraction)?

Recording muscle activity during fast isotonic single-joint

movements reveals a typical electromyographic (EMG) pattern
usually termed the triphasic pattern (figure 23.3). The beginning of
the agonist EMG burst is usually the first detectable event
accompanying fast voluntary movement. It precedes the first
detectable kinematic changes by a few tens of milliseconds. The
initial agonist burst is accompanied by relatively low coactivation of
the antagonist muscle and is followed by an antagonist EMG burst
during which the agonist is relatively quiescent. The antagonist burst
can be followed by a second burst in the agonist. At the final
position, there is usually a visible increase in both agonist and
antagonist tonic muscle activity. The triphasic pattern is obvious
during very fast movements. A decrease in movement speed leads
to less pronounced EMG bursts, and eventually, slow movements
are accompanied by a sustained agonist activity with a low
coactivation of the antagonist.
Changes in other task parameters lead to a variety of reproducible
changes in the triphasic pattern. Typically, the subjects are asked to
make movements “as fast as possible,” or “at different speeds,” or
“being both fast and accurate.” Then movement amplitude, load,
target size, or several of these parameters are changed. As a result,
regularities can be seen in certain parameters describing the
triphasic pattern. These parameters may include the duration of the
EMG bursts, their peak values, the integrals of the bursts, the delay
between two successive bursts, or the level of muscle coactivation at
the end of the movement.
Figure 23.3 The triphasic electromyographic (EMG) pattern begins with a burst
of activity in the agonist muscle (triceps), followed by an antagonist burst (biceps),
which sometimes is followed by a second agonist burst. Note that the first agonist
burst starts several tens of milliseconds prior to joint trajectory.
Adapted by permission from M.L. Latash and J.G. Anson, “What Are Normal Movements in
Atypical Populations,” Behavioral and Brain Sciences 19, no. 57 (1996).

1. An increase in the velocity of movements over a constant

amplitude against a constant load (figure 23.4) leads to an
increase in the rate of EMG rise, peak value and area of the
first agonist burst (biceps and brachioradialis curves during
elbow flexions illustrated in figure 23.4), a decrease in the
delay before the antagonist burst, and an increase in the
antagonist burst peak amplitude and area (Corcos et al.
1989). The duration of the first agonist EMG burst is
sometimes reported to increase and sometimes to stay
constant. These differences are likely to reflect different
methods used for assessing burst duration rather than any
fundamental differences in EMG patterns. The level of final
 coactivation of agonist and antagonist muscles increases with
movement velocity.
2. An increase in movement amplitude without changes in
external load and instructions concerning movement velocity
(e.g., “as fast as possible” or “at the same speed”) leads to
relatively uniform rates of agonist EMG rise, a higher and
longer first EMG burst with a corresponding increase in its
area (biceps and brachioradialis curves in figure 23.5), longer
delays before the antagonist burst, and inconsistent changes
in the antagonist burst amplitude and duration (Gottlieb et al.
1989a,b, 1990). An increase in movement amplitude for
relatively small movements leads to an increase in the
antagonist burst, while further increase may lead to a
decrease in the antagonist activity (Gottlieb et al. 1989a,b,
1990).
3. An increase in inertial load without changing movement
amplitude or instructions concerning movement velocity (e.g.,
“move as fast as possible,” figure 23.6) leads to a higher and
longer agonist EMG activity, although without obvious
changes in the rate of the EMG rise, a longer delay before the
antagonist burst, and no apparent changes in the antagonist
burst characteristics (Gottlieb et al. 1989b). Final coactivation
of agonist and antagonist muscles increases with inertial load.
Figure 23.4 Typical changes in the triphasic EMG pattern during movements
over a constant amplitude, against a constant external load, but at different
speeds.
Adapted by permission from G.L. Gottieb, D.M. Corcos, and G.C. Agarwal, “Organizing
Principles for Single-Joint Movements: A Speed-Insensitive Strategy,” Journal of
Neurophysiology 62 (1989): 342-357. ©1989 The American Physiological Society.

Variations in more than one of the three major task parameters

(amplitude, velocity, and load) lead to combined effects on the EMG
bursts, as, for example, in experiments when the subjects were
required to perform movements of different amplitude within the
same movement time (Gielen et al. 1985; Shapiro and Walter 1986).
Figure 23.5 Typical changes in the triphasic EMG pattern during movements
over different amplitudes, against a constant external load, and under an
instruction to move “as fast as possible.”
Adapted by permission from G.L. Gottieb, D.M. Corcos, and G.C. Agarwal, “Organizing
Principles for Single-Joint Movements: A Speed-Insensitive Strategy,” Journal of
Neurophysiology 62 (1989): 342-357. ©1989 The American Physiological Society.
Figure 23.6 In this figure, the inertial load was changed while movement
amplitude and the instruction to be “as fast as possible” were preserved.
Adapted by permission from G.L. Gottieb, D.M. Corcos, and G.C. Agarwal, “Organizing
Principles for Single-Joint Movements: A Speed-Insensitive Strategy,” Journal of
Neurophysiology 62 (1989): 342-357. ©1989 The American Physiological Society.

PROBLEM 23.5
What will happen with the first agonist burst and the antagonist
burst if the subject is performing movements of different
amplitudes within the same movement time?

PROBLEM 23.6
What kind of an EMG pattern do you expect to see if a subject is
performing a fast isotonic movement to a target and quickly back
to the initial position?
23.5 EMG Patterns During
Single-Joint Isometric
Contractions
Two types of isometric contractions have usually been studied: step
and pulse contractions. Step contractions require the subject to
increase joint torque up to a certain level, while pulse contractions
require also returning quickly back to the initial level of joint torque
(commonly, zero net torque). Other instructions have been used to
specify the time of the torque increase, accuracy constraints, and the
like—similar to what has been done in studies of single-joint isotonic
movements. Fast isometric contractions are accompanied by
triphasic EMG patterns similar to those observed during fast isotonic
movements (Ghez and Gordon 1987; Gordon and Ghez 1987;
Corcos et al. 1990). However, the second delayed agonist burst is
more frequently absent. Bursts of the EMG activity become smaller
and poorly defined with a decrease in the rate of torque increase,
and slow increases in joint torque are accompanied by a tonic
increase in the agonist EMG and a smaller increase in the antagonist
EMG.
Figure 23.7 Typical changes in the triphasic electromyographic (EMG) pattern
during isometric contractions with different rates of torque rise.
Adapted by permission from D.M. Corcos, G.L. Gottieb, G.C. Agarwal, and B.P. Flaherty,
“Organizing Principles for Single-Joint Movements: IV. Implications for Isometric
Contractions,” Journal of Neurophysiology 64, (1990): 1033-1042. ©1989 The American
Physiological Society.

Similar measures were used for relating parameters of isometric

contractions and isotonic movements to changes in the EMG
patterns. Let us describe regularities in muscle activation patterns
starting from the step experiments.
1. An increase in the rate of torque rise keeping the final torque
level constant leads to an increase in the rate of EMG rise,
peak value and area of the first agonist burst (see biceps and
brachioradialis in figure 23.7), no apparent changes in the
delay before the antagonist burst, and an increase in the
antagonist burst amplitude and area (Corcos et al. 1990).
There are no obvious changes in the duration of either burst
or the level of final coactivation of agonist and antagonist
muscles with the rate of torque rise.
 2. An increase in the final torque level without changes in
instructions concerning the rate of torque rise (e.g., “as fast as
possible,” figure 23.8) leads to relatively uniform rates of
agonist EMG rise, higher agonist and antagonist burst
amplitudes, higher integrated EMG activity for both muscles,
higher final levels of both agonist and antagonist EMGs, and
inconsistent changes in the duration of both bursts (Corcos et
al. 1990).
During pulse contractions, EMG patterns become more phasic,
with better-defined bursts, better pronounced second agonist burst,
and a lower level of muscle cocontraction at the final state (Corcos et
al. 1990). In this case, an increase in pulse torque amplitude without
changes in the rate of torque increase leads to a longer first agonist
burst, a delay in the antagonist burst, and inconsistent changes in
the antagonist burst amplitude and area (figure 23.9). An increase in
the rate of torque rise without changing pulse amplitude leads to
faster-rising first agonist and antagonist EMG bursts.
Figure 23.8 Typical changes in the triphasic electromyographic (EMG) pattern
during isometric contractions with the same rates of torque rise but to different
target values of torque.
Adapted by permission from D.M. Corcos, G.L. Gottieb, G.C. Agarwal, and B.P. Flaherty,
“Organizing Principles for Single-Joint Movements: IV. Implications for Isometric
Contractions,” Journal of Neurophysiology 64, (1990): 1033-1042. ©1989 The American
Physiological Society.

23.6 Dual-Strategy Hypothesis

Changes in the EMG patterns during simple movements have been
used as the basis for a number of hypotheses on the neural control
of such movements. Some of these hypotheses assume that EMGs
are reliable indices of control signals within the central nervous
system reflecting processes of voluntary motor control. An
alternative group of hypotheses is based on an assumption that
EMGs are generated with equally important contributions from two
sources: central control signals and the activity in peripheral reflex
loops. Analysis of EMG patterns is an important tool for both groups
of hypotheses. One of the most developed hypotheses within the
first group is the dual-strategy hypothesis (Gottlieb et al. 1989a,b).
The variety of EMG findings in single-joint experiments creates an
impression of total chaos. An attempt to introduce an order (a
classification) into this array of data was undertaken in the form of
the dual-strategy hypothesis. The basic idea of this hypothesis is
rather straightforward: A person can perform movements “at the
same speed” or “at different speeds.” Movements performed “at the
same speed” are assumed to be controlled using a speed-insensitive
strategy, while movements performed “at different speeds” are
assumed to be controlled using a speed-sensitive strategy. Note,
however, that actual movement speed is a time function that
depends, in particular, on external loading conditions. So, the
instruction to move “at the same speed” over different amplitudes or
against different inertial loads makes no sense if understood directly:
Both peak speed and average speed will vary under such
instructions. It is interesting, however, that naïve subjects have no
problems following such vaguely formulated instructions, which
suggests that there exists a central variable at the control level
associated with the notion of movement speed.
Figure 23.9 A typical EMG pattern during an isometric pulse contraction.
Adapted by permission from D.M. Corcos, G.L. Gottieb, G.C. Agarwal, and B.P. Flaherty,
“Organizing Principles for Single-Joint Movements: IV. Implications for Isometric
Contractions,” Journal of Neurophysiology 64, (1990): 1033-1042. ©1989 The American
Physiological Society.

When using instructions with the word “speed,” the dual-strategy

hypothesis implies not the actual movement speed (or its measure,
such as average speed or peak speed) but an internal variable that
is used by the brain to vary movement speed. For example, if a
subject is asked to perform movements “as fast as possible,” he or
she will probably use the highest available value of this internal
variable related to movement speed. Actual speed, however, may be
very different if, for example, movements are made against different
inertial loads (see figure 23.6).
This hypothesis was based originally on an idea that humans
control movements by specifying commands to the α-motoneuronal
pools of the agonist and antagonist muscles that define the EMG
patterns of these muscles. The commands were approximated as
rectangular pulses—the excitation pulses—with modifiable height
and duration. The speed-insensitive strategy was associated with
using the same height of the excitation pulses and varying their
duration (figure 23.10a). The speed-sensitive strategy was
associated with modifications in the height of the excitation pulses
with or without changes in their duration (figure 23.10b).

Figure 23.10 An illustration of changes in the excitation pulse during movements

performed based on the speed-insensitive (top) and speed-sensitive (bottom)
strategies.

Unfortunately for this version of the hypothesis, the α-

motoneuronal pools receive not only descending signals from the
brain but also signals from peripheral receptors that induce reflex
changes in their activity. These changes are reflected in the EMG
patterns. The activity of the peripheral receptors depends on actual
changes in muscle length, joint angles, and tendon forces. Imagine
that a person uses the same signals from the brain to control a
couple of movements (figure 23.11). During the second movement,
the inertial load is unexpectedly increased four-fold. Apparently,
movement speed will drop. This will be reflected in changes in the
activity of virtually all the peripheral receptors. As a result, the reflex
effects of the receptors on the α-motoneurons will change, leading to
a change in the EMG patterns. This example illustrates that EMG
patterns are not reliable indices of central control signals.

Figure 23.11 Schematic illustration of kinematic and EMG patterns for two
movements, both performed over the same amplitude and under the same
instruction to be “as fast as possible.” The inertial load was, however, four times as
high during the second movement. Note the difference in movement kinematics,
which will obviously be reflected in different reflex effects on both agonist and
antagonist α-motoneurons.

PROBLEM 23.7
Can one consider EMG patterns during isometric contractions as
pure reflections of hypothetical central commands? Why?

The dual-strategy hypothesis was rather successful in describing

the regularities in the first agonist EMG burst during single-joint
movements (Gottlieb et al. 1989b; Corcos et al. 1989, 1990). This is
not surprising, because during fast movements, the first agonist
burst is commonly about 100 ms long, so there is little time for the
reflexes to exert significant effects on the muscle activation levels.
The problems of the dual-strategy hypothesis became obvious when
the same approach was used for delayed events, like the antagonist
burst, or muscle contractions with significantly different kinematics, in
particular comparing isotonic movements and isometric contractions.
However, the basic idea of classifying hypothetical motor commands
into two categories may be applied without assuming exclusive
central control over the activity of α-motoneurons (and EMGs). Then,
it helps to introduce order into a variety of experimental findings
performed in different laboratories.

Figure 23.12 According to the equilibrium-point hypothesis, neural control of a

joint can be viewed as the process of specifying stretch reflex thresholds to the
agonist (λAG) and antagonist (λANT) muscles. An equivalent pair of variables can
be used corresponding to the joint d when the resultant moment is zero (reciprocal
command or r-command) and to the range of joint angles when both muscles are
activated (co-activation command or c-command).
23.7 Single-Joint Movements
Within the Equilibrium-Point
Hypothesis
Within the equilibrium-point hypothesis (see chapters 20 and 21), the
neural control of a joint with one kinematic degree of freedom can be
described with two control variables corresponding to the stretch
reflex thresholds for the agonist and antagonist muscles, λAG and
λANT, respectively. Alternatively, two different variables can be
introduced (see figure 23.12; see also Feldman 1980, 1986), known
as reciprocal command (r-command) and coactivation command (c-
command). The former command corresponds to joint angle where
the total moment of force equals zero. In figure 23.12, the force–
length characteristics of the opposing muscles are assumed to be
symmetrical. Then the r-command corresponds to the midpoint
between λAG and λANT. The c-command defines the joint angle range
where both muscles are active (i.e., the range between λAG and
λANT). In terms of mechanics, changes in the c-command lead to
variations in the slope of the joint torque–angle characteristic shown
as the thick line in figure 23.12.
The control of single-joint movements can be described as time
functions λAG(t) and λANT(t), or, alternatively as r(t) and c(t). To
describe the typical triphasic EMG patterns, one has to take into
account the velocity sensitivity of the primary spindle endings, which
contribute to muscle activation via the stretch reflex loop. This can
be done by assuming that this factor changes the effective λ values,
which we will label as λ* (Abdusamatov and Feldman 1986). In fact,
the velocity sensitivity of the primary spindle endings is only one of
the factors that make λ* different from the centrally defined λ. Other
factors include reflex effects from other muscles, including those
crossing other joints and the dependence of λ on the history of
neuronal activity. This was expressed formally as an equation:
where μ is a coefficient reflecting the effects of velocity (V) sensitivity
of the primary spindle endings, ρ reflects the effects of other reflex
loops, and ƒ(t) reflects history-dependent changes in the stretch
reflex threshold (Feldman and Latash 2005). Let us consider a
simplified version:

Figure 23.13 illustrates the trajectory of the agonist muscle during

a fast change of its λ to a smaller value on the length–velocity plane
(phase plane). This plane can be divided into zones, according to
equation 23.3, corresponding to the muscle being silent (to the left of
the dashed slanted line) and activated (to the right of that line). Let
us assume for simplicity that, when the muscle is in its activation
zone, its activation level depends on its distance from the line
separating the two zones. The muscle trajectory is illustrated as a
thick dashed line. Note that muscle activation (due to the initial shift
in λAG) leads to its shortening (i.e., negative velocity), which
contributes to a drop in the muscle activation. If movement velocity is
high, the muscle enters the silent zone (this is the end of the first
agonist burst) and then can re-emerge (second agonist burst) before
settling in the final state. A similar picture can be drawn for the
antagonist muscle showing the emergence of the antagonist burst.
Figure 23.13 Velocity sensitivity of the primary endings in muscle spindles can
lead to a sequence of muscle activation bursts without preprogramming these
bursts. If the controller shifts the command to the agonist from λ1 to λ2, the
following movement will shorten the agonist and stretch the antagonist. This will
lead to cessation of the first agonist burst and the emergence of the antagonist
burst. Further, the agonist can enter the activation zone one more time, leading to
the second agonist burst. The slanted lines show the border between the
activation zone (EMG > 0) and the silence zone (EMG = 0) for the agonist muscle.

One of the main problems of the equilibrium-point hypothesis is

the lack of tools for measuring relevant control variables such as λAG
and λANT or the r-command and c-command. Several studies tried to
compute these variables based on the performance of subjects in
typical single-joint tasks assuming a simplified model of joint
behavior (Latash and Gottlieb 1991, 1992; Latash 1992). During very
fast movements, the reconstructed patterns of the r-command
showed nonmonotonic shapes (the N-shaped trajectories) with a
transient peak in the c-command during the movement. These
results have been a point of controversy because of the simplicity of
the underlying model (see Gribble et al. 1998).
Figure 23.14 The speed-insensitive and speed-sensitive strategies are
associated with changes in the r-command at a constant rate (top) and at varying
rates (bottom).

Within the equilibrium-point hypothesis, the speed-insensitive

strategy may be associated with a shift in the r-command at a
constant rate over different amplitudes (figure 23.14a). The speed-
sensitive strategy can be associated with shifts of the r-command at
different rates (figure 23.14b; Latash 1993). There is also a transient
increase in the c-command during the movement, which is an
important factor for both accelerating the movement and braking it
efficiently (reviewed in Feldman 2015). Note also that the
equilibrium-point hypothesis does not view the control of movements
and isometric contraction as qualitatively different. The main
difference is in the external loading conditions, which lead to different
EMG time profiles due to the action of reflex loops. This does not
mean, of course, that the control patterns are always identical. They
can be modified by the person based on experience to achieve the
most effective and comfortable performance.
CHAPTER 23 IN A NUTSHELL
Investigation of relations among task
and performance parameters or
variables is a common method of
studying a complex system. Fast
single-joint movements are typically
accompanied by a triphasic EMG pattern
seen in the agonist–antagonist muscle
pair, a nearly symmetrical bell-shaped
velocity, and a double-peaked
acceleration. There is a set of
reproducible relations among task
parameters and parameters of the
triphasic pattern. Such relations have
been extensively studied in isotonic
and isometric loading conditions. The
dual-strategy hypothesis introduces a
classification of all movements into
two groups, with and without explicit
or implicit control over movement
time. Its original version implies
independent central control over the
total presynaptic input into
motoneuronal pools; thus, it does not
take into account the reflex effects
on EMG patterns. The equilibrium-point
hypothesis introduced two basic
commands, the r-command and the c-
command, for the control of single
joints. It can account for the
triphasic EMG patterns based on the
velocity sensitivity of muscle spindle
endings.
 Chapter 24

Multijoint Movement

KEY TERMS AND TOPICS

reaching movement
interaction torques
interjoint reflexes
spinal mechanisms of joint coordination
supraspinal mechanisms
multijoint synergy
referent coordinate
equilibrium trajectory

We are going to discuss now how a unidirectional movement of a

limb from a certain initial to a certain final position is organized. Such
movements may be considered basic components of virtually the
entire human everyday motor repertoire. Pointing at an object,
picking up a cup of tea, and hitting a nail with a hammer all involve
components that may be considered single, unidirectional reaching
movements.
Let us start by introducing a notion of the working point—that is,
a point whose trajectory is most directly related to successful
execution of a motor task. When pointing at, picking up, or
manipulating objects, this point may be located on the fingertips or
somewhere on the palm. When kicking a football, the working point
is somewhere on the tip of the shoe. This point need not even be in
permanent direct contact with our body. When Michael Jordan of the
Chicago Bulls threw a basketball, the task was to get the working
point (the basketball) into the basket. In this example, the working
point is in direct contact with the player’s hand only during the initial
segment of its trajectory. In most computer games, the working point
is an image of a fighter plane or a superman on the screen, and the
player controls it by moving the joystick or by pressing keys, without
any direct mechanical contact with the working point.
If you ask a naive person to make a movement with the dominant
hand from a comfortable position in the external space to another
comfortable position, this movement will demonstrate certain
reproducible features (figure 24.1):
1. The trajectory of the endpoint will be nearly straight.
2. The velocity profile of the endpoint will be bell-shaped, nearly
symmetrical.
3. The acceleration of the endpoint will demonstrate two peaks.
4. Trajectories, velocities, and accelerations of individual joints
may demonstrate different properties, including reversals in
the direction of motion.
If the subject of this mental experiment is asked to repeat the
same movement several times, there will be high variability of the
individual joint trajectories and lower variability of the endpoint
trajectory. These observations date back to the classic studies of
blacksmiths by Bernstein in the 1920s, as discussed in chapter 22.
Thus, the most reproducible feature of the trajectories of multijoint
reaching movements is the lack of perfect reproducibility, which is
addressed as variability. Variability is a curse of motor control studies
that do not investigate it explicitly. Reducing variability is frequently
the goal of training procedures in both everyday life (including the
training of athletes or the rehabilitation of patients with motor
disorders) and laboratory environment. On the other hand, variability
by itself is a fascinating phenomenon that obeys its own laws and
demonstrates consistent relations among task and performance
parameters (reviewed in Newell and Corcos 1996; Davids et al.
2005). Any theory that aspires to explain how multijoint movements
are controlled must have an explanation for the phenomena of the
variability of natural human movements.

Figure 24.1 A natural reaching movement to a target is characterized by a nearly

straight trajectory of the endpoint (the working point, WP) and a bell-shaped
velocity profile.

PROBLEM 24.1
What can you conclude from the fact that the variability of
individual joints is higher than the variability of the tip of a tool
moved by the hand?

24.1 Two Issues With

Controlling Natural Reaching
Movements
During reaching movements, the working point is typically located on
the hand, and controlling its trajectory in a body-centered reference
frame requires moving a multilink chain connecting the hand to the
body. If the body moves, and you would like to control the trajectory
of the working point in the external, Cartesian space, the situation
becomes even more complicated. Two major sources of complexity
emerge when a researcher makes the step from single-joint motion
to the more realistic multijoint reaching. The first source of
complexity is related to the more complex biomechanics of the
controlled peripheral system—related in particular to the important
roles of new forces that are typically ignored during single-joint
movement analysis, such as moments of forces emerging in joints
due to motion of other joints of the kinematic chain. These include
biarticular muscles that provide links between adjacent joints, thus
introducing one more important factor into joint interactions. There
are also new neurophysiological factors playing roles during
multijoint movements, such as interjoint and interlimb reflexes.
During single-joint movements, muscles crossing the joint and
external forces coming from the environment (such as the external
load and gravity) define the joint torque and the mechanics of joint
movement. During the movement of a multijoint limb, there is one
more source affecting joint torque, namely interaction torques,
which are due to the motion of other joints of the same limb.
Consider a very simple two-joint system (figure 24.2). The following
equations describe torques (T) in each of the joints (Zatsiorsky
2002):
Figure 24.2 A simple two-joint system moving in a plane. Torques in each of the
two joints depend on motion in the other joint (see equations in the text).

In these equations, I1,1 is the moment of inertia of the chain about

joint α1; I2,2 is the moment of inertia of the second link about joint α2;
I1,2=I2,1 are coupling inertial terms; D is the inertial resistance of the
chain to the centrifugal and Coriolis forces; and G stands for gravity
torque components. In each of the two equations, there are terms
that depend on motion of the other joint. These are interaction
torques.
Several studies have suggested that the ability to deal with
interaction torques is important for the production of accurate natural
movements (Hollerbach and Flash 1982; Sainburg et al. 1999;
Ketcham et al. 2004). In particular, one study compared arm
movements during a horizontal hand motion resembling cutting a
loaf of bread in healthy persons and in a person with large-fiber
peripheral sensory neuropathy (Sainburg et al. 1995). This rare
condition leads to a complete loss of signals from proprioceptors and
receptors sensitive to touch (see chapter 6) in certain areas of the
body (in particular, in the arm) without affecting motor pathways. The
study has shown that, when the person suffering from this type of
neuropathy could not see the hand action, the hand trajectory
became curved, and the curvature could be explained by the inability
of that person to compensate for interaction torques that emerged
during the action.
Studies have suggested that the ability to stabilize the trajectory of
the working point during fast multijoint arm movements in conditions
of changed interaction torques is different between the dominant and
nondominant arms (Bagesteiro and Sainburg 2002, 2003). The
dominant arm of right-handed persons shows a much better ability to
produce smooth, straight trajectories to a target in such conditions.
In contrast, the nondominant arm produces more curved trajectories
that, however, allow the working point to reach the target more
accurately than in the dominant arm.
More recent studies of the effects of hand dominance on reaching
movements (reviewed in Sainburg 2005; Mutha et al. 2012) have
suggested that movements of the dominant arm (and the associated
neurophysiological structures and circuits) are specialized for
predictive control of fast actions. In contrast, movements of the
nondominant arm are specialized for steady-state performance.

PROBLEM 24.2
Give examples from everyday life of the different roles of the
dominant and nondominant arms.

The second source of complexity is the famous Bernstein’s

problem of overcoming excessive degrees of freedom that was
discussed in chapter 22. During virtually all voluntary movements,
the number of kinematic degrees of freedom for the limb, which can
be associated with the number of independent axes of joint rotation
summed over all the joints of the limb, is higher than the number of
variables needed to execute a motor task or to describe its
execution. The latter number is frequently three, corresponding to
the three-dimensional space where we happen to live. Figure 24.3
illustrates Bernstein’s problem for a three-joint limb performing a
planar reaching task to a target. Note that an infinite number of joint
configurations (three are illustrated in figure 24.3) can correspond to
the final position of the working point.

Figure 24.3 There are an infinite number of limb configurations (combinations of

joint angles) that can correspond to a fixed position of the tip (working point) of a
three-joint limb on a plane.

24.2 Interjoint Reflexes

A variety of reflex responses to a stimulus applied to a limb induce
contractions of muscles acting at different joints of the limb or even
muscles of another limb. Among the best known examples are the
flexor reflex and the crossed extensor reflex described in chapter 18:
a pinch applied to the surface of an animal’s paw, or an electrical
stimulation of a skin nerve, can induce bursts of activity in virtually all
the major flexor muscles of the limb. Simultaneously, a burst of
activity in extensor muscles of the contralateral limb can be seen.
These responses come at a relatively long latency (between 50 and
90 ms), suggesting their polysynaptic nature. They originate from the
activation of afferent fibers of relatively small receptors (nociceptors,
free endings, and secondary muscle afferents) united under the
name flexor reflex afferents (FRA).
In an intact animal or human, a change in the length of a muscle
of a limb necessarily leads to joint motion and, therefore, to a change
in the length of other muscles crossing the same joint. In addition,
the presence of biarticular muscles makes purely single-joint motions
impossible; that is, a movement in a joint leads to a change in the
length of all the muscles crossing the joint, including the biarticular
muscle(s), which may lead to a motion in an adjacent joint (figure
24.4). In animal experiments, however, it is possible to separate the
distal tendon of a muscle from its place of attachment and to apply
controlled changes in muscle length without changing the length of
other muscles. Moreover, in such studies one can independently
manipulate muscle force and length.
A series of such experiments performed by T. Richard Nichols and
his colleagues (Nichols et al. 1993; Bonasera and Nichols 1996;
Nichols 2002, 2018) has demonstrated a complex pattern of the
reflex effects from force-sensitive receptors across the muscles of a
cat hindlimb. The well-known pattern of these reflexes, described in
chapter 18, involves inhibition of α-motoneurons controlling the
agonist muscle and disinhibition of α-motoneurons controlling the
antagonist muscle induced by the afferent fibers of the Golgi tendon
organs. Nichols and colleagues showed, however, that reflex
interaction among apparent agonists (three heads to the triceps
surae muscle) could be less “classical” and more asymmetrical. On
top of those responses, there are reflex effects on muscles crossing
other joints of the limb, such as the quadriceps and hamstrings.
These patterns are sensible from the point of view of biomechanics
and could provide the basis for organizing coordinated patterns of
muscle activity (synergies; see chapter 22) participating in functional
multijoint limb movements.
Figure 24.4 A change in the length of muscle M1 will lead to a motion in joint J1.
As a result, the length of the biarticular muscle M2 will also change, leading to a
motion in joint J2.

PROBLEM 24.3
You want to perform a single-joint movement in a joint that is
crossed by a uniarticular muscle and a biarticular muscle. What
kind of reflex effects would be helpful for such an action between
peripheral receptors in the uniarticular muscle and α-motoneurons
innervating the biarticular muscle?

24.3 Multijoint Coordination by

the Spinal Cord
There is a variety of exciting studies in multijoint motor control using
animal preparations. Let us look closely at one group, studies of the
wiping reflex in the spinal frog (i.e., in a frog preparation with the
spinal cord surgically separated from the brain). This choice is
defined by the extremely intriguing performance of such a seemingly
simple object (a spinal frog) during such a complicated motor task
(wipe a stimulus off the back with a hindlimb; try it and you will see
that this is not easy!). After spinalization, impulses from the brain
cannot reach segments of the spinal cord caudal to the place of
transection. On the other hand, the spinal neuronal mechanisms
controlling hindlimb movements remain intact.
 If an experimenter places a stimulus (a small piece of paper
soaked in an acid solution) on the back of a sitting spinal frog (figure
24.5), the frog, after a certain latent period, performs a series of
coordinated movements wiping the stimulus off the back and,
sometimes, throwing it away from the body (reviewed in Berkinblit et
al. 1986). Wiping of the same skin area in successive cycles could
be performed in different directions, and orientation of the foot
relative to the stimulus (the attack angle) could also change. If the
stimulus is placed on the ipsilateral forelimb, accurate wiping
movements are observed even if the position of the forelimb relative
to the body changes (Fukson et al. 1980). So the spinal cord “knows”
where the limbs are!

Figure 24.5 A spinal frog can remove an irritating stimulus from its back with a
coordinated movement of the ipsilateral hindlimb. The wiping is performed as a
sequence of movements in different directions.
Adapted by permission from M.L. Latash and M.T. Turvey, Dexterity and Its Development
(Mahwah, NJ: Springer and Lawrence Erlbaum, 1996), 286. Reproduced with permission of
The Licensor through PLSclear.
 A series of experiments was performed investigating the effects of
unexpected perturbations on the wiping (reviewed in Latash 1993).
In one series, a loose thread loop was placed on the hindlimb,
preventing movements in the knee joint beyond a certain limit so that
the maximal knee joint excursion was about 5°. The frog was able to
remove the stimulus from its back during the first attempt! Then, the
knee was released, and a cast was placed preventing movements in
the next (more distal) joint. The frog once again wiped the stimulus
at the first trial! Next, a lead bracelet was placed on the distal part of
the hindlimb; the weight of the bracelet was similar to the weight of
the hindlimb itself. The frog still was able to wipe the stimulus
accurately.
These experiments show that control signals, even at the spinal
level, are not formulated in terms of contractions of individual
muscles or even movements in individual joints. Otherwise, loading a
segment or blocking a joint would have led to grossly inaccurate
movements. Second, they imply the existence of very fast
corrections of movement patterns that are presumably built into the
sequence of control signals for wiping generated by the spinal cord.
Taken together, the observations point to the neural control being
organized at the level of the working point of the hindlimb (its toes)
directed toward the target. The ability of the frog to wipe the stimulus
off in the presence of major perturbations is an example of the
dynamical stability of the toe trajectory, suggesting that the spinal
cord is able to organize kinematic synergies stabilizing salient
variables (see chapter 22 and later, in section 24.7).
Further investigations of the organization of hindlimb movements
in the spinal frog used electrical stimulation of spinal cord structures
(Giszter et al. 1993; Bizzi et al. 1995). These studies have shown
that the mechanical effects of such stimulation could be described
with a relatively small number of basic force fields known as motor
primitives. Some of the primitives were associated with converging
force fields to a particular equilibrium of the hindlimb’s endpoint in
the body-centered space. Other fields moved the endpoint to an
anatomical limit, which can be a viewed as a consequence of
specifying an equilibrium point beyond the anatomical reach of the
hindlimb. One can question the physiological adequacy of the
electrical stimulation used because it likely activated numerous
groups of interneurons with various functions. However, in general,
these findings point at two important conclusions. First, control of the
hindlimb by the spinal cord uses the equilibrium-point principle (see
chapter 21). Second, control may be simplified by manipulating with
a small number of primitives rather than with a larger number of
muscles (see the discussion of synergies in chapter 22).

24.4 Supraspinal Mechanisms

Earlier, several groups of studies were discussed that demonstrated
dependences between the activity of a population of neurons in
various brain structures (including the primary motor cortex) and the
direction of voluntary limb movement (chapters 8-10). These studies
are in good agreement with the general idea that the central neural
mechanisms manipulate parameters related to movement of the
working point (commonly, the endpoint of a limb) rather than
parameters related to the involvement of individual muscles or joints.
There are other studies suggesting that the activity of supraspinal
structures during voluntary movements is related to characteristics of
the endpoint trajectory. In particular, patterns of neuronal activity in
the supplementary motor area, motor cortex, putamen, and red
nucleus have suggested relations to characteristics of voluntary
movement at a rather abstract kinematic level, which are compatible
with the general idea of the neural control of working-point trajectory
during multijoint movements. In particular, it has been shown that
patterns of activity of cortical cells can encode the trajectory
independently of required muscle forces and patterns of muscle
activation (Georgopoulos et al. 1986).
James Houk suggested that neuronal activity in the red nucleus
can participate in motor control by encoding properties of feedback
reflex systems in the spinal cord (Gibson et al. 1985; Houk and
Gibson 1987). Earlier, Houk’s group reported that the activity of
individual red nucleus neurons could be related to the initiation,
speed, and amplitude of voluntary movements. These are the basic
variables assumed to be used in limb endpoint control. Later, a more
general model of control of endpoint position was suggested based
on interactions between the cerebellar and rubral nuclei. This model
is based on positive feedback loops between the nucleus
interpositus and magnocellular red nucleus and nucleus reticularis
tegmenti pontis (figure 24.6). These loops were assumed to be
under an inhibitory control from the Purkinje cells in the cerebellum.
Releasing of the feedback loops by switching off sets of Purkinje
cells has been assumed to lead to the emergence of adjustable
pattern generators, leading to a movement.
Most studies in animals used brief electrical (or transcranial
magnetic stimulation, TMS) pulses applied to various brain
structures to investigate their role in the organization of movements.
This method has certain advantages, such as the ability to measure
the latencies of motor reactions that provide information about the
involved neural circuits. However, this method can also be viewed as
rather nonphysiological because, during natural actions, brain
structures are more likely to produce long episodes of activity
leading to smooth, longer-lasting movements. Experiments in
monkeys have shown that long-lasting (a few seconds) stimulation of
the motor-related cortices leads not to jerky movements of a
particular anatomical area (as suggested by the classic Penfield
homunculus) but to coordinated multijoint movements that could
involve several effectors (Graziano et al. 2005). These apparently
involuntary actions looked like segments of natural everyday actions
by the monkey taken out of context. For example, they could involve
an arm motion toward the head and a simultaneous rotation of the
head toward the arm as if trying to place a morsel of food into the
mouth. These observations suggest that the activity of neurons in the
cortical areas may be related to complex, purposeful multijoint
actions.
Figure 24.6 An interaction between the red nucleus and cerebellum is assumed
to participate in the generation of “motor programs” by turning off sets of Purkinje
cells, thus disinhibiting positive feedback loops shown in the scheme.

24.5 Neural Control Variables

for Multijoint Movements
What is controlled during multijoint movements? At present, one can
suggest only hypothetical answers to this question. But let us first try
to answer another question: What variables cannot be control
variables? Hopefully all the previous material has persuaded the
reader that the following list does not contain viable candidates for
the vacant positions of neural control variables for multijoint
movements:
1. forces produced by individual muscles,
2. activation patterns of individual muscles,
3. torques in individual joints, and
4. rotations in individual joints.
 PROBLEM 24.4
Why cannot levels of muscle activation be control variables for the
production of multijoint movements?

In order to be able to ensure accurate control over the trajectory of

a working point, the central nervous system should use control
variables that are related to functionally significant external variables,
such as coordinates of the working point or force vectors generated
at the working point. These variables are what we care about during
multijoint movements because they are directly related to
performance, while individual muscle forces and joint rotations are
not so important as long as they do not lead to awkward limb
configurations and postural problems. We can call these essential
variables following the term coined by Israel Gelfand and Michael
Tsetlin half a century ago (1966). The fact that working point
trajectory is the most reproducible performance variable provides
indirect support for this view. If we now remember that our muscles
have spring properties, it becomes apparent that a working point has
these properties as well; that is, if you keep the values of neural
control variables unchanged, the coordinates of a working point in
space will depend on the external force vector.
Perform the following simple experiment. Ask a friend to push with
a fist against your palm somewhere in front of his or her body (figure
24.7a). Then ask the friend not to intervene voluntarily, and push his
or her fist with your palm with smooth and quick pushes. You will feel
that the working point (the fist) resists attempts to move it from the
original position in a spring-like fashion as if it were a ball resting on
the bottom of a pit (figure 24.7b).
Figure 24.7 When you ask a friend to keep a hand in position while pressing
against your hand, small variations of the force applied by your hand will lead to
small displacements of your friend’s hand, as if (lower drawing) the hand resided
on the bottom of a potential well on an elastic membrane.

So the neural command of the subject in this experiment did not

define the position vector or the force vector of the fist because
these both changed with changes in the external force (all the EMGs
would change as well, if you placed electrodes on involved muscles).
Rather, the neural command defines spring parameters of the
system that can be expressed with a spatial equilibrium vector
(similar to the situation with the equilibrium length of a regular,
unidimensional spring) and properties of the force field in the vicinity
of that equilibrium coordinate generated by all the arm muscles
(these properties, for a single muscle, can be associated with just
one number, apparent stiffness). There is an indirect link between
this approach and neurophysiological observations by Donald
Humphrey (1982), who identified two groups of neurons in the motor
cortex of the monkey: the stimulation of the neurons of the first group
caused movement in a joint, while the stimulation of the neurons of
the other group caused cocontraction of agonist–antagonist muscle
pairs, thus modifying joint apparent stiffness. This is exactly what
one would expect from hypothetical cells controlling equilibrium
coordinates and a force field in the vicinity of that coordinate.

Figure 24.8 An illustration of the force-control and equilibrium-point control

approaches. A ball on an elastic membrane can be moved from point A to point B
either by applying precomputed force (top drawing) or by pressing with a finger on
the membrane close to point A and moving the finger to point B (bottom drawing).
No computation and no direct contact with the ball (exerting forces on the ball) is
required in the latter case.

One can visualize this approach with the following metaphor:

Imagine a ball on an elastic membrane (figure 24.8). The ball will be
in an equilibrium on the bottom of a potential well created by its
weight in point A. If you want to move the ball to a new point B, you
can either calculate and apply patterns of forces directly to the ball
(F(t) in the top drawing in figure 24.8) or change the shape of the
elastic force field—for example, by pressing near the ball with a
finger and then moving the finger to a new, desired position (X(t) in
the bottom drawing in figure 24.8). Note that in the second case,
possible errors in the estimation of the mass of the ball, as well as
possible small perturbations during the ball movement, will not
change the final position: The ball will eventually end up in the same
spot—that is, on the bottom of a newly created potential well. This
property is called equifinality. If the subject of this mental
experiment applies precomputed force patterns directly to the ball,
the same error factors may lead to different ball trajectories and
different final positions. Similar approaches have been developed by
Anatol Feldman and Mindy Levin (1995; reviewed in Feldman 2015),
who consider the neural control of multijoint movements with shifting
positional frames of reference, and by a group including Emilio Bizzi
and Ferdinando Mussa-Ivaldi (Bizzi et al. 1992; Shadmehr et al.
1993), who propose that multijoint movements are controlled with
combinations of force fields applied to the endpoint of a limb.
The idea of control with referent spatial coordinates for important
body parts, such as the working point of a limb, has been developed
actively over the years. Consider the concept of equilibrium
trajectory and the associated hypothesis, and then discuss
hierarchical control of natural multijoint reaching movements.

PROBLEM 24.5
Where is the referent spatial coordinate for the endpoint of a limb
when pressing with constant force against a rigid object?

24.6 Equilibrium-Trajectory
Hypothesis
The equilibrium-trajectory hypothesis, originally suggested by
Tamar Flash and Neville Hogan (1985), is a natural generalization of
the single-joint equilibrium-point hypothesis (see chapters 20 and 21)
to multijoint movements. It implies that the central nervous system
can shift an image of the working point along a desired trajectory
expressed in external Cartesian coordinates. During a movement,
this virtual trajectory is always ahead of the current actual position of
the working point, and this disparity produces active forces driving
the working point (figure 24.9). Note that the actual trajectory of a
working point may be different from the virtual trajectory due to a
number of factors, including the dynamic properties of the limb and
changes in the external forces (for example, changes in the effects
of gravity during movements that change the orientation of limb
segments). The equilibrium-trajectory approach avoids the
computational problem of inverse dynamics and inverse kinematics,
since muscle forces are not calculated by the central nervous system
but emerge as results of shifts of centrally defined equilibrium
coordinates of the working point, which is done in the external
Cartesian space.

Figure 24.9 The control of a reaching movement can be seen as specifying a

time series of the equilibrium trajectory (open circles) of the endpoint of the limb.
The actual trajectory (filled circles) will lag behind because of a number of factors,
including the inertia of the limb.

The equilibrium-trajectory hypothesis has been combined with an

optimization approach called the minimum jerk principle (see chapter
22). According to this approach, the central nervous system
generates trajectories corresponding to a minimum of an integral of a
function of jerk of the limb endpoint. This approach has led to a
number of theoretical predictions, including straight paths of the
working point trajectories, smooth, unimodal velocity profiles (bell-
shaped velocity profiles), invariancy of the trajectory under
translations, rotations, and its scaling with speed or amplitude
(Morasso 1981, 1983). A number of these predictions correspond to
experimental observations. In particular, nearly straight paths and
bell-shaped velocity profiles have been observed for various initial
and final working point locations in external space and for
movements at different speeds corresponding to the predictions
concerning invariancy of the working point trajectory.

24.7 Hierarchical Control With

Spatial Referent Coordinates
When a referent coordinate (RC) is specified for the endpoint of a
limb that is different from its actual coordinate, active force is
generated in the direction of the RC, and the endpoint moves. This
endpoint movement can only happen as a result of joint rotations
within the limb, which can only be produced by the activation of
muscles, which, according to the equilibrium-point hypothesis
(chapter 21), is a consequence of the differences between the
values of muscle length and threshold of the tonic stretch reflex (λ).
Figure 24.10 illustrates a sequence of transformations leading from
RC specified at the highest, task-specific level of this control
hierarchy to the r-command and c-command pairs for each of the
involved joints, and to λ values for each of the involved muscles.
Note that the number of involved joints is typically mechanically
redundant or abundant in terms of motor control (chapter 22).
Similarly, the number of muscles acting at each axis of joint rotation
is also abundant.
Figure 24.10 The process of control of a whole-body movement is associated
with a sequence of transformations of control variables specified at the task level
(referent coordinate, RCTASK) to those at the limb level (RCLIMBS), joint level
(RCJOINTS), and muscle level (RCMUSCLES). At the muscle level, RC is equivalent
to the λ–stretch reflex threshold. Note that the number of variables increases along
this sequence of transformations.

Figure 24.10 suggests that the generation of commands to

muscles during reaching multijoint movements is always associated
with a sequence of few-to-many mappings. Each such mapping
seemingly involves a problem of motor redundancy because it
requires specifying a larger number of output variables based on a
smaller number of inputs (constraints). As a result, reaching
movements typically can be performed using different combinations
of joint rotations. For example, persons who have suffered a stroke
involve their proximal (commonly less affected) joints, including the
trunk, to a larger degree than healthy persons (Cirstea and Levin
2000; Levin et al. 2002). Relative joint involvement has been
addressed using optimization methods (see also chapter 21) based
on cost functions involving psychological concepts, such as comfort
(Cruse and Bruwer 1987; Bruwer and Cruse 1990), and more
complex cost functions with costs associated with moving each of
the joints (Rosenbaum et al. 1993).
Analysis of the activity in neuronal brain populations (Feldman
2019) allowed researchers to link them to two components of RC at
the task-specific level. The first one is related to the target spatial
coordinate (three-dimensional) for the limb endpoint, and the second
one reflects the coactivation zone about this target coordinate. In
other words, the neuronal populations seem to encode the r-
command and c-command at the level of the extremity, similar to
those commands introduced for the control of a single joint (Feldman
1986; see also chapter 21). These insights have been indirectly
supported by observations of intentional changes in the ellipse of
stiffness at the limb endpoint, reflecting its resistance to small
changes in the external force (Flash 1987). Humans could only
intentionally change the dimensions of the stiffness ellipse, but not
its orientation. In other words, they modified a single parameter to
accomplish this task: the c-command at the limb endpoint level.

24.8 Multijoint Synergies

As illustrated by Bernstein in his classic study of blacksmiths
(Bernstein 1930), the average trajectory of the limb endpoint across
repeated attempts shows less variability than the trajectories of
individual joints (confirmed for pointing movements by Jaric and
Latash 1999). These observations suggest the existence of multijoint
synergies (see chapter 22) stabilizing the endpoint trajectory. Indeed,
the hierarchy illustrated in figure 24.10 suggests that such synergies
can be seen at different levels by the covaried involvement of
individual joints and muscles.

PROBLEM 24.6
What kind of a multijoint synergy can be expected during pressing
by the endpoint against a stationary object with a required force
vector?
 Kinematic multijoint synergies stabilizing important variables have
been documented for a range of limb movements, including
reaching, pointing, aiming, and moving an object to be thrown at a
target (Scholz et al. 2000; Reisman and Scholz 2003; Tseng et al.
2003; Yang and Scholz 2005; Mattos et al. 2011; Hasanbarani and
Latash 2020). Multi-muscle synergies stabilizing the salient endpoint
variable during actions by a multijoint limb have also been
documented (Krishnamoorthy et al. 2007).
Figure 24.11 illustrates the time profiles of the two variance
components quantified in the joint configuration space, affecting and
not affecting the orientation of the barrel of the handheld infrared
pistol with respect to the infrared-sensitive target. In other words,
one of the components reflected joint variance within the
uncontrolled manifold (VUCM; see chapter 22), and the other one
reflected joint variance orthogonal to the UCM (VORT). Note that VUCM
> VORT from the very beginning of the movement (i.e., the central
nervous system of the subject stabilized the pistol barrel orientation
even when it pointed at a large angle from the target). Such
strategies—stabilizing salient variables during the whole movement
duration—have been reported in other studies, such as those
involving throwing a Frisbee and a basketball (Yang and Scholz
2005; Hasanbarani and Latash 2020).
Figure 24.11 also implies the existence of RC-based synergies—
for example, stabilization of the task-related variables by covarying
involvement of the r-command and c-command at the level of the
endpoint or at the level of individual joints, or of λs at the level of
muscles. Such studies have not yet been performed during reaching
movements, but RC-based synergies have been documented for
finger-pressing tasks (Ambike et al. 2016; Reschechtko and Latash
2017).
Figure 24.11 Multijoint synergies stabilizing important variables during reaching,
pointing, and shooting tasks are reflected in higher inter-trial variance along the
corresponding uncontrolled manifold (VUCM) as compared to variance along the
orthogonal to the UCM direction (VORT).

CHAPTER 24 IN A NUTSHELL
Reaching movements to a target are
characterized by nearly straight
endpoint trajectories, nearly
symmetrical bell-shaped velocity
profiles, and two-peak acceleration
profiles. Movement trajectories vary
across trials at the same motor task.
Trajectory of the working point
(endpoint) demonstrates lower
variability than individual joint
trajectories, suggesting the existence
of multijoint synergies. Interaction
torques play an important role during
multijoint movements. Interjoint and
intermuscle reflexes provide patterns
of joint interaction that may benefit
everyday movements. The spinal cord
can organize multijoint targeted
movements and associated synergies
stabilizing endpoint trajectory.
Generalization of the equilibrium-
trajectory hypothesis views the
control of a multijoint limb as a
process of shifting of a referent
coordinate for the endpoint. Exact
control variables for multijoint
movements are unknown; they are not
muscle force patterns, patterns of
muscle activation, or joint
trajectories. The motor cortex, the
cerebellum, and the red nucleus are
likely to be very important for the
control of voluntary reaching
movements.
 Chapter 25

Postural Control

KEY TERMS AND TOPICS

vertical posture
role of vestibular system
role of vision
preprogrammed reactions
anticipatory postural control
postural synergy
effects of vibration on posture

In biomechanics, the term posture relates to joint configuration

(Zatsiorsky 2002). This definition, however, is not very useful if one
wants to study postural control in healthy persons or postural
disorders in neurological patients. Indeed, postural control commonly
refers to one’s ability to keep either vertical posture in the field of
gravity or another kinematic variable relatively unchanged (stable)
under changes in external forces and during voluntary actions
(Latash and Zatsiorsky 2016). If a clinician states that a patient
displays impaired postural control, this statement does not mean
problems with keeping specific joint configurations but rather with
keeping certain important kinematic variables within acceptable
ranges or being able to avoid falling during everyday actions.
 The issue of postural control has two aspects that have both
common features and differences. First, postural control may refer to
maintaining the position of a part of the body with respect to either
the body itself or the environment (or an external object moving in
the environment). Consider the following examples of this aspect of
postural control:
A figure skater maintaining a beautiful arm configuration while
sliding over the skating rink
A waiter navigating through a restaurant with a loaded tray on
the right hand
A person holding a nail with one hand while hitting it with a
hammer
The other aspect, which will be discussed in this chapter, refers to
maintaining the position of the whole body or of its center of mass
with respect either to the environment or to a particular direction in
the environment, such as the direction of gravity. Examples of this
aspect involve such common activities as sitting, standing, and
walking. Examples of maintaining postural control for the whole body
include
Sitting while reading this textbook
Standing in line at a bus stop or coffee shop
Walking to class

25.1 Vertical Posture

The fact that human beings are able to maintain vertical posture
looks like a miracle. One would have a hard time imagining a
mechanical system that is less stable in the field of gravity.
Sometimes, during analysis of the neural control of vertical posture,
the human body is modeled as an inverted pendulum (figure 25.1),
which is not easy to equilibrate, especially in the presence of
external perturbations and changes in its orientation with respect to
the field of gravity.
Figure 25.1 The human body in the field of gravity can be modeled as an
inverted pendulum, which is inherently very unstable.

However, the problem is much more complicated due to the

presence of a number of joints along the axis of the pendulum:
Potentially, each joint can collapse. In physics, the stability of a static
mechanical system in the field of gravity requires that the projection
of its center of mass (COM) fall within the area of support (figure
25.2). The area of support for humans is relatively small (on the
order of one square foot), which requires fine-tuning of the
interaction between movements in different joints along the human
body in order to maintain the equilibrium.
Figure 25.2 The human body has a number of joints along its vertical axis.
Maintaining equilibrium in the field of gravity requires the projection of the COM to
fall within the area of support.

PROBLEM 25.1
Can one maintain postural equilibrium if the projection of the COM
falls beyond the limits of the area of support?

Another miraculous phenomenon is that humans can perform limb

movements without falling down. Assessments of the transient
forces acting on body segments during natural limb movements
suggest that they are more than sufficient to destroy the fragile
postural equilibrium. And, finally, there are no words adequate to
express the profound awe experienced by motor control
physiologists when they see a walking person. A multilink inverted
pendulum can walk and run and even maintain vertical posture after
a slip or a stumble!
These fantastic features of the hypothetical system controlling
vertical posture did not escape the attention of Nikolai Bernstein,
who formulated a number of problems and approaches that still
remain in the center of contemporary research. Bernstein suggested
that programming of a voluntary movement must include two distinct
components: the first related to the movement itself, and the second
related to maintenance of the vertical posture. Bernstein also
considered maintenance of the vertical posture to be an illustration of
his concept of synergies (see chapter 22), coordinated combinations
of neural commands to a number of elements, such as joints and
muscles, leading to a desired common goal (e.g., not falling down).
Later in this chapter, we will consider examples of postural synergies
based on the definition presented in chapter 22.
Let us start the analysis of control of the vertical posture with a
seemingly very easy task—to stand quietly and do nothing. Quiet
standing is associated with one of the poorly understood phenomena
in motor control—that is, postural sway.

25.2 Postural Sway

When a person tries to stand still, he or she never stands truly still.
There are unavoidable changes in different mechanical
characteristics of the vertical posture. The trajectories of two of these
characteristics are commonly addressed as postural sway: those of
the COM of the body and those of the center of pressure (COP).
COP is defined as the point of application of the resultant vertical
reaction force acting from the support surface on the body (figure
25.3). Different characteristics of sway have been studied, including
mean velocity, the area covered by the sway trajectory over a fixed
time interval, and more sophisticated measures that address the
temporal structure of the sway (Collins and De Luca 1993; Riley et
al. 1997; Oullier et al. 2006).
COP shifts change the moment of force acting on the body and
can change the body’s orientation with respect to gravity. COP shifts,
therefore, may be viewed as the means of moving COM. Typically,
COP shows considerably larger shifts during quiet stance than COM
because of the large inertia of the body. Most studies used
characteristics of the COP trajectory to describe and quantify
postural sway.

Figure 25.3During quiet standing, both the COM and the COP show
spontaneous migration known as postural sway.

Figure 25.4 illustrates typical postural sway (COP migration) of a

person standing with (a) eyes open and (b) eyes closed. It is obvious
that closing the eyes leads to a substantial increase in the sway
magnitude. The lower graph shows the time profile of postural sway
in the anterior–posterior direction. Sway looks like an irregular
process of a relatively modest amplitude, up to 1 cm.
Figure 25.4 Postural sway increases when the subject closes his or her eyes.
 The aforementioned relation between shifts of the COP and COM
has been used to partition postural sway into two components.
Zatsiorsky and Duarte (1999, 2000) have suggested that postural
equilibrium is maintained with respect to a moving, rather than a
stationary, reference point. They developed a method that
decomposes the sway into two processes termed rambling and
trembling (figure 25.5). Rambling represents migration of the
reference point, with respect to which the equilibrium is instantly
maintained, while trembling represents the COP oscillation about the
rambling trajectory. According to the rambling-trembling hypothesis,
the former reflects supraspinal control processes, while the latter is
defined primarily by the peripheral mechanical properties of the
postural system and spinal reflexes. In more intuitive terms, the body
oscillates about a position, which itself migrates for reasons that are
not well understood.

Figure 25.5 An illustration of decomposition of postural sway (COP trajectory)

into two components, rambling and trembling.
Reprinted by permission from L. Mochizuki, M. Duarte, A.C. Amadio, et al., "Changes in
Postural Sway and its Fractions in Conditions of Postural Instability," Journal of Applied
Biomechanics 22 (2006): 51-60.
 There has been an ongoing debate on the role of muscle stiffness
in postural stabilization during quiet stance. First, muscle stiffness is
an ill-defined term in motor control (see chapter 5 and Latash and
Zatsiorsky 1993), which usually refers to the generation of muscle
force against stretch and is assumed to be proportional to the
magnitude of stretch up to a certain length and at a given level of
muscle activation. Some authors claimed that ankle muscle stiffness
alone was sufficient to maintain upright posture (Winter et al. 1996,
1998), while others claimed that this hypothesis was not supported
by data (Morasso and Sanguineti 2002; Casadio et al. 2005; Loram
et al. 2005).
There are two polar views on the role of postural sway in balance.
One view is that sway has no functional role, representing “noise”—
that is, a reflection of imperfect functioning of the neural control
system that cannot avoid generating some sway (Kiemel et al.
2002). As mentioned earlier, the view that something in human
bodies is imperfectly designed does not seem like a very attractive
starting point. Therefore, we prefer the alternative view that sway is a
consequence of purposeful processes within the central nervous
system, possibly reflecting a search for the limits of stability (Riccio
1993; Riley et al. 1997).

PROBLEM 25.2
Based on the two views on postural sway, predict what will
happen with sway if a person balances on a board with a narrow
beam glued to its bottom. Will the sway be bigger or smaller than
during standing on the floor? Why?

Proper balance during standing requires the integration of

information from different sources, including the vestibular system,
visual information, and proprioceptive information. Let us start with
the role of the vestibular and ocular systems in postural control and
then discuss the contribution of signals originating from
proprioceptors in urgent corrections of posture.
25.3 Role of the Vestibular
System
The sense of balance during standing is one of the least prominent
in our consciousness. Humans become aware of it only in situations
when balance is seriously endangered. One of the important sensory
systems involved in the neural control of balance is the vestibular
system of the brain and inner ear, which provides signals related to
the relative orientation of the head with respect to the direction of the
field of gravity. A detailed description of the vestibular system can be
found in chapter 13.
One can separate the dynamic and the static functions of the
vestibular system. The dynamic function is principally mediated by
the receptors in the semicircular ducts; it allows humans to track
head rotations in space and plays an important role in the reflex
control of eye movements. The static function is mediated mostly by
the hair cells in the utricle and the saccule. It enables humans to
monitor the absolute head position in space and plays an important
role in postural control. The cell bodies of neurons innervating
vestibular receptors are located in the Scarpa’s ganglion. These
neurons are bipolar. One branch of each axon goes to the peripheral
receptors (hair cells), while the other branch travels in the eighth
cranial nerve and terminates in the brainstem.
The effects of signals from the vestibular system on postural
control are mediated by the vestibular nuclei (figure 25.6). These
include the lateral vestibular nucleus (also known as Deiters’
nucleus), the medial vestibular nucleus, the superior vestibular
nucleus, and the inferior vestibular nucleus. In addition to signals
from vestibular receptors, the Deiters’ nucleus receives inputs from
the cerebellum and from the spinal cord. Neurons of the dorsal
portion of the Deiters’ nucleus send their axons into the lateral
vestibulospinal tract, which terminates ipsilaterally in the central
horns and has a profound facilitatory effect on both α- and γ-
motoneurons innervating limb muscles. This tonic input is likely to
play a major role in producing the background activity of the
antigravity muscles.

Figure 25.6 Vestibular nuclei include the lateral vestibular nucleus (also known
as Deiters’ nucleus), the medial vestibular nucleus, the superior vestibular nucleus,
and the inferior vestibular nucleus.

PROBLEM 25.3
Which cerebellar cells project onto the Deiters’ nucleus? Can
these projections be excitatory or inhibitory?

The effects of sensory signals from the vestibular system on

postural control are not trivial. Indeed, on the one hand, signals from
the vestibular system inform the central nervous system about the
position and motion of the head in space, which may or may not be
relevant for the task of standing. Note that a healthy person can
easily perform rather fast head movements without losing balance.
On the other hand, vestibular disorders commonly lead to postural
problems and, in severe cases, to an inability to maintain vertical
posture. One of the hypotheses is that the vestibular system
participates in the creation of a reference frame within which sensory
signals of other modalities, such as vision and somatosensation, are
evaluated with respect to the task of keeping vertical posture
(Maurer et al. 2006; Mergner 2007).
Electrical stimulation of the vestibular system (commonly known
as galvanic vestibular stimulation) can lead to visible effects on
whole-body posture during standing, namely to steady body lean.
Recent studies have suggested that galvanic vestibular stimulation
produces deviations in the referent body orientation in the
environment, resulting in muscle activation and postural changes
(Zhang et al. 2018). This interpretation views descending signals
along the vestibulospinal tracts as contributors, acting in parallel to
other descending systems and defining the overall referent body
configuration (see chapter 21).

25.4 Role of Vision

Vision provides one of the most reliable sources of information for
the human brain. When visual information comes into conflict with
information from another modality, people tend to “believe” their
eyes, not the other source of information. This happens, for example,
when a person looks at his or her limb during high-frequency, low-
amplitude vibration of a muscle within the limb. Without visual
information, the vibration often leads to strong illusions of limb
motion, sometimes leading to sensations of anatomically impossible
joint positions (Craske 1977; Lackner and Taublieb 1983; Roll et al.
1989). If one looks at the limb, however, the illusions become much
less pronounced and commonly disappear.
The system for the control of vertical posture is also strongly
dependent upon visual information. For example, all possible indices
of postural stability become worse if the subject is standing with the
eyes closed (see figure 25.4). In particular, there is an increase in
body sway during quiet standing, larger deviations of the COM in
response to postural perturbations, and larger postural deviations
induced by the vibration of postural muscles, which we discuss
somewhat later.
Movement of the visual background is well known to produce
illusions of body movement of the observer (reviewed by Oullier et
al. 2006). For example, if a person is sitting in a train that has
stopped at a station, and the train on the next track starts to move
slowly, the person has the feeling that the stopped train is moving in
the opposite direction. If a person stands and looks at a screen that
is displaying a certain pattern, an accelerated movement of the
pattern toward the person induces a sensation of moving forward,
leading to a sway of the body backward (figure 25.7; see Van Asten
et al. 1988; Dijkstra et al. 1994; Ravaioli et al. 2005). Similarly, a
movement of the pattern from the person induces body sway
forward.
Figure 25.7 If a standing subject looks at a screen that displays a certain pattern,
an accelerated movement of the pattern toward the subject induces a sway of the
body backward.

Usually, there are limits to the extent to which the human brain
“believes” any one of the sources of information used for postural
control. Thus, an illusion created by a flow of information along one
of the channels is kept down or may even be eliminated by
information coming from other sources. In order to create a really
strong illusion, one needs to manipulate synchronously all three
major sources of information related to body position and orientation:
vestibular, visual, and proprioceptive. Disney World uses such
methods in their simulation rides during which the viewers are
seated in comfortable chairs that can be tilted synchronously with
changes of an image projected on the large screen.

25.5 Role of Proprioception

One of the major sources of information about the role of signals
from proprioceptors in postural control is observations of postural
disturbances when proprioceptive signals are distorted. Some of the
effects of high-frequency, low-amplitude muscle vibration were
discussed earlier, including the tonic vibration reflex and changes in
other reflex responses (chapter 19). Effects of muscle vibration may
also be observed at the level of postural control. They are related to
the unusually high level of activity of muscle spindle endings induced
by the vibration. Recall that muscle vibration applied to a muscle
tendon can “drive” virtually all the spindle primary endings of the
muscle (i.e., induce at least one action potential from each primary
ending per vibration cycle) (Brown et al. 1967; Matthews and Stein
1969). The central nervous system may be misled by this
proprioceptive information and interpret it as signaling an increase in
muscle length.
If the vibration is applied to a postural muscle, an illusory increase
in its length is further interpreted as a change in the body’s
orientation. This illusory body sway is compensated by an actual
change in body position (Lackner and Levine 1979; Hayashi et al.
1981). For example, vibration of the Achilles tendon leads to central
overestimation of the length of the triceps muscles (an illusory body
sway forward), which is corrected by an actually observed body
sway backwards. These effects have been termed vibration-
induced fallings (VIFs) (Eklund 1969). They are very strong if the
subject is standing with eyes closed. Vibration of the Achilles tendon
may even lead to actual falling backwards or taking a protective step.
If the subject opens his or her eyes, the effects of vibration are
attenuated and may disappear.
Similar effects on body posture can be observed during vibration
of other postural muscles as well as muscles whose contribution to
postural control may not be so obvious. For example, vibration of
neck muscles induces similar illusions with respect to the position of
the head (Lund 1980). Illusory deviations of the head may further
lead to vestibular illusions, and ultimately, to postural adjustments
similar to VIFs observed during the vibration of leg muscles.
The importance of somatosensors for postural control has also
been illustrated by observations of difficulties with standing in certain
groups of patients who have problems with the somatosensory
modalities. In particular, the late stages of untreated syphilis lead to
a state known as tabes dorsalis, degeneration of the dorsal
conducting pathways in the spinal cord. These patients are unable to
keep vertical posture with their eyes closed. Some severe cases of
diabetes are associated with a loss of signals from proprioceptors in
distal areas of the legs, also leading to major problems with postural
control, which are particularly evident during standing with closed
eyes. Finally, patients with a rare condition known as large-fiber
peripheral neuropathy (sometimes imprecisely called “deafferented
persons”) cannot perform any functional movements without
continuous visual control, and this applies to the task of standing
(Sanes et al. 1985; Sainburg et al. 1995).

25.6 Anticipatory Postural

Adjustments
We have already noted that voluntary arm movements performed by
a standing person are sources of postural perturbations (figure 25.8).
There are two major sources of postural perturbations associated
with fast voluntary movements. First, a change in the body geometry
leads to a change in the coordinates of the COM (see figure 25.8),
which may move outside the area of support and thus needs to be
corrected. Second, during arm movements, inertial forces (shown as
reactive torque, TR, in the shoulders in figure 25.8) and mechanical
joint coupling create torque changes in numerous joints of the body,
including those involved in postural control.
 Therefore, voluntary movements, particularly fast ones, are almost
always associated with changes in the activity of postural muscles.
Some of these changes occur prior to the movement itself and have
been termed anticipatory postural adjustments (APAs). Their
assumed role is to minimize perturbations of the vertical posture that
would otherwise be induced by the movement (Bouisset and Zattara
1987; Massion 1992). Accordingly, one may expect APAs to produce
mechanical effects that would oppose the expected perturbation due
to an intended limb movement or another action, such as picking up
a heavy object or releasing a load from extended arms.
Electromyographic consequences of APAs can be observed as
changes in the activity of postural muscles, while their mechanical
consequences are seen as displacements of the body’s COP. These
adjustments are apparently prepared by the central nervous system
before the actual perturbation occurs (i.e., in a feedforward manner).
As a result, the mechanical consequences of APAs are frequently
suboptimal, such that a residual postural perturbation takes place.

Figure 25.8 A fast arm movement by a standing person is a source of strong

postural perturbation because of the joint coupling and changes in the location of
the COM. A fast shoulder flexion creates reactive torques (TR) that act to push the
body backward.
 Figure 25.9 shows a typical pattern of APAs associated with a fast
bilateral shoulder flexion (arm movement forward). Note that
changes in the background activity of postural muscles occur prior to
the first visible increase in the activity of the prime mover (i.e., a
muscle that is assumed to initiate the required arm movement), the
deltoid muscle in this case; this time instant is marked with the arrow
in figure 25.9. These changes lead to a COP displacement. One can
also see in that figure later, corrective reactions in the activity of
postural muscles, which begin about 100 ms after the surge in the
deltoid EMG.
The properties of APAs, in particular their magnitude, depend on
three major factors: the magnitude of an expected perturbation, the
characteristics of the motor action associated with the perturbation,
and postural stability. The first factor is almost trivial: If a person
anticipates a larger perturbation, larger APAs are generated as
compared to a similar task associated with a smaller perturbation
(Aruin and Latash 1996).
Figure 25.9 A typical pattern of postural adjustments associated with a fast
bilateral shoulder flexion. Changes in the background activity of postural muscles
occur prior to a visible increase in the activity of the “prime mover” (anterior deltoid,
ΔA in the left lower panel). These changes lead to a displacement of the COP
(ΔCP, left middle panel) and movement of major leg joints (left upper panel). Later,
corrective reactions in the activity of postural muscles occur.
Reprinted by permission from M.L. Latash, A.S. Aruin, I. Neyman, and J.J. Nicholas,
“Anticipatory Postural Adjustments During Self Inflicted and Predictable Perturbations in
Parkinson’s Disease,” Journal of Neurology, Neurosugery,and Psychiatry 58 (1995): 326-
334. ©1995 BMJ Publishing Group.

The second factor is less trivial: The neural controller scales the
magnitude of APAs with the magnitude of an action that the person
uses to trigger a perturbation even if the magnitude of the
perturbation does not depend on that action. Studies have
demonstrated that, if an unusually minor action triggers a large
perturbation, the amplitude of the adjustment may be scaled with
respect to the magnitude of the action, even if the perturbation is
standard and fully predictable (Aruin and Latash 1995). For example,
if a person shoots a rifle, the rebound commonly leads to a strong
postural perturbation. One needs to be an experienced marksman to
be able to generate APAs that would compensate for this predictable
and relatively standard perturbation. This dependence of APAs on
action characteristics may result from everyday experience, which
suggests that more vigorous actions are typically associated with
more vigorous perturbations.
The third factor reflects the dependence of APAs on postural
stability and on background whole-body activity. APAs are reduced in
conditions of both very stable and unstable vertical posture (Nardone
and Schieppati 1988; Nouillot et al. 1992; Aruin et al. 1998). If a
person performs an action associated with a postural perturbation
while walking or swaying, APAs in anticipation of a standard
perturbation are modulated within the cycle of the whole-body action
and may even show reversals of their direction (Hirschfeld and
Forssberg 1991; Krishnamoorthy and Latash 2005). This happens
when APAs themselves may potentially destabilize posture—for
example, if their action is associated with COP shift toward a
dangerously close edge of stability.

PROBLEM 25.4
Suggest an example from everyday life when postural
adjustments are scaled with respect to an action rather than with
respect to an expected perturbation.

25.7 Corrective Postural

Reactions
There are several lines of defense against unexpected or
uncompensated postural perturbations (table 25.1). The first one is
peripheral elasticity of muscles, tendons, and other tissues. Any
displacement of a joint creates elastic forces resisting the
displacement. Since the elastic properties of a muscle depend on its
level of activation (see chapter 4), the central nervous system can
modulate the efficacy of peripheral elasticity in counteracting
perturbations by adjusting the level of cocontraction of muscle pairs
acting at major postural joints. This method of control is sometimes
known as using preflexes (Prochazka et al. 2000).

Table 25.1 Lines of Defense of Vertical Posture

Typical time
Mechanism delay Important features
Anticipatory postural < 0 Based on a prediction of a
adjustments perturbation
Muscle–tendon elasticity Zero Can be modulated
Monosynaptic reflexes 30 ms Poorly controlled
Polysynaptic reflexes 50 ms Low gain
Preprogrammed reactions 70 ms Approximate correction
Voluntary action 150 ms Late!

The second line of defense is the stretch reflex with its phasic
(mostly monosynaptic) and tonic (mostly polysynaptic) components
(chapters 17 and 18). The stretch reflex also demonstrates
viscoelastic properties and contributes to damping external
perturbations, although at a certain, relatively short, reflex delay.
However, these two mechanisms are not enough to ensure
equilibrium of the body in the field of gravity. The next defensive
mechanism has a longer delay and belongs to the group of
preprogrammed reactions (see chapter 19). It is more powerful and
more flexible than the first two mechanisms. In particular, in certain
situations, preprogrammed responses can be seen in muscles
whose length is unaffected or even decreased by an external
perturbation (Marsden et al. 1979; Nashner et al. 1979; Nashner and
Cordo 1981). Such posture-stabilizing reactions can be seen, for
example, in arm muscles when a person stands in a bus and the bus
unexpectedly starts moving.
 In laboratories, unexpected platform rotations or translations are
commonly used as sources of controlled postural perturbations.
Preprogrammed reactions to such perturbations are frequently
described as combinations of muscle activation patterns specific for
a given perturbation. The earliest of these reactions appear at
relatively short latencies, less than 80 ms, suggesting their
preprogrammed rather than voluntary nature. They are supposed to
be triggered by multimodal sensory inputs with an important
contribution from proprioceptive, visual, and vestibular receptors.
Some of these reactions seem to be rather general—for example,
coactivation of agonist–antagonist muscle pairs stabilizing a postural
joint irrespective of the direction of a perturbation. Other reactions
are specific to the type and direction of a perturbation.
There are certain preferred patterns of corrective postural
reactions. In young, healthy subjects standing on the force platform,
a relatively slow forward translation of the platform induces a body
sway backward, leading to an increase in the background activity of
the ventral muscles (such as the tibialis anterior, rectus femoris, and
rectus abdominis). On the other hand, a backward translation of the
platform results in a body sway forward and an increase in the
background activity of the dorsal muscles (such as soleus, biceps
femoris, and erector spinae). These responses occur at a delay of
about 80 ms in a distal-to-proximal order. This pattern of muscle
activation and the accompanying kinematics have been called the
ankle strategy (Horak and Nashner 1986).
When healthy subjects stand on a surface fitted with support that
is narrow in the anterior–posterior direction or when the platform is
translated very quickly, the order of muscle recruitment is reversed to
proximal-to-distal, known as the hip strategy (Horak and Nashner
1986; Woollacott and Shumway-Cook 1990). The hip strategy is also
seen in elderly persons in conditions when young persons typically
show the ankle strategy (Woollacott et al. 1988).
Because of the geometry of the human body, an ankle movement
leads to a larger horizontal displacement of the COM as compared to
a hip movement of the same amplitude. On the other hand, if a
larger or a smaller-than-optimal joint rotation is generated, the ankle
strategy is likely to lead to a bigger error in the displacement of the
COM and consequently to a higher possibility of losing balance.
Thus, the ankle strategy seems to be more effective and challenging,
while the hip strategy trades efficacy for safety. There are also
intermediate patterns of corrective postural adjustments that may
involve the knee joint (Allum et al. 1989).

PROBLEM 25.5
Can one see anticipatory and corrective postural adjustments in
arm muscles? Give examples.

There are clear differences in the mode of control and function

between the APAs and corrective postural responses. APAs are
initiated by the subject; the later, compensatory reactions are
triggered by sensory feedback signals. APAs try to predict postural
perturbations associated with a planned movement and minimize
them, while compensatory reactions deal with actual perturbations of
balance that occur due to suboptimal efficacy of the APAs or their
absence if the perturbation is unexpected. So, both anticipatory and
compensatory reactions are assumed to be prepared in advance
(i.e., preprogrammed), and they differ in their relative timing with
respect to the limb movement and method of triggering, feedforward
or feedback.

PROBLEM 25.6
What changes in anticipatory and corrective postural reactions do
you expect to observe if one and the same perturbation is applied
many times to a standing person?

PROBLEM 25.7
When a car makes a turn at a high speed, the upper bodies of the
driver and the front seat passenger deviate in opposite directions,
 toward each other or from each other. Why? If a car makes a left
turn, in what directions will the upper bodies of the driver and the
passenger tilt?

If a person performs a motor task several times in a row while

standing, the magnitudes of APAs and preprogrammed reactions
typically show negative covariation: Smaller APAs are associated
with larger preprogrammed reactions (Santos et al. 2010;
Kaewmanee et al. 2020). These observations suggest that these two
basic mechanisms of postural control are based on shared sources
of relevant information.

25.8 Postural Synergies

Nikolai Bernstein introduced the notion of postural synergy as a
combination of control signals to a number of muscles with the
purpose to ensure the stability of a limb or of the whole body in
anticipation of a predictable postural perturbation or in response to
an actual perturbation. According to this definition, postural
synergies may be considered as building blocks used by the central
nervous system to construct meaningful control signals to numerous
joints and muscles. The existence of synergies has been supposed
to take part of the computational load off the shoulders of the central
nervous system. A number of studies explored the activation
patterns in large groups of leg and trunk muscles during responses
to perturbations in different directions (Krishnamoorthy et al. 2003;
Ting and Macpherson 2005; Ting and McKay 2007; Danna-dos
Santos et al. 2007). These studies have shown that, indeed, muscles
formed robust groups with parallel scaling of activation levels
addressed as modes, modules, or synergies.
We have already discussed a somewhat different definition and
function of synergies (chapter 22). According to that definition,
synergies are created in the space of elemental variables to stabilize
important performance variables of the system. For postural tasks,
such performance variables may be associated with the location of
the COM, the location of the COP, trunk orientation, and head
orientation, for example. Elemental variables depend on the selected
level of analysis. For example, if one considers joint kinematics,
postural synergies may be searched for in the space on individual
joint rotations (Scholz et al. 2000; Latash et al. 2002b). For example,
when a person stands up from a sitting posture (sit-to-stand action;
Scholz and Schöner 1999), joint angles have been shown to covary
across trials such that they stabilized the COM trajectory in the
horizontal direction but not in the vertical direction. In another study,
joint angle covariation has been shown to stabilize both horizontal
COM trajectory and trunk orientation during voluntary whole-body
sway movements when the subjects were required to change COP
while moving between visual targets “as quickly and accurately as
possible” (Freitas et al. 2006).
Since the end of the 19th century, researchers have agreed that
the brain does not control muscles one by one but unites them into
groups (Hughlings Jackson 1889). In a number of studies, principal
component analysis, non-negative matrix factorization, and similar
methods have been used to identify such groups (Saltiel et al. 2001;
Ting and Macpherson 2005; chapter 22). These groups can be
viewed as elemental variables (muscle modes or modules) that are
covaried by the controller to ensure postural stability. A number of
studies have also analyzed such muscle-mode synergies and have
shown that they can be assembled in a task-specific manner to
stabilize such variables as COP trajectory (Krishnamoorthy et al.
2003; Wang et al. 2005; Danna-dos-Santos et al. 2007).
A traditional view has been that postural synergies form a
separate group of motor commands that can be mixed and matched
with the motor commands for intentional actions. In other words,
humans are assumed to have two pockets in the brain; they search
one for a command for a planned voluntary movement and they
search the other for an appropriate anticipatory or corrective reaction
(a postural synergy) (figure 25.10). An alternative view is that there is
only one pocket and that separation of peripheral motor patterns into
“action” and “postural” is done by researchers but not by the brain
(figure 25.11). According to the latter view, any movement, even the
one that looks very local, involves many more joints than those
apparently involved in the movement. Thus, changes in the activity
of postural muscles become not additions to a control process for
action but an inherent part of it.

Figure 25.10 According to the traditional view, two motor commands are
generated by the central nervous system based on the task and external force
fields. One is related to the “focal” movement, while the other provides postural
stability. They target different subsets of joints.
Figure 25.11 A motor command is generated by the central nervous system
based on the task and external force fields. It is later transformed to commands to
individual joints. Some of them may have an apparently postural function, while
others may be “focal.” This classification, however, is made by the researchers, not
by the brain.

CHAPTER 25 IN A NUTSHELL
Human vertical posture is inherently
unstable. Quiet stance is associated
with postural sway. Sway depends
strongly on sensory signals; it
increases in the absence of vision and
decreases when a person touches a
stable point in space. Voluntary limb
movements generate postural
perturbations because of joint
coupling and changes in body geometry.
Postural stabilization is helped by a
number of mechanisms, including APAs,
peripheral elasticity of muscles and
tendons, muscle reflexes,
preprogrammed postural corrections,
and voluntary corrections. Postural
synergy is a relation among elemental
variables, such as joint rotations or
recruitment of muscle groups, with the
purpose to stabilize salient
performance variables such as body
orientation in space. The vestibular
apparatus plays a major role in
postural equilibrium; it consists of
peripheral receptors sensitive to
acceleration and vestibular nuclei.
Other sensory systems, such as vision
and proprioception, also contribute to
postural control. Changes in visual or
proprioceptive information can lead to
postural disturbances.
 Chapter 26

Locomotion

KEY TERMS AND TOPICS

motor programming approach to locomotion
central pattern generator
spinal locomotion
gaits
corrective stumbling reaction
dynamic systems approach
step initiation

Locomotion is probably the most common everyday activity of

higher animals, including humans. It is defined as a motor action
during which the location of the whole body in the environment
changes. Lower animals do not locomote, and as a result their
activity is limited to a small area in close proximity to the body.
Locomotion is a great invention of the evolution that expanded the
horizons of our remote ancestors, allowing them to use qualitatively
new strategies to search for food or to escape potential dangers. We
might say that the emergence of locomotion during evolution led to
the emergence of a whole new class of motor problems that animals
confronted in everyday life, leading to the emergence and
development of new systems of neural control involving nearly all of
the central nervous system in contemporary higher animals. If it were
not for locomotion, humans would probably have continued to wait
for a piece of food to emerge by pure chance close enough to their
tentacles to be captured and placed into their mouths, or trembled
with fear as they witnessed a larger tentacle slowly approaching their
helpless bodies.

26.1 Two Approaches to

Locomotion
There are many types of locomotion, such as crawling, flying,
swimming, hopping, walking, and running. In this chapter, we
consider mostly walking and running as the two most common
modes of locomotion used by humans. An analysis of studies on the
possible organization of the central neural structures that control
locomotion clearly reveals two highly influential, competing
philosophical approaches to the control of voluntary movements. The
first approach is related to the concept of motor programming (see
chapter 21) and is based, in the case of locomotion, on the notion of
a central pattern generator (CPG). A CPG is a hypothetical neural
structure that explicitly generates a rhythmic neural activity reflected
in rhythmic peripheral motor patterns. Further, this activity is
transformed into a rhythmic muscle activity leading to a rhythmic
behavior, such as locomotion. Note that rhythmicity is probably one
of the most common and basic features of locomotor behaviors.

PROBLEM 26.1
Some animals can participate in different kinds of locomotion,
such as swimming, walking, hopping, crawling, and flying. How
many CPGs do these creatures have?

The competing approach considers rhythmicity as an emergent

feature of a possibly nonrhythmic neural activity and an interaction of
the peripheral apparatus (including its connections with the central
nervous system) with the environment. This approach has been
termed the dynamic systems approach. We are going to argue that
both approaches have elements of truth but not the whole truth, so
they are not really in competition but can naturally be reconciled.

26.2 Central Pattern Generator

In the beginning of the 20th century, two views on the neural
organization of locomotion were in competition. One of the views
originated from the seminal works of Sir Charles Sherrington (1910)
on muscle reflexes. In particular, Sherrington studied the flexor reflex
and the crossed extensor reflex (see chapter 18). The apparently
reciprocal, and potentially alternating, pattern of the reflex
movements in the two hindlimbs resembled a building block for
walking. Sherrington concluded that locomotion was a motor pattern
produced by alternating reflex responses: When a paw touches the
ground and accepts the weight of the animal, the contact produces
paw stimulation, leading to the flexor reflex in that extremity and the
crossed extensor reflex in the contralateral one. The reflexes
produce movement, leading to contact with the ground of the
contralateral extremity, which accepts the weight of the animal and
induces the flexor response in that extremity and the crossed
extensor reflex in the first one. And so on. Further, Sherrington
generalized this conclusion by suggesting a hypothesis that any
voluntary movement was a result of modulation of parameters of a
few basic muscle reflexes. These ideas formed the foundation of the
equilibrium-point hypothesis (Feldman 1966, 1986) described in
chapters 20 and 21.
The alternative view on locomotion was developed by a student of
Sherrington, Graham Brown (1914). He suggested that the rhythmic
motor pattern during locomotion was produced by a specialized
neural network, a CPG, which could produce rhythmic activity even
in the absence of muscle reflexes. Graham Brown proved that
locomotion was possible after limb deafferentation—that is, in the
absence of reflexes or other feedback effects originating from the
limb’s proprioceptors.
The notion of CPG originally represented rather simple schemes
consisting of sets of neural connections that could, by themselves,
generate rhythmic activity (a very simple example is presented in
figure 26.1). In those simple schemes, a CPG is supposed to involve
three types of neural cells. Neurons of the first type (N1) provide
output for the executive apparatus (for example, agonist and
antagonist motoneurons of muscles involved in locomotion). These
neurons also excite neurons of the second type (N2), which inhibit
N1 neurons controlling the antagonist muscle groups. It is assumed
that N1 neurons fatigue quickly and turn off after a brief period of
high activity. Imagine that one group of N1 neurons become very
active and produce agonist muscle activation. They will inhibit the
antagonist N1 group via N2. After a period of time, N1 of the first
group will fatigue and turn off, and antagonist N1 will be released
from the inhibition and become active. Then they will fatigue. This
cycle will continue until an external influence turns both N1 groups
off. The input from the brain (neurons N3) provides tonic excitation
input into both agonist and antagonist N1. This input can also modify
relations among the N1 and N2 groups (not shown in the figure),
thus controlling parameters of the behavior produced by this simple
system.
Figure 26.1 The “two half-center” CPG. Neurons N1 receive excitation from
higher neural centers (“brain”) and project on α-motoneurons. Neurons N2 mediate
reciprocal inhibition.

PROBLEM 26.2
Imagine that the input from the brain in figure 26.1 is completely
absent. Which output will the network produce?

The simple scheme illustrated in figure 26.1 is usually referred to

as the “two half-center model.” More complex schemes allow CPGs
to generate different locomotion patterns and even produce discrete
actions (such as leaps) in addition to more traditional rhythmical
actions. Neuronal networks of CPGs have been deciphered for
relatively lower animals only—for example, for the lamprey (Orlovsky
et al. 1999). The actual neuronal networks of CPGs in higher
animals, including humans, are unknown.
The position of a CPG within the whole movement-producing
system is illustrated in figure 26.2. It is supposed that the CPG is
under the control of a smart “higher center” that makes a decision on
where to locomote, when, and how. The CPG also receives inputs
from peripheral sensors (in particular, visual receptors, vestibular
receptors, and proprioceptors) and possibly other structures within
the central nervous system. In particular, afferent inputs into a CPG
may bring about changes in the pattern of its activity, leading, for
example, to changes in gait (from walking to trotting to galloping).
Such changes can also be induced voluntarily (i.e., by changing the
input from the “higher center” shown in figure 26.2).

Figure 26.2 A CPG can be driven both by descending signals from “higher
centers” and by afferent information. Ultimately, the CPG leads to changes in the
patterns of activation of α-motoneuronal pools.

Note that the notion of a CPG may be applicable not only to

locomotion but also to other types of rhythmic activity such as
breathing, scratching, and mastication. Experimental support for the
idea of a CPG came from studies of locomotor-like rhythms (and
rhythms resembling other activities) generated by the central
nervous system of animals whose movements were suppressed by
agents such as curare (a poison suppressing neuromuscular
transmission that was originally used by hunters in Central and
South America), as well as studies of rhythms generated by isolated
preparations (reviewed in Grillner and Wallen 1985; Orlovsky et al.
1999). In those studies, parts of the central nervous system have
been proven to be able to generate rhythmic changes in neural
activity.

26.3 Locomotor Centers

A groundbreaking series of experiments were performed in the late
1960s by the group of Mark Shik, Grigory Orlovsky, and Fedor
Severin in Moscow, Russia (Shik et al. 1967; Shik and Orlovsky
1976). These researchers applied electrical stimulation to the
reticular formation of the midbrain (mesencephalon) of decerebrate
cats with a constant frequency and amplitude. Within the reticular
formation they found certain areas whose stimulation led to the
emergence of rhythmic locomotor-like movements of the cat’s limbs.
The frequency of the locomotion was not explicitly related to the
frequency of the stimulation (which was about 30 Hz), so a CPG in
the spinal cord was assumed to be activated by the descending
signals induced by the stimulation. An increase in the amplitude of
the stimulation could lead to speeding-up of the locomotor
movement and, at a certain level, to a change in the gait—for
example, from walking to trotting. Changing the location of the
stimulating electrode allowed these researchers to delineate an area
where stimulation could induce locomotion (the mesencephalic
locomotor region, figure 26.3).
Later, similar studies involving the stimulation of structures within
the upper segments of the spinal cord allowed researchers to track
down the locomotor-related area to cervical spinal segments,
revealing the locomotor strip (Selionov and Shik 1984; Kazennikov
and Shik 1988; figure 26.3). To test whether rhythmical feedback
from peripheral proprioceptors was crucial for maintenance of the
rhythmical pattern, neuromuscular transmission was blocked by
curare (Shik et al. 1966). In those animal preparations, stimulation of
the mesencephalic locomotor region led to rhythmical locomotor-like
patterns of activity in muscle nerves (recorded in the ventral roots of
the spinal cord at an appropriate level) in the absence of peripheral
muscular activity.
Figure 26.3 An illustration of the mesencephalic locomotor area and the
locomotor strip.

26.4 Spinal Locomotion

Locomotion in spinal preparations (i.e., in animals with the spinal
cord surgically separated from the more rostral structures of the
central nervous system) has been known since the classic studies by
Graham Brown and Sherrington in the beginning of the 20th century.
If the spinal cord of an animal is cut acutely, a locomotor pattern is
typically observed for a few seconds—for example, a chicken whose
head has just been cut off runs and even tries to fly. These
observations are commonly interpreted as a release of the activity of
spinal CPGs from the tonic descending inhibitory influence.
A chronic spinal animal does not display locomotion in the
absence of external stimuli or special influences. However, if the
limbs of a spinal cat are placed on a treadmill, movement of the
treadmill at a constant speed at first leads to the limbs being dragged
by the friction force, and then locomotor-like stepping of the limbs
can be observed. Changing the speed of the treadmill leads to
associated changes in the stepping frequency and to shifts in the
gait from walking to trotting, and then to galloping. Stepping cycles
can also be observed in response to certain neuroactive substances,
such as L-DOPA, which is better known as a drug often used to treat
Parkinson’s disease (see chapter 37). On the other hand, locomotor
movements of the hindlimbs can also be observed in a spinalized
animal with all the dorsal roots of the spinal cord innervating the
hindlimbs cut (i.e., without any afferent inflow) (Goldberger 1977;
Atsuta et al. 1991). Such movements are, however, grossly impaired
(Shik et al. 1966; Grillner and Zangger 1975). These observations
prove that the spinal cord is capable of producing locomotor-like
rhythmical activity even without any sensory feedback. Further
experiments, particularly by the Swedish research group led by Sten
Grillner (reviewed in Grillner 1975; Grillner and Wallen 1985), have
demonstrated that individual CPGs exist for each limb. During
normal locomotion, all the individual limb CPGs are coordinated so
as to produce a coherent interlimb pattern.
It can be concluded from the described data that CPGs for
locomotion exist in the spinal cords of mammals and can produce
coordinated locomotor-like activity in response to both descending
stimulation and peripheral input. This activity is insufficient by itself to
produce meaningful locomotion for a number of reasons. First, the
animal needs to know where to locomote (i.e., it needs signals from
visual or other receptors carrying information about the environment,
attractive or dangerous objects, etc.). Second, locomotion is always
intimately tied to control of posture in the field of gravity (chapter 25).
In all the experiments with spinal locomotion, the animals were
suspended with a system of belts. As a result, they did not need to
support their own weight or worry about losing balance. Without the
supporting belt system, spinal animals collapse. Third, normal
locomotion is always associated with perturbations (e.g., stepping on
an uneven area of the surface) that may require urgent corrections.
These corrections will be described in one of the next sections.

PROBLEM 26.3
Are there spinal CPGs for vertical posture? Give reasons for your
answer.

26.5 Spinal Control of

Locomotion in Humans
While the notion of spinal CPGs for locomotion in quadrupeds, such
as cats and dogs, had been accepted for a long time, researchers
were not sure whether similar spinal CPGs existed in higher apes
and humans. Attempts to induce locomotion after a complete spinal
cord transection using the same means as in experiments on cats
and dogs were unsuccessful until relatively recently. It was only at
the end of the 20th century that a few studies described locomotor
patterns in apes and humans following complete spinal cord injury
(Vilensky et al. 1992; Calancie et al. 1994; Shapkov et al. 1995).
Newborn babies show alternating leg movements that resemble
stepping in response to mechanical stimulation of the soles of the
feet (Cooke and Thelen 1987; Thelen and Cooke 1987) or when they
are supported in the air and the feet touch a moving treadmill
(Dominici et al. 2011). These early primitive locomotor-like
movements disappear by the third month of life and then reemerge
in the form of crawling. The emergence of independent walking has
been associated with complete myelinization of the neural tracts and
maturation of brain structures (Thelen 1986, 1995; Sutherland et al.
1988).
The existence of a spinal locomotor CPG at a lower thoracic–
upper lumbar level in humans has been suggested based on
observations in patients with spinal cord injury (reviewed in
Shapkova 2004). In certain conditions, these patients can
demonstrate involuntary stepping-like leg movements while they are
unable to produce such movements voluntarily. In another series of
observations, stepping movements were induced in persons with
spinal cord injury by electrical stimulation applied to the spinal cord
at the lower thoracic–upper lumbar level. The stimulation induced
alternating activity in the muscles of the two legs with the movement
frequency different from the frequency of the stimulation. In all those
studies, the patients were supine, and the legs were suspended in
the air with a system of belts.
Figure 26.4 illustrates such a movement pattern. The large vertical
lines seen across the EMG channels reflect the electrical stimuli
applied over the spinal cord. There are clear alternating bursts of
muscle activity and cyclic changes in the joint angles. This means
that the stimulation provided a nonspecific input into the neural
structures, which then generated the rhythmic pattern. Changing the
strength and the frequency of the stimulation could change the
pattern of the induced movement in a way similar to the effects
observed by Shik, Orlovsky, and Severin in their cat experiments
described earlier. Sometimes the stimulation induced rhythmic
movement in only one leg, or even in only one of the major leg joints.
These observations suggest that the CPG for locomotion can be
based on a hierarchy of CPGs for each extremity and maybe even
for each joint.
Figure 26.4 An illustration of alternating muscle activity induced by electrical
stimulation of the spinal cord in a patient with no voluntary movements in the lower
extremities. Note the cyclic motion of the legs at a frequency different from the
frequency of the stimulation.
Reprinted by permission from E. Yu Shapkova, “Spinal Locomotor Capability Revealed by
Electrical Stimulation of the Lumbar Enlargement in Paraplegic Patients” in Progress in
Motor Control, 3rd ed., edited by M.L. Latash (Champaign, IL: Human Kinetics, 2004), 53-
290.

26.6 Gait Patterns

If the paws of a spinal cat are placed on a moving treadmill, its
hindlimbs demonstrate a stable phase relation during induced
locomotion. At low speeds of the treadmill, the hindlimbs will be out
of phase with each other. As the speed of the treadmill increases,
the hindlimbs will preserve the phase relation up to a certain speed,
and then the relations among individual limb kinematics will change
abruptly as the animal switches from walking to trotting, and from
trotting to galloping (figure 26.5).
Figure 26.5 An illustration of the three major gaits in a quadrupedal animal.

PROBLEM 26.4
What are, from your point of view, the most important limiting
factors for the maximal speed of locomotion?

These and some other observations suggested the application of

ideas from the area of dynamic systems to processes controlling
locomotion. According to this approach, oscillatory behaviors can
demonstrate changes in the stability properties of their kinematics
and phase transitions in response to changes in one or more input
variables. In particular, a change in the descending signals and in
the peripheral information (as in the experiments by Shik and his
colleagues) may lead to the emergence of a new, stable solution for
the system, leading to a new gait pattern.

26.7 Dynamic Pattern Generation

The alternative approach to locomotion (as well as to the generation
of other movements) has been pioneered by Scott Kelso, Peter
Kugler, Michael Turvey, and Gregor Schöner (Kugler and Turvey
1987; Kelso and Schöner 1988; Schöner and Kelso 1988) and is
termed the dynamic systems approach or dynamic pattern
generation. According to this approach, the system for movement
production—including the central neural structures, the effectors and
their connections with the central structures, and environmental
forces and sources of sensory information—can be modeled with
rather complex, nonlinear differential equations. The term nonlinear
means that the response of a system described by this equation to
an input signal may change disproportionally to changes in the input
signal. Such equations cannot typically be solved analytically. When
applied to motor systems, these equations can describe rather
complex behaviors including, in particular, oscillations and changes
in relative coordination. Note that oscillations are typical features of
locomotor movements, while changes in relative limb coordination
describe changes in gaits.
This approach has shown impressive success in its descriptions
of certain features of motor coordination, including interlimb and
interjoint coordination. There are two major views on this approach.
The first one accepts it as the only correct view on voluntary
movement production in humans, the view that ties together events
of the inanimate world and biological phenomena into a single
coherent scheme. The alternative view considers this approach to be
another example of mathematical modeling—that is, as an example
of trying to address biological problems with tools that have been
developed for other areas of science. It is no surprise that a complex
equation can model complex behavior better than a simple equation.
A major question is whether the equation has biological relevance,
whether its parameters can be assigned physiological meaning. Until
now, parameters of equations used within the dynamic systems
approach have been selected rather arbitrarily in order to make sure
that the model produces the desired coordination patterns.
Let us use figure 26.6 to illustrate the major difference between
the motor programming (or CPG) approach and the dynamic
systems approach (see Turvey and Carello 1996). Figure 26.6a
illustrates control of locomotor movements (swimming) of a fish from
the motor programming (or CPG) view. The “fish homunculus”
controls all the details of fish movement patterns, as with a
marionette. Figure 26.6b illustrates the same fish without any
supreme homunculus but with numerous links connecting its
elements (connections with external variables are also implied; note
the open eyes of the fish!), which is in line with the dynamic systems
view. These links are supposed to give rise to the equations
mentioned, potentially leading to complex behavioral patterns.
Figure 26.6 An illustration of (a) the motor programming approach, (b) the
dynamic pattern generation approach, and (c) a combination of the two. Note that
part a does not involve coordination, while part b lacks control.

Note that figure 26.6a lacks coordination or, more precisely, all the
details of coordination are delegated to the ultimate controller; in
other words, they are assumed to be preplanned by the supreme
omnipotent homunculus. This is not very attractive for a number of
reasons, in particular because, as we know, similar coordination
patterns can be observed in spinal animals that apparently lack any
homunculus. Besides that, assigning all the details of coordination to
a smart “black box” does not solve the problem but rather
emphasizes our inability to deal with it.
Figure 26.6b illustrates a much more appealing approach to
coordination, which can emerge without any supreme problem
solver, but it lacks the element of control. This fish will never be able
to change its behavior based on its own will, only in response to
signals from the environment. Remember that in chapter 21 we
discussed the differences between the physiology of initiative and
the physiology of reflex-type movements and came to the conclusion
that behavior cannot be based exclusively on reactions to external
stimuli. Thus, this image is also unsatisfactory.
Figure 26.6c shows a “hybrid fish” that retains all the coordinative
links among its elements but also has an independent descending
signal, generated by its upper neural structures, that can be used for
movement initiation or modification even if the environment does not
dictate it. This descending signal represents one of the important
parameters in the equations that describe the coordination.
Bernstein’s principle of initiative states that this input cannot be
reduced to reactions to external stimuli. At this time, its nature
remains mysterious. This may not sound very scientific, but
unfortunately, there seem to be no viable alternatives to the scheme
shown in figure 26.6c. Coordination and control can and must
coexist to allow both active central generation of meaningful
movements and adjustment of coordination to changes within the
body and the environment.

PROBLEM 26.5
In classic experiments by Scott Kelso, when the subjects tapped a
rhythm with two index fingers, an increase in the tapping
frequency could lead to an involuntary switch from an out-of-
phase regime into an in-phase regime. Try to interpret these
observations based on the three approaches illustrated by the
three parts of figure 26.6.

Within the idea of control with spatial referent coordinates (RCs,

described in chapter 24 and reviewed in Feldman 2015), locomotion
begins with specifying time-varying RCs for the whole body in the
external space (RCBODY). Then a sequence of few-to-many
(abundant; see chapter 22) transformations leads to the emergence
of RCs at the level of individual limbs, joints, and muscles (figure
26.7). These signals produce muscle activations, which also depend
on the time-varying actual muscle length. Within this scheme, CPGs
may be viewed as mechanisms providing for the transformations
from RCBODY to lower-level RCs. This scheme remains hypothetical
because of the lack of reliable tools to record RC(t) patterns in real
time.
Figure 26.7 The CPGs are positioned between the task-related upper level of
the control hierarchy (RCBODY) and the level generating commands to muscles
(RCMUSCLES).

26.8 Step Initiation

Obviously, in order to walk, one has to make the first step. The issue
of step initiation is on the border of postural control and locomotion.
Prior to stepping, a person maintains vertical posture. To produce a
step, at least two conditions should be met. First, the stepping foot
must be unloaded. Second, the body has to start moving in the
direction of the planned step; commonly this means moving forward.
The first condition can be satisfied by shifting the weight of the body
to the supporting foot. The second condition requires changes in the
reactive forces and moments acting on the body.
A typical pattern of the center of pressure (COP) shifts during
preparation to stepping is illustrated in figure 26.8 (see also Elble et
al. 1994; Lepers and Brenière 1995). COP shifts start more than 0.5
s before the stepping leg loses contact with the ground. In the
mediolateral direction, the COP shifts toward the stepping foot and
then reverses and shifts toward the supporting foot. At the same
time, the COP also shifts backwards. This latter COP shift produces
a moment of the reactive force that tends to rotate the body forward.
These postural adjustments in preparation for stepping may be
viewed as obligatory, dictated by the aforementioned mechanical
requirements of step initiation. Sometimes they are addressed as
anticipatory postural adjustments (APAs, reviewed in Massion 1992;
see chapter 25) to stepping. However, the relatively long duration of
these adjustments does not allow them to be considered to be
results of purely feedforward control processes. In this respect, they
differ from APAs described in the previous chapter. In addition, the
functions of APAs differ from those of postural adjustments to
stepping. The former try to counteract expected postural
perturbations, while the latter adjust the body to facilitate future
planned action. It may be more appropriate to address those
adjustments with a different term, such as early postural adjustments
(Klous et al. 2011; Krishnan et al. 2011).
Figure 26.8 Typical shifts of the COP in preparation for making a step. The COP
shifts backwards and toward the supporting foot (after a transient deviation toward
the stepping foot).
Reprinted by permission from Y. Wang, V.M. Zatsiorsky, and M.L. Latash, “Muscle
Synergies Involved in Shifting Center of Pressure During Making a First Step,”
Expermimental Brain Research 167 (2005):196-210. © 2005 Springer.

PROBLEM 26.6
Imagine that a person stands on a board placed on a very narrow
support that does not allow the COP to shift in the anterior–
posterior direction. How can a person initiate a step in such
conditions?
26.9 Corrective Stumbling
Reaction
A particular pattern of automatic, reflex-like responses was observed
during cat locomotion associated with overcoming an unexpected
obstacle, the corrective stumbling reaction (see chapter 19). This
pattern could be observed during a weak mechanical stimulation of
skin areas of the paw or of the leg (even with an air puff) or during
short episodes of electrical stimulation of skin nerves or dermatomes
(Duysens and Pearson 1976; Forssberg et al. 1975, 1977). Similar
patterns have also been seen in humans (Lisin et al. 1973).
The application of any of these stimuli during the swing phase
gave rise to a flexor reaction with the hindlimb transfer over a
hypothetical obstacle (figure 26.9b). The same stimulation applied
during the stance phase could give rise to an extensor reaction
(figure 26.9a). The latency of these reactions was higher than the
latency of monosynaptic reflexes and lower than the voluntary
reaction time. The functional appropriateness of these reactions and
the relative independence of the stimulus suggest that they are in
fact preprogrammed responses of a mechanism responsible in
everyday life for the compensatory reactions during stumbling (see
chapter 19).
Figure 26.9 The corrective stumbling reaction involves a coordinated adjustment
of the walking pattern in response to an external stimulus (mechanical or electrical)
applied to one of the feet (or paws). It involves different changes in the muscle
activation patterns and different kinematic consequences, depending on the phase
of the step cycle when the stimulus is delivered.

PROBLEM 26.7
When you suddenly step on a nail while walking, preprogrammed
reactions are generated (e.g., jumping). What kind of reaction
would you expect to see in a person stepping on a nail while
crossing an abyss on a narrow plank?

One can hypothesize that the execution of any functionally

important motor task is associated with preprogramming of certain
corrective motor patterns, depending on the task and on possible
perturbations that can occur during task execution. These
preprogrammed patterns provide for the very quick initiation of
compensatory responses to the perturbation. Because these motor
reactions are prepared by the central nervous system prior to an
actual perturbation, they always lead to rather crude approximate
corrections that can be further corrected with a voluntary action.
Locomotion is one of the most commonly used movements in
everyday animal and human activity. As such, the mechanism of
locomotor movements is well protected with a set of preprogrammed
corrections that can be triggered by appropriate proprioceptive
stimuli.

PROBLEM 26.8
Give a couple of examples of other tasks that are associated with
a system of preprogrammed corrections.

CHAPTER 26 IN A NUTSHELL
Locomotion is an activity during which
the location of the whole body of the
animal in the environment changes.
Electrical stimulation of an area in
the reticular formation of the
midbrain and also of areas in the
cervical spinal cord can induce
locomotion and lead to changes in
gait. Peripheral stimulation of the
legs as well as certain drugs can lead
to gait initiation and changes in a
spinal animal preparation. Locomotion
can be viewed as a result of activity
of a CPG that is under control of a
higher executive structure and can
change its activity in response to
changes in peripheral information.
Alternatively, locomotion can be
viewed as an emerging pattern within a
complex system that involves central
neural structures, peripheral organs,
and interactions with the environment.
There is evidence for the existence of
spinal CPGs for locomotion in humans.
To initiate a first step, the COP
shifts show a characteristic pattern
that unloads the stepping leg and
creates a moment of reactive force
that rotates the body forward. Reflex
responses of limb muscles to
mechanical stimulation of a paw or to
an electrical stimulation of an
afferent nerve in a limb depend on the
phase of locomotion. This pattern is
known as the corrective stumbling
reaction.
 Chapter 27

Prehension

KEY TERMS AND TOPICS

hand muscles
muscle compartments
cortical representations
finger interaction
multidigit synergies
grasping
principle of superposition

The human hand is uniquely designed to perform a wide variety of

motor actions. Despite the impressive recent progress in
engineering, the dexterity and versatility of the human hand are
unmatched by artificial manipulators. Compared to the human hand,
robotic hands look clumsy and inept.
The motor function of the hand is inseparable from its sensory
function. The hands are used both to explore objects by touch,
including dynamic touch (i.e., welding handheld objects), and to
manipulate them by applying adequate forces and moments of
forces. This combined sensory–motor function of the hand is known
as prehension.
Both the peripheral design of the hand and its central neural
control are crucial for the hand’s ability to manipulate objects. In the
next few sections, we will discuss the specific features of the
peripheral and neural factors that form the basis for dexterous
manipulation. We will also consider the hand as a particular example
of a redundant system that is controlled using synergies (see chapter
22).

27.1 Hand Joints and Muscles

The hand is a complex anatomical structure with rich kinematic
properties. There are 27 bones in the human hand, wrist, and
forearm; 14 phalangeal bones in fingers, five metacarpal bones in
the palm area of the hand, and eight carpal bones in the wrist. These
bones build four major joint groups in the hand: distal
interphalangeal joints, proximal interphalangeal joints,
metacarpophalangeal joints, and carpometacarpal joints.
During typical accurate manipulations, humans apply forces to
handheld objects with fingertips. From the point of view of
mechanics, the design of the hand combines both serial and parallel
chains (figure 27.1). Individual digits can be viewed as serial
mechanisms, with several joints linking the fingertip to the wrist.
Such mechanisms are redundant in kinematic tasks that typically
have fewer task constraints than the number of the joints. However,
in force-producing isometric tasks, serial mechanisms are over-
constrained because a vector of force (including moment of force!) at
the endpoint defines unambiguously all the joint torques (Zatsiorsky
2002). In contrast, when several fingers grasp a rigid object, they
represent a parallel mechanism that is over-constrained in
kinematics, because movement of one finger requires movement of
all other fingers to keep contact with the object. However, the parallel
mechanism is redundant in statics because an infinite number of
combinations of finger forces can produce a required total force (Li et
al. 1998; Burstedt et al. 1999).
Figure 27.1 The design of the human hand includes both serial and parallel
chains. R, S, and U stand for revolute, spherical, and universal joints respectively.
Reprinted by permission from J.K. Shim, Rotational Equilibrium Control in Multi-Digit Human
Prehension (Ph.D. Dissertation, 2005). © 2005, J.K. Skim.

The muscular apparatus of the hand is rather complex. It involves

intrinsic muscles with the bellies lying within the hand, and
extrinsic muscles with the bellies outside the hand, in the forearm
(figure 27.2). The intrinsic flexor muscles (INT) are digit-specific in
their apparent action—that is, their distal tendons attach to proximal
phalanges of one finger only. However, they also attach to a
tendinous structure that forms the extensor mechanism, a network
of passive elastic tissues that produce extending action in the distal
finger joints. As a result, when a person tries to press against a stop
with the proximal phalanx of a finger, intrinsic muscles produce the
required focal action and also contribute to extension of all the
fingers in distal phalanges.
Figure 27.2 There are intrinsic and extrinsic hand muscles. The tendons of the
extrinsic flexor muscles (FDP and FDS) attach at the distal and intermediate
phalanges respectively, while the tendons of the INT attach at the proximal
phalanges (they also contribute to the extensor mechanism not shown in the
drawing).

Each of the extrinsic hand muscles has four distal tendons that
attach to individual fingers. In particular, the flexor digitorum
superficialis (FDS) sends four tendons to the four fingers of the
hand. These tendons attach at the intermediate phalanges (figure
27.2). The other major flexor, flexor digitorum profundis (FDP), has
four tendons that attach to the distal phalanges of the four fingers.
This design of the muscular hand apparatus may be viewed as
contributing to the interdependence of forces and movements
produced by individual fingers. For example, if you try to flex the ring
finger, other fingers will also show some flexion movement. However,
this interdependence is not absolutely obligatory; professional
musicians can learn to override it and produce much more
independent finger actions (Slobounov et al. 2002).
Extrinsic hand muscles are commonly viewed not as single
physiological structures but as combinations of several muscle
compartments (Jeneson et al. 1990; Serlin and Schieber 1993).
The idea of compartments implies the existence of subgroups of
motor units that produce forces transmitted primarily to one of the
four tendons and lead to motor effects in one of the four fingers. This
idea allows for both independent and combined finger action. It has
received support in a study that showed that force response by a
finger to a single transcranial magnetic stimulus (TMS) applied over
the primary motor area of the cortex depends strongly on the
background force produced by this particular finger and shows only
weak dependence on the forces produced by other fingers of the
hand (Danion et al. 2003a).

PROBLEM 27.1
How will the force response of a finger to a single TMS pulse
applied over the primary motor cortex depend on the background
force produced by that finger?

27.2 Cortical Representations

of the Hand
The human hand has a highly developed system of cortical control.
The primary motor cortex (M1) and the corticospinal tract are crucial
for the proper functioning of the hand. Individual neurons in the
corticospinal tract produce both excitatory and inhibitory effects on α-
motoneurons of hand muscles on the other side of the body. The
excitatory effects can be direct, resulting from a single excitatory
synapse of a corticospinal fiber on the target α-motoneuron. The
inhibitory effects are always mediated by at least one interneuron.
A lesion of the primary motor cortex or of the corticospinal tract
leads to dramatic weakening of hand muscles and loss of dexterity.
In particular, humans become less able to move one finger at a time;
rather, an attempt to move a finger leads to motion of all the fingers
of the hand. In nonhuman primates, the effects of lesions of these
neural structures are less dramatic, and the recovery is quicker and
more complete.
Since the classic works by Penfield (see chapter 8), both motor
and sensory cortical areas have been viewed as maps of body parts
with disproportionally large representations of the hands and digits.
Later, however, the development of new, more accurate methods of
assessment of cortical representations has resulted in much less
structured, mosaic pictures with no clear borders between
anatomical areas (Schieber 2001; Schieber and Santello 2004).
Moreover, these representations have been shown to change under
a variety of conditions, including injury and practice (Merzenich et al.
1984; Cohen et al. 1991a,b; Recanzone et al. 1992; Classen et al.
1998).
Two phenomena characterize motor projections from cortical
areas to the segmental spinal apparatus and muscles, and sensory
projections from peripheral structures to cortical areas. The
phenomenon of convergence (figure 27.3a) refers to projections
from two different sources that affect the same target neurons. The
phenomenon of divergence refers to projections from a single
neuron (or from a group of neurons) to several targets (figure 27.3b).
In the case of hand representation in the primary motor area, signals
from neurons distributed over a relatively wide cortical territory can
lead to responses in the same hand muscle and movement of the
same digit (convergence), while stimulation of a single cortical
neuron can lead to motor effects in a number of muscles of the arm,
including hand muscles and more proximal muscles that control
elbow and shoulder motion (divergence). This organization suggests
that the activity of cortical neurons is more likely to be related to
multijoint and multidigit actions than to the activation of individual
muscles.
Figure 27.3 (a) The phenomenon of convergence refers to projections from two
or more different sources that affect the same target neurons. (b) The
phenomenon of divergence refers to projections from a single neuron to several
targets.

Studies using brain mapping techniques have shown major

reorganizations of cortical maps following the amputation of a digit
(Merzenich et al. 1984) or specialized training such as reading
Braille (Pascual-Leone et al. 1995; Pascual-Leone 2001). This
phenomenon, referred to as neural plasticity, is likely to mediate
such processes as motor learning and motor rehabilitation. Neural
plasticity has also been demonstrated in experiments when subjects
were trained for a relatively short time period, under 2 h (Classen et
al. 1998; Latash et al. 2003). These observations suggest that the
connections within the central nervous system are continuously in
the process of rewiring themselves.

PROBLEM 27.2
 What kinds of changes in responses of a muscle to a standard
TMS stimulus applied over the primary motor cortex can be
expected after a prolonged practice of this muscle or of its
antagonist?

27.3 Indices of Finger

Interaction
When a person tries to press with a finger against a stop, other
fingers show an increase in force as well (Kilbreath and Gandevia
1994; Li et al. 1998; Zatsiorsky et al. 1998). This phenomenon has
been termed finger interdependence or enslaving. Patterns of
enslaving show both differences and similarities across persons.
Typically, these effects are stronger between pairs of neighboring
fingers, and they show a degree of symmetry—that is, if finger X
enslaves finger Y strongly, finger Y is likely to enslave finger X
strongly. The index finger usually shows the least enslaving effects,
while the ring finger is typically the least independent—that is, when
a person tries to press with the ring finger, other fingers of the hand
show very high enslaved forces. The magnitude of enslaving can
vary within a rather wide range across individuals. It changes with
such factors as specialized practice (in particular, in professional
musicians), age, and motor disorders (Li et al. 2003; Shinohara et al.
2003b; Chiang et al. 2004).
When a person tries to press as strongly as possible with several
fingers of a hand simultaneously (i.e., produce maximal voluntary
contraction force [MVC]), the peak forces that are reached by
individual fingers are smaller than when this person tries to press as
strongly as possible with one finger at a time (Ohtsuki 1981;
Kinoshita et al. 1996; Li et al. 1998). This phenomenon has been
termed force deficit. Force deficit increases with an increase in the
number of fingers explicitly involved in the task. A typical data set for
one-finger and multifinger tasks is shown in table 27.1. These data
illustrate both enslaving (production of force by nontask fingers in
one-finger tasks) and force deficit (smaller forces of fingers in the
four-finger task).

PROBLEM 27.3
What changes of enslaving and force deficit would you expect to
see in the right and left hands of a trained right-handed cellist?

The thumb is a special digit of the human hand. Its muscular

apparatus is different from that of the fingers; in particular, the thumb
does not share muscles with any of the fingers. Its kinematic
repertoire is also quite different from that of the fingers. It allows the
thumb to act in parallel with the fingers (for example, when a person
pushes against a heavy door) and in opposition to them (when
holding a glass with a drink). However, a study of indices of interdigit
interaction has not shown significant differences between digit pairs
that include and that do not include the thumb (Olafsdottir et al.
2005b). The thumb seems to interact with fingers similarly to how
fingers interact with each other. Anthropological evidence suggests
that the long thumb flexor muscle, the flexor pollicis longus, evolved
from being the fifth compartment of the multidigit FDP (Marzke
1992). This might be the origin of some of the similarities in indices
of thumb–finger interaction as compared to indices of finger–finger
interaction.

Table 27.1 A Typical Data Set of Forces Recorded in Trials

with MVC
Task/Fingers I M R L
I 49 10.5 5.5 2.5
M 10 40 13 4
R 9 16.5 30 10.5
L 6 7 15 25
IMRL 33 27 22 16
The data are in newtons. Numbers in bold represent the forces produced by
the master (instructed) fingers, and the others are the forces produced by
slave (noninstructed) fingers. The bottom row shows finger forces in the
four-finger MVC test. Note that each finger shows a force deficit in the
four-finger test.

The phenomena of enslaving and force deficit have been

combined into a single scheme that assumes that the brain sends
hypothetical commands—finger modes—reflecting the desired
(intended) involvement of individual fingers. These commands are
attenuated, with a gain factor reflecting the number of intentionally
involved fingers leading to the phenomenon of force deficit. They are
also multiplied by an enslaving matrix showing how a small change
in the intended command to a finger leads to force changes in all
four fingers of the hand (Danion et al. 2003b):

where F is a four-dimensional finger force vector, n = the number of

explicitly involved fingers, |E| = the enslaving matrix, m = a four-
dimensional mode vector, and T = signs of transpose.
Finger interdependence gets contributions from both the
peripheral design of the hand and the organization of cortical neural
control, in particular from the overlapping cortical representations for
individual fingers and the aforementioned phenomenon of
divergence (reviewed in Schieber and Santello 2004). The
phenomena of finger interdependence are likely to reflect patterns of
finger force covariation that are beneficial for everyday tasks—
multifinger synergies. In particular, patterns of enslaving have been
viewed as minimizing changes in the pronation–supination moment
with changes in pressing finger forces (Zatsiorsky et al. 2000).

27.4 Multifinger Synergies in

Pressing Tasks
Multidigit action is a very attractive object to study motor synergies
because of a few factors. First, these synergies are learned by
humans over their lifetimes and may be expected to show
reproducible behaviors across a variety of tasks. Second, these
synergies are obviously very important for everyday activities. Third,
understanding how such synergies are organized may help in the
development of strategies for the treatment of hand function
disorders. Finally, forces and moments of forces applied by fingertips
during manipulations can be relatively easily recorded and analyzed.
In chapter 22 on motor synergies, we discussed two features of
synergies, sharing and error compensation (ensuring stability with
respect to salient performance variables). Both features have been
studied using multifinger force production tasks. When a person is
asked to press with the four fingers of a hand and produce a slowly
increasing force, individual finger forces scale together such that
each finger produces more or less the same percentage of the total
force at all times (figure 27.4). Typically, the index and middle fingers
produce about 60% of the total force, while the ring and the little
finger produce the remaining 40%. Such stable sharing patterns of
finger forces have been related to the principle of minimization of
secondary moments (Li et al. 1998; Zatsiorsky et al. 1998, 2000).
This principle suggests that finger forces are shared to minimize the
total moment in pronation or supination created by the reactive
forces about the longitudinal axis of the forearm.
Figure 27.4 When a person produces a ramp profile of the total force while
pressing with four fingers on force sensors, individual finger forces scale
proportionally over the whole total force range.
Reprinted by permission from Z.-M. Li, M.L. Latash, and V.M. Zatsiorsky, “Force Sharing
Among Fingers as a Model of the Reduncancy Problem,” Expermimental Brain
Research,119 (1998): 276-286. © 1998 Springer.

It is natural to assume that, when a person is asked to produce a

certain steady-state value or slowly changing pattern of the total
force, there will be multifinger synergies stabilizing the total force–
time profile. The uncontrolled manifold hypothesis (chapter 22)
allows for testing this hypothesis in an experiment. However, to
approach this problem, an important step has to be taken. As
mentioned in the previous section, individual finger forces are not
independent of each other because of the phenomenon of enslaving.
It is possible to observe in experiments patterns of finger force
covariation, which reflect not task-specific control strategies
(synergies) but rather task-independent relations among finger
forces. To avoid this problem, one has to discover another set of
elemental variables that can be manipulated by the central nervous
system one at a time.
 To address this problem, a notion of finger modes has been
introduced (Zatsiorsky et al. 1998; Danion et al. 2003b; figure 27.5).
The idea behind this notion is that the neural controller manipulates
not finger forces but hypothetical neural variables that lead to the
production of forces by all the fingers of the hand (due to enslaving).

Figure 27.5 An illustration of the idea of finger modes. The controller

manipulates independent variables (modes) corresponding to desired involvement
of fingers in a multifinger task. Each mode leads to force production by all four
fingers. The thickness of arrows reflects a typical enslaving pattern. The box “force
deficit” reflects the drop in force when more fingers are involved in the task
explicitly.

Analysis of covariation of finger modes has shown the existence

of two multifinger synergies in accurate total force production tasks
(Latash et al. 2001; Scholz et al. 2002). One of the synergies, as
expected, stabilized the total force. The other synergy stabilized the
total pronation or supination moment produced by the reactive
forces. In those studies, the production of moment was not part of
the task. Hence, a conclusion has been drawn that moment-
stabilizing synergies reflect the everyday experience of humans that
typically places much more stringent constraints on errors in the
moment of force than in the total pressing force. For example, when
taking a sip from a glass, the grasping force only has to be over the
slipping threshold and under the crushing threshold, while the total
moment has to be controlled precisely to avoid spilling the contents
of the glass (figure 27.6).

Figure 27.6 To take a sip from a glass, the grip force (FGRIP) has to be within a
rather wide range, over the slipping level and below the crushing level. The
moment of forces (M) should be controlled precisely to avoid spilling the contents
of the glass.

PROBLEM 27.4
Suggest a natural task in which the production of accurate finger
forces would be more important than the production of accurate
moments of forces.

27.5 Grasping
During object manipulation, the hand can be used to produce
different types of grasps (Schlesinger 1919; Napier 1956), from a
precision grip (holding a small object between the tips of the index
finger and the thumb) to a power grip (using a wrench). Kinematic
analysis of grasping commonly uses the notion of grip aperture, the
distance between the tips of the thumb and the index finger. When a
hand reaches for an object, grip aperture increases up to about 60%
to 70% of the duration of the reaching movement, and then the digits
start to close to land on the object. The maximum grip aperture
depends on the properties of the object to be grasped, including its
size, weight, and texture. Grasping actions strongly depend on
sensory information, including visual information and signals from
somatosensory receptors in the muscles, tendons, joints, and skin.
There is an alternative view on grasping that views it as a
combination of two pointing movements of the opposing digits
directed at certain landing points on the object (Smeets and Brenner
1999).
Brain imaging studies (as well as more direct studies of the brain
activity in monkeys) have suggested the importance of three cortical
areas for grasping: the primary motor cortex, the premotor cortex,
and the parietal cortex. Studies of the neural activity in the ventral
premotor cortex of the macaque monkey have shown the existence
of neurons that discharge both during the execution of hand actions
and during the observation of the same actions made by others:
mirror neurons (Rizzolatti et al. 2001; Rizzolatti and Craighero
2004). Of course, one has to keep in mind that the actions were “the
same” at a very high, abstract, topological level: At the level of
mechanics, such as joint rotations and muscle forces, the actions
could not be the same because of the obvious differences in the
anatomy and physiology of the monkey and the experimenter.
Indirect evidence suggests that such neurons exist in humans as
well; in particular, it has been shown that grasp observation is
accompanied by significant activation of the middle temporal gyrus of
the cortex, including that of the adjacent superior temporal sulcus
(Brodmann’s area 21) and the caudal part of the left inferior frontal
gyrus (Brodmann’s area 45) (Fadiga et al. 2005).

PROBLEM 27.5
 Suggest a functional explanation for the existence of the mirror
neurons.

Grasping is associated with two groups of synergies. First, there

are kinematic synergies that stabilize the grasp aperture. For
example, if a person tries to grip a small object between the index
finger and the thumb, a mechanical perturbation applied to one of
the digits produces quick adjustments in the motion of both digits
such that the time profile of the grip aperture remains relatively
unchanged (Cole and Abbs 1987). Besides, there are dynamic
synergies that stabilize the mechanical action of the hand on the
grasped object. If a person holds an object steadily, there are
fluctuations in the forces applied by individual digits. The fluctuations
of the forces produced by the thumb and the combined force
produced by the opposing fingers tend to be in-phase, while
fluctuations of forces of a pair of fingers acting in parallel are more
likely to be out-of-phase (Santello and Soechting 2000). The former
relation makes sure that the net force applied to the object is small,
while the second relation prevents large variations in the total force
applied by the set of fingers opposing the thumb.
Within the theory of control with spatial referent coordinates (RCs;
see chapter 21), the production of grip force is associated with
setting a value of referent aperture, analogous to RC (AREF in figure
27.7; see also Pilon et al. 2007). Normal force produced by the
fingertips results from the difference between the actual and referent
apertures. If the object suddenly disappears, the opposing digits
would move to AREF (white dashed images in figure 27.7). During
quick object manipulation, changes in RC are seen for all
components of the action, including referent aperture as well as RCs
in external spatial coordinates driving the object movement, and also
in angular coordinates associated with object rotation (Latash et al.
2010).
Figure 27.7 Gripping an object involves specifying referent coordinates for the
opposing digits that are located inside the object. As a result, the referent aperture
(AREF) is smaller than the actual one (AACT).

27.6 Prehension Synergies and

Principle of Superposition
Prehension synergies have been defined as conjoint changes in
forces and moments of forces (sometimes termed elemental
variables) produced by a set of digits on a handheld object that
stabilize the overall mechanical action of the hand on the object
(reviewed in Zatsiorsky and Latash 2004). If an object is grasped
using the prismatic grasp with all five digits of the hand (figure 27.8),
there are constraints of statics that have to be met for the object to
be in an equilibrium. For a relatively simple case of an external load
acting vertically (along the direction of gravity) and an external
moment acting in the plane of the grasp (a vertical plane that
contains all the points of digit contacts with the object), three
equations have to be satisfied:
where the subscripts th, i, m, and r refer to the thumb, index, middle
and ring fingers, respectively; the superscripts n and t stand for the
normal and tangential force components, respectively; L is load
(weight of the object), and coefficients d and r stand for the moment
arms of the normal and tangential force with respect to a preselected
center, respectively.
To simplify the analysis of multifinger grasps, a hierarchical control
scheme has been suggested that includes two levels (figure 27.9;
Arbib et al. 1985; MacKenzie and Iberall 1994). At the upper level,
the total mechanical effect is distributed between the thumb and a
virtual finger—an imagined digit with mechanical action equal to the
combined action of all four fingers. At the lower level, the action of
the virtual finger is distributed among the fingers of the hand. A
number of studies have supported the idea of such two-level
hierarchical control (Shim et al. 2003, 2005).
Figure 27.8 Holding an object with a prismatic grasp involves thumb action
opposing the four fingers.

Figure 27.9 Hierarchical control of the hand involves two levels. At the first level,
mechanical actions of the thumb and the virtual finger (VF) are defined. At the
second level, the mechanical action of the VF is distributed among the actual
fingers.
Reprinted by permission from J.K. Shim, Rotational Equilbrium Control in Multi-Digit Human
Prehension. (Ph.D. Dissertation, 2005). © 2005, J.K. Skim.
 Experimental studies of variations in the individual finger forces
across repetitions of the same task have shown that humans indeed
satisfy the three equations (Shim et al. 2003). However, they do so in
a particular way. This is possible because the number of variables in
the equations is more than three; that is, the system is abundant.
Note that the third equation contains variables that are also parts of
the first two equations. So, one could expect that a change in the
rotational action (equation 27.4) would lead to changes in
components that produce adjustments in grasping action and net
force generation (force production, equations 27.2 and 27.3).
Nevertheless, experiments show that humans organize the
elemental variables into two groups. Within one group, the variables
covary to stabilize the total force acting on the object. The other
group unites variables that covary to stabilize the rotational effect
produced by the hand on the object. So there seem to be two
prehension synergies, force-stabilizing and moment-stabilizing. The
two synergies seem to be relatively independent, so a change in one
of the variables (for example, the total moment) requires adjustment
only in the corresponding synergy without affecting the other (force-
stabilizing) synergy.
These observations fit the principle of superposition suggested
in robotics for the control of artificial grippers (Arimoto et al. 2001).
The principle states that skilled actions can be decomposed into
several elemental actions that are controlled independently by
separate controllers. Such a decoupled control in robotics has been
demonstrated to reduce the total computation time. Human
experiments suggest that the central nervous system may also use
the principle of superposition to organize the control of two elemental
actions: (1) grasping a handheld object with an adequate force and
(2) producing a rotational action on the object.

PROBLEM 27.6
A change in the force of a digit always produces a change in the
moment of force about any axis with respect to which the force
vector has a nonzero lever arm. How can control of the grip force
be decoupled from control of the moment of forces?

CHAPTER 27 IN A NUTSHELL
The human hand combines both parallel
and serial kinematic chains. Digit
action is controlled by extrinsic,
multidigit muscles and intrinsic, more
digit-specific muscles. The extrinsic
muscles are composed of compartments
that show a high degree of
physiological independence. Cortical
hand representations are mosaic
without clear borders between
representations of anatomically close
areas. Finger interactions are
reflected in unintended movements and
force production by fingers
(enslaving) and lower forces produced
by fingers in multifinger tasks as
compared to single-finger tasks (force
deficit). These interactions have a
strong neural component. They show
similar patterns among fingers of a
hand and between the thumb and the
fingers. Multifinger synergies in
pressing tasks commonly show patterns
stabilizing the pronation and
supination moment. These patterns may
be conditioned by everyday experience
that places strict accuracy
constraints on the rotational hand
action. During manipulation of a
handheld object, multidigit synergies
show patterns that stabilize both the
grip force and the total moment of
force applied to the object. The
existence of these two synergies
conforms to the principle of
superposition suggested in robotics.
Problems for Part V
Self-Test Problems
1. A subject is performing a series of very fast elbow flexions
over 40° against a constant external load. Movement time is
200 ms. Unexpectedly, in one trial, the movement is
completely blocked. Draw time patterns of the biceps and
triceps EMGs based on central programming of EMG patterns
and on equilibrium-point control ideas.
2. A person sits with the right forearm, hand, and index finger
vertical and pointing up. The hand is fully supinated. The
person makes a very fast elbow extension movement over
about 90° in response to a visual signal. Movement time is
250 ms. Describe (or draw) EMG time profiles for the following
muscle pairs: elbow flexor-extensor, shoulder flexor-extensor,
and wrist flexor-extensor (assume that all muscles are
uniarticular).
3. A student is standing quietly. Unexpectedly, a friend pushes
them from behind. Describe all the mechanisms that help the
student not to fall down.
4. While walking barefoot, a person steps unexpectedly on a
sharp object. Describe muscle activation patterns that help
this person to avoid falling down. What differences in muscle
activation patterns would you expect in the other leg?
5. High-frequency, low-amplitude muscle vibration applied to the
back of the neck of a person, who is standing with his or her
eyes closed and the arms hanging by the sides, can lead to a
deviation of the body posture from the vertical. Why? In what
direction would you expect the body to move?
6. A person makes a series of quick reaching movements to a
target while holding a 1 kg load in one hand. In one trial, the
load unexpectedly becomes 3 kg. The person is instructed not
to correct actions voluntarily. Describe changes in the
grasping and reaching components of the action and their
typical time delays. Where will the movement terminate?
For Those Addicted to Multiple-Choice Tests
You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. During a very fast voluntary isotonic movement
a. the length of the muscle fibers does not change
b. antagonist burst occurs a few tens of milliseconds prior to
the agonist burst
c. damping properties of muscles and tendons help
accelerate the movement
d. muscle fiber stiffness is higher than in the relaxed state
e. none of the above
Why?
2. You saw a huge bag and expected it to be very heavy. You
lifted it quickly with the right arm, but the bag happened to be
very light (empty), and you nearly lost your balance. What
phenomena caused the problem with the balance?
a. elastic properties of muscles and tendons
b. monosynaptic reflexes from muscle spindles
c. monosynaptic reflexes from Golgi tendon organs
d. anticipatory postural adjustments
e. preprogrammed responses
Why?
3. A person moves a handheld object rhythmically in a vertical
direction at a very high speed (accelerations over 1 g). Which
of the following can be expected?
a. Grasping force will show rhythmic changes at the
frequency of the motion.
b. Grasping force will remain constant.
c. Grasping force will change at twice the motion frequency.
d. Grasping force will be modulated based on proprioceptive
feedback.
 e. All of the above are correct.
Why?
4. When a person performs a series of very fast voluntary
reaching movements
a. joint trajectories are ideally reproduced
b. many muscle pairs show triphasic activation patterns
c. muscle reflexes do not contribute to muscle activation
d. muscle force patterns are ideally reproduced
e. all of the above
Why?
5. The locomotor central pattern generator in an animal with the
spinal cord surgically separated from the brain at a cervical
level
a. helps the animal to maintain the posture needed for
locomotion
b. can be driven by stimulation of certain brain structures
c. can lead to changes in locomotion patterns in response to
sensory information
d. is unable to produce rhythmic patterns of activity
e. all of the above
Why?
 Part VI

Sensorimotor Integration for

Perception and Action
 Chapter 28

Kinesthetic Perception

KEY TERMS AND TOPICS

sensory signals
Weber-Fechner law
efferent copy
referent coordinate
vibration-induced illusions
sense of effort
stability
iso-perceptual manifold
action–perception coupling

Each healthy person is typically aware of the position of the

segments of his or her body in space and in relation to each other
and also of forces at the interface with the environment. This group
of perceptual phenomena is united under the name of kinesthesia.
It allows humans to perform accurate movements without continuous
visual control, to adjust patterns of neural control variables in cases
of movement errors, to perform motor tasks that require multijoint
and multilimb coordination, and so on. This chapter deals with the
basic mechanisms that bring about kinesthetic perception.
28.1 Sensation and Perception
Two terms have been used to describe the neural consequences of
signals produced by peripheral sensory endings. The first one refers
to patterns of activity of those endings, their dependence on salient
physical variables (for example, muscle length and velocity for
primary spindle endings, force for Golgi tendon organs, and
acceleration for the hair cells in the vestibular apparatus),
transmission of those signals within the nervous system, subjective
estimation of their magnitude, and some of their effects on body
functions. These effects are known as the sensory function or
sensation. If sensory signals lead to a change in important aspects
of the intrinsic image of the body location or configuration, or of the
subjective picture of the world, these effects are known as
perception. Somewhat imprecisely, one can also say that sensation
refers to local effects of sensory signals, in close proximity to the
sensory ending, while perception refers to effects at the level of
effectors, up to the whole body, which house many types of sensory
endings.
For example, sensory signals from receptors sensitive to painful
stimuli can wake the person up or produce a withdrawal action from
a hot object before they are consciously perceived, and perception of
pain may come later. The border between sensation and perception
is not 100% clear. For example, muscle reflexes are typically viewed
as being produced by sensory signals without invoking perception.
On the other hand, the concept of action–perception coupling (see
section 28.9) within the field of ecological psychology (Gibson 1979;
Turvey 2007) uses the word perception in expressions such as direct
perception and perception–action coupling to address changes in
action produced by sensory signals, which are not mediated by an
update of the representation of the body in the environment.

PROBLEM 28.1
 Are muscle reflexes, including monosynaptic ones, examples of
perception–action coupling?

One of the problems with studies of sensation and perception is

the lack of tools for their objective measurement. Indeed, since
percepts (unitary phenomena of perception) happen within the body,
produced by unknown processes in unknown neurophysiological
structures, one can only measure their indirect effects on behavior.
Two main groups of methods have been used to quantify percepts.
The first group involves using psychophysical scales: individuals’
subjective reports on a scale that typically ranges from zero to 100
corresponding to lack of a percept and its highest possible
magnitude, respectively. For example, a person can be told to
measure experienced pain on such a scale with 100 corresponding
to the maximal pain imaginable. It is also possible to use scales
expressed in actual physical units, such as degrees (to report joint
angles), meters (to report distances between effectors), and newtons
(to report forces). It is obvious that comparing reports on such scales
across individuals is very complicated. Even within an individual,
subjective scales change with time and experience.
Another method involves reproducing a kinesthetic percept with
another action using the same effector or another effector. For
example, a person can be asked to report the perceived joint angle
by matching it with the contralateral extremity or by reproducing it
with the same extremity after a brief time interval. A variation of this
method involves matching the perceived effector position or
configuration by adjustment of lines on the computer screen. The
method of matching with the contralateral extremity has obvious
drawbacks related to imperfect symmetry of the body and
differences in the control of the dominant and nondominant limbs
(see chapter 8). Reproduction of a perceived variable with the same
effector involves memory, which is also imperfect. As discussed later
in this chapter (section 28.6), the two methods, verbal report on a
psychophysical scale and matching the percept, may involve
different neurophysiological circuits and result in qualitatively
different outcomes.

28.2 Weber-Fechner Law

Studies of the effects of changes in the magnitude of physical stimuli
on their perceived magnitude were performed by two German
psychologists of the 19th century, Ernst Heinrich Weber (1795-1878)
and Gustav Theodor Fechner (1801-1887). In particular, Fechner is
credited with introducing the term psychophysics to address the field
of studies of how humans perceive magnitudes of physical variables.
Weber studied just-noticeable differences in the magnitude of
physical variables (i.e., the smallest change in a variable [ΔV] that
can be reliably perceived). He discovered that ΔV changed
proportionally with the baseline magnitude of the variable (V):

where k is a constant.
For example, you may be able to distinguish a weight of 2 N from
another weight when it differs from the first one by 0.2 N. If the
original weight is now 4 N, you would need a difference of 0.4 N to
be able to make such a distinction. This law works across sensory
modalities (e.g., for brightness of light, loudness of sound, intensity
of pain, etc.). It fails at very low and very high magnitudes of the
physical variables but is valid within a broad midrange of
magnitudes.
Fechner offered a generalization of Weber’s law and stated that
the intensity of our sensation (S) increases as the logarithm of an
increase in the relevant physical variable:

where V0 is some baseline magnitude of the variable. From this

equation, it follows that a change in sensation is proportional to the
relative change in the magnitude of the variable: ΔS = ΔV/V, which
forms the basis for Weber’s equation for just-noticeable differences.
The Weber-Fechner law always fails at low intensities, near and
below the absolute detection threshold, and often also at high
intensities, but it may be approximately true across a wide middle
range of intensities.
The Weber-Fechner law was further developed by Stanley Smith
Stevens (1906-1973), who described an empirical relationship
between an increase in the magnitude of a physical stimulus and the
perceived increase in the sensation magnitude:

where α is a constant, which depends on stimulus modality. This

constant ranges from 0.3-0.5 for brightness, to 1.5-1.7 for muscle
force and weight of objects, and to 3.5 for electric shock. The
Stevens law can be deduced mathematically from the Weber-
Fechner logarithmic function.

28.3 Ambiguity of Sensory

Information
In chapter 6, we considered the properties of some of the peripheral
receptors sensitive to such physical variables as muscle length,
velocity, force, joint angle, pressure on skin, and so on. All these
variables may be used to get kinesthetic information. However, more
careful analysis of the properties of all these receptors shows that
the situation is rather complicated. In particular, the mechanical
design of the human muscles and tendons seems to make the
information coming from receptors of each modality dependent on
factors other than physical variables salient for kinesthetic perception
(table 28.1). Let us consider once again the properties of the most
important proprioceptors; in this chapter, we will view their signals
not as peripheral components of the reflex machinery, but as
sources of kinesthetic information.
Muscle spindles and articular receptors look like perfect
candidates for providing information about the position of segments
of the body, while Golgi tendon organs seem to be perfect
detectors of muscle force. Recall that muscle spindles are small
structures scattered throughout the parent muscle in parallel to the
power-producing (extrafusal) muscle fibers. Muscle spindles contain
two types of sensory endings, primary and secondary spindle
receptors. Primary spindle endings are sensitive to both muscle
length and velocity, while secondary endings are sensitive only to
muscle length.

Table 28.1 Main Proprioceptors and Their Features

Receptor What it measures Complicating factors
Primary Muscle fiber length Muscle fiber length differs from
spindle ending and velocity “muscle + tendon” length.
Secondary Muscle fiber length Sensitivity is modulated by γ-
spindle ending motoneurons.
Golgi tendon Tendon force Joint moment of force (torque)
organ depends on the lever arm. Tendon
force is length- and velocity-
dependent.
Articular Joint angle Poor sensitivity in the midrange of
receptors joint rotation.
Sensitive to joint capsule tension
and joint inflammation.
Cutaneous and Skin displacement Depend on many factors.
subcutaneous and pressure
receptors

Up to this point, things seem to be straightforward. Recall,

however, that muscle spindles have a special system of innervation,
that of γ-motoneurons, which changes the sensitivity of both primary
and secondary endings. A number of experiments have
demonstrated that γ-motoneurons are commonly activated together
with α-motoneurons (α-γ coactivation) in natural actions that lead to
muscle activation and the development of muscle force and
movement. This means that the level of activity of a spindle ending
depends not only upon actual muscle length (and velocity, for
primary endings) but also upon the level of muscle activation. For
example, consider a muscle in isometric conditions (figure 28.1). For
simplicity, let us assume that the average level of firing of its spindle
endings is proportional to muscle length.

Figure 28.1 During isometric force production, signals from γ-motoneurons

change in parallel, thus affecting afferent information from spindle endings. These
changes can potentially be misinterpreted as reflecting muscle length changes.

Now imagine that there is an increase in the level of muscle

activation. Since the conditions are isometric, there will be no joint
movement. However, muscle activation is accompanied by an
increase in the activity of the γ-system. So, we may expect to see an
increase in the activity of spindle endings, which can be interpreted
by the central nervous system as a movement even though no
movement takes place. On the other hand, if the joint is released
(nonisometric conditions) and the muscle is allowed to shorten, the
increase in the level of activity of the γ-motoneurons may lead to the
same average level of activity of spindle receptors at different values
of muscle length, which has been observed during movements in
isotonic conditions (Hulliger et al. 1982). Since humans do not
misjudge joint position at different levels of muscle activation, there
must be a mechanism that allows accurate assessment of joint
configuration in these conditions.
 Another complicating factor is muscle and tendon elasticity. If a
muscle is relaxed, the stiffness of its fibers is smaller than tendon
stiffness. If the muscle is activated, its fibers show considerably
higher stiffness than tendon stiffness. Imagine once again that a
muscle is being activated in isometric conditions (figure 28.2).
Although the total length of the “muscle plus tendon” complex is
unchanged, the length of muscle fibers will decrease during the
activation, while the length of the tendon will increase. In order to
accurately assess the position of a limb segment, the central
nervous system needs to have information on the length of the
“muscle plus tendon” complex rather than on the length of individual
muscle fibers. However, the level of firing of muscle spindle
receptors depends upon the length of muscle fibers and not upon the
length of the tendon.

Figure 28.2 In isometric conditions, muscle fiber length changes (shortens with
force increase) while the “muscle plus tendon” length remains unchanged.

What about articular receptors? In order to get reliable information

about joint angle, it would be desirable to have a receptor whose
level of activity is related to joint angle. At first glance, articular
receptors seem to fulfil this requirement. However, if only the slowly
adapting receptors are considered—that is, those receptors that
maintain their level of activity for a long time when there is no joint
motion—one can notice that they cover mostly the extremes of the
range of joint motion and that they are rather sparse in the middle
(see chapter 6). This would be a good feature to detect when the
joint is approaching the physiological limits of its rotations, but it is
not very helpful if the task is to perceive joint position somewhere in
its midrange. Moreover, articular receptors change their level of firing
when the tension of the articular capsule changes. This happens, in
particular, when muscles acting at the joint change their force. In
isometric conditions, there are likely to be changes in the activity of
articular receptors with a change in the level of muscle activation,
even though no movement occurs due to changes in the tension of
the articular capsule. So, for both articular and spindle receptors, the
central nervous system has to have information about the level of
muscle activation or muscle forces in order to estimate joint position.
Golgi tendon organs are nearly ideal force sensors that are
located in series with muscle fibers. They are not sensitive to muscle
length and do not have central innervation. However, calculating the
joint moment of force from signals from Golgi tendon organs is not
trivial because the relation between muscle force and joint moment
of force changes depending upon joint angle due to changes in the
lever arm (figure 28.3).
Figure 28.3 Signals from Golgi tendon organs reflect muscle force magnitude
(F). Moment of force (M) depends on both force and lever arm (L). In different joint
positions, the lever arm differs. As a result, the same signals from Golgi tendon
organs correspond to different moment of force magnitudes.

PROBLEM 28.2
Does muscle and tendon elasticity affect signals from Golgi
tendon organs? Do these receptors always accurately monitor
tendon force?

So we end up with a seemingly pessimistic conclusion that there

are no peripheral sensory endings that signal functionally significant
posture and movement variables such as joint moments of force
(torques) and angles. Remember, though: If a system within the
human body looks imperfectly designed, it is likely that we have
overlooked or misinterpreted something.
Accurate kinesthetic perception is likely to emerge with the
participation of signals from various sensors, and this design has
more positive features than potential problems. For example, if
articular receptors were the only source of positional information and
were absolutely reliable, inflammation of a joint (unfortunately rather
common) would have caused a major deterioration of kinesthetic
perception. If there is a mixture of kinesthetic information
represented in receptors of different types, the system becomes less
susceptible to a disorder of one of the receptor systems. Actually,
even if a joint is replaced with an artificial one, position sense in that
joint virtually does not suffer (Grigg et al. 1973; Skinner et al. 1984;
Wada et al. 2002).
It is time now to turn to the role of motor command in kinesthetic
perception. This issue has also been described in chapter 14 on the
control of eye movements. Briefly, if you displace your eye by
pressing on it with a finger, you perceive a shift of the environment,
whereas an active eye movement does not lead to a similar false
perception. This observation suggests that efferent signals to
extraocular muscles affect perception, an important insight that we
will consider in the next section.

28.4 Afferent and Efferent

Components of Perception
The idea of the motor process being an important contributor to
perception was formalized using the concept of efferent copy (also
known as efference copy; von Holst and Mittelstaedt, 1950). At about
the same time, a related concept was developed by Sperry (1950)
under the name of corollary discharge. Von Holst and Mittelstaedt
viewed sensory signals from muscles as the sum of two
components, exafference and reafference. Exafference represented
changes in signals along afferent fibers induced by a change in the
external force field. Reafference represented changes in afferent
signals induced by a voluntary movement.
Figure 28.4 (a) In the original scheme, a copy of signals from α-motoneurons to
an effector (“efference copy,” EC) is used to predict changes in afferent signals
from the effector—reafference (ReA). The difference between the predicted and
actual ReA is used to drive perception and to correct signals from the α-
motoneurons. (b) In this image, the muscle is shorter than that in part a. This
means that signals from length-sensitive endings in its spindles should be larger in
the upper drawing. If the muscle is relaxed in both positions, its efferent copy is the
same: zero. This means that the ΔA signal cannot be zero in both drawings,
leading to muscle activation.

According to the original definition, a copy of signals from α-

motoneurons to an effector was used to predict changes in afferent
signals from the effector—reafference (figure 28.4a). If the prediction
was matched by the reafference, no correction to the ongoing motor
process was introduced. In cases of a mismatch between the
reafference and predictions based on the efferent copy, these signals
were used to adjust the ongoing motor process. Note that the
scheme in figure 28.4a assumes that efference copy signals and
afferent signals from sensory endings are somehow converted into
commensurable units—a very nontrivial assumption! Exafference
was always assumed to affect kinesthetic perception, while
reafference affected it only if it deviated from the prediction.
The scheme in figure 28.4 was able to account for a number of
experimental observations, but it has been challenged (Feldman
2009). Indeed, this scheme suggests that, following a voluntary
movement to a new joint position, muscles cannot be relaxed.
Consider two joint positions in figure 28.4b. In the bottom drawing,
the muscle is shorter than in the upper drawing. This means that
signals from length-sensitive endings in its spindles should be larger
in the upper drawing. But if the muscle is relaxed in both positions,
its efferent copy is the same—zero! So, there is no way the changes
in afferent signals can be balanced by signals related to the efferent
copy. As a result, this scheme predicts that there will be nonzero
muscle activation in obvious contradiction to everyday experience.
The inadequacy of the original scheme in figure 28.4 does not
mean that the basic idea of the motor control process taking part in
kinesthetic perception is wrong. An alternative scheme has been
proposed based on the theory that the control of movement of an
effector can be adequately described as time changes in the spatial
referent coordinate (RC) for that effector (see Feldman 2009, 2016;
chapter 21).
Figure 28.5 illustrates the scheme for the combined perception of
muscle length and force. Note that setting a value of RC (λ for a
muscle) defines combinations of force and length values that can be
achieved by the effector at steady states—points on the force–length
characteristic. Deviations from λ along the muscle characteristic lead
to a parallel increase in signals from a number of sources, including
muscle spindle endings (because length is increasing), Golgi tendon
organs (because force is increasing), and α-motoneurons (because
muscle activation is increasing). Any of those signals can be used to
identify a point along the force–length characteristic (i.e., values of
both muscle force and length). Indeed, a number of studies have
shown that spindle endings can play the role of force sensors and
Golgi tendon organs the role of length sensors (Watson et al. 1984;
Luu et al. 2011; Proske and Gandevia 2012).
In other words, there are two elements necessary to measure any
variable: a point from which to measure and a measuring device.
The efferent process defines RC, which serves as the first element,
and signals from sensory endings serve as the other element. A
copy of the RC used in the perceptual process can be termed the
efferent copy. The scheme illustrated in figure 28.5 has been used
to interpret a variety of perceptual phenomena, including linked
perception of force and coordinate, kinesthetic illusions, and
phantom perception after an amputation (reviewed in Feldman 2009;
Latash 2019).

Figure 28.5 The scheme for the combined perception of muscle length and
force. Setting a value of RC (λ for a muscle) defines the force–length
characteristic. The efferent signal (EFF) defines a referent point, while signals from
afferents (AFF) define deviation of the muscle from RC. Deviations from λ along
the characteristic lead to a parallel increase in signals from a number of sources,
including muscle spindles, Golgi tendon organs, and α-motoneurons.

Consider now the physiological pathways and structures involved

in the transmission of sensory information. Signals from a peripheral
receptor travel along the peripheral end of the T-shaped axon of a
ganglionic neuron and then are transmitted along the central end of
the sensory axon into the spinal cord (figure 28.6). Sensory axons
from receptors of different modalities ascend into supraspinal
structures within the ipsilateral dorsal column and make synapses in
the dorsal column nuclei at the medullar level. The axons of neurons
within these nuclei cross the midline and ascend further in the
contralateral side of the body to the ventral posterior nucleus of the
thalamus in the medial lemniscus pathway. The thalamus is a major
relay transmitting sensory information to the sensory cortical areas.
Thalamocortical projections are made via a large bundle of fibers
called the internal capsule.
There are two major types of thalamic nuclei. Relay nuclei
process either a single sensory modality or information from a
distinct part of the body; they project to a specific region of the
cerebral cortex and receive recurrent input from the same cortical
region. Diffuse-projection nuclei mostly deal with transmitting afferent
inputs related to the functioning of the limbic structures and
modulation of the thalamus’ own activity. Their projections are more
widespread. Somatosensory perception, which is presently in the
focus of our attention, is based on transmission via a relay,
ventroposterolateral (VPL) nucleus.

Figure 28.6 A schematic of the afferent signal transmission along the dorsal
columns of the spinal cord, via the lemniscus, to the ventroposteriorlateral (VPL)
nucleus of the thalamus.

Thalamic inputs terminate in the parietal cortex, in the

somatosensory areas (areas 1, 2, 3a, and 3b). There they create
sensory maps that look like distorted images of the body with
disproportionate representations of the face, tongue, hand, and
particularly, the thumb. Information processing within the cortex is
supposed to lead to conscious perception and also to participate in
using kinesthetic information for purposes of control of voluntary
movements.

28.5 Vibration-Induced
Kinesthetic Illusions
High-frequency muscle vibration (e.g., at 100 Hz) creates a large
number of motor and sensory effects mediated primarily by the very
high induced activity of the primary spindle endings. These sensory
endings are highly sensitive to velocity and can be driven by
relatively low-amplitude vibration (e.g., with the peak-to-peak
amplitude of 1 mm applied to the belly of a muscle or to its tendon);
this means that the primary endings generate one or several action
potentials in response to each cycle of vibration (Matthews and Stein
1969). Motor effects, including the tonic vibration reflex, were
discussed in chapter 18. In this section, we will discuss the effects of
vibration on kinesthetic perception.
Most commonly, muscle vibration leads to kinesthetic illusions
corresponding to perceiving joint positions associated with longer
length of the muscle subjected to vibration (Goodwin et al. 1972; Roll
and Vedel 1982; see figure 28.7). Such illusions can be very strong
and can even lead to a perception of anatomically impossible joint
configurations (Craske 1977). Interpretations of such vibration-
induced kinesthetic illusions (VIKIs) have linked it to the unusually
high activity of primary spindle endings. The velocity sensitivity of
these endings probably plays an especially important role. Indeed,
sometimes humans report very high velocities of illusory movements
produced by muscle vibration while illusions of joint displacement are
relatively modest, much smaller than what could be expected from
the reported joint velocity (Sittig et al. 1985).
Figure 28.7 High-frequency muscle vibration can lead to misperception of the
position in a joint crossed by the muscle. In this illustration, the person is trying to
match the elbow joint positions under the vibration applied to the left elbow flexor.
Trying to match the joint position of the left elbow (αL) commonly leads to showing
a joint position in the right elbow (αR) that corresponds to longer elbow flexors.

If vibration is applied to a muscle involved in a postural task—for

example, to the Achilles tendon during standing—humans feel that
the body is pulled backwards and show significant body deviations
from the vertical, which sometimes require making a step (Eklund
and Hagbarth 1967). Such vibration-induced fallings (VIF) have
also been interpreted as consequences of the effects of vibration on
primary spindle endings. Indeed, the unusually high activity levels of
the primary spindle endings in the ankle plantarflexors is interpreted
by the brain as elongation of that muscle group, which, under normal
conditions, is associated with body sway forward. This illusory
perception is corrected by an actually observed deviation of the body
in the opposite direction (see figure 28.8).
Figure 28.8 Vibrating the Achilles tendon of a standing person leads to body
deviation backwards (VIF).

PROBLEM 28.3
Imagine that two vibrators are placed over the triceps surae and
tibialis anterior. What effects would you expect to observe in a
standing person?

A number of features of VIKIs have remained puzzling. For

example, the direction of a VIKI could reverse with changes in
sensory signals of other modalities and correspond to shortening of
the muscle subjected to vibration (Feldman and Latash 1982).
Besides, some people report very strong VIKIs, while others feel only
mild illusions. Some illusions are seen consistently in only a small
proportion of the general population. These include, in particular, the
Pinocchio effect (Burrack and Brugger 2005). This illusion is
sometimes reported if a person is asked to touch lightly the tip of the
nose with the index finger of a hand and then vibration is applied to
the elbow flexor group (e.g., over the biceps). The high activity level
of the spindle endings in the elbow flexors suggests to the subject
that the elbow is being extended. On the other hand, the contact
between the fingertip and the nose is not broken. The combination of
the two factors leads, in some persons, to reporting an illusion of
nose elongation.

28.6 Distorted Efferent Copy

and Preconceptions
Figure 28.5 suggests that there may be two causes of kinesthetic
illusions, those associated with distorted RCs and distorted sensory
signals. The direct effects of vibration of the primary spindle endings
correspond to the latter factor. However, a number of studies have
reported VIKIs and other perceptual phenomena that suggest using
a distorted RC during the creation of percepts (i.e., the efferent copy
used in the perceptual process does not match the actual RC)
(reviewed in Feldman 2009; Latash 2021b). This scheme also
suggests that illusions of position may be accompanied by illusions
of force since perception of these two variables is intimately linked
by the force–length characteristic. This prediction has been
confirmed experimentally (Cafarelli and Kostka 1981; Reschechtko
et al. 2018).
Another important aspect of perception is that it is strongly
affected by preconceptions about the external world and one’s body.
A number of well-known visual illusions are based on such
preconceptions, but they also play a major role in kinesthetic
perception. Consider, for example, a simple joint with two muscles,
flexor and extensor (figure 28.9). During natural movements, when
one muscle shortens, the other muscle necessarily stretches.
Imagine now that a vibrator is placed on the flexor leading to a
perception of its elongation. Since sensory signals from the extensor
are undistorted, the effects of vibration should lead to an impossible
percept: the flexor stretches but the extensor does not shorten,
which suggests that body integrity is violated. The brain refuses to
accept this disallowed percept and replaces it with another one,
which may differ across subjects and conditions, leading to the
aforementioned high interperson variability of perceptual effects of
muscle vibration.

Figure 28.9 (a) Vibrating a flexor muscle can lead to misperception of the length
of this muscle without a change in perception of the extensor muscle. (b) This
should lead to perception of an anatomically impossible situation. (c) Subjects
typically report an intermediate percept corresponding to length changes in both
muscles, although only one of them is subjected to vibration.

28.7 Sense of Effort

Effort is not a well-defined concept. By definition, it is subjective and
cannot be measured directly. Its origin has been viewed as primarily
central, although signals from peripheral sensory endings can affect
reported effort (Gandevia et al. 1980; Sanes and Shadmehr 1995;
Luu et al. 2011). Sometimes, sense of effort is confused with sense
of force, but these are conceptually different senses. Unlike sense of
effort, sense of force is an internal representation of an objective
physical variable that can be measured directly. As it follows from
figure 28.5, sense of force emerges, coupled with sense of position,
based on two neural signals, RC reflecting the efferent process and
the activity of sensory endings reflecting deviation of the effector
from RC along both spatial and force axes.
 Sense of effort can be associated with RC. Indeed, when human
subjects were asked to compress cylinders with different springs
inside with the index finger and the thumb and to match the forces,
the subjects matched referent apertures between the hands (an
example of RC in this particular task), while both force and actual
aperture differed (van Doren 1995, 1998). These observations were
interpreted as the subjects trying to match efforts leading to RC
matching. However, a number of recent studies have shown that,
under certain conditions, RC-related signals participating in the
perceptual process represent distorted copies of those participating
in the motor process (Cuadra et al. 2020, 2021).
This is particularly striking when a person is asked to produce a
constant force against a stop (in isometric conditions) and then
cocontract agonist–antagonist muscles without changing the force
magnitude. Typically, subjects show an unintentional force increase,
which can be expected from a change in the c-command (see
chapter 21) and resulting change in the apparent stiffness of the
effector (figure 28.10). However, when asked to report force change
verbally, the subjects claim that the force dropped somewhat!
Moreover, when the subjects are asked to match the force with the
contralateral effector, they show force increase, typically even larger
than the actual one.

PROBLEM 28.4
How would you interpret the reports of somewhat decreased force
under muscle coactivation in figure 28.10?

These observations suggest that humans have a relatively poor

sense of force and cannot easily incorporate muscle coactivation into
their efferent copy. This conclusion is corroborated by everyday
experience. Try to press against a stop strongly. Remember what
you are doing. Now try to match the activity of the antagonist muscle!
Most naïve humans cannot do this and assume that the antagonist
was silent (which is not true; see chapter 23).
 Another important conclusion from these observations is that
“perceiving to report” (as reflected in verbal reports) and “perceiving
to act” (as reflected on force matching) are likely to involve different
neural pathways. This idea is similar to that of two neural streams
participating in the processing of visual information (Goodale et al.
1991; Goodale and Milner 1992), which is discussed in more detail in
chapter 30.

Figure 28.10 If a person is asked to cocontract the agonist–antagonist muscles

without changing the net force on the environment, a change in the c-command is
expected to lead to larger apparent stiffness of the effector (k2 > k1). If no change
in the r-command takes place, force should increase (in spite of the instruction!);
see the open and filled circles. This is what is observed in experiments.

28.8 Stability of Percepts

To perform successful and accurate movements, our perception of
the body and its interactions with the environment has to be stable.
This is not a trivial feature. Indeed, during various actions, all the
relevant afferent and efferent neural signals change. Imagine, for
example, that you cocontract the agonist and antagonist muscles
acting at the elbow joint without moving it. This is easy to do without
looking at the arm. Signals from all relevant sensory endings will
change. In particular, signals from spindle endings will reflect the
changes in the activity of γ-motoneurons (because of the
phenomenon of α-γ coactivation). Signals from Golgi tendon organs
will change with muscle force, and signals from articular receptors
will change because of the induced change in the joint capsule
tension. How do we know that the joint did not move? The fact that
we correctly perceive the joint as stationary suggests that all the
changes in the numerous relevant signals, afferent and efferent, are
constrained to a subspace, which our brain associates with
perception of the unchanged joint angle. This subspace has been
termed the iso-perceptual manifold (IPM) (Latash 2018b).
This concept suggests that changes in afferent and efferent
signals can either lead to changes in percepts (perceptually
nonequivalent) or keep percepts unchanged (perceptually
equivalent). In the previous example of muscle cocontraction,
changes in the signals did not move outside the respective IPM (i.e.,
they were perceptually equivalent).

28.9 Perception–Action Coupling

Intimate links between sensory and motor processes have been
known for a long time. Indeed, spinal reflexes lead to changes in
muscle activation and movement mechanics induced by signals from
sensory endings. On the other hand, any motor process leads to
sensory changes due to changes in body configuration and force at
the interface with the environment and also to activation of the
system of γ-motoneurons that change the sensitivity of spindle
endings to muscle length and velocity (see chapter 6). In addition to
the aforementioned sensory–motor links, there are also links
between actions and percepts (i.e., higher-order consequences of
both afferent and efferent processes). These include both the effects
of action on perception and the effects of sensory signals on action.
This is logical because perception and action processes are
functionally intertwined: We use perception to plan and perform
actions, and we use actions to achieve desired percepts.
Some of these effects are seen without the subject’s being aware
of them. For example, when a person stands in a dark room and
watches a pattern of dots on the screen, motion of the dots toward
the person or away from the person entrain postural sway (Dijkstra
et al. 1994; see chapter 25). Similar effects of sway entrainment are
seen when a person touches a touch pad and the pad starts to move
very slowly (Jeka et al. 1998).
Strong effects of visual perception on action stability have been
shown in experiments when the subjects were asked to rotate two
handles hidden under a desk, which produced motion of two dowels
seen by the subject on top of the table (Mechsner et al. 2001). When
handle rotation and dowel rotation were in the same direction, in-
phase motion of the two handles was very stable, and the out-of-
phase motion was less stable. When one of the handles was
connected to its dowel in such a way that clockwise handle motion
led to counterclockwise motion of the dowel, the in-phase dowel
motion remained more stable, although it corresponded to out-of-
phase hand motion.
There are also examples of the effects of action on perception.
Imagine, for example, that two patterns of dots on the screen are
alternating at a regular pace (figure 28.11). When a person watches
this screen, he or she typically perceives that the dot pattern rotates
clockwise or counterclockwise. This illusion is commonly unstable,
and one reports changes in the direction of dot rotation. If now this
person imagines holding a steering wheel, rotating this imaginary
wheel drives the illusion: The dots start to move in the direction of
wheel rotation.
The tight coupling between action and perception has been
incorporated into the framework of ecological psychology (Gibson
1979; Turvey 1990, 2007). An example of this coupling is the
synchronization of movements between two persons watching each
other (Schmidt et al. 1990; Schmidt and Turvey 1994). There have
been attempts to link action–perception coupling phenomena to the
system of mirror neurons in the ventral premotor cortex, dorsal
premotor cortex, and intraparietal cortex that are active in humans
during observation of an individual performing an action and during
performance of the same action.

Figure 28.11 Two patterns of dots are alternating at a regular pace (black circles
and gray circles; imagine that both sets are of the same color). When a person
watches this screen, he or she typically perceives that the dot pattern rotates
clockwise or counterclockwise, commonly with changes in the direction of dot
rotation. If the person imagines holding a steering wheel, rotating this imaginary
wheel drives the illusion: The dots start to move in the direction of wheel rotation.

CHAPTER 28 IN A NUTSHELL
Sensation of physical variables
follows the Weber-Fechner law and the
Stevens law. Perception of body
configuration and forces at the
interface with the environment is
known as kinesthesia. Kinesthetic
perception is based on an interaction
of efferent (motor) and afferent
(sensory) signals. Signals from
peripheral sensory endings deliver
ambiguous information that requires
further processing to produce
kinesthetic percepts. These signals
are transmitted via the dorsal spinal
column to nuclei in the medulla and
then, via the lemniscus, to the
thalamus and further to somatosensory
areas in the parietal cortex.
According to one theory, efferent
signals participating in perception
(efference copy) define a referent
spatial coordinate, and sensory
signals define deviations from this
coordinate. This theory links the
perception of forces and coordinates
into a single scheme. Artificial
changes in sensory signals (e.g., by
muscle vibration) lead to complex
patterns of kinesthetic illusions as
associated phenomena, such as postural
deviations. The stability of percepts
suggests the existence of a stable
subspace (iso-perceptual manifold,
IPM) in the combined afferent–efferent
space. Coupling between perceptual and
motor processes takes place at
different levels, from muscle reflexes
to complex actions.
 Chapter 29

Multisensory Integration

KEY TERMS AND TOPICS

cross-modal stimuli
sensory weighting
sensory noise
optimal integration
sensory reference frames
superior colliculus

During most activities of daily living, multiple sensory systems are

simultaneously stimulated. For example, when we change channels
while watching television, our visual system provides information
about the location of the hand, the remote control, and the television.
At the same time, the proprioceptive system relays to the nervous
system the current state of the muscle and the arm location in space.
Finally, the tactile system senses the physical properties of the
buttons on the remote control and the changes in pressure at the
fingertips to ensure that the correct button is sufficiently depressed.
Multisensory integration is the consolidation of sensory information
from simultaneously experienced individual sensory modalities into a
single percept that allows us to act on the world.
29.1 Spatial Multisensory
Integration for Limb Motor
Control
The visual system plays a key role in the planning and execution of
all goal-directed movements and provides visuospatial information
with high acuity. Other sensory systems, such as the proprioceptive
or the auditory system, conduct sensory information at faster
temporal time scales, but they don’t provide the same level of spatial
acuity as the visual system. Together, these cross-modal stimuli
often provide abundant sensory information about the same events
(e.g., both the proprioceptive and visual systems sense where the
limb is in space during a pointing movement), but the differential
spatial resolutions can create perceptual ambiguity. Multisensory
integration can then be restated as a problem of exploitation of
abundant sources of information, in which the brain combines
sensory information from different senses to get more reliable
perceptual estimates (Stein and Stanford, 2008).
Multisensory integration arises from interactions amongst different
sensory systems. The interaction evokes a stronger neural response
in multisensory neurons compared to the individual component
stimuli. Multisensory neurons are those neurons that respond to
cues from more than one sensory modality or whose responses to
one modality are significantly altered by the presence of a stimulus
from another modality (Stein and Meredith 1993). In many instances,
this integration is necessary because the sensory information
provided by an individual sensory modality is often spatially
ambiguous. Combining cues from multiple sensory sources can
improve the localization of limb location in space (see figure 29.1).
The accuracy of a reaching movement improves when sensory
information from both vision and proprioception are combined (van
Beers et al. 1999; Sarlegna and Sainburg 2009). Though vision
provides accurate feedback about limb location in space,
proprioception is critical for controlling other aspects of limb
mechanics, such as intersegmental dynamics.

Figure 29.1 The neuromotor hierarchy is organized in a way that integrates

visual and proprioceptive sensory information in an optimal fashion for limb motor
control. The ellipses represent the variance of visual and proprioceptive
localization. Vision is not only more precise than proprioception, but vision and
proprioception both exhibit direction-dependent variance. They also have different
biases (i.e., if multiple trials of movements under purely visual and proprioceptive
conditions were performed, the means of the two conditions would be in different
locations). The different locations of the two ellipses reflect these biases. The solid
ellipse indicates direction- and variance-dependent weighting of the two
modalities. The dashed circle indicates only variance-dependent weighting.
Optimal integration implies both direction-dependent and variance-dependent
integration of sensory information in the two modalities.
Reprinted by permission from R. J. van Beers, A.C. Sittig, and J.J. Denier van der Gon,
“Integration of Proprioceptive and Visual Position-Information: An Experimentally Supported
Model,” Journal of Neurophysiology 81, no. 3 (1999).

In their study, van Beers and colleagues asked participants to

make reaching movements to match target positions in different
directions on a tabletop with their unseen left hand under the table.
Participants performed the experiments in three different conditions:
proprioception (the target was the participant’s right hand that they
could not see), vision (a visual target was presented), and
proprioception-vision (target was the participant’s right hand that
they could see). The authors measured errors in the different
conditions and concluded that multisensory integration between the
visual and proprioceptive systems occurs as a direction-dependent
sensory weighting of the information in the two modalities.

PROBLEM 29.1
In the mammalian nervous system, many types of multisensory
neurons have been found that respond to a combination of visual,
auditory, vestibular, and somatosensory inputs. This allows many
biological events to be registered by more than one sense, and
animals can use these senses synergistically. Suggest one or two
good hypotheses as to why the nervous system would develop
both unisensory and multisensory neurons.

The idea of sensory weighting can best be understood in terms of

a simple formulation. Let’s imagine we are making a reaching
movement toward a target. The image of the hand on the retina
provides a visual estimate (XV), and the muscle spindles and joint
receptors provide a proprioceptive estimate of the hand location (XP).
It is hypothesized that the nervous system weights and integrates
them into a single estimate of the hand position. Let’s consider that
the weight of the proprioceptive system is WV. Then the integrated
optimal estimate of XVP is given by multiplying the weights and the
estimates of each modality and then performing a linear summation:

Here, the weights for vision and proprioception should add up to

1. The weighting of each sensory input is considered to be inversely
proportional to the associated sensory noise levels in the individual
sensory modality. By noise here we simply imply a measure of
performance dispersion (such as variance or standard deviation).
Thus, the weights are inversely related to the variances of the
individual signals. In equation 29.1, typically WV would be greater
than 0.5 because people rely more on vision than on proprioception
(Ernst and Banks 2002). Also note that this equation for the optimal
estimate assumes that there is no cross talk between the sensory
modalities. In other words, the proprioceptive and visual systems are
assumed to operate independently.
Psychophysical studies have shown that multisensory integration
between vision and other sensory modalities occurs in a statistically
optimal fashion. Statistical optimality implies that the nervous system
weights each signal based on its reliability before linearly combining
them. Optimal integration has been shown in visuo-proprioceptive
(van Beers et al. 1999), visual-auditory (Kording et al. 2007), visual-
haptic (Ernst and Banks 2002), and visual-vestibular (Fetsch et al.
2009) paradigms. Other researchers have also proposed that the
CNS processes available sensory information in a near-optimal
fashion to minimize reaching errors (Kording and Wolpert 2004).
These experiments suggest that the mechanisms of multisensory
integration optimally combine information from different sensory
modalities to provide the nervous system with more accurate
estimates of limb and body location and orientation in space.
Consistent with the optimal integration paradigm, the dominant role
of vision can be reversed if proprioceptive feedback becomes more
reliable (van Beers et al. 2002; Tagliabue and McIntyre 2014). How
this optimality is achieved in the nervous system is not exactly clear,
though multiple hypotheses have been proposed (Berniker and
Kording 2011; Todorov and Jordan 2002). Some of the most
prominent models are based on Bayesian decision theory (reviewed
in Chandrasekaran 2017). Bayesian decision theory is a statistical
approach to studying optimal integration as a statistical inference
process performed by the brain. Whether the brain itself performs
Bayesian statistics is debatable and has been discussed in many
review articles (Knill and Pouget 2004; Loeb 2012; Kording 2014).
Sensory information in individual modalities is weighed depending
on not only the levels of noise, but also prior expectations. For
example, if we are hiking on uneven uphill terrain during daytime, our
visual, proprioceptive, and vestibular systems provide congruent and
consistent information, and our nervous system might scale the
sensory information based purely on their levels of noise. However, if
we were hiking during nighttime, our nervous system would
downregulate the weight of the visual system and upregulate the
weights of the vestibular and proprioceptive systems. Similarly, if we
are walking on slippery surfaces, our nervous system might
selectively upregulate only the vestibular system to preferentially
respond to any sensory stimulation that might indicate that the head
is experiencing strong acceleration or is not parallel to gravity.

PROBLEM 29.2
The theory of Bayesian integration suggests that the nervous
system integrates multisensory information based on its relative
reliability. What does optimal integration mean to you? What
would suboptimal integration look like? Could you relate this to
any real-world experience or example?

29.2 Temporal Multisensory

Integration for Limb Motor
Control
When we make reaching movements toward objects in our
peripersonal space, we often don’t look directly at those objects.
Peripersonal space is the immediate space surrounding the body
within arm’s reach. For example, we may be writing an email on the
computer and reach out to grab a cup of coffee without even looking
at the coffee cup. This is made possible by the proprioceptive
system estimating where the limb is in space and continuously
integrating that information with the approximate location of the cup
obtained through peripheral vision or based on memory. While the
limb is reaching toward the target, if the integration of sensory
information suggests that the limb might not reach the desired target
(cup), then corrections will have to be made to bring the limb back to
the target location.
Humans perform this action with effortless ease, but it is in fact
extremely demanding from a motor control standpoint. Specifically,
integrating information from visual and proprioceptive modalities
while the movement is underway requires the nervous system to
account for differential axonal conduction and central processing
delays in the two sensory modalities. Axonal conduction in the
proprioceptive system of mammals is much faster than in the visual
system. Temporal processing of proprioceptive and visual stimuli has
been determined by experimentally perturbing the limb during
reaching movements, either mechanically using a motor
(proprioception) or by visually shifting the feedback of the hand or
the target location (vision). The temporal latencies of the corrective
actions made in response to the perturbations provide an estimate of
how afferent information is processed by a particular sensory
system.
When the arm is mechanically perturbed, the muscle spindles
relay to the nervous system the current length of the muscle and the
speed at which the muscle is changing its length. After the
perturbation has been applied, the stretched muscles exhibit multiple
responses at different latencies. The short-latency reflexes occur at
approximately 20 to 50 ms after the perturbation in the upper
extremity muscles (see chapters 17-19). This response is mediated
primarily by spinal circuits. The long-latency response occurs later, at
around 50 to 100 ms in the muscles of the upper extremity. The long-
latency responses exhibit task-dependent modulation (Marsden and
Rothwell 1984; Scott 2012; Lewis et al. 2005). By task-dependent,
we mean that the muscle responses scale not only with the
amplitude and direction of perturbations but also with instruction to
the subject (see chapter 21). Part of the long-latency response is
mediated by a transcortical pathway that includes the primary motor
cortex, dorsal premotor cortex, supplementary motor area, posterior
parietal cortex, and cerebellum (reviewed in Forgaard et al. 2021).
These rapid responses are made possible by the fast axonal
conductance in the myelinated Ia afferent fibers and involve both
ascending and descending pathways.
In contrast, when visual perturbations are applied during reaching
movements (this is typically done using augmented reality), it takes
the nervous system longer, about 150 to 220 ms, to initiate corrective
motor responses (reviewed in Smeets et al. 2016, Franklin et al.
2016, Sarlegna and Mutha 2015). Augmented reality provides an
interactive experience that combines computer-generated visual
elements in real time to probe sensorimotor function, and it has been
used widely in the field of motor control to probe the neural control of
movement. These differential sensorimotor delays in the visual and
proprioceptive systems have minimal consequences for most
activities of daily living (that tend to be relatively slow), but they could
be disadvantageous for movements that require fast reactions.
During fast movements, the more rapidly the hand moves, the more
difficult it is for the sensorimotor system to localize its instantaneous
location in space.

Figure 29.2 The visual and proprioceptive systems process sensory information
with different delays. The proprioceptive system is faster but has higher levels of
sensory noise.
 So how does the nervous system reconcile temporal delays
across vision and proprioception that can differ by tens of
milliseconds (see figure 29.2)? This question was addressed by
Crevecoeur and colleagues in an elegant study (Crevecoeur et al.
2016). The authors applied either mechanical perturbations to the
hand or visual perturbations to the hand position feedback. In some
conditions, they applied both perturbations. Their main outcome
variables of interest were the long-latency reflexes (M3; also see
chapter 19), saccadic reaction times, and limb kinematics. They
used saccadic eye movements as proxies for estimating participants’
hand motion and showed that participants relied more on
proprioceptive feedback than on visual feedback to estimate hand
motion. They also showed that vision mostly influenced fine motor
control near the end of movement, whereas limb motor corrections in
the early stages of the movement appeared to be primarily driven by
proprioceptive feedback. These results suggest that the nervous
system accounts for the differential temporal conduction delays in
different sensory modalities for limb motor control.
Figure 29.3 Participants make reaching movement from the solid dark circle to
the solid light circle. During the movement, a new target is presented (dashed light
circle), but the hand is constrained to be within the “force channel” (shown as a
dashed rectangle). Within this channel, the robot applies a strong elastic and
viscous force to prevent the hand from escaping the channel. Inset shows the
experimental system used in these studies.

It has also been suggested that the fastest feedback corrections

are mediated by a single sensory modality, the proprioceptive
system, whereas visual estimates of target and hand positions are
used later, after about 200 ms, for the online control of reaching
(Franklin et al. 2016). In the study done by Franklin and colleagues,
participants were asked to make forward reaching movements while
holding a robotic manipulandum. The target of the reaching
movement and the hand feedback were jumped to different
locations. In some of the trials, participants were prevented from
making corrective movements by forcing the hand to be constrained
within a force channel. A force channel is implemented using a
robotic manipulandum that constrains the hand to move in a straight
line; any deviation away from the line is prevented by the robot (see
figure 29.3). When participants made corrective movements, they
applied a limb force against this channel that was recorded. The
authors showed that the early applied limb force against the channel
walls depended on the relative jump of the target and the hand
cursor feedback instead of the absolute difference vector. After about
200 ms, the changes in target and hand positions were combined
appropriately to produce a motor response that suggested that the
true difference vector between the target and the hand was being
used for motor corrections. Consistent with the results of Crevecoeur
and colleagues (Crevecoeur et al. 2016), these results suggest that
early responses are mediated by a fast approximation within a single
sensory modality, whereas later responses involve integration of
information from different sensory modalities.

PROBLEM 29.3
The flash-lag effect occurs when a nonmoving object is quickly
flashed directly next to a moving object. Though they are
presented in the same location at the same time, the two stimuli
are perceived to be displaced from one another. Specifically, in
this visual illusion the nonmoving object is perceived as “lagging”
the moving object. What do you think would cause this illusion?

29.3 Coordinate Frames for Limb

Motor Control
A major problem for multisensory integration is that task-relevant
information must be referenced to the external world. In other words,
it is not just important to learn a tennis backhand swing, but the ball
must also land within the court perimeter. Thus, our movements
must defer to the constraints imposed by the external world.
However, movement-related neural signals directed from the motor
cortex (M1) must specify the task in terms of limb effectors (i.e.,
agonist–antagonist pairs). But how does the nervous system know
that a particular spatiotemporal sequence of agonist–antagonist
activation would result in a successful performance? This would
require some form of mapping between the external world that is
sensed by multiple sensory modalities, such as the visual and
auditory systems, and the joint- and muscle-based maps used by the
motor system.
Vision provides coordinate information about the extrinsic world in
a retinotopic reference frame. That information is then used to plan
spatiotemporal characteristics of movements toward target locations.
In contrast, proprioception provides intrinsic information about joint
and limb configuration, as well as the state of the muscle, and plays
an important role in transforming a spatial goal plan in extrinsic
coordinates into intrinsic referent coordinates for the limb to move
from point A to point B (see figure 29.4).
Figure 29.4 Different sensory reference frames (RF) involved in visually guided
upper limb movements. The visual and proprioceptive systems encode information
in different reference frames that have to be aligned for making visually guided
movements.

These sensory systems send information to the parietal cortex in

different reference frames, and that makes the parietal cortex the
most likely location for multisensory integration. In contrast to the
visual system that encodes information in a retinotopic reference
frame (Gardner et al. 2008; Pouget et al. 2002), the somatosensory
system encodes tactile information in a head-based reference frame
(Avillac et al. 2005), and the proprioceptive system provides sensory
information about muscle length and velocity in a joint-centered
reference frame (Gardner and Costanzo 1981). Thus, to make goal-
oriented movements, the nervous system must transform the
sensory information from the external world for that information to be
rendered useful for the limb motor system and oculomotor system.
The proposed solution to this problem is that the nervous system
establishes a shared reference frame from the different sensory
modalities. To that end, a prominent hypothesis in the field of
multisensory integration is that of sensorimotor coordinate frames
(Avillac et al. 2005; Batista et al. 2007). This hypothesis proposes
that at some stage of planning for reaching movements, visual
information about the hand position and the target location is
encoded in the same spatial frame of reference to compute a
difference vector between the target and the hand (Beurze et al.
2006; Oostwoud et al. 2017; Shadmehr and Wise 2005). Then the
motor system uses that difference vector to release a reaching
movement toward the target (see figure 29.5).

Figure 29.5 The difference vector between the target and the hand (XTarget-Hand)
is first transformed from a retinotopic coordinate frame to a joint (shoulder-based)
coordinate frame. The crosshairs show the current location of the gaze. The
dashed lines show the locations of the targets in an eye-based coordinate frame.
The dotted lines show the locations of the targets in a shoulder-based coordinate
frame.

If reaching movements are indeed based on a difference vector

between the target and the hand location, then the first challenge for
the nervous system is to align the sensory information in different
modalities into a common reference frame. If reaching movements
are being performed while standing upright with a forward trunk lean,
then the vestibular system will also be involved. The vestibular
system encodes information in a head-centered reference frame
(reviewed in Angelaki and Cullen 2008). This ensemble of different
reference frames for different sensory modalities can create sensory
conflict and perceptual ambiguity and can make motor planning and
execution of reaching movements challenging.
The nervous system could solve this problem in one of two ways.
First, it could devote substantial neural resources to align all afferent
sensory information into a single reference system. The advantage
of using such a scheme would be that the neural input into the motor
system would be processed and simplified. The disadvantage would
be that significant neural resources would have to be devoted to
transform sensory information from multiple modalities to determine
the relative location of the hand and the target. These neural
transformations would take and prolong reaction times. Alternatively,
the nervous system could use multiple reference frames for motor
planning and execution. This would imply that the nervous system
would determine the relative locations between the hand and the
target in multiple coordinate frames and continuously compare them
to modify and correct movement trajectories if they deviate from the
desired trajectories.
There is stronger evidence to support the view that reaching and
pointing movements are planned in multiple reference frames in
large cortical networks spanning the occipital, parietal, and frontal
areas (McGuire and Sabes 2009; Beurze et al. 2006, Berniker et al.
2014). It has also been proposed that time-consuming neural
coordinate transformations could be avoided if neuronal activity were
organized as a multilayer recurrent neural network that mapped
sensory input into multiple coordinate frames for motor planning and
execution (reviewed in Pouget et al. 2002).

PROBLEM 29.4
The parietal cortex stores object locations for reaching in a
retinotopic reference frame. So if a saccade is made during an
ongoing reaching movement, the location of the object would be
remapped in a retinotopic reference frame. Would that affect the
sensorimotor transformations associated with the planning of
 reaching movements? Would that affect the trajectory of the
reaching movement?

29.4 Postural Balance and

Motion Perception
When a sound and a spatially incongruent visual stimulus
accompany each other, the sound is incorrectly perceived to be
emanating from the same location as the visual stimulus. This is
called the ventriloquism effect (Warren et al. 1981). The
ventriloquism effect can be explained as a phenomenon in which a
sensory modality with higher spatial acuity (vision) dominates over
other sensory modalities (such as the auditory system) with lower
spatial acuity. Though the dominant effect of vision in the
ventriloquism effect may seem counterproductive, in most other
situations encountered in daily life, vision actually contributes to
resolving peripheral ambiguity between different sensory systems.
For example, the otolith organs sense both translational and
gravitational accelerations, and their firing rates are ambiguous to
the two types of motion. The nervous system resolves the ambiguity
by combining otolith signals with either signals from the semicircular
canals (chapter 13) or visual cues (Green and Angelaki 2010).
Multisensory integration is also important for balance control.
Visual, somatosensory, and vestibular inputs interact continuously in
the control of upright balance. There is a considerable degree of
functional overlap, or redundancy, between the three sensory
systems (Mergner et al. 2003). For example, we can ascertain
whether we are standing upright by judging the tactile pressure on
the soles of the feet and the tension in the lower extremity muscles
(somatosensory), by judging the static vestibular input to the otolith
organs (vestibular), or by judging whether tall structures such as
buildings and trees appear upright (visual). Despite this overlap,
each sensory system specializes in unique aspects of function that
require that all three systems work together to determine whether the
body is upright or not. Specifically, the proprioceptive system cannot
distinguish whether the head rotates with respect to the trunk, or the
trunk rotates with respect to the head. So if a person slips on ice or if
their head is passively tilted (this happens when we are standing in a
bus and it suddenly accelerates or decelerates), and if the relative
joint angles between the head and the trunk are the same, then the
proprioceptive system will signal the same information to the nervous
system. Only the vestibular system can detect head movement
(translation and rotation) with respect to the gravity vector. Thus, the
vestibular system cannot be tricked into believing that a slip is
occurring if the head orientation is parallel with respect to gravity. In
the absence of vestibular signaling, the nervous system will not be
able to distinguish between a slip and a passive head tilt. Thus,
these two systems will have to work together to maintain upright
balance (see figure 29.6).
In contrast to the vestibular system, the visual system can be
easily deceived into believing that one is actually in motion (self-
motion), even when there is no motion. This phenomenon is called
vection, and it is experimentally created using the moving room (or
the moving wall) paradigm. The experimenter moves a virtual wall
unbeknownst to the participant while the participant performs a
difficult postural balance task (e.g., standing on a narrow beam).
This causes the participant to lose their balance. In real life, vection
is experienced when we are in a train and we see the train next to us
suddenly move. The visual system creates a temporary perception of
motion in these conditions that is eventually overridden by the
proprioceptive and vestibular systems.
Usually, there are limits to the extent to which the human brain
“believes” any one of the sources of information used for postural
control. Thus, an illusion created by a flow of information in one of
the sensory systems can be regulated or may even be eliminated by
information coming from other sensory sources. To create a strong
illusion, one needs to manipulate synchronously all three major
sources of information related to body position and orientation:
vestibular, visual, and proprioceptive. Theme parks use such
methods in their simulation rides during which the viewers are
seated in comfortable chairs, which can be tilted synchronously with
changes of an image projected on a large screen.
One way we can think about this is that the in-built redundancy in
the sensory systems is used by the nervous system to confirm an
ambiguous sensory signal. If the other sensory systems do not
confirm the source of the sensory signals in one system, the nervous
system might trigger voluntary responses to further investigate the
source of signals observed in one system. For example, if the visual
system signals motion, we might look at surrounding areas or at our
toes to confirm whether the body is actually stationary or moving. For
these situations, multisensory processing would be critical to resolve
sensory ambiguities and conflict.
One influential idea for postural control, coordinate frames, and
multisensory integration between the vestibular, vision, and
proprioceptive systems has come from the work of Thomas Mergner
and colleagues (Mergner and Rosemeier 1998). Their model
proposes that the nervous system, through a series of vestibular-
proprioceptive and visual-proprioceptive transformations, learns to
use a global reference frame for multisensory interactions that
develops over a lifetime of experience of inertial and gravitational
forces.

PROBLEM 29.5
Humans can walk on flat surfaces with their eyes closed for short
periods of time, but it is almost impossible to do tightrope walking
with eyes closed. Why is that, and what does that tell us about
multisensory integration and conduction in the visual and
vestibular sensory systems?
Figure 29.6 Multiple sensory modalities contribute to maintaining postural
equilibrium. At any given instance, self-motion signals from the visual system,
proprioceptive information from the neck and lower limb, and vestibular information
are simultaneously integrated. This helps us resolve perceptual ambiguities in real
time, especially when the sensory receptors receive conflicting information from
different sensory modalities. Resolving these conflicts is necessary to be able to
walk in a stable fashion on uneven terrain.

29.5 Neural Correlates of

Multisensory Integration
There are three main sites for multisensory integration in the central
nervous system: the thalamus, the cerebral cortex, and the superior
colliculus. Almost all sensory information (with the exception of
olfaction) reaches the cerebral cortex via the thalamus. The
thalamus has dedicated nuclei that are topographically organized
based on sensory modality, input source, and receptive fields. A
large portion of the sensory input into the thalamus is unisensory.
These principal sensory nuclei are the lateral geniculate (vision),
medial geniculate (audition), and ventroposterior (somatosensation)
nuclei. In contrast, the multisensory nuclei receive converging
information from multiple modalities. These nuclei include the
pulvinar and the suprageniculate nucleus. The cortex receives
multisensory input from the thalamus through thalamocortical and
corticothalamocortical projections.
In the cerebral cortex, the main multisensory area is the superior
temporal sulcus, where visuo-auditory and visuo-somatosensory
integration occur (Bruce et al. 1981; Hikosaka et al. 1988). The
parieto-insular vestibular cortex is located in the mid-posterior
Sylvian fissure and is another multisensory area that is involved in
the integration of vestibular and visual signals (Frank and Greenlee
2018). Multisensory areas contain different types of neuronal
subtypes (bimodal, unisensory) and receive most of their input in the
supragranular layers. As a consequence, multisensory cortical areas
have a reduced layer 4 because they do not receive much input from
the thalamus. Afferent information in multisensory areas converges
from unisensory areas that represent different sensory modalities.
The superior colliculus (SC) in the midbrain is a multisensory area
that mediates fast orienting movements of the eyes, head, and neck
toward salient objects in space by integrating visual, auditory, and
somatosensory spatial information (Meredith and Stein 1993).
Neurons in the SC synthesize and integrate information from
different sensory modalities. The cells in the superficial layers are
primarily dedicated to the visual modality and mainly receive input
directly from the retina, but they also receive input from the visual
cortex and the pretectum. The majority of cells in the deeper layers
respond to combinations of visual, tactile, and auditory information
and are connected to a range of sensorimotor areas of the brain.
The neural responses to multisensory stimuli in the SC are
enhanced when the stimuli in different senses have the same spatial
receptive fields (spatial summation). The responses are also
enhanced when stimuli occur at the same time (temporal
summation). Finally, when at least one of the two stimuli is by itself
only weakly effective in exciting a multisensory neuron, the
combined neural response is enhanced. This is helpful in very
special situations. Imagine you are at party at a friend’s place where
teenagers are playing catch with a baseball. While you are enjoying
a conversation with your friend, a teenager accidentally throws the
ball toward you. You briefly see the ball in your peripheral vision, but
that moving stimulus does not evoke a strong enough neural
response for you to take evasive action. Now if the teenagers also
scream to warn you as the ball is approaching your head, the visual
and auditory neural responses would undergo superaddition and
produce a very strong neural response in the SC. That would make
you aware of the impending danger and force you to reorient your
head to avoid collision with the ball. The superaddition here implies
that the multisensory sum is greater than the arithmetic sum of its
unisensory parts.
CHAPTER 29 IN A NUTSHELL
Multisensory integration is the
process of integrating and
synthesizing sensory information from
different modalities. The nervous
system integrates sensory information
from different modalities in a
statistically optimal fashion. In
other words, while making sensory
judgments, the nervous system combines
sensory information from different
modalities based on their
reliabilities or the levels of sensory
noise in those systems. Multisensory
integration between the
proprioceptive, visual, vestibular,
and somatosensory systems subserves
balance and posture stabilization. The
central nervous system has dedicated
cortical and subcortical regions for
processing multisensory stimuli. In
particular, the parietal cortex,
thalamus, and superior colliculus are
important neural structures for
multisensory integration.
 Chapter 30

Visual Perception and Action

KEY TERMS AND TOPICS

visuomotor coordination
visual perception
magnocellular cells
parvocellular cells
dorsal and ventral streams
motion processing
form processing
oculomotor control
eye–hand coordination

Imagine a game of American football where the quarterback is trying

to throw the football to a wide receiver while evading charging
linebackers who are concealed behind a wall of players from both
sides who are all moving in different directions. To throw the ball
accurately to the intended receiver, the quarterback must exhibit
nimble visual perception and judgment, predictive motion
processing, and precise visuomotor coordination. The amount of
sensory processing that occurs within those few seconds between
the snap and the throw is remarkable. The skills exhibited by a
talented quarterback in that short time span epitomize the
physiological limits of communication between the visual system and
the limb motor system in the service of skilled motor actions. In this
chapter, we will define the neurophysiological bases of the diverse
visual sensory processes that occur within the nervous system
during the execution of complex visuomotor skills.

30.1 Two Visual Streams

One of the most significant findings of the 20th century in the field of
visual neuroscience was the discovery of the two parallel visual
streams (Goodale and Milner 1992; Mishkin et al. 1983). Though
recently some overlap in function between the two streams has been
discovered, the two streams operate relatively independently
(Farivar 2009). The support for the independence of these streams
has come from many behavioral studies, including reach-to-grasp
motor control studies. One line of investigation has used the
Ebbinghaus illusion. This illusion is an optical illusion in size
perception, where a circle surrounded by smaller and larger circles
appears to be larger or smaller. But if participants make reach-to-
grasp movements, their grasp apertures are not affected by this
illusion. So the perceptual system is duped into misjudging the
relative sizes of objects, but the motor system is not.
The dorsal stream supports visuospatial and visuomotor functions,
and it consists of the primary visual cortex (V1), the extrastriate
middle temporal (MT) and medial superior temporal (MST) areas,
and the posterior parietal cortex (PPC). It allows us to locate objects
in space and to interact with those objects, but it does not let us
identify those objects. This stream will enable the quarterback to see
where different players are on the field, how far they are from him
and each other, and how fast and in which direction they are moving,
but it won’t let him differentiate his teammates from his opponents.
That would require the quarterback to differentiate colors of jerseys
and recognize faces (hidden under the helmets). Color, face
recognition, and other perceptual functions are performed by the
ventral visual stream in areas V4, V8, and the fusiform face area,
respectively. The ventral stream consists of V1, extrastriate areas
V2, V4, and V8, and the inferior temporal cortex. The fusiform face
area is a region in the inferior temporal cortex, and it plays an
important role in face recognition. These two streams work together
during almost all activities of daily living to help us perceive the
external world as well as act in it.

30.2 Magnocellular and

Parvocellular Ganglion Cells
and Streams
The retinal ganglion cells in primates (including humans) can be
divided into two broad categories of cells based on their overall
appearance, neural connections, and electrophysiological features.
Magnocellular (M) ganglion cells have larger cell bodies and
dendritic trees and larger-diameter axons than the parvocellular (P)
ganglion cells. M ganglion cells also have larger receptive fields, and
their axons conduct information faster in the optic nerve than the P
ganglion cells. However, the P ganglion cells represent about 90% of
the total population of ganglion cells. The M ganglion cells only
account for 5% to 10% of the overall population.
Besides morphological differences, The M- and P-type ganglion
cells also differ in how they respond to visual stimuli. P ganglion cells
can encode and transmit information about color, but the M ganglion
cells cannot. The response of an M cell to visual stimulus is phasic
(transient), whereas the response of P cells is more tonic, and it
continues as long as the visual stimulus is active. These features
allow M cells to detect motion and P cells, with smaller receptive
fields, to detect shape, color, and form.
M and P ganglion cells terminate in different layers of the lateral
geniculate nucleus (LGN). As was discussed in chapter 14, there are
six layers in the LGN, and each layer receives input from only one
eye. In addition, these layers also differ with respect to the size of
the cells they receive input from. The M cells terminate in the
magnocellular layers (layers 1 and 2) of the LGN, and the P cells
terminate in the parvocellular layers (layers 3-6). The information
from these layers remains separate in the early stages of cortical
processing. Signals from these layers get mixed in V1. Thus,
information derived from M and P cells are called the magnocellular
and parvocellular streams, respectively. The dorsal and ventral
visual pathways are driven predominantly by input from the
magnocellular and parvocellular layers of the LGN, respectively.
As stated earlier, the dorsal stream is involved in the guidance of
manual actions and recognizing visuospatial relationships among
objects. Damage to the dorsal stream can cause optic ataxia
(inability to use visual information to make goal-directed limb
movements), akinetopsia (inability to perceive motion), and spatial
neglect (loss of spatial awareness in the workspace contralateral to
the side of the lesion). In contrast, the ventral visual stream is
responsible for object, shape, color, and face recognition. Damage to
this stream causes visual form agnosia (inability to recognize visually
presented objects despite intact elementary sensory functions),
prosopagnosia (deficits in face perception), and achromatopsia
(cortical color blindness).

PROBLEM 30.1
How is cerebral achromatopsia different from the color blindness
discussed in chapter 14?

30.3 Motion Processing in the

Cortex
Whether it is an animal trying to detect the motion of a prey in the
wild, or a quarterback trying to connect a pass to a teammate, the
ability to detect and process motion is critical for the success of
many animal endeavors. Thus, animals need dedicated neural
circuitry to process motion quickly and accurately. The magnocellular
pathway serves this purpose by processing the movement of objects
in space. Surprisingly, in primates, directionally selective ganglion
cells are absent in the retina (Ilg 2008). It is only in the thalamus that
the magnocellular and parvocellular streams begin to diverge, and
the first directionally sensitive cells are recorded in V1.
These cells from V1 provide input to two areas of the brain, the
MT and MST cortices, the main motion-processing cortical areas.
The MT area is in the posterior bank of the superior temporal sulcus.
Neurons in MT respond to visual motion in a direction-selective
manner. MT is considered to be the gateway to motion perception
because it is the first cortical visual area that is largely dedicated to
motion processing. To test how MT neurons respond to motion,
researchers present, as visual stimulus, a group of dots moving
coherently located within the receptive fields of MT neurons (see
figure 30.1). If most of the dots move in the preferred direction of a
neuron, the firing rate of that MT neuron increases above the
baseline values. If the stimulus moves in the opposite direction, the
firing rate decreases below the baseline values.

Figure 30.1 Large-field visual motion elicits strong neural activity in neural areas
MT and MST. (left) Large-field visual motion of random dots where all dots are
moving in the same direction to the left (100% coherence). (right) All dots are
moving randomly in different directions. In typical experiments, 3% to 50% of the
dots move in the same direction and the rest move randomly.
 MST is also located in the superior temporal sulcus. MST neurons
receive input directly from MT. There are two types of neurons in
MST. First, neurons with smaller receptive fields are important for the
execution of smooth-pursuit eye movements. Recall that smooth-
pursuit eye movements are slow eye movements (maximum speed
of 80º/s to 100º/s) that are involved in tracking the motion of moving
stimuli. Second, MST neurons with larger receptive fields are
involved in the perception of optic flow and self-motion. In addition,
MST also receives input from parietal areas that are involved in
processing vestibular signals. This helps with distinguishing motions
that arise from external sources versus those that arise from self-
motion. Thus, MST plays an important role in distinguishing different
sources of motion, self-generated and externally generated motion.
Damage to MT/MST (through injury) or temporary inactivation
(through experimental manipulation) causes deficits in visual motion
perception. This can range from mild deficits, such as reduction in
the accuracy of direction discrimination at small spatial
displacements (Rudolph and Paternak 1999) to more pronounced
disorders, such as akinetopsia. A quarterback who had this deficit
would see an opposing linebacker moving toward them as “jumping”
from one stationary position to another, instead of running toward
them in one continuous motion.

30.4 Color, Object, and Face

Recognition in the Ventral
Stream
Before reaching and manipulating objects, we must first look at those
objects and identify them. When we walk on the street and identify a
familiar face, we wave at the person. Identification of objects and
faces is fundamental to many actions we perform during daily living.
Functional neuroimaging studies have revealed that many areas in
the ventral stream are involved in form processing (i.e., the
recognition and identification of colors, shapes, and faces). V1
projects to the extrastriate cortical areas, which are located anterior
to the visual cortex, for higher-order visual processing. Some of
these extrastriate areas, such as V4 and V8, are involved in color
processing (Goddard et al. 2011; Hadjikhani et al. 1998). The lateral
occipital cortex (LOC) is involved in shape processing and object
recognition (Grill-Spector et al. 2001). This region is located on the
lateral bank of the fusiform gyrus. While general features of object
shape are processed in LOC, it appears that LOC responds to
almost any shape. More anterior areas along the ventral stream in
the temporal cortex respond to different recognizable objects.
Specifically, it appears that objects are represented along the ventral
stream by many different groups of neurons that are clustered near
each other (Reddy and Kanwisher 2006). Finally, the fusiform face
area is an important node along the ventral visual stream in the
inferior temporal cortex that is involved in both object and face
recognition (Kanwisher et al. 1997).
It is likely that both upstream extrastriate areas, with smaller
receptive fields, and downstream areas in the temporal cortex with
larger receptive fields act in concert to distinguish different shapes,
objects, and faces. The size of the receptive field of V2 is about 1.5º,
of V4 about 5º, and of the inferior temporal cortex about 12º
(reviewed in Kravitz et al. 2013). Because of the larger receptive
fields, downstream neurons of the ventral stream can process more
complex features of object shape and form (see figure 30.2).
There is also new neuroanatomical evidence that shows that
white matter tracts connect the dorsal and ventral visual streams.
These tracts are assumed to play an important role in relatively
complex motor skills, such as reach-to-grasp and lifting movements
(Budisavljevic et al. 2018). It appears that the reach and lift
components of the movement are mediated primarily by the dorsal
stream, whereas shaping the fingers to grasp objects of different
sizes and shapes requires shape processing in the ventral stream
(see section 30.5).
Figure 30.2 Information flow along the ventral visual stream and the size of the
receptive fields of neurons obtained from studies done on primates. IT =
inferotemporal cortex.
Adapted from D.J. Kravitz, K.S. Saleem, C.I. Baker, L.G. Ungerleider, M. Mishkin, “The
Ventral Visual Pathway: An Expanded Neural Framework for the Processing of Object
Quality,” Trends in Cognitive Science, 17, no. 1 (2013):26-49.

PROBLEM 30.2
Which other activities of daily living are critically dependent on
simultaneous neural processing in both the dorsal and ventral
visual streams?

30.5 Feedforward and Feedback

Control for Reach-to-Grasp
Movements
Adult humans make reach-to-grasp movements accurately and
smoothly during many activities of daily living. These fundamental
movements seem simple to learn and easy to perform, but the
prolonged process of iterative and trial-and-error–based learning
observed in infants, and the difficulties associated with relearning the
movement after a neurological insult, underscore the visuomotor and
visuospatial complexities the nervous system overcomes to make
these movements possible. Reach-to-grasp movements require
dynamic interactions between the two visual streams and the motor
areas in the frontal cortex. The synchronized interactions between
these areas to produce stable motor output develop under visual
guidance and over sustained periods of time (weeks to months).
When children first learn to make reach-to-grasp movements, they
alternate their gaze between the target and their hand (reviewed in
Corbetta et al. 2018). This causes children to make many small,
discrete submovements instead of one smooth motion to make a
reach-to-grasp movement. These submovements reflect a
predominant closed-loop feedback process that dictates how
movement is continuously controlled and adjusted during the early
stages of motor learning. Here, by closed-loop we imply a movement
that is made entirely under visual supervision with fixations on the
target and the hand and constant corrections of limb trajectories. As
children gradually become more and more proficient in making these
movements, the control transitions to incorporate open-loop
feedforward control where they only look at their hand (or target) at
the beginning of the movement, and then only make corrective
adjustments if they believe they would miss the target during the
movement.
Humans typically employ a combination of both feedforward and
feedback processes during reaching movements. This was first
shown by R.S. Woodworth (Woodworth 1899) in an elegant
experiment where he asked participants to make horizontal back-
and-forth movements with a pencil on paper secured to a drum
rotating at a constant speed. Participants performed the movements
either between lines that were a fixed distance apart or that matched
the amplitude of their movement to a previous attempt. Woodworth
measured the spatial accuracy of the movement as well as the
spatiotemporal trajectories. He noticed that the initial movement was
ballistic and stereotyped, but as the pencil approached the target,
the movement slowed down and showed spatiotemporal
discontinuities. Woodworth concluded that the initial ballistic phase
was under feedforward control that brought the hand close to the
target. This process is likely completed before the beginning of the
movement. When the hand gets closer to the target, it is then under
feedback control to improve performance accuracy. This model of
speed–accuracy relations, which involves a feedforward and a
feedback-based component in the control of upper limb reaching
movements, is known as Woodworth’s two-component model.
Woodworth’s model has withstood the test of time and numerous
subsequent investigations. These feedforward and feedback
processes involved in eye–hand coordination for reaching and
pointing movements are regulated by the dorsal visual stream. For
feedforward control, visual acquisition of the target is important to
make comparisons between hand and target locations. This also
involves coordination and transformation of both target and hand
location within the same reference frame (also discussed in chapter
29). For feedback control, it is important to continuously compare
both efferent and afferent information between different motor (ocular
and limb) and sensory systems (limb proprioceptive and vision) while
the movement is underway. This involves constant comparison
between the movement goal and the current state of the limb to
minimize movement deviations away from the intended goal (see
section Eye and Hand Coordination During Movement). These
comparisons are likely performed by the posterior parietal cortex.

30.6 Neural Structures Involved

in Oculomotor Control
During reaching and pointing movements, humans first make a
saccade to the target or hand location. Making a successful saccade
requires interactions between several cortical and subcortical
oculomotor structures. Normal visual perception necessitates the
proper functioning of the oculomotor system to control eye position
and its movement to focus on objects of interest. The anatomically
closest neural structures to the saccade execution are the saccadic
burst generators (SBGs), two sets of nuclei of the reticular formation
that directly drive the ocular motoneurons arising from the
oculomotor, abducens, and trochlear nuclei (Girard and Berthoz
2005). The job of the SBGs is to produce eye movements based on
instructions from higher-level cortical structures (see figure 14.12).
The superior colliculus (SC), a midbrain area, along with the frontal
eye fields (FEF), a premotor structure in the frontal cortex, are the
main structures involved in sending saccadic commands to the SBG.
The SC integrates sensory information from the visual,
somatosensory, and auditory systems, as well as cognitive signals
that carry information on attention and motivation. It is both a
sensory and motor structure and is the final common pathway for
saccadic eye movements. The SC sends a motor command to the
gaze centers in the brainstem (paramedian pontine reticular
formation and the medial longitudinal fasciculus) to rotate the eyes.
The SC receives input from the prefrontal cortex, the FEF, PPC,
and the basal ganglia. The prefrontal cortex provides input to the SC
based on the level of visual attention. The FEF provides input to the
SC to select targets for saccades from several available potential
goals and at the same time to suppress reflexive saccades to
nonessential targets (e.g., if we are looking for a glass of water on a
table, we would saccade to the glass and inhibit saccades to other
irrelevant objects). The input from the lateral intraparietal area (LIP)
in the parietal cortex determines the spatial location of the saccade,
as well as the latency of the saccade (Ipata et al. 2006). Finally, the
input from the basal ganglia prevents unwanted saccades by
inhibiting the SC. Voluntary saccades are made possible by
disinhibition arising in the caudate nucleus of the basal ganglia,
which is itself activated by signals from the frontal cortex.
The output from the SC to the brainstem is modulated by the
cerebellum. The cerebellum plays an important role in coordination
and timing of movements. Indeed, if the caudal fastigial nucleus in
the cerebellum is lesioned, saccades are inaccurate, slow, and
abnormally variable in size and speed. The input from the
cerebellum also contributes to saccadic gains (ratio of saccadic
amplitude and target distance). Lesions of the oculomotor vermis
cause hypometric leftward and rightward saccades (reviewed in
Robinson and Fuchs 2001).
During most activities of daily living, there are numerous items
within our visual field that compete for attention. But our nervous
system filters out most of the irrelevant information and directs our
attention to those spatial locations where there is maximum
information available that is relevant to the task at hand. For
example, if we are driving to a new place, we are focused on the
directions provided by the GPS (or road signs), while ignoring any
other visual stimulation that might be competing for attention. Thus, it
is not surprising that the brain has many dedicated cortical and
subcortical areas dedicated for saccadic eye movements and visual
attention.
The neural regions involved in smooth-pursuit eye movements are
somewhat different from those involved in saccadic eye movements
(see chapter 14 for more details). These regions are shown in figure
14.12. The goal of the smooth-pursuit system is to match the velocity
of the eyes to the moving target in space. Smooth-pursuit eye
movements also involve MT and MST, two motion-sensitive areas in
the brain that process object motion. The FEF play an important role
in initiating pursuit eye movements. In addition, cerebellum neurons
in the vermis and flocculus transmit eye velocity signals to the
brainstem areas. Currently, it is unclear if the basal ganglia play a
role in pursuit eye movements, but that is an active area of
investigation.

PROBLEM 30.3
If your friend threw a ball toward you from a distance to catch,
which eye movements would you be making to track the ball?
30.7 Roles of Two Visual
Streams in Eye–Hand
Coordination
Before limb movements are initiated, the visual system must select
an action goal from among many viable alternatives. For example,
the quarterback could throw to any of the teammates who might be
free to receive the ball. But a good quarterback would also assess
which of those teammates have the maximum chance of carrying the
ball farthest downfield. Thus, the expectation of reward or outcomes
also determines which action we select in a situation. The cortical
areas represent targets for action (dorsal visual stream) and
perception (ventral visual stream) and act with the reward centers
(basal ganglia) to execute complex movements. The frontoparietal
cortical areas along the dorsal stream facilitate automatic control of
action, and the inferotemporal areas along the ventral stream also
contribute to conscious judgments of those actions.
The PPC is a critical node in the dorsal visual stream responsible
for the automatic control of eye and hand movements. The main role
of the PPC is target localization for static targets (reaching
movements) and rapid spatial updating of target locations for
dynamic targets (interception movements). Specific areas within the
intraparietal sulcus (IPS) serve as interfaces between the sensory
and motor systems for controlling arm and eye movements
(reviewed in Grefkes and Fink 2005; Vesia 2012). Motor commands
for reaching movements are prepared based on input from both
reach-related regions (such as the parietal reach region) and eye
movement–related regions (such as the LIP) in the parietal cortex.
These regions communicate with frontal areas involved in limb motor
planning (premotor cortex) and eye movements (FEF). Once
movement is initiated, the frontoparietal circuits are simultaneously
activated (Naranjo et al. 2007) to minimize visuomotor errors through
continuous comparisons between target locations and hand
trajectory (Franklin et al. 2016; Binsted et al. 2010).
The role of the ventral visual stream in eye-hand coordination is
not entirely clear. But it is likely that it plays a role in grasping during
reach-to-grasp movements. If we are trying to grasp a cup versus a
pen with a precision grip, the grasp component of the movement
would be different in the two cases. The precision grip is facilitated
by canonical neurons in the ventral premotor cortex. These neurons
discharge during motor execution and in response to the
presentation of three-dimensional objects (Rizolatti et al. 1988).
These neurons receive extensive input from the lateral parietal areas
and are considered to encode important object dimensions that allow
us to calibrate our digits to the dimensions of the objects. The
canonical neurons most likely receive input from the ventral visual
stream through the ventrolateral prefrontal cortex (reviewed in
Kravitz et al. 2013; Thorpe and Fabre-Thorpe 2001). This pathway
from the ventral visual stream to the frontal cortex and ventral
premotor cortex would act in a feedforward fashion to shape the
grasp aperture in preparation for contact with the object of interest
(see figure 30.3).

30.8 Eye and Hand Coordination

for Movements Starting From
Rest
When humans make a reaching or a pointing movement to a target,
a saccadic eye movement is initiated before the onset of limb
movement. The exact reason for why the eyes move before the limb
is not clear, but the most common hypothesis is that the larger limb
inertia causes a delay in movement initiation. During a reaching
movement, humans also anchor their gaze to the target location until
the hand reaches the target (Neggers and Bekkering 2000), though
this mostly happens in discrete reaching tasks and not in continuous
natural tasks (Hayhoe and Ballard 2005; Singh et al. 2018). In
continuous tasks, look-ahead saccadic eye movements are made to
acquire multiple target features and locations.

Figure 30.3 The neural network involved in judging object shape to plan and
execute the finger movements for the grasp component of the reach-to-grasp
movement. IT = inferotemporal cortex; PFC = prefrontal cortex; PMd = dorsal
premotor cortex; M1 = primary motor cortex.
Adapted from S.J. Thorpe and M. Fabre-Thorpe, “Neuroscience - Seeking Categories in the
Brain,” Science, 291, no. 5502 (2001), https://doi.org/10.1126/science.1058249. PMID:
11253215.

In discrete reaching tasks, whether the eye movement is initiated

before or after the limb movement, the gaze almost always arrives at
the target before the limb reaches the target. Then the steady ocular
fixation on the target provides high spatial resolution of the target
location during an ongoing movement. Target foveation improves
reaching kinematics and spatial accuracy of the reach; if the gaze
deviates from the target location, the reach accuracy is degraded.
An interesting set of results shown in many studies is that eye and
hand movements influence each other’s kinematic profiles (reviewed
in Binsted et al. 2010). For example, saccadic peak velocity is faster
when the saccade is also accompanied by an arm movement
(Snyder et al. 2002). This suggests that reaching toward a target
under visual control may necessitate a common integration of both
eye and hand movements. This is not to say that eye and hand
movements cannot be adaptively decoupled. Most healthy humans
can easily continue looking at their computer monitors while reaching
out to pick up a cup of coffee. In contrast, neurological impairments,
such as damage to the PPC, can cause the arm to obligatorily follow
the eye movements even when the person makes a conscious
decision not to do that. This phenomenon has been termed
magnetic misreaching by Carey and colleagues (Carey et al.
1997).

Figure 30.4 The Müller-Lyer illusion causes a perceptual difference in the

apparent length of a line as a consequence of how the arrowheads are oriented.
The line on the bottom appears to be longer, although both lines are the same
length.

The hypothesized common integration of eye and hand

movements may occur through a common motor command, which
might manifest as different spatiotemporal trajectories of eye and
hand movements because of the different inertias. This is reflected in
the coupling between the reaction times of eye and hand
movements. This has been called the common command
hypothesis. Though the simplicity of this hypothesis is attractive, this
hypothesis does not withstand stricter scrutiny. Trial-by-trial analyses
of eye and hand movements show little correlation of the endpoint
location of the eye and hand positions. Furthermore, using visual
illusions, such as the Müller-Lyer illusion (see figure 30.4),
researchers have shown that eye and hand movements are biased
differently when making movements to the Müller-Lyer target
(Binsted et al 2001). The Müller-Lyer illusion is a geometrical illusion
that consists of two horizontal line segments that are perceived to
have different lengths depending on whether the arrowheads point
inward or outward. The hand movement was accurate, even though
the eye tended to undershoot or overshoot the target position. These
results sow serious doubts about the tenability of the common
command hypothesis as a global descriptor of how motor commands
for eye and limb movements are generated.
The common command hypothesis is also untenable because
ocular and limb manual reaction times and performance depend on
instructions provided to the participants. If spatial accuracy for the
limb is emphasized as a primary requirement for eye and limb
movements to peripheral targets, then eye movement is initiated
before the limb movement. In contrast, if speed is emphasized for
the limb movement, then hand movement is initiated before the eye
movement (Binsted et al. 2010). This suggests that the nervous
system can flexibly couple and decouple eye and hand movements
based on task requirements. This flexibility allows humans to
perform both fine and gross motor tasks that require different
spatiotemporal coupling between eye and hand movements. So if we
are putting a thread in a needle, the eye and hand movements will
be yoked together. But if we are changing a radio channel while
driving a car, we can focus on the road while the hand turns the
knob.

30.9 Eye and Hand Coordination

During Movement
The previous segment focused on eye–hand coordination when both
the eye and hand movements were initiated from rest. However,
humans often respond to environmental exigencies by changing a
movement while the movement is already under way. For example,
we may be reaching for a cup of coffee when we suddenly realize
that the hand might bump against a book. We will then change the
hand movement trajectory to avoid the book. This can be easily done
without bringing the hand back to its original resting place. In sports,
players routinely change their movement plans during a movement.
The ability to change motor plans during an ongoing movement is
also under the control of the PPC.
Typically, online movement correction is studied in the laboratory
by jumping the spatial location of the target or by moving the actual
feedback of the hand cursor on an augmented-reality display (see
figure 30.5). In these experiments, participants move a cursor shown
on a visual display to a target shown on the same display. The
induced error is detected by the visual system, specifically the
posterior parietal cortex, before a corrective limb motor response is
initiated. The posterior parietal cortex updates target location relative
to the current location of the hand (Prablanc et al. 2003).

Figure 30.5 Typical task design for online feedback control experiments.
Participants start from rest. Then a peripheral target (shown on the right) comes
on, and the participant initiates a movement toward the target. While the
movement is underway, a secondary target comes on and the participant aborts
the previous motor plan and moves toward the new target. Inset shows the entire
experimental setup.

The role of the PPC for online feedback correction was first
studied by Mountcastle and colleagues. They showed that certain
neurons in the PPC only fired during reaching movements to a target
(Mountcastle et al., 1975). These neurons did not fire either when
the target was presented, or if the arm moved toward a nonvisual
target. The neurons fired only when the arm moved toward a visual
target and stopped firing once the arm reached the target. This
suggests that the neurons encode the error signal between the arm
and the target. Correction of a hand trajectory toward a displaced
target during reaching occurs without conscious control (Goodale et
al. 1986) and is often referred to as the “automatic pilot.” These
automatic corrections during target movements occur at a short
latency (within 150 ms) (Day and Lyon 2000; Kadota and Gomi
2010) compared to the typical reaction times observed for reaching
movements when the movement is initiated from rest (~200-250 ms).
The involvement of the PPC in online corrections has been further
corroborated by a study done by Pisella and colleagues (Pisella et
al. 2000). They tested a stroke patient with bilateral lesions to the
parietal cortex. The patient had optic ataxia and accurately pointed
to stationary targets, but if the target was displaced at movement
onset, her ability to correct an ongoing movement was severely
impaired. This study suggests that the parietal cortex plays a central
role for online feedback control to regulate reaching movements in
response to sudden and unexpected changes in target location.

PROBLEM 30.4
Rapid online control is quantified in many studies that focus on
developmental coordination disorders (DCDs). Many children with
DCDs can make simple reaching movements with relative ease,
but they exhibit impairments in the ability to make rapid online
corrections. Even healthy adults sometimes experience difficulties
executing movements that involve rapid online control. Why would
movements that involve rapid online control be “harder” to learn
and execute than simple reaching movements?
30.10 Eye and Hand Coordination
While Intercepting Moving
Targets
Intercepting and catching a moving projectile are fundamental motor
skills that require precise timing and coordination between the eye
and the hand. The moving projectile is tracked with smooth-pursuit
eye movements, and then the limb is directed toward the point of
impending collision. The main difference between reaching and
interception movements is that in contrast to saccadic eye
movements that are made during reaching, interception movements
primarily engage smooth-pursuit eye movements. Tracking moving
objects enables continuous high-acuity vision, but smooth-pursuit
eye movements are limited to a maximum speed of 80º/s to 100º/s
(Meyer et al., 1985). Thus, for stimuli that move faster than that
speed, the oculomotor system launches catch-up saccades to
compensate for position and velocity errors.
During catching or interception action in sports, players track the
moving object with a combination of eye and head movements until
the hand is aligned with the moving object at the point of interception
(reviewed in Fooken et al. 2021). Often there is a short phase of
smooth tracking observed right before the ball is intercepted,
suggesting that keeping the eye on the object is important for
accurate prediction of object motion (Hayhoe et al. 2012; Land and
McLeod 2000). If the eye movements are constrained (e.g., when
participants are asked to fixate at a point instead of freely pursuing
the moving projectile), their performance is degraded. The most
likely explanation for this is that object speed during visual fixation is
overestimated, leading to interception movements ahead of the
moving object.
It is currently also believed that the nervous system uses
efference copies of oculomotor commands associated with smooth-
pursuit eye movements to make predictions about object motion and
coordinating eye and hand movements for interception. The reliance
on both retinal projections of the object image and the extraretinal
oculomotor signals might be more important for smooth eye–hand
coordination when the object motion is more unpredictable (e.g.,
intercepting a curving soccer kick or catching a fly) (see figure 30.6).
The studies on manual interception with human participants and
primates have highlighted the important role played by the MT/MST
and FEF areas and the parietal cortex in the precise spatiotemporal
coordination of both eye and limb movements (Merchant et al. 2004,
2006; Zago et al. 2009). Lesions in the parietal cortex, FEF, and
MT/MST cause deficits in smooth-pursuit eye movements (Dursteler
and Wurtz 1988; Heide et al. 1996) and consequently in eye–hand
coordination.
MT and MST project to the premotor cortex, especially the FEF
and the dorsal premotor cortex. These projections likely are involved
in continuous transformation of motion signals from a projectile into
motor commands to move the limb to the right location in space to
make contact with the projectile. Simultaneously, the nervous system
also prepares a postural response to absorb the projectile’s
momentum.
Evidence for the transformation of motion signals to continuous
modification of limb posture has come from a study by Selen and
colleagues (Selen et al. 2012). In their experiment, participants
viewed a dynamic random dot display and indicated their decision
about the direction of the dot movement by moving a handle to one
of two targets. The random dot (figure 30.1) display engages neural
area MT/MST, which is also involved in tracking moving objects. The
experimenters perturbed the arm at random times during decision
formation. The long-latency reflex gains of the M3 component (also
see chapter 18) were modulated by the strength and duration of
motion of the random dots, reflecting the accumulated evidence in
support of the evolving decision. Their results support the existence
of a sensorimotor process that continuously transforms motion
signals for limb motor control. Furthermore, this suggests that the
motor system might continuously set the referent joint configurations
for catching movements based on the sensory output from area
MT/MST.

Figure 30.6 It has been hypothesized that oculomotor extraretinal signals are
used as a feedforward input to form predictions and guide eye and hand
movements. Humans rely predominantly on retinal signals when the object moves
predictably. Unpredictable motion of a visual stimulus causes stronger reliance on
extraretinal motor commands associated with eye movements.
Part a adapted from J. Fooken, P. Kreyenmeier, and M. Spering, “The Role of Eye
Movements in Manual Interception: A Mini-Review,” Vision Research, 183 (2021): 81-90.

CHAPTER 30 IN A NUTSHELL
The visual and oculomotor systems must
work in synchrony for normal visual
perception. The visual system has
dedicated streams and specialized
neurons for visuomotor and perceptual
functions. The dorsal and ventral
visual streams facilitate visuomotor
and perceptual function, respectively.
The oculomotor system governs eye
movement kinematics and directs the
fovea to static target locations or to
follow moving objects. For reaching
movements, target foveation allows the
visual system to extract information
about finer details of the object and
localize targets in space with high
spatial resolution. The parietal
cortex, specifically the dorsal visual
stream, plays a critical role in
spatial planning of movements, as well
as online correction of movement
trajectories. For interception
movements, target foveation is
important for accurate prediction of
object motion.
Problems for Part VI
Self-Test Problems
1. Suggest examples of kinesthetic illusions induced by distorted
or inappropriate efferent copy signals.
2. You participate in an experiment where a researcher draws
one of four different weights of 1, 2, and 4 lb from a box and
places it on your left hand. She then takes another weight
drawn from a different box that could weigh anywhere from
0.5 to 4 lb in steps of 0.1 lb and places it on your right hand.
She then asks you to judge if the weight on the right hand is
heavier or lighter than the weight on the left hand and by how
much (approximately). She keeps drawing the weights till you
report that both weights are approximately the same. The
same process is repeated multiple times with different weights
on the left hand. She then compiles a score of how off you
were (in pounds) for each of the weights placed on the left
hand. Are you likely to be more accurate for the 1 lb weight or
the 4 lb weight? Why?
3. Tasks such as throwing a basketball free throw require
dynamic multisensory integration. First, the visual system
must localize the hoop with respect to the body. Then the
proprioceptive system judges where the joints are with respect
to the body and the force applied by the limb. If a person has
suffered damage to the proprioceptive system at the cervical
level, how would they be able to perform a free throw?
4. The reference frames in which sensory stimuli from different
modalities are encoded are different from each other and from
those of motor effectors. For example, visual information is
encoded in a retinotopic reference frame, but proprioceptive
information is encoded in a joint-based reference frame.
Neurons in the parietal cortex transform sensory signals into a
common reference frame to guide movements. If neurons that
transform visual and vestibular signals into a common
reference frame were to be inhibited selectively with a drug,
what kind of impact would we expect to see on motor
 performance during the execution of complex motor skills that
require posture stabilization and fine motor control?
5. The sensation of illusory self-motion is called vection. We
sometimes experience an instance of this sensation when a
car parked right next to ours suddenly moves. In contrast, this
sensation is almost never experienced when we ride a bicycle
and the bicycle stopped next to ours suddenly moves. What
might explain this discrepancy?
6. How would the absence of canonical neurons affect reach-to-
grasp actions? What would we expect to see with the
movement of the digits?

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. Which sensory receptors are primarily targeted during
vibration-induced kinesthetic illusions?
a. primary spindle endings
b. secondary spindle endings
c. Golgi tendon organs
d. Merkel disks
e. articular receptors
Why?
2. The nervous system weights individual sensory modalities
based on how spatially reliable the information in each
modality is. Let’s assume that proprioception, vision, and
vestibular sensations are all weighted equally when walking
on a nonslippery, even surface during daytime. How would
this weighting change while walking on a slippery, uneven
surface during nighttime?
a. Vision would be downregulated and proprioception would
be upregulated.
 b. Vision would be upregulated and proprioception would be
upregulated.
c. Proprioception and vestibular input would be
downregulated.
d. Proprioception and vestibular input would be upregulated.
e. Vision would be downregulated, and proprioception and
vestibular input would be upregulated.
Why?
3. Which sensory systems play an important role for referencing
task-relevant sensory information to the external world?
a. proprioceptive system
b. visual system
c. vestibular system
d. both b and c are correct
e. a, b, and c are correct
Why?
4. Which area plays the most important role in online feedback
control during reaching movements?
a. prefrontal cortex
b. parietal cortex
c. premotor cortex
d. temporal cortex
e. motor cortex
Why?
5. Which neural areas likely play a minimal role in manual
interception tasks?
a. MT/MST
b. frontal eye fields (FEF)
c. premotor cortex
d. parietal cortex
e. temporal cortex
Why?
 Part VII

Emerging, Evolving, and

Adapting Movements
 Chapter 31

Fatigue

KEY TERMS AND TOPICS

muscle mechanisms of fatigue
spinal mechanisms of fatigue
supraspinal mechanisms of fatigue
adaptations during fatigue
atypical fatigue

In the next three chapters, we consider the neurophysiological

mechanisms of movements that may be viewed as suboptimal, while
persons showing such movements may be healthy. The reasons for
such apparent motor imperfections may differ. For example, babies
are still developing, typically or atypically, and learning how to use
their bodies, while elderly persons have to deal with unavoidable
consequences of aging in the neuromuscular system. In the fourth
chapter, we discuss the topic of motor learning, which in our opinion
also reflects adaptations and emerging new movement patterns. We
begin with a state that leads to temporarily impaired motor
performance at any age and at any level of skill: This is fatigue.

31.1 Fatigue and Its

Contributors
Human beings demonstrate a striking difference from machines in
their reaction to prolonged use. Machines basically deteriorate with
functioning, and the best machines are those that do not require
repair for a long time. The situation with the “human machine” is the
opposite, at least on a certain time scale. The longer a human being
practices a certain activity, the better he or she performs it. Humans
not only avoid becoming worse with exercise but, quite the opposite,
they become stronger and quicker and show higher endurance and
dexterity, particularly with respect to the type of activity that has been
performed. This feature of living organisms has been termed
exercisability.
However, over relatively short time intervals, humans can
demonstrate machinelike behavior; their performance declines, and
this decline may spread to other types of activities. This
phenomenon is commonly called fatigue. Among most apparent
causes of physical fatigue, we can identify shortage of energy
sources for muscle work, and the inability of the circulatory system to
quickly remove the products of muscle metabolism (lactic acid being
probably the best-known one). However, fatigue is a complex
phenomenon that may involve factors at different levels that
contribute to the overall decline in performance. These may include
a decrease in the ability of muscle fibers to generate force,
a decrease in the efficacy of neuromuscular synapses,
changes in the activity of certain peripheral receptors leading to
changes in their reflex effects,
changes in the patterns of firing (recruitment patterns) of α-
motoneurons,
changes at any level of the hypothetical process of generation of
a motor command, and
psychological factors including, in particular, motivation.
Some of the changes may be adaptive. This means that they
represent responses of the body to direct effects of exercise that
attenuate negative effects of fatigue on the neuromuscular function.
 For the purposes of our discussion, we will address all changes
that happen in the neuromuscular synapse and in the muscle as
peripheral, and all changes that occur earlier (in the α-motoneurons
and other neurons within the spinal cord, and in the brain) as central.
There are rather straightforward methods for studying the relative
contribution of peripheral and central factors in fatigue. In particular,
if a short episode of electrical stimulation is applied to a muscle or to
its motor nerve, a direct muscle contraction can be induced. If a
fatigued muscle demonstrates contractions in response to direct
electrical stimulation with characteristics that are different from those
in the same muscle in a nonfatigued state, apparently all these
differences may be attributed to peripheral factors (Bigland-Ritchie et
al. 1986). On the other hand, if the response of a muscle to direct
electrical stimulation is unchanged, while its force during voluntary
activation decreases, the decrease in force may be attributed to
central factors.
For example, imagine that a standard brief electrical stimulation is
applied to the quadriceps muscle at the beginning and after 1 min of
a steady-state contraction at different levels of knee joint torque
ranging from 25% (which is not expected to induce fatigue) to 100%
(which is expected to lead to fatigue) of the maximal voluntary
contraction (MVC) level. A drop in the muscle response to the
stimulation with time at high levels of joint torque would suggest that
the muscle loses its ability to respond with fatigue—that is, that
peripheral factors play a major role in fatigue. Figure 31.1 illustrates
the ratio between the magnitudes of the responses to the late (after
1 min) and early stimuli (Latash et al. 1994). Note a drop in the ratio
in subjects who honestly tried their best to keep the torque high at all
times (Good Subjects) and an increase in the ratio in a subject who
only pretended to be tired (a Malingerer). In this latter subject, the
apparent “fatigue” (the drop in the voluntarily produced muscle force)
was apparently of a central nature. Despite the existence of such
seemingly direct methods, the relative contribution of central and
peripheral factors to fatigue is still very much under discussion.
Figure 31.1 The ratio between the magnitudes of the responses to a standard
direct electrical stimulation of the quadriceps muscle late (after 1 min) and early
into a steady-state contraction trial at different levels of the MVC torque the subject
could produce. Note a drop in the ratio for most subjects (shown as Good
Subjects) reflecting peripheral fatigue and its increase in a Malingerer who
pretended to be unable to maintain the required torque level.
Adapted by permission from M.L. Latash et al., “Combining Electrical Muscle Stimulation
with Voluntary Contraction for Studying Muscle Fatigue,” Archives of Physical Medicine and
Rehabilitation 75 (1994): 29-35. ©1994 The American Congress of Rehabilitation Medicine
and the American Academy of Physical Medicine and Rehabilitation.

31.2 Muscular Mechanisms of

Fatigue
Some of the fatigue-induced changes occur within a muscle. These
changes have been studied mostly in animal preparations in which
fatigue was induced by a prolonged direct electrical stimulation of a
muscle. Among these changes are the following:
Slowing of conduction velocity of the muscle action potential
leading to a decrease in its amplitude and an increase in its
duration when recorded with surface electrodes (Fuglevand et al.
1993; figure 31.2). Eventually, the action potential may stop to
 propagate completely. Similar effects have been shown in
experiments with an increase in the extracellular K+
concentration. So, the efflux of K+ ions during muscle action
potentials has been assumed to contribute to this effect.

Figure 31.2 An action potential recorded in a fatigued muscle by surface

electrodes has a smaller amplitude, longer duration, and slower conduction speed
than in a nonfatigued muscle. (a) Nonfatigued muscle. (b) Fatigued muscle.

Alteration of the excitation threshold of muscle fibers to action

potentials coming in motor axons and with external stimulation
(Spira et al. 1976; Kugelberg and Lindegren 1979), which may
contribute to the aforementioned slowing of the conduction of
action potentials.
Slowing of the relaxation phase after a twitch contraction
(Bigland-Ritchie et al. 1983). This may lead to a two- to threefold
increase in the time from the peak of a contraction to the time
when the force drops to 50% of the peak value (figure 31.3).
Possible mechanisms for this phenomenon include the slowing
of Ca++ removal following a decrease in the ATP concentration
and changes in the time course of cross-bridge detachment after
Ca++ ions are removed (Fitts et al. 1982).
Figure 31.3 A twitch contraction of a fatigued muscle is characterized by a
longer relaxation phase than in nonfatigued muscle. (a) Nonfatigued muscle. (b)
Fatigued muscle.

After a brief tetanic stimulation, twitch contraction force is

commonly increased for a short time. This effect is called
posttetanic potentiation (PTP; McComas 1977; Belanger et al.
1983; Vandervoort et al. 1983). The combined effects of PTP
and fatigue have been described after a longer tetanic
contraction (Rankin et al. 1988; Grange and Houston 1991. The
twitch peak becomes smaller and is prolonged mostly due to the
prolongation of the relaxation phase (see figure 31.3).

PROBLEM 31.1
Why does the slowing of conduction lead to a decrease in the
amplitude and an increase in the duration of a potential recorded
 from the muscle surface?

Restoration of muscle characteristics after a fatiguing tetanic

stimulation may take minutes and even hours.

31.3 Spinal Mechanisms of

Fatigue
Maximal voluntary contraction force declines with prolonged
contraction, and changes in the MVC force have been used as the
most common indices of fatigue. The drop in MVC muscle force is
accompanied by a decrease in the α-motoneuronal excitability and a
decrease in the firing frequency of individual motor units. Individual
motor units differ in their ability to maintain a constant level of firing
during prolonged contractions (see chapter 7). Basically, smaller,
slower motor units are less fatigable and are able to maintain a
constant level of firing throughout a prolonged contraction. Larger,
faster motor units are more likely to show a decrease in the level of
firing and even to fail to maintain a sustained level of activity.
According to the Henneman principle, larger motor units are likely to
recruit last and to derecruit first during voluntary changes in muscle
force.
Both recruitment order and modulation of the discharge rate can
be changed by fatigue; in particular, the variability of the discharge
frequency increases under fatigue. However, the basic size principle
holds in fatigued muscles as well. This principle also works for the
pattern of derecruitment during a prolonged, fatiguing contraction at
a constant level: Larger motor units will show a decrease in the
frequency of firing and derecruitment, while the induced drop in
muscle force will be compensated by recruitment of new motor units
or a change in the firing pattern of already recruited motor units.

PROBLEM 31.2
 Prolonged fatiguing contraction is commonly accompanied by
tremor at a frequency of 4 to 6 Hz. Can you suggest a mechanism
for this phenomenon?

A fatigue-induced decline in an electromyogram (EMG) has been

shown to be central to the neuromuscular junction and is
hypothesized to result mainly from the autogenic reflex inhibition of
the α-motoneuronal pool (inhibition originating from receptors within
a muscle and affecting motoneurons innervating the same muscle).
In particular, a decrease in the H-reflex amplitude has been
observed in fatigued muscles. One hypothesis on the origin of the
presumed effects favors a decline in the spindle afferent firing rate
during isometric contractions as the primary cause of the lower
response of α-motoneurons to the standard stimulation via the Ia-
afferent input, which leads to the H-reflex during standard reflex
testing (Hagbarth et al. 1986; Macefield et al. 1991). An alternative
hypothesis suggests that this inhibition originates from small
afferents of the groups III and IV (including free nerve endings)
reacting to the products of muscle metabolism (Bigland-Ritchie et al.
1986; Woods et al. 1987). According to this hypothesis, these small
afferents cause increased presynaptic inhibition of Ia afferents at
their synapses with α-motoneurons.

PROBLEM 31.3
A decrease in apparent joint stiffness has been reported during
fatigue. Suggest a mechanism for this finding.

If a person maintains a constant level of isometric force, there is a

gradual increase in the average level of interference EMG recorded
by surface electrodes. This increase is mostly due to the recruitment
of new motor units, which compensates for a decrease in the
contribution of fatigued motor units to total muscle force (Garland et
al. 1994; Christova and Kossev 1998). The discharge frequency of
motor units changes only slightly. Fatigue leads to a change in the
relation between the level of muscle activation and the magnitude of
joint torque the muscle produces. Under fatigue, a muscle preserves
a close to linear relation between its level of activation and
generated torque, but the slope of this relation changes considerably
(figure 31.4) such that the muscle can show a three- to fourfold
higher level of activation while producing the same torque in a
fatigued state.

Figure 31.4 The two lines illustrate two relations between the level of muscle
activity (normalized EMG) and isometric muscle torque (also normalized). Under
fatigue (dashed line), the muscle shows much higher activation levels for the same
torque levels as compared to the prefatigue state (solid line).

Changes in different components of muscle reflexes to stretch

were described in fatigued muscles. In particular, monosynaptic
reflexes (in particular, the H-reflex) are suppressed (Hagbarth et al.
1995; Nicol et al. 2003; Kalezic et al. 2004). Reports on changes in
the tonic stretch reflex have been ambiguous, ranging from a drop in
its gain in fatigued muscles to basically unchanged reflex responses
to muscle stretch (Marsden et al. 1976a,b; Kirsch and Rymer 1987;
Zhang and Rymer 2001). Conflicting results have been reported for
the long-latency, M2-M3 reflexes (also called preprogrammed or
triggered reactions; see chapter 19), including no changes in the
magnitude of these reactions, their suppression, and their facilitation.
For example, a typical sequence of responses to a quick mechanical
perturbation of a joint under fatigue has been reported to show a
significant reduction in the M1 response (the earliest response with
strong contribution by the monosynaptic stretch reflex), smaller
changes in the M2 response, and an increase in the M3 response
(Windhorst et al. 1986; Balestra et al. 1992; Duchateau et al. 2002;
figure 31.5).

Figure 31.5 A schematic illustration of changes in the three muscle responses to

quick stretch (the upper trace) under fatigue. Note the very strong suppression of
the short-latency response (M1), smaller suppression of the medium-latency
response (M2), and facilitation of the long-latency response (M3). The lines before
fatigue (solid) and under fatigue (dashed) are shifted with respect to each other
along the vertical axis for better visualization.

These observations have several implications. First, they support

the idea that the effects of fatigue are mediated by presynaptic
influences on α-motoneurons that affect different reflex loops
differently, rather than by postsynaptic influences that would be
expected to show qualitatively similar influences on all inputs into the
α-motoneuronal pool. Besides, these observations suggest that the
M2 and M3 responses involve different neural loops. An increase in
the M3 response can be seen as a reflection of its transcortical
nature. Indeed, to produce the same baseline activation level by a
fatigued muscle requires stronger descending input into the
corresponding α-motoneuronal pool to compensate for the expected
lower contribution of reflex loops (e.g., due to their presynaptic
inhibition). As a result, a larger baseline activation level of cortical
neurons is expected to lead to their increased excitability and to
cause an increase in responses to transcortical loops.

31.4 Supraspinal Mechanisms of

Fatigue
Basically, changes in all the supraspinal structures involved in the
production of voluntary movements can contribute to the decrease in
muscle force during fatigue. In general, prolonged, fatiguing
contractions in humans are accompanied by a gradual increase in
the activity of cortical neurons, which may be followed by a decrease
in this activity (Gandevia 2001; Liu et al. 2002; Benwell et al. 2005),
as well as by an increase in the readiness potential (see chapter 8;
Barthel et al. 2001).
Experiments have demonstrated changes in the excitability of
neurons in the primary motor area during prolonged muscle
contraction (Ljubisavljevic et al. 1996; Taylor et al. 2000; Taylor and
Gandevia 2001). However, these changes could vary from subject to
subject and across studies. So at present, it is impossible to identify
supraspinal structures or mechanisms that can be blamed for playing
a particularly important role in a fatigue-induced drop in voluntary
muscle contraction force.
Fatigue leads to changes in indices of finger interaction, such as
enslaving and force deficit (see chapter 27) during force-production
tasks (Danion et al. 2000, 2001). These changes have been shown
not to be limited to actions that rely on muscles directly involved in
the fatiguing exercise; effects of fatigue are also seen in other
actions performed by the hand. For example, when subjects
performed a fatiguing exercise by pressing with the fingertips, the
extrinsic flexor muscles were expected to fatigue, while intrinsic
muscles were expected to produce much lower forces and not show
signs of fatigue. However, similar drops in the maximal voluntary
force were observed after fatigue when the subjects pressed with the
fingertips and when they pressed with the proximal phalanges. For
the latter action, intrinsic hand muscles act as the prime movers.
These observations suggest that fatigue-induced changes involve
neural circuits that participate in the coordination of muscle groups
underlying multifinger synergies (Singh et al. 2010). This hypothesis
is described in more detail in section 31.5.

31.5 Adaptive Changes During

Fatigue
First, it is necessary to define what an adaptive change is. This term
can be used with respect to all secondary changes that occur in an
organism as a reaction to some phenomenon—for example, to
fatigue. Some of these changes, however, may be forced upon the
system and may not be helpful in counteracting the undesired effects
of the original cause. For example, if a stone hits a glass window, the
glass may break. This is apparently not an adaptive reaction.
However, if the glass can change its mechanical properties in
response to an impact—for example, to become viscous and to
absorb the energy of the stone without breaking—this would be a
useful adaptive reaction. Unfortunately, in humans and animals, the
question of the usefulness of a reaction may not have an
unambiguous answer. For example, if a person cocontracts many
limb muscles and stiffens all their joints in a certain situation, this
reaction looks suboptimal if one considers energy expenditure.
However, the reaction may serve a purpose, for example, if an
external perturbation is expected, and the person wants to avoid
changes in the limb posture. So we will consider all secondary
reactions within the human body to be adaptive if they are not
obviously forced upon the body. In some cases, it will be possible to
classify a reaction as being useful (adaptive) or detrimental
(maladaptive); in many cases, however, this is impossible to do with
certainty.
A number of adaptive mechanisms to fatigue have been
suggested. In particular, prolongation of the relaxation phase of the
muscle twitch contraction (figures 31.3 and 31.4) may be considered
adaptive and useful because it does not allow muscle force to drop
quickly when the ability of the muscle to generate a new contraction
is impaired or when action potentials are generated by α-
motoneurons at a lower frequency. A negative correlation between
prolongation of the relaxation phase and firing frequency of individual
motor units has been reported (van Groeningen et al. 1999). This
finding makes sense because the prolongation of the relaxation
phase helps generate smooth muscle contractions in conditions of a
low-frequency motoneuronal drive.
Synchronization of motor unit discharges is another adaptive
mechanism. It leads to an increase of muscle force, although it may
eventually lead to burstlike activity and to a sawtooth-like tetanus
rather than a smooth contraction. Synchronization of motor units
occurs not only in fatigue (Bigland and Lippold 1954; Kadefors et al.
1968; Sato 1982) but also in motor disorders that are characterized
by reduced muscle force (paresis) due to damage to the central
nervous system (Larsson 1975; Lindstrom et al. 1985; Latash 1988).
In conditions of an increased synchronization of motor unit
discharges, the spectrum of an EMG recorded by surface electrodes
demonstrates a shift toward the range of lower frequencies (figure
31.6).
Figure 31.6 Spectrum of an EMG recorded by surface electrodes in a fatigued
muscle shows a shift toward low frequencies when compared to the spectrum of
an EMG recorded in a nonfatigued muscle. (a) Nonfatigued muscle. (b) Fatigued
muscle.

PROBLEM 31.4
Why does the EMG spectrum shift to lower frequencies in
conditions of excessive motor unit synchronization? Can you
suggest conditions under which a spectral shift toward high
frequencies might be expected?

The central nervous system apparently makes use of the

redundancy of the motor apparatus at different levels. For example,
if there are several muscles that can contribute to joint torque in a
certain direction, their relative contributions can be rotated during a
prolonged fatiguing contraction so that fatigued muscles can get a
rest period without a drop in total joint torque (Sjogaard et al. 1986,
1988). Similar effects can be seen at the level of motor unit
recruitment when, during a prolonged contraction, a group of motor
units can be derecruited (turned off) and substituted with recruitment
of another group of motor units; at a later time, the second group
may take a rest while the first one will resume firing (Jensen et al.
2000). These phenomena are sometimes referred to as muscle
rotation and motoneuron rotation.
Reactions to fatigue illustrate the benefit of having multi-element
systems involved in everyday actions. This abundance of elements
has room for adaptive reactions of motor synergies, which helps
mitigate the potential effects of fatigue on accuracy of performance.
In particular, fatigue of a muscle group does not exert major effects
on whole-arm actions due to adjustments in contributions of other
muscle groups (Côté et al. 2002, 2008). Fatigue of a finger leads to
an increase in variability of its force, but this is not reflected in
variability of total force produced by a set of fingers (Singh et al.
2010). Similar adaptive changes have been seen in multi-muscle
synergies during whole-body tasks following fatigue of foot
dorsiflexors (Singh and Latash 2011) and in multidigit synergies
during prehensile tasks following the fatigue of a finger or of the
thumb (Singh et al. 2012, 2013).

31.6 Abnormal Fatigue

Patients with multiple sclerosis (see chapter 40) often demonstrate
an abnormal sense of tiredness, out of proportion to the degree of
daily effort or degree of disability, which represents a major disabling
factor (Freal et al. 1984; Krupp et al. 1988, Monks 1989). This sense
is commonly addressed as fatigue, although patients with multiple
sclerosis regard it as a fundamentally different experience than
ordinary fatigue. However, when tested in laboratory conditions
requiring a prolonged, fatiguing contraction of one muscle group,
patients with multiple sclerosis show a faster drop in muscle force,
suggesting a change in “ordinary” muscle fatigue as well. A study of
the relative contributions of central and peripheral factors to muscle
fatigue in multiple sclerosis has demonstrated an apparent role for
the central factors that was not seen in healthy subjects (Latash et
al. 1996). This is an expected result because the primary underlying
cause of all the symptoms in multiple sclerosis is demyelination of
fast-conducting axons within the central nervous system.
Much attention has been drawn to a state that is characterized by
exhaustion and an inability to be involved in virtually any kind of
activity involving even minimal motor effort. This condition has been
termed chronic fatigue syndrome (reviewed in Johnson et al.
1999; Jason et al. 2005). This syndrome has an unknown etiology,
with a possible role played by an episode of viral infection or by a
central neurological disorder. It is unclear whether changes in
muscle fatigue contribute to this condition or whether it has an
absolutely different, central nature. Unfortunately, there is no
treatment.

CHAPTER 31 IN A NUTSHELL
Fatigue is a complex phenomenon that
receives contributions from peripheral
and central neural and psychological
factors. Fatigued muscles show a
slower conduction velocity of action
potentials and prolonged twitch
contractions, primarily due to the
prolongation of the relaxation phase.
Fatigue-related reflex changes, in
particular the suppression of the H-
reflex, are likely to be mediated by
changes in the activity of small
receptors of groups III and IV,
leading to increased presynaptic
inhibition. Medium-latency
preprogrammed reactions show a modest
decline, but long-latency reactions
can increase under fatigue, suggesting
their transcortical nature. During
voluntary contractions, there is an
increase in the EMG level required to
keep force unchanged. Motor units show
an increase in the synchronization of
their firing patterns with fatigue.
Prolonged, fatiguing contractions in
humans are accompanied by a gradual
increase in the activity of cortical
neurons. There are adaptive changes to
fatigue at different levels, including
adaptive changes in motor synergies.
An unusual sense of fatigue (e.g.,
chronic fatigue and fatigue in
multiple sclerosis) is likely to be
mostly of a central nature.
 Chapter 32

Effects of Aging

KEY TERMS AND TOPICS

sarcopenia
reinnervation
changes in reflexes
changes in the sensory function
excessive coactivation
changes in reaching movements
changes in posture and gait
changes in motor synergies

Aging is associated with a variety of changes in the human body.

These changes happen even in the absence of any apparent
disorders that commonly accompany the process of aging, such as,
for example, Parkinson’s disease, stroke, arthritis, and peripheral
neuropathies. Some of them will be discussed in later chapters in
part 8 of the textbook. In this chapter, we will focus on healthy aging
—that is, on changes with age that are seen across the otherwise
healthy population.
32.1 General Features of
Movements in Elderly Persons
Aging leads to changes in many aspects of voluntary movements
that cover all the major motor actions forming the everyday
movement repertoire. These differences can be seen in postural
tasks, during locomotion, during reaching arm movement, during
manipulation of handheld objects, and during handwriting. Some of
the differences in the movements of elderly people can be attributed
to age-related changes in peripheral structures, in particular, in
muscle properties. There are also contributions from changes in the
central nervous system, some of which may be viewed as imposed
by aging (in particular, by the death of neural cells), while others may
represent adaptive changes that mitigate the effects of age on motor
performance.
In general, the elderly produce movements that are not as fast,
smooth, accurate, or resistant to perturbations as those of younger
persons. In particular, one of the well-established features of
movements in elderly persons is the slowness in movement initiation
and in movement execution (Stelmach et al. 1987, 1988; Welford
1984). In particular, the longer reaction times and movement times
can be seen in elderly persons across a range of tasks of differing
complexity. These age-related differences persist with modification of
such task parameters as movement extent, direction, and the arm
that was supposed to move could be specified or not specified. The
difference in movement speed persists even in the simplest
movements, such as single-joint flexion performed in a self-paced
manner (Buchman et al. 2000; figure 32.1). Fast movements
performed by older persons typically show a change in their
kinematic structure with a longer deceleration phase and an
increased incidence of corrective adjustments (Pratt et al. 1994).
Another universal feature of movements performed by elderly
persons is the increased unintended force production—in particular,
an increase in the cocontraction of agonist–antagonist muscle pairs
acting at a joint.

Figure 32.1 Single-joint voluntary movements of elderly persons are typically

slower than in younger persons.
Adapted from Narici, Bordini, and Cerretelli (1991).

PROBLEM 32.1
Increased muscle cocontraction apparently does not contribute to
the net torque acting at a joint and may be seen as wasteful. Can
you suggest benefits from using increased muscle cocontraction
levels?

Some of the differences in the movements performed by elderly

people are likely to reflect changes in the generation and processing
of sensory information. In particular, elderly persons tend to rely
more on visual feedback and show more pronounced disturbances
of movements when they try to perform them with their eyes closed.
When a person tries to produce the same action several times, all
movement characteristics vary among the repeated attempts—a
feature called variability. Elderly persons tend to show higher
variability, but this increase is very much task-dependent. The
difference in the measures of variability between elderly and young
persons tends to be higher for tasks that require relatively low forces
(Cole et al. 1999; Burnett et al. 2000; Enoka et al. 2003). There are
also more subtle changes in motor variability with age, reflected not
only in the magnitude of variability but also in its structure
(Vaillancourt and Newell 2003; Vaillancourt et al. 2003). In particular,
indices of complexity of movement patterns (for example,
approximate entropy) become smaller with age, leading to a
counterintuitive combination of increased variability and increased
predictability of movement patterns.

32.2 Changes in Muscles and

Motor Units
Aging is associated with sarcopenia, an unavoidable decline in
muscle mass, along with a loss of both voluntary and electrically
evoked strength (Winegard et al. 1997). This decline seems to begin
at the age of 50 to 60 y, and it varies greatly among individuals
(Narici et al. 1991; figure 32.2). Muscles lose both cross-sectional
area and number of fibers, particularly the fast-twitch fibers
(Kirkendall and Garrett 1998; Bemben 1998). There are also
changes in the mechanical properties of peripheral tissues. In
particular, with aging, connective tissue has been shown to replace
contractile proteins (Zimmerman et al. 1993). As a result, the elderly
demonstrate increased apparent muscle stiffness (McDonagh et al.
1984) and reduced tendon compliance (Tuite et al. 1997).
Figure 32.2 Changes in the number of motor units in a foot muscle (extensor
digitorum brevis) with age. Note the accelerated decrease in the number of motor
units after age 50.
Adapted from M.V. Narici, M. Bordini, and P. Cerretelli (1991).

The number of both cortical neurons, whose axons form the

corticospinal tract (Eisen et al. 1996), and α-motoneurons declines
with age (Campbell et al. 1973; Roos et al. 1997). This loss becomes
apparent after age 60. High-threshold motor unit atrophy is
particularly pronounced (Owings and Grabiner 1998). As a result of
the death of α-motoneurons, groups of muscle fibers become
denervated; that is, they lose excitatory neural inputs. This is
accompanied by the processes of sprouting and reinnervation,
when the axons of remaining α-motoneurons grow additional
terminal fibers, which make synapses on some of the vacated
muscle fibers. Figure 32.3 illustrates this process when a large motor
unit (MU1) is lost and its muscle fibers are partly reinnervated. This
leads to an increase in the innervation ratio: the average number of
muscle fibers per motor unit (for units N2 and N3 in figure 32.3).
Some of the denervated muscle fibers do not get reinnervated; they
lose their contractile properties and ultimately turn into connective
tissue (the light muscle fiber in figure 32.3).
Figure 32.3 An illustration of the processes of denervation and reinnervation.
The death of a neuron N1 leads to denervation of a group of motor fibers (dashed
lines). Axons of neurons N2 and N3 show sprouting (solid, thick lines) and
reinnervate some of the orphan fibers. The size of the N2 and N3 motor units
increases.

As a result of the processes of denervation and reinnervation,

there are fewer motor units, but they are on average larger in size
and slower. The typical relationship between motor unit size and
fatigability (see the size principle in chapter 7) tends to break down,
and larger motor units become as fatigable as smaller ones (Luff
1998). These changes are particularly pronounced after age 60.

PROBLEM 32.2
Suggest reasons why, in elderly persons, motor unit size and
fatigability stop showing the typical relationship in which larger
motor units are more fatigable.

The lack of smaller motor units may be blamed for the poor
control of low forces, high force variability, and poor smoothness of
movements because the recruitment and derecruitment of larger
motor units leads to larger changes in the total muscle force. Figure
32.4 illustrates the high variability in the average frequency of action
potentials in motor units of the first dorsal interosseous muscle when
the subjects were producing 5% of their maximal voluntary force;
note also the larger fluctuations in the force level in the elderly
(Laidlaw et al. 2000). Despite the described changes in motor unit
properties, changes in fatigability with age are ambiguous when
quantified at the level of motor performance. There are reports of no
changes, an increase, and a decrease in fatigability in the elderly
(Chan et al. 2000; Bilodeau et al. 2001; Allman and Rice 2002).

Figure 32.4 The subjects tried to produce a constant force at 5% of their

maximal voluntary force level by the first dorsal interosseous muscle. Note the
larger variability in both the force level and frequency of action potentials of motor
units in elderly persons.
Reprinted by permission from D.H. Laidlaw, M. Bilodeau, and R.M. Noka, “Steadiness is
Reduced and Motor Unit Discharge is More Variable in Old Adults,” Muscle Nerve 23:
(2000): 600-612. ©John Wiley & Sons, Inc.

With age, strength becomes a limiting factor in certain everyday

activities, such as rising from a chair (Hughes et al. 1997). There are
controversial reports regarding possible differential losses of force in
different muscle groups with age. In particular, some authors
reported significant differences between the force loss in the upper
versus the lower extremity muscles (Grimby et al. 1982) and
between proximal and distal muscles (Nakao et al. 1989; Shinohara
et al. 2003a). Other studies, however, failed to confirm these results
(e.g., Viitasalo et al. 1985).

PROBLEM 32.3
Suggest reasons for the different effects of aging on proximal and
distal muscles.

Aging leads to a decline in hand strength and a loss of manual

dexterity, which affects many of the activities of daily living (Boatright
et al. 1997; Giampaoli et al. 1999; Hughes et al. 1997; Rantanen et
al. 1999; Francis and Spirduso 2000). This is associated with the
described changes in the neuromuscular apparatus, such as a drop
in the number of motor units, an increase in the size of the motor
units, and a general slowing down of their contractile properties
(Doherty and Brown 1997; Duchateau and Hainaut 1990; Kamen et
al. 1995; Kernell et al. 1983; Owings and Grabiner 1998). As you will
see in section 32.7, there are also changes in multidigit synergies
that stabilize hand function during prehension (chapter 27).

32.3 Muscle Reflexes in Elderly

Persons
There is a general drop in the magnitude and latency of responses to
sensory (or electrical) stimuli with age. The magnitude of the H-reflex
(chapter 17) in the elderly is decreased as compared to that of
younger persons (Vandervoort and Hayes 1989). This change can
partly be due to peripheral factors such as a decrease in the muscle
response to excitation. In fact, a similar drop in the direct M-
response with age has been reported. It suggests that the proportion
of motor units activated during H-reflex testing may be relatively
unchanged.
Studies of the tendon tap reflex (T-reflex) have shown a difference
in both the magnitude and latency of the T-reflex between the elderly
and young persons. The drop in the magnitude of the T-reflex was
similar to that described for the H-reflex. The longer latency of the T-
reflex suggests a change in the conduction velocity along the axons
(Henderson et al. 1980) or a slower reaction of spindle sensory
endings to the tap in elderly persons.

PROBLEM 32.4
How can the conduction speed of action potentials change with
age? Suggest a mechanism. How can the speed of reaction of a
sensory ending change with age?

Polysynaptic reflexes (see chapter 18) are also reduced and

delayed in the elderly; the mechanisms of this reduction are
generally unknown, but it is likely to get a contribution from changes
in muscle excitability. Although voluntary reactions to sensory stimuli
are not commonly viewed as reflexes, let us mention in this section
the well documented increase in the simple reaction time with age
(Warabi et al. 1986). Larger brain areas are activated in the elderly in
simple reaction time tasks; these include both the ipsi- and
contralateral sensorimotor cortex, the basal ganglia, and the
cerebellum (Mattay et al. 2002). It is possible that these findings
reflect adaptive patterns used by the central nervous system to
optimize performance given the changed state of the brain
structures.

32.4 Changes in Sensory

Function
Aging is associated with a progressive loss in many of the human
sensory functions. In particular, there is a decline in visual acuity and
in the ability to focus gaze on near objects. There is an age-related
loss of vestibular receptors and cutaneous receptors (in particular,
Meissner corpuscles; Mathewson and Nava 1985). There is also a
loss of the number of sensory neurons innervating peripheral
sensory receptors: peripheral neuropathy (see chapter 35). These
changes may be blamed for deteriorated tactile discrimination
typically seen in elderly persons.

32.5 Muscle Activation Patterns

During Fast Movements
During fast voluntary movements, elderly persons show typical
triphasic patterns of EMG activity in the agonist–antagonist muscle
pairs (chapter 23) similar to those seen in younger persons but with
a relatively larger coactivation of the muscles (Seidler-Dobrin et al.
1998; Buchman et al. 2000; Klein et al. 2001). The increased
coactivation seems to be a universal finding across many tasks
performed by the elderly (figure 32.5). It sometimes involves not only
the explicitly involved muscles but also their neighbors. For example,
studies of the first dorsal interosseus muscle have shown excessive
coactivation of the second palmar interosseus in addition to
coactivation of an antagonist (Spiegel et al. 1996). Studies of
postural responses (see section 32.6) have also shown increased
coactivation in elderly persons (Tang and Woollacott 1998).
Figure 32.5 Indices of muscle coactivation are higher in elderly persons than in
young persons. The indices were quantified for elbow muscles in percent of the
maximal muscle activity.
Reprinted by permission from C.S. Klein, C.L. Rice, and G.D. Marsh, “Normalized Force,
Activation, and Coactivation in the Arm Muscles of Young and Old Men,” Journal of Applied
Physiology 91 (2001): 1341-1349. With permission from American Physiological Society.

32.6 Changes in Posture and

Gait
A reduced ability to control vertical posture and gait is seen in most
elderly persons. Impaired postural and gait control potentially leads
to more frequent episodes of stumbling and falling—a major factor
contributing to morbidity and mortality in the elderly. It limits the
ability of elderly persons to exercise and, as a result, may contribute
to other age-related changes in the neuromuscular system.
During quiet standing, elderly persons typically show increased
postural sway (Maki et al. 1990; Melzer et al. 2004; Fujita et al.
2005). Many factors can potentially contribute to this change. These
include the aforementioned decrease in muscle strength with age; as
a result, it takes muscles longer to reverse a deviation of the body
from a particular set point. Elderly persons may also show reduced
perception of stability limits and an atypical selection of adequate
sources of sensory information used for postural stabilization (Hay et
al. 1996; Redfern et al. 2001; Teasdale and Simoneau 2001). Both
these factors may also lead to larger body deviations as compared to
younger persons.
Anticipatory postural adjustments (APAs) represent an important
feedforward posture-stabilizing mechanism (chapter 25). Recall that
young people show changes in the activity of postural muscles about
100 ms prior to an action associated with a postural perturbation.
Elderly persons show APAs that occur later—that is, closer in time to
the action initiation—and are of a smaller magnitude (Inglin and
Woollacott 1988; Woollacott et al. 1988; Rogers et al. 1992).

PROBLEM 32.5
Reduced APAs are seen in young persons when they stand on a
board with a reduced support area. Suggest a common
explanation for this observation and the reduced APAs in the
elderly.

Recall that when an actual perturbation of posture happens, its

effects are counteracted by several mechanisms. The first is the
muscle and tendon elasticity. As mentioned earlier, muscles and
tendons are stiffer in the elderly. This could be expected to lead to
the generation of larger forces to joint deviations caused by a
perturbation. These forces depend on the background level of
muscle activation that varies during quiet standing; as a result, they
can offer varying resistance to external perturbations. One method of
increasing the resistance provided by the peripheral tendon and
muscle elasticity (and damping!) is to coactivate agonist–antagonist
muscle pairs acting at the major leg and trunk joints. Indeed,
agonist–antagonist coactivation leads to higher apparent joint
stiffness (i.e., its smaller deviations in response to changes in
external moment of force). Increased coactivation during standing is
typical of older persons as well as of other groups of persons with
impaired postural control (reviewed in Latash 2018a). This method of
increasing joint resistance to perturbations is, however, questionable
when applied to whole-body tasks such as standing: Studies of the
effects of voluntary muscle coactivation in young persons have
shown worse postural stability associated with increased levels of
coactivation (Yamagata et al. 2019, 2021).
In contrast to the increased contribution of peripheral muscle
elasticity and damping, monosynaptic and polysynaptic reflexes are
smaller and are delayed in the elderly (see section 32.3). This
potentially limits their contribution to postural responses to external
perturbations.
The next line of defense is the preprogrammed responses
(chapters 19 and 25). There are several typical patterns of muscle
responses to perturbations that are similar to those experienced by a
standing passenger when a bus starts to move. If a young person
stands in comfortable conditions and the perturbation is not too
strong, the person is likely to demonstrate the ankle strategy
described in chapter 25. Under similar conditions, elderly persons
are more likely to show the hip strategy (Horak et al. 1989). The
magnitude of these responses is decreased in elderly persons, while
their delay is prolonged (similarly to what happens with muscle
reflexes). There is also a marked cocontraction of agonist–
antagonist muscle groups acting at major postural joints in response
to postural perturbations (Tang and Woollacott 1998; Woollacott et
al. 1988).
Preferred walking speed declines after about 60 y of age; elderly
people take shorter steps but at about the same frequency (Samson
et al. 2001; Laufer 2005). They exert smaller forces on the ground,
which is expected from slower gaits. Reasons for the slower gait
may include muscle weakness, particularly weakness of the ankle
plantarflexor group, and poor balance. Slower gaits are associated
with smaller reaction forces acting on the body from the supporting
surface. Therefore, they are associated with smaller postural
perturbations. Besides, walking slowly allows a person to gain extra
time to prepare for an obstacle or another unexpected complication
that may emerge in the course of locomotion. It has been shown that
walking speed in the elderly correlates with the level of fitness. So
exercise seems to be an effective way to avoid changes that lead to
the drop in walking speed.
Changes in muscle strength with age are likely to affect other
activities that require high forces. In particular, maximal running
speed shows a decline after age 30 and continues to drop over the
life span (figure 32.6).

Figure 32.6 A schematic representation of changes in muscle strength and

maximal running speed with age. Both show a visible drop at the age of about 60
yrs.

32.7 Hand Function in Elderly

Persons
Some of the aforementioned differences between motor patterns in
the elderly and young persons may be related not to changes in
muscle properties, transmission speed of signals within the central
nervous system, composition of motor units, and other apparently
detrimental factors that occur with age but to the coordination of the
many elements that are commonly involved in natural motor tasks.
The problem of coordinating such redundant sets of elements was
discussed in chapter 22. There, a notion of synergies has been
introduced as task-specific neural organizations of signals to
elements that stabilize certain important features of motor
performance.
Studies of prehensile tasks have shown that elderly persons tend
to coordinate the action of their digits differently from younger
persons. In particular, when lifting a small object, elderly persons are
likely to show grip forces that are nearly twice as high as the grip
forces produced by younger persons (Cole 1991; Gilles and Wing
2003). There may be a good reason for the elderly to apply such
apparently excessive forces. Aging is typically associated with
increased tremor and higher variability of movement patterns
(Galganski et al. 1993; Enoka et al. 2003). Both these factors
contribute to poorly controlled inertial forces that may be acting on
the handheld object. Applying higher grip forces seems like a
sensible strategy to ensure that, even if an unexpected inertial force
emerges, the increased safety margin will prevent the object from
slipping from the hand. On the other hand, the application of higher
grip forces may be related to the impaired sensory function of the
hand: Higher forces are produced to induce comparable sensory
effects. Studies (Cole et al. 1998, 1999) have, however, challenged
the hypothesis that the decline in the ability of older persons to grip
and lift objects is solely due to their impaired tactile sensitivity. There
are also age-related changes in properties of the skin leading, in
particular, to lower friction between the skin and the grasped object.
This factor can also contribute to the larger grip forces applied by the
elderly.
Age-related decline in muscle strength is larger in intrinsic hand
muscles than in extrinsic hand muscles (Shinohara et al. 2003a).
Note that, during natural tasks, forces produced by these muscle
groups should be balanced. For example, when a person applies
forces with the fingertips to a handheld object, the extrinsic deep
flexor (flexor digitorum profundis) is the focal force generator, while
intrinsic muscles balance moments of forces in the
metacarpophalangeal joints (chapter 27). Disproportionate changes
in muscle strength with age are likely to lead to the necessity to
adjust the muscle synergies elaborated by the brain during the
lifetime. Several studies have indeed shown that multidigit synergies
stabilizing the total force and the total moment of forces (rotational
action) applied to an object are weaker in the elderly than in younger
persons (Shinohara et al. 2004; Shim et al. 2004; figure 32.7).

Figure 32.7 Indices of force-stabilizing (force–control) and moment of force-

stabilizing (moment–control) synergies in young and elderly persons who
produced isometric tasks during pressing with the fingertips (DP) and proximal
phalanges (PP). Note the generally lower values of the index in elderly persons,
particularly during force production at PP, where intrinsic muscles are expected to
be primary force generators.
Reprinted by permission from M. Shinohara, J.P. Scholz, V.M. Zatsiorsky, and M.L. Latash,
“Finger Interaction During Accurate Multi-Finger Force Production Tasks in Young and
Elderly Persons,” Experimental Brain Research 1556 (2004): 282-292. © 2004 Springer.
 PROBLEM 32.6
When elderly persons manipulate handheld objects, they typically
show larger forces applied by fingers that produce moments of
forces with respect to the thumb contact point directed not against
the external moment but in the same direction. Suggest a
functional explanation for this finding

Studies of the performance of elderly subjects in handwriting tasks

(Contreras-Vidal et al. 1998) have suggested that the spatial
coordination of fingers and wrist movements declines with age, while
the control of force pulses applied by individual digits may be
preserved. These observations point at problems with coordinating
force pulses produced by individual digits—that is, problems with
multidigit synergies.

32.8 Changes in Motor Synergies

A number of studies explored age-related changes in synergies that
stabilize potentially important performance variables across tasks
ranging from multidigit force and moment of force production
(Olafsdottir et al. 2007; Kang et al. 2019) to multijoint pointing (Verrel
et al. 2012), and to whole-body stepping (Wang et al. 2017). Most of
these studies have reported a drop in the synergy indices in older
persons.
In addition to impaired stability, healthy older persons show an
impairment in their ability to prepare for planned actions. These are
reflected, in particular, in anticipatory synergy adjustments (ASAs;
see chapter 25), which reflect the attenuation of synergies stabilizing
a performance variable in preparation for actions that require a
change in that performance variable. ASAs are shorter and smaller
in the elderly (Olafsdottir et al. 2008), which may contribute to their
reduced ability to initiate actions quickly.
Note that ASAs represent an aspect of feedforward control. Other
phenomena representing feedforward control are also impaired in
the elderly. These include the shorter and smaller APAs (reviewed in
Massion 1992) and grip force adjustments in preparation for
manipulation tasks (Inglin and Woollacott 1988; Kinoshita and
Francis 1996).

32.9 Adaptive Changes in Motor

Patterns
In the previous section, some of the apparent changes in motor
behavior with age have been considered as potential adaptive
changes in the organization of motor control that may only look
atypical but in fact may be optimal for everyday activities. Along
similar lines, other features of motor behavior typical of the elderly
may be viewed as adaptive. In particular, the excessive
cocontraction of agonist–antagonist muscle pairs may look
suboptimal and wasteful. However, the same cocontraction may look
appropriate if the person wants to avoid major joint deviations if an
unexpected perturbation occurs; cocontraction leads to an increase
in apparent joint stiffness. The documented deficit of the elderly in
reflex responses, preprogrammed responses, and APAs makes
muscle cocontraction one of the few remaining strategies to mitigate
the effects of unexpected perturbations.
Similarly, the slowness of movements may be viewed as being
imposed by the lack of muscle strength or by a deliberate control
strategy to avoid high contact forces and to provide more time for
corrections of movements if something unexpected happens.
Unexpected changes in the external conditions of movement
execution happen rather frequently; they may be particularly
dangerous for a person with an impaired control of balance. Slowing
down seems like a reasonable adaptive strategy.

32.10 Effects of Training

Effects of training have been documented in many studies of elderly
subjects. In particular, strength training has been shown to lead to
higher forces and lower antagonist coactivation (Sipila and
Suominen 1995; Tsutsumi et al. 1997; Izquierdo et al. 2003). Since
muscle cross-sectional area shows only minor enlargements in the
process of training, neural adaptations are likely to play a major role
in bringing about these effects (Hakkinen et al. 1998). Specialized
training has also been shown to improve the impaired ability of
elderly persons to control pinch force accurately (Ranganathan et al.
2001). Maintaining physical fitness and muscle strength can lead to
lower fluctuations in force production. Research suggests that 8 wk
of strength training is sufficient to induce significant strength and
endurance changes in elderly persons (Izquierdo et al. 2003). It
remains to be seen whether an increase in strength in an elderly
person can lead to an improvement of motor synergies and patterns
of everyday movements.
Strength training of particularly affected muscle groups, such as
hand muscles, has been shown to lead to better performance not
only of strength-related tasks but also of tasks that require accuracy
(Olafsdottir et al. 2008). These changes have been associated with
improved indices of multifinger synergies in accurate force
production tasks.

CHAPTER 32 IN A NUTSHELL
Aging leads to losses in muscle mass,
strength, and the number of α-
motoneurons. Processes of denervation,
sprouting, and reinnervation lead to a
smaller number of motor units that
are, on average, slower and larger
than in younger persons. There is an
impairment of the sensory functions
with age. Reactions to sensory
stimuli, from monosynaptic reflexes to
voluntary reactions, are typically
slower and smaller in magnitude.
Control of vertical posture and gait
declines with age, leading to more
frequent falls. This decline is
reflected in the larger postural sway,
smaller and delayed anticipatory
postural adjustments, and smaller
preprogrammed reactions to
perturbations. Movement patterns in
elderly persons are characterized by
slowness and excessive cocontraction
of agonist–antagonist muscle pairs.
Movements are less smooth and more
variable. Prehensile tasks are
associated with excessive grip forces
and weaker multidigit synergies
stabilizing the hand action. Some of
the features of movements in elderly
persons may be viewed as reflecting
adaptive strategies by the central
nervous system. Training can increase
the strength and endurance of elderly
persons as well as accuracy in force-
producing tasks and the associated
synergies.
 Chapter 33

Typical and Atypical

Development

KEY TERMS AND TOPICS

motor milestones
maturation of the brain
emerging motor patterns
exploration
development of synergies
Down syndrome
Asperger’s disorder
developmental coordination disorder
autism

In this chapter, certain features of movements during typical and

atypical development are described and tentatively linked to
neurophysiological structures. Unfortunately, neurophysiological
mechanisms of human development are all but unknown.

33.1 Humans at Birth

At birth, different animals show a broad variety of abilities to function
in the environment. Some animals can show independent activity at
birth (referred to as precocial), while others are rather helpless
(referred to as altricial). In human newborns, sensory functions seem
to be more mature at birth than motor functions. There is, however, a
significant difference in the rate of maturation of sensory systems of
different modalities. The kinesthetic and vestibular systems are
basically mature by the time of birth, while the visual system is still
developing. In particular, the optic nerve axons are not fully
myelinated, and visual acuity is poor. Overall, human newborns are
rather helpless. This is at least partly due to the incomplete process
of development of the central nervous system (CNS).
At birth, the human brain weighs about 300 g, 25% of its weight in
an adult. The increase in the brain weight is accompanied by an
increase in the number of neurons, their size, the myelination of their
axons, and the growth of glial cells. The cerebral hemispheres are
formed by the time of birth but are not fully functional, partly due to
the incomplete myelination of neural tracts. Myelination of both
sensory and motor axons begins before birth and continues for about
6 mo after birth. It proceeds in a cephalocaudal direction—that is,
starting from the brain and moving toward the tail end of the spinal
cord. So the conduction pathways, both descending and ascending,
are not fully functional until the middle of the first year of life.

PROBLEM 33.1
What would you expect from incompletely myelinated fibers, an
inability to conduct action potentials or slow conduction of action
potentials?

Another important factor in the motor abilities of newborns is the

state of their muscles and reflexes. By the time of birth, there are
about 40% of slow-twitch fibers and 45% of fast-twitch fibers in the
total. The relative distribution of different fiber types does not reach
typical adult steady state until age 3. During that time, the muscle
fibers grow in both length and width. In parallel to muscle changes,
there are also changes in reflexes. Newborns show a set of reflexes
called primitive reflexes, such as the sucking reflex. There are
atypical patterns of monosynaptic reflexes; in particular,
monosynaptic excitation of α-motoneurons of an antagonist muscle
produced by signals from primary spindle endings (Myklebust 1990;
Myklebust and Gottlieb 1993).

33.2 Motor Milestones During

Typical Development
There are two basic attitudes toward motor development. The first
operates with such notions as stages of development, milestones or
timetables, or even ticking of the biological clock. The term motor
milestone refers to the ability of a baby to perform certain motor
tasks such as holding the head up, sitting independently, walking,
and reaching for objects. The remarkable orderliness of the
appearance of milestones during typical development has led some
researchers to conclude that development follows a fixed sequence
and timetable.
Figure 33.1 illustrates a typical sequence of milestones and the
ages when they are expected to be reached by typically developing
babies. Actually, this figure is a major simplification because each of
these milestones represents a complex behavior whose components
may also emerge in a particular sequence. For example, to be able
to walk, a baby first has to be able to turn over, to sit, and to stand.
To be able to grasp an object, a baby has first to be able to identify it,
to reach for it, to touch it, and to preshape the hand.
Figure 33.1 A typical sequence of milestones achieved over the first 2 y of life.
The horizontal bars show typical time intervals when particular milestones are
achieved.

PROBLEM 33.2
How can conduction speed change with age? Suggest a
mechanism. How can the speed of reaction of a sensory ending
change with age?

33.3 Exploration and Emergent

Motor Patterns
Research suggests that there is no fixed timetable in the emergence
of apparent milestones (Thelen et al. 1993, 1996). A series of
elegant studies by the group of Esther Thelen have produced
evidence that milestones are not a consequence of maturation of the
CNS that follows a fixed timetable but a result of changing
interactions between constraints imposed by the organism and by
the environment. Thelen and her colleagues showed, in particular,
that such rhythmic behaviors of babies as kicking, rocking, arm
waving, and banging typically precede the emergence of new
functional motor patterns, such as crawling. She viewed those
behaviors not as staged phenomena or “noise” but rather as signs of
exploration by the baby of the physical properties of his or her own
body and of the environment.
This alternative approach to motor development has its roots in
the dynamic systems approach to motor control (see chapters 21
and 26) and the ecological psychology approach that dates back to
the classic works by Gibson (1979). In particular, within this
approach, any action is considered to be a result of interaction
between perceptual and motor processes. Such perception–action
couplings between patterns of sensory signals and patterns of
control variables to effectors are discovered by the babies during
seemingly purposeless and chaotic movements. As a result,
functionally appropriate, stable behaviors emerge. Within this
framework, the variability of motor patterns in babies is viewed as
functionally important and exploratory (reviewed in Savelsbergh et
al. 2006).

PROBLEM 33.3
Suggest other examples in which motor variability may be viewed
as exploratory and functionally useful.

33.4 Development of Motor

Synergies
As described in chapter 22, motor synergies are characterized by
two features: organizing elements into stable groups, and ensuring
covariation in the group involvement to stabilize salient performance
variables. A series of recent studies of muscle activation patterns
during walking have documented the process of the gradual
formation of muscle groups during the development of human babies
(Dominici et al. 2011; Ivanenko et al. 2013; Sylos-Labini et al. 2020).
A newborn baby can show stepping leg movements if held in the air
with the feet touching a moving treadmill. During these movements,
two typical muscle groups can be seen in the activation of the leg
and trunk muscles. Their time patterns resemble two of the modes
seen in grown-ups during walking. However, their timing is not well
defined, and they show excessive agonist–antagonist cocontraction.
In toddlers, the number of modes increases to three and then four,
which is typical of adult walking, and their timing becomes more
reproducible across consecutive steps. Even in preschool children,
the muscle activation bursts are not yet well defined and tend to be
prolonged compared to typical adult-like patterns of muscle
activation. These developmental changes are reflected in patterns of
activation of motoneuronal pools across the lumbar segments of the
spinal cord (Ivanenko et al. 2013).
The crude grouping of muscles in newborns suggests that this
phenomenon reflects the functioning of a locomotor central pattern
generator (CPG) in the spinal cord with possible contributions from
reflex loops. Further refinement of the muscle activation patterns
could reflect maturation of the CPG circuitry, the descending control
of the CPG from supraspinal structures, modulation of reflex
projections, changes in movement mechanics, and possibly other
factors. So far, the relative role of these factors in the observed
patterns of muscle activation has been unknown. Observations of
nearly unchanged muscle activation modes during walking by
children with Duchenne muscular dystrophy (Vandekerckhove et al.
2020) have suggested, however, that factors at the muscle level are
unlikely to play a decisive role.
Several studies used the method of analysis of the other feature
of synergies—stability of salient performance variables—using the
framework of the UCM hypothesis (Scholz and Schöner 1999; see
also chapter 22) to explore developmental changes in this feature. A
study of reaching analyzed at the level of joint kinematics has
suggested a search for synergies during mid-childhood (5-10 y)
(Golenia et al. 2018). This was reflected in the steeper decline in the
component of variance affecting salient performance variables as
compared to the variance component along the UCM. A study of
multifinger coordination during force production has documented a
close to linear increase in the synergy index between 4 and 12 y of
age (Shaklai et al. 2017). Overall, maybe not very surprisingly, the
stability of performance improves with age as reflected by respective
synergy indices.

33.5 Down Syndrome

Down syndrome is a chromosomal disorder that affects both mental
and physical development; it was originally described in 1866 by a
British physician J.H.L. Down as “mongolism.” The prevalence of this
condition (its rate per unit of population) is between 1:600 and
1:1,000 live births. The majority of children born with Down
syndrome have three copies, rather than two copies, of chromosome
21, a condition known as trisomy-21. This extra chromosome results
in significant changes in the architecture and functioning of the brain
at birth, and it continues to affect brain development throughout
childhood. Down syndrome is also associated with changes in the
anatomy of the body. In particular, persons with Down syndrome, on
average, have relatively low height and shorter extremities, and they
show a tendency toward obesity.
There are many nonmotor problems associated with Down
syndrome. The most important are congenital heart disease,
obstruction of the intestinal tract, and increased susceptibility to
infection. Progress in medical care—particularly in newborn heart
surgery—and programs of special education have led to a dramatic
increase in the life expectancy for persons with Down syndrome,
from under 10 yrs in the 1940s to over 60 yrs at present. This
progress has also led to the emergence of new challenges, in
particular those related to the functioning of adults and elderly
persons with Down syndrome (Connolly 2001). After age 40,
persons with Down syndrome are likely to show signs of Alzheimer’s
disease, which also affects persons without Down syndrome, but
typically at a much later age.
Problems with motor coordination are just a fraction of all the
challenges that persons with Down syndrome may face during their
lifetimes. However, coping with poor motor coordination is arguably
the most frequently encountered problem in everyday life.
Babies with Down syndrome typically show slow development of
the motor function. They take more time to reach motor milestones
such as reaching for objects, crawling, and sitting independently
(Ketcheson et al. 2017). A study by Haley (1986) has shown that
postural reactions and motor milestones are closely coupled in both
typically and atypically developing children. The influence of visual
information on postural sway has been studied using a moving room
paradigm. In this paradigm, the walls move independently of the floor
and give the perception of self-movement. In standing young
children, movement of such a room increases postural sway, leading
in some cases to a complete loss of balance. Infants with Down
syndrome demonstrate much larger postural sway increases under
these conditions, suggesting a greater reliance of the postural
control system on visual input (Butterworth and Cichetti 1978). As
children grow, balance control and stability shift from a dependence
on visual feedback to an increased reliance on the proprioceptive
and vestibular systems. This shift is thought to occur about at 6 to 7
y of age in typically developing children (Shumway-Cook and
Woollacott 1985). A study using the moving room paradigm with
older children with Down syndrome has shown that their larger
postural responses persist at least up to the age of 10 (Wade et al.
2000).
A number of studies tried to relate motor deficits, including the
delayed reaching of motor milestones, in persons with DS to their
hypotonia and weakness (Rarick et al. 1976; Morris et al. 1982; Cioni
et al. 1994). Hypotonia is a poorly defined term related to another
misnomer, muscle tone, which is commonly used to reflect an
impression of joint resistance to motion imposed by another person
(we will return to this notion in chapter 36). Hypotonia (or low tone)
reflects an impression of low resistance of joints to external motion—
that is, low apparent stiffness. It may result from a change in the
properties of the peripheral structures such as muscles and tendons
as well as from changes in reflex pathways. Movements by persons
with Down syndrome are sometimes described as clumsy. The word
“clumsiness” is used to indicate movements that look different from
and less efficient than those observed in the general population. Two
major components of clumsiness in Down syndrome include
slowness of the movements and the decreased ability to rapidly
respond to the changing environment. The latter factor can be seen
in laboratory studies as a deficit in preprogramming (see chapter 19)
and longer reaction time (the time from the presentation of a stimulus
to a motor reaction). Variability in various aspects of motor
performance is increased in persons with Down syndrome (which is
also typical of a number of motor disorders), and they lack adequate
adaptation to changes in sensory information.

PROBLEM 33.4
Can the increased motor variability in young adults with Down
syndrome be viewed as exploratory and functionally useful? Why?

Discrete, single-joint movements of individuals with Down

syndrome are typically slow and frequently consist of several distinct
submovements. In some trials, movement kinematics may be
characterized by a bell-shaped velocity profile that is typical of fast
movements of persons without Down syndrome (see chapter 23).
Other trials in the same series, however, can demonstrate irregular
trajectories with visible bumps and possible reversals of movement
direction accompanied by multiple bursts of activity in the agonist
and antagonist muscles (figure 33.2). Despite being slow and
“clumsy,” persons with Down syndrome are typically very accurate in
achieving the prescribed target.
Persons with Down syndrome are commonly described as taking
more time to make decisions. During their lifetime, the CNS of such
a person accumulates experiences that allow it to predict that
unexpected changes in external conditions occur rather frequently.
Based on this experience, the CNS may be reluctant to facilitate
motor commands that lead to very fast movements in order to have
more time for evasive actions or for movement corrections in
response to a change in the environment (perturbation) or to
attenuate potentially damaging effects of the perturbation.
Recall that preprogramming during unidirectional single-joint
movements in unimpaired control subjects usually involves a
reciprocal pattern of muscle activation (chapter 19); that is, an
unexpected loading leads to an increase in the agonist activity, while
an unexpected unloading leads to a decrease in the agonist activity
with a possible increase in the activity of the antagonist. Subjects
with Down syndrome frequently demonstrate a coactivation pattern
of preprogramming that involves an increase in the activity of both
agonist and antagonist muscles irrespective of the direction of a
perturbation (figure 33.3). Should this difference be considered a
sign of an inability of the system of preprogramming to produce more
typical patterns, or is this a sign of a preferred strategy for a changed
CNS?
If the reciprocal strategy is used, preprogramming an increase in
the activity of a “wrong” muscle group can lead to exacerbation of
the effects of the perturbation. The coactivation strategy is more
universal in the sense that it stiffens the joint and thus leads to
attenuation of the effects of perturbations independently of the
perturbation direction. On the other hand, it is always suboptimal,
since it cannot, in principle, lead to total compensation of the effects
of perturbation. This may be the reason why this strategy has not
been seen in highly practiced typical persons who prefer to use the
more effective, although more challenging, reciprocal strategy.
Apparently, this strategy is within a safety zone established by their
unimpaired CNS.
Coactivation could represent the consequence of an impaired
mechanism of preprogramming or it could represent a safety-catch
imposed by the CNS to allow movement to be controlled within the
constraints of its impaired operating capacity. Novices in the early
stages of acquiring a new motor skill frequently demonstrate greater
than optimal levels of coactivation, which appear to increase stability
and reduce the likelihood of error. This excessive coactivation
commonly disappears after the skill is well learned. Coactivation is
also typical of movements of patients with Parkinson’s disease and
of the healthy elderly. Thus, it may well be that muscle coactivation
reflects active intervention by the CNS rather than its inability to use
“more normal” patterns of muscle activation. In the reproducible and
friendly conditions of the laboratory, these internal restrictions may
be lifted, leading to virtually normal performance in motor tests. In
particular, subjects with Down syndrome who are well-practiced
frequently demonstrate a mixture of reciprocal and coactivation
patterns of preprogramming in different trials within the same block
of trials.
Figure 33.2 Two elbow flexion movements performed by a person with Down
syndrome. One trial (solid lines) is characterized by a bumpy trajectory and
numerous irregular EMG bursts. The other trial (dashed lines) is much smoother
and does not show many irregular EMG bursts.
Reprinted by permission from M.L. Latash and D.M. Corcos, “Kinematic and
Electromyographic Characteristics of Single-Joint Movements by Individuals with Down
Syndrome,” American Journal of Mental Retardation 96 (1991): 189-201.
Figure 33.3 A schematic illustration of muscle reactions to an unexpected
change in external load during a task of holding a steady joint position. (a) A
typical reciprocal pattern of EMG changes of an agonist and antagonist muscle
pair to an unexpected loading and unloading. (b) A coactivation pattern more
typical of EMG reactions in persons with Down syndrome.

The weight of the cerebellum has been reported to be lower in

persons with Down syndrome compared to the general population
(Crome and Stern 1967; Molnar 1978; Bellugi et al. 1990). There
have also been reports of changes in the cerebellum in other groups
of atypical persons (Berntson and Torello 1982; Keele and Ivry 1990)
whose movements may be viewed as “clumsy.” These observations,
taken together with the involvement of the cerebellum in synergy
organization (chapter 22; figure 33.4), suggest that the formation of
synergies presents more serious problems for persons with Down
syndrome than for other people. Synergies are assumed to be the
main biological mechanism for controlling a large number of
elements with a relatively few command signals. In particular, they
are supposed to simplify the control of numerous muscles and joints
involved even in rather simple movements of the human body. If this
mechanism is malfunctioning, the everyday motor repertoire, which
all people from the general population take for granted and perform
without any visible effort, may start to pose major problems, affecting
such automatic movements as walking, standing, or reaching for an
object.

Figure 33.4 An illustration of possible changes in motor patterns resulting from

an impairment of the cerebellar involvement in synergy organization.
Reprinted by permission from M.L. Latash, Motor Coordination in Down Syndrome: The
Role of Adaptive Changes, In Perceptual-Motor Behavior in Down Syndrome, edited by D.J.
Weeks, R. Chua, and D. Elliot (Champaign, IL: Human Kinetics, Inc., 2000), 199-223.

33.6 Effects of Practice in

Persons with Down Syndrome
Prolonged practice of single-joint movements leads to a striking
improvement in the performance of persons with Down syndrome:
The movements become much faster and smoother without a
decline in their accuracy (figure 33.5). This improvement can be
transferred to different distances, and different initial and final
positions. The question of whether an improvement acquired in a
standardized laboratory environment may benefit everyday
movements performed in much less reproducible conditions remains
open. There are enough reasons to be pessimistic. When the CNS
of an atypical person for the first time encounters an unpredictable
situation (which occur frequently in everyday life), it may quickly
return to the old, reliable, and safe patterns. It is possible, however,
that practice with an element of uncertainty may be successful in
persuading the CNS that it can reconsider its control strategies and
shift to more effective, albeit more challenging, modes of control.
Figure 33.5 Prolonged practice can lead to a dramatic improvement in the
performance of a simple motor task (fast elbow flexion movements) by persons
with Down syndrome. With practice, peak velocities increased nearly twofold (gray
bars, DS), close to the level of performance of control subjects. Persons without
Down syndrome (control) also benefited from practice (black bars) but to a much
lesser degree.
Adapted by permission from M.L. Latash, “Motor Control in Down Syndrome: The Role of
Adaptation and Practice,” Journal of Developmental and Physical Disabilities 4, (1992): 227-
261. © 1992 Springer.

A study of the effects of practice on finger synergies in persons

with Down syndrome has shown a dramatic improvement in the
indices of synergies in a matter of 2 d (Latash et al. 2002a). In this
study, persons with Down syndrome practiced accurate force
production tasks while pressing with the four fingers of the dominant
hand on individual force sensors. Prior to practice, these persons
showed predominantly positive covariation of finger forces across
trials. That is, if a finger in a particular trial produced more than its
expected share of the total force, other fingers were also likely to
produce more force. As a result, the total force showed large
deviations from its prescribed pattern. This type of finger
coordination may be termed a “fork strategy”: Imagine that you grab
the handle of a fork, turn it upside down, and press with the four
prongs of a fork on force sensors (figure 33.6). If at some point in
time you press more strongly, all the prongs will produce higher
forces, showing positive force covariation. This strategy of using the
fingers does not take advantage of the flexibility of the hand’s design
that enables it to produce higher forces with some fingers and lower
forces with others.

Figure 33.6 If a person tries to produce a constant force level by pressing with
four fingers, two strategies of finger force covariation are possible. (a) A “fork
strategy” when finger forces covary positively. (b) A flexible strategy when the
finger forces covary predominantly negatively. Deviations of the total force are
smaller in the right graph.

PROBLEM 33.5
Suggest a benefit of the “fork strategy” as compared to more
typical patterns seen in persons without Down syndrome.

After 2 d of practice, persons with Down syndrome learned to

perform the task better (more accurately). They also started to show
more typical patterns of negative covariation among finger forces
that apparently helped them to stabilize the total force (figure 33.7).
Interestingly, persons with Down syndrome who used a variety of
motor tasks during practice showed greater beneficial effects of
practice, while those persons who practiced only the main task
showed a smaller improvement. These findings of the more
beneficial effects of variable practice as compared to blocked
practice have also been supported in other studies of persons with
Down syndrome (Edwards and Elliott 1989; Welsh and Elliot 2000).

Figure 33.7 Prior to practice, persons with Down syndrome typically showed a
“fork strategy” (negative values of the synergy index). After 2 d of practice, the fork
strategy is replaced by a more typical negative covariation of finger forces (positive
values of the synergy index) that help stabilize the total force. Persons who used
variable practice (thin solid line) showed greater improvement than those who
used blocked practice (thin dashed line).
Reprinted by permissions from M.L. Latash, N. Kang, and D. Patterson, “Finger
Coordination in Persons with Down Syndrome: Atypical Patterns of Coordination and the
Effects of Practice,” Experimental Brain Research 146 (2002): 345-355. © 2002 Springer.

33.7 Autism
Autism was first described in 1943 by Dr. Leo Kanner. It is estimated
that babies with autism are born at a rate of 1 out of 250 live births.
One and a half million Americans are estimated to have autism.
Autism is a spectrum disorder. This means that it is characterized
by a spectrum of signs that may not necessarily be linked to a single
neurophysiological or other mechanism. Persons with autism
commonly show resistance to change, distress for unclear reasons,
difficulty in mixing with others, lack of responsiveness to words, and
difficulty in verbal expression, sometimes reflected in repeating
words. Their motor behavior may be characterized by such dissimilar
features as physical overactivity or extreme underactivity. They may
show stereotypical, repeated movements and sustained odd play.
Their gross and fine motor skills may show very uneven
development.
Asperger’s disorder is a neurobiological disorder named after
the Austrian physician Hans Asperger, who in 1944 described a
pattern of behaviors in young boys who had normal intelligence and
language development, but who also exhibited autistic-like behaviors
and marked deficiencies in social and communication skills. Persons
with Asperger’s disorder are similar to persons with autism in their
inability to regulate social interaction by using nonverbal behaviors
such as body posture and gestures, eye contact and facial
expression. They show a preoccupation with stereotypical behaviors,
routines, and rituals and have problems making social contacts and
developing peer relationships. Unlike persons with autism, those with
Asperger’s disorder show no significant delay in language, in
cognitive development, or in the development of self-help skills,
adaptive behaviors (other than in social interaction), and curiosity
about the environment in childhood.
Babies with both autism and Asperger’s disorder often show
abnormal features of motor development, including lack of protective
reflexes when falling and delayed development of walking
(Teitelbaum et al. 1998, 2004). There is increasing evidence that
both disorders are associated with cerebellar abnormalities, more
pronounced in autism, including reduced cerebellar gray matter and
neuronal loss in the cerebellar nuclei (Courchesne 1997; Palmen et
al. 2004; McAlonan et al. 2005). As discussed earlier in the section
on Down syndrome, this may be related to problems in assembling
multi-element synergies. Here, under synergies, we imply not only
acts of motor coordination but a more general notion of elements
that are brought to work together. The role of elements can be
played by words—then, one observes difficulties with such things as
verbal expression and agrammatisms. The role of elements can also
be played by persons—then, problems with communication and
interacting with others may be expected.
Autism is characterized by a specific feature that suggests that
persons with autism have difficulties perceiving others as “other
beings just like me” who can have their own views and ideas. This
feature is revealed in the false belief test (Muris et al. 1998; Pilowsky
et al. 2000; Grant et al. 2001). In this test, a person is shown a
sequence of pictures with two kids and a cake. One of the kids puts
the cake into the refrigerator to make sure that it does not get
spoiled. Then, this kid leaves the room. The other kid thinks that the
cake is to be eaten in a few hours and it makes sense to warm it up.
So, the second kid takes the cake from the refrigerator and puts it
into a cabinet. Then, the second kid leaves the room, and, after
some time, the first kid enters. The question is: “Where will the first
kid look for the cake?” Giving the right answer (“In the refrigerator”)
implies that the person taking the test understands that these
personalities may have wrong views of the world (false beliefs)
based on outdated information. Children with autism have been
shown to fail this test at ages when typically developing children
pass it.

33.8 Developmental Coordination

Disorder
Developmental coordination disorder (DCD) is a childhood
disorder characterized by poor coordination and clumsiness
(reviewed in Gillberg and Kadesjo 2003). It is more common in boys
than in girls. DCD is a rather vague grouping, based mainly on
observable behaviors that might be the outcome of a range of
underlying problems. So there may be a number of possible
subgroups of persons who are diagnosed as having DCD. Roughly 1
out of 20 school-age children have some degree of DCD. Children
with this disorder may trip over their own feet, run into other children,
have trouble holding objects, and have an unsteady gait. These
children have difficulties in mastering gross motor coordination skills
such as crawling, walking, jumping, standing on one foot, and
catching a ball and fine coordination tasks such as tying shoelaces.
Some children also demonstrate expressive speech problems.
In tests with accurate pointing, children with DCD show lower
accuracy than their age-matched control peers. Both typically
developing children and those with DCD increase the scatter of the
endpoint final positions with target size; however, children with DCD
show much larger scatter across different targets (Smits-Engelsman
et al. 2003).
Children with DCD show developmental delays in sitting up,
crawling, and walking; deficits in handwriting and reading; and
problems with gross and fine motor skills; and in general they may
appear “clumsy.” They show problems with postural control that
become obvious in challenging tasks, such as one-leg standing
(Geuze 2005). In a variety of motor tasks, they show atypical
features that should look familiar to readers. In particular, during fast
movements, these persons show increased levels of coactivation of
agonist–antagonist muscle pairs, while during manipulation of
handheld objects, they show higher grip forces and increased safety
margins (Pereira et al. 2001; Rainor 2001; Zoia et al. 2005). The
commonalities of these features with those demonstrated by the
healthy elderly and, as we will see in future chapters, by persons
with different motor disorders, suggest that these features are indeed
adaptive to deficiencies in motor performance that may have
different causes. Just like in other cases of atypical development
described in this chapter, motor abnormalities in DCD have been
related to a cerebellar dysfunction (O’Hare and Khalid 2002; Ivry
2003).

CHAPTER 33 IN A NUTSHELL
Human babies are born with an
incompletely developed central nervous
system. The process of myelinization
continues over the first 6 mo of life.
Motor development typically follows a
progression along a set of milestones.
The idea of a fixed timetable has been
challenged by proponents of a view
that motor development results from
changing interactions between
constraints imposed by the organism
and by the environment. Persons with
Down syndrome have an extra chromosome
in the 21st pair. This leads to both
motor and nonmotor consequences. Motor
consequences can be summarized as
“clumsiness,” which gets contributions
from longer reaction times, slower
movements, and higher variability.
Persons with Down syndrome show
preference for cocontraction patterns
of muscle activation in a variety of
motor tasks. Practice leads to
dramatic changes in motor patterns of
these persons. They become faster in
simple single-joint motor tests. Their
multifinger synergies show shifts
toward patterns observed in persons
without Down syndrome. Autism and
Asperger’s disorder are characterized
by features that may be partly
summarized as a poor ability to form
synergies; this impairment may be
causally related to cerebellar changes
in these states. Development
coordination disorder is a delay in
motor coordination that is also
potentially linked to a cerebellar
dysfunction.
 Chapter 34

Motor Learning

KEY TERMS AND TOPICS

practice
learning and adaptation
memory
motor skill
neural plasticity
stages of motor learning
learning synergies

In this chapter, we consider changes in performance following

multiple attempts at solving a motor task (practice). Performance
changes may be accompanied by changes in multiple parts of the
body, from muscles to spinal circuitry, and to different brain
structures. These changes may be more or less transient (lasting
from seconds to the lifetime), limited to the effectors directly involved
in the task or seen in other effectors up to the whole body, seen only
during performance of the practiced task or in task variations, or
even in seemingly unrelated tasks. In describing the effects of
practice, we will try, when possible, to link such effects to specific
mechanisms of motor control and changes in neurophysiological
structures and loops.
 We are not going to consider negative effects of motor practice on
performance—for instance, those due to effects of fatigue (described
in chapter 31) or even damage to bones, joints, ligaments, and
muscles. We will also not consider psychological factors such as
attention, motivation, and alertness, which go beyond the scope of
this book. Since this book is not on motor learning, we will not
discuss in detail such important issues as learning curves,
interference among tasks, and specific practice for strength training
and different skills. These topics are covered in the following books:
Bernstein 1947/2020, 1996; Schmidt and Lee 2020; Zatsiorsky et al.
2021.

34.1 Adaptation, Learning, and

Memory
It is common knowledge that solving a motor task multiple times can
lead to improved performance. Some of these improvements do not
last long, especially when a novel task is presented over a relatively
brief time interval. For example, wearing prismatic glasses (see
chapter 30), which distort the visual image of the environment, at first
leads to inaccurate reaching movements because target locations
are perceived wrongly. With repetitive reaching, movements rather
quickly become more and more accurate. These changes, however,
quickly disappear after the glasses have been taken off. Such
transient, easily reversible changes in performance are known as
adaptation. If practice is followed by longer-lasting changes in
performance and, for example, its effects can be seen the next day,
the effects of practice are called learning.
There is not a strict border between adaptation and learning
because their definitions rely on such subjective characteristics as
“easily reversible” and “longer-lasting.” To complicate the situation,
long-lasting changes in peripheral structures, such as muscles,
induced by strength training exercises are commonly known as
adaptations (Zatsiorsky et al. 2021). So, to claim learning effects,
one has to ascertain that the effects of practice are seen in
neurophysiological structures, with or without changes in muscles.
However, this may also be insufficient. For example, long-lasting
asymmetry of monosynaptic reflexes can be induced using operant
conditioning methods following numerous repetitions (Wolpaw 1987;
Wolpaw and Carp 1993; see section 34.6). Such effects can be
called adaptation or learning.
There is little argument that the effects of motor learning involve
mechanisms of memory. This is another central concept without an
unambiguous definition. For how long should the effects of an event
be detectable to quantify as memory? For example, would you use
this word to describe a piece of playdough, which remains deformed
after you squeeze it? What about an old photo that has faded to a
state when one can hardly recognize the objects, or an old floppy
disk that has no drive to read its contents? Is a nail in the wall a
memory of the painting that used to hang from that nail? We will
consider two types of memory, those that do not require a key to be
recollected and those that do.

PROBLEM 34.1
When you look at a great painting, you feel emotions reflecting the
ideas of the artist who died a long time ago. Can we view the
painting as a memory of those ideas?

34.2 Muscle Memory

In physics, the term hysteresis means that a physical system can
display different behaviors depending on its immediate history. There
are effects of hysteresis in elements of the human motor system, in
particular in muscles (Partridge 1965; Joyce et al. 1969; Gottlieb and
Agarwal 1988). For example, imagine that an isolated muscle is
being stimulated at a constant rate by an external stimulator (figure
34.1). Muscle length is changed very slowly (such that we can ignore
the dependence of muscle force on velocity), and muscle force is
measured. If the muscle is stretched, its force will change along
curve-1 in figure 34.1, while if the muscle passes through the same
values of length while being shortened, its force changes along
curve-2. This means that muscle force depends not only on its length
but also on its history.

Figure 34.1 An isolated muscle is being stimulated at a constant rate and

strength. Muscle force (F) will depend on muscle length (L). During slow stretching
and slow shortening, the muscle will display different dependences F(L) shown by
the two curves, curve-1 and curve-2. This behavior is called hysteresis.

Another example of changes in muscle properties that depend on

its immediate history is the catch phenomenon. It represents a
transient increase in muscle response to a standard neural signal
following a brief episode of very strong muscle contraction. This
phenomenon can be seen, for example, for about 1 to 2 s following a
brief, strong electrical stimulation of a muscle leading to its tetanic
contraction. This phenomenon has been shown to reside in the
muscle itself rather than in neural structures or neuromuscular
synapses (Burke et al. 1970, 1976).
 PROBLEM 34.2
Give a couple of examples of hysteresis from everyday life.

There are several classifications of memory. One of the basic

classifications is that into declarative and nondeclarative memory.
The former is also known as explicit memory and the latter as implicit
or procedural memory. Declarative memory generally refers to
memories of events, objects, persons, and other things. This is a
memory of “what.” In contrast, nondeclarative memory is about “how
to.” In particular, it involves motor skills and habits, conditioned
reflexes, and other phenomena that are particularly close to the topic
of this book. Further, we will consider only examples and
hypothetical mechanisms of nondeclarative memories that have
direct relations to motor learning.

34.3 Habituation of Reflexes

In the beginning of the 20th century, it was known that repetitive
activation of spinal reflexes causes them to become smaller or to
require larger stimuli. This effect has been termed habituation.
Habituation may last considerable amounts of time; thus, it is closer
to learning than to adaptation. Habituation can be seen in human
experiments, such as with the startle reaction (a phasic motor
reaction, seen primarily in extensor muscle groups, to an
unexpectedly loud auditory stimulus), which decreases in magnitude
when the stimulus is applied again and again.
A habituated response can be restored or recovered after a period
of rest or if another, typically strong stimulus is presented to the
animal or person. This phenomenon is called dishabituation. Even in
the absence of habituation, a strong stimulus can enhance a
response to a standard stimulus. This phenomenon is called
sensitization.
Habituation, dishabituation, and sensitization are all examples of
nonassociative, nondeclarative learning. Nonassociative learning
occurs when the central nervous system learns certain properties of
a stimulus—that is, it adjusts its response when the properties of the
stimulus are repeated.

34.4 Conditioned Reflexes

One of the subtypes of nondeclarative memories is associative
learning that involves creating a relation between two stimuli.
Typical examples include classical conditioning and operant
conditioning. Classical conditioning involves a special group of
phenomena termed conditioned reflexes.
It is known that having food in one’s mouth leads to salivation,
which is particularly strong if the food is dry. Food stimulates the
mucous membrane of the mouth; this stimulation is transferred by
sensory nerves to the salivatory brain center (salivatory nuclei in the
pontine tegmentum), which reacts to the stimulation with a command
to the salivary glands. This phenomenon occurs even in the smallest
baby animals. Such inborn mechanisms are termed inborn reflexes.
Famous Russian physiologist I.P. Pavlov discovered that, if a hungry
dog each day hears a bell or a whistle, or sees a bulb of a certain
color turning on, or something else half a minute prior to feeding,
after many days of repeating this pattern, the animal gradually starts
to salivate not when it gets the food, and not even when it sees the
food, but when the additional signal is turned on. It is clear that, in
such a case, the experimenter witnesses the birth of a new version
of the salivation reflex elaborated with artificial means. This reflex
reflects an enrichment of the personal experience of the dog. These
reflexes were termed conditioned in contrast to the inborn reflexes.
Pavlov suggested a theory of brain functioning based on an
interaction among inborn and conditioned reflexes. Pavlov’s theory
of conditioned reflexes considers the brain as a purely reactive
organ whose behavior is defined by environmental stimuli, while
human (or animal) behavior is seen as the process of “equilibrating”
the body with the environment. Maintaining homeostasis in a
changing environment is indeed crucial for both individual and
evolutionary survival. However, higher animals are not purely
reactive systems but are actively exploring the environment,
formulating their needs, and trying to satisfy them, rather than
waiting for appropriate stimuli for conditioned reflexes from the
environment. Activity is the driving force of the functioning of the
human brain.
This was understood by the middle of the 20th century by
Bernstein, who formulated the goals of behavior as overcoming the
environment rather than equilibrating the body with it. Bernstein
created the physiology of activity, a whole new field of study that
tries to explain behavior based on internal needs and goals of an
animal (or a human), which are commonly in conflict with stimuli from
the environment (Bernstein 1947/2020, 1966).
Bernstein created the physiology of activity largely based on his
earlier studies of the process of learning labor and athletic
movements. He found out that, in the process of learning, the
variability of movement trajectories and other characteristics was not
eliminated. Movements did not become identical or machinelike,
although the ultimate motor outcome became highly reproducible.
For example, during such movements as shaving with a sharp razor
blade or shooting at a target, the success of the ultimate outcome
depends on precision within fractions of millimeters or angular
seconds. Only the high variability of the behavior of individual
elements of well-learned movements allows subjects to achieve such
amazing accuracy in conditions of unexpected forces from the
environment and varying intrinsic body states. Therefore, memory
traces of such movements cannot represent “motor programs” such
as combinations of patterns of muscle forces, muscle activation
levels, or joint trajectories. It is safe to say that skill formation is an
active process by the central nervous system, the process leading to
discovery of dynamically stable solutions with respect to salient
performance variables (see chapter 22).

PROBLEM 34.3
 Pavlov built “towers of silence” where his dogs were deprived of
any stimuli except those used for conditioned reflexes. Predict the
behavior of the dogs based on the theory of conditioned reflexes
and Bernstein’s idea that all actions are initiated from within the
central nervous system.

34.5 Operant Conditioning and

Learning Spinal Reflexes
Another typical procedure used in animal memory studies is operant
conditioning. During operant conditioning, an animal is rewarded
(with a small portion of a favorite food) for correct behavioral
responses. The response may be under apparent control by the
animal (for example, choosing a correct turn while running in a
maze) or it may not (for example, during spinal reflex responses).
However, even the simplest monosynaptic spinal reflexes can show
stable changes and store memory traces in operant conditioning
experiments, as shown in the ingenious studies by Jonathan Wolpaw
and his group (Wolpaw 1987; Wolpaw and Carp 1993). Achieving
such effects may require thousands of repetitions, but the animal’s
central nervous system eventually learns to modify an apparently
uncontrollable phenomenon, such as the amplitude of a
monosynaptic response. Similar effects can be observed in humans
who require considerably fewer repetitions (Thompson and Wolpaw
2015).

PROBLEM 34.4
What physiological mechanisms could be involved in changes of
the amplitude of the monosynaptic responses in Wolpaw’s
experiments?
34.6 Short-Term and Long-Term
Memory
Two overlapping categories of memory are identified in humans. The
first one is called short-term memory. Its effectiveness usually lasts
a few minutes or hours. Special, memory consolidation processes
are invoked to explain the process of conversion of short-term
memories into long-term memories, which can last for years, up to
a lifetime (figure 34.2). There are many clinical cases, mainly
different types of brain trauma, leading to a memory loss or
deficiency. In particular, some patients lack the ability to consolidate
short-term memory. They demonstrate normal ability to maintain
short-term memories, but lose them when their attention is
distracted. There are also cases when a trauma leads to a loss of
memories related to events that happened for some time period prior
to the trauma (retrograde amnesia) or to a loss of memories related
to events that happened for some time period immediately following
the trauma (anterograde amnesia).
Figure 34.2 Processing of a sensory stimulus may lead to the creation of a short-
term memory trace in parallel with the production of an effector (motor) output.
Short-term memory can be consolidated into a long-term memory. Both short-term
and long-term memories can participate in processing future incoming sensory
stimuli and forming an effector response.

The process of consolidation depends on the context in which a

memorized event occurs as well as on the number of repetitions of
presenting the event. For example, highly emotional events have
higher chances of being consolidated into long-term memory,
sometimes even after a single presentation. More boring things, like
the contents of the present book, require repetition and effort to be
remembered.
If memory is considered as an emergent property of many
neurons that pertain to many neuronal structures, it is obvious that
searching for a neuronal or synaptic location for a certain memory is
doomed to failure. In classic experiments by Karl Lashley, a learned
motor task in rats was preserved even after the surgical removal of
up to 98% of the cortical areas related to the task. Moreover, it did
not matter which 2% of the areas were spared by the surgical
procedure. These observations have suggested that memory is a
distributed phenomenon.
 There was a theory, advanced in the 1930s by the great
neuroscientist Rafael Lorente de No, that short-term memory was
based on reverberating circuits—that is, chains of neurons—that
give rise to a certain pattern of activity sustained in time and in
space. This theory has not been confirmed, however, in experiments
that failed to find repeated bursts of activity in anatomically identified
closed loops.
An alternative theory suggests that short-term memory results
from modifications of neuronal ion channels and mediator receptors
leading to changes in the presynaptic mediator release mechanisms
or in the postsynaptic sensitivity to the mediator. This view is based
on numerous observations of synaptic changes that may last from a
few hundred milliseconds to several hours or even days (Linden
1996; Hodgson et al. 2005; Nicoll and Schmitz 2005). Longer-lasting
synaptic changes have been described, including long-term
potentiation, long-term depression, and activity-dependent
presynaptic facilitation. These changes may be relevant to
information storage over time; however, their relation to human
memory is rather speculative.
Other theories of short-term memory are based on assumptions of
the creation of new neuronal pathways or the synthesis of new
proteins. Protein synthesis has also been implicated in the process
of consolidation of short-term memories into long-term memories.
With respect to motor memory, the neural mechanisms are
unknown due, in particular, to the fact that the researchers do not
know how hypothetical control variables (Bernstein’s engrams) are
represented and stored in the CNS. It is commonly assumed that
motor memory resides in synapses. Note that using synapses for
storing long-term memory looks rather noneconomical: Using a
synapse to remember an event makes it occupied and useless for
future memories. As a result, numerous “disposable synapses” are
required to memorize a single event. Even the astronomical number
of synapses within the CNS may be insufficient if such a crude and
straightforward mechanism is used. More likely, memorizing an event
requires a pattern of activity in complex neuronal formations whose
organization and neuronal composition may vary.
 One of the most ubiquitous features of human voluntary
movements is their variability. Even the most skilled movements
demonstrate different kinematic profiles, different muscle force
patterns, and different muscle activation patterns in successive trials.
The variability of patterns of skilled movements suggests that there
is a comparable variability in the afferent signals from proprioceptive
receptors. Hence, one is safe to assume that virtually all the neurons
in the body demonstrate somewhat different patterns of activity
during repetitions of even the simplest motor task. This is probably
the strongest argument that a control pattern for a skilled movement
cannot be represented as a combination of activity of a number of
individual neurons induced by stable changes in individual synapses.

34.7 Adaptation to Unusual

Force Fields
Studies of motor adaptation and learning have used the method of
applying unusual external force fields, such as force fields where the
force direction was orthogonal to the movement direction, and the
force magnitude was proportional to the speed (figure 34.3;
Shadmehr and Mussa-Ivaldi 1994; reviewed in Shadmehr and Wise
2005; McNamee and Wolpert 2019). Under such conditions,
naturally straight trajectories observed during reaching (see chapter
24) become curved. After some practice, the trajectories become
straight again in the presence of the force field. And after the force
field has been turned off, the trajectories show an aftereffect: They
become curved in the opposite direction. These typical stages are
illustrated in the four panels of Figure 34.3.
Figure 34.3 (a) Natural reaching movements are straight. (b) When an unusual
force field is turned on with the force proportional to speed and directed orthogonal
to the velocity vector (central drawing), trajectories become curved. (c) After some
practice, the trajectories become straight again in the presence of the force field.
(d) After the force field has been turned off, the trajectories show an aftereffect:
They become curved in the opposite direction.

The most common interpretation of the effects of adaptation to

such force fields invokes the notion of internal models (see chapter
21). It assumes that the central nervous system develops a
representation of the unusual force field and modifies muscle
activations to compensate for its effects on the effector. The
cerebellum has been most commonly considered to be the likely side
for the formation of such hypothetical internal models (Wolpert et al.
1998; Kawato et al. 2003; Imamizu and Kawato 2012; Herzfeld and
Shadmehr 2014).

PROBLEM 34.5
An alternative explanation of returning to straight trajectories with
practice is using high muscle coactivation, which reduces the
kinematic effects of the force field. Present arguments for or
against this explanation.
 Within the equilibrium-point hypothesis (chapters 20-21), changes
in neural commands with practice are described in terms of control
trajectories, such as the time changes in the tonic stretch reflex
threshold, λ(t) for a single muscle or in the referent coordinates of
the involved effector, RC(t). Currently, rules that result in adjustments
of control trajectories during practice are unknown. An elegant
solution has been suggested (Gribble and Ostry 2000) that takes
advantage of the fact that RC is expressed in actual kinematic
variables. If one wants to move an effector along a certain trajectory
and observes a deviation from the desired path (figure 34.4a), the
RC trajectory (dashed lines in figure 34.4) is adjusted in the next trial
following a simple rule (figure 34.4): The difference between the
actual and desired trajectories (34.4b) is added to the previous RC(t)
function (figure 34.4c). Using this simple rule has been shown to
lead to a very quick adaptation to novel conditions taking 2 to 3 trials
without any computations related to the complex transformations of
neural and mechanical variables involved in producing the task
mechanics.
Figure 34.4 (a) Referent coordinate trajectory (RC) is similar to the actual
trajectory (TR) in shape in the absence of force field. (b) When an unusual force
field is on (central image, see figure 34.3), TR becomes curved. (c) The deviation
from the desired trajectory is perceived by the subject and used to correct RC.
This results in a straight TR after just one trial. (d) After the force field has been
turned off, the trajectories show an aftereffect: They become curved in the
opposite direction.

34.8 Motor Skills

Motor skills are another example of implicit, nondeclarative
memories. There are many classic books on motor learning,
including a book by Bernstein, On Dexterity and Its Development
(Bernstein 1996), as well as more recent textbooks. Unfortunately,
neurophysiological mechanisms of motor learning are not known,
although recent studies of the phenomenon of neural plasticity
promise progress in this area (see chapter 41).
Probably, the first and most important question related to motor
learning is, “What is stored in memory?” Commonly, a poorly defined
term motor program is invoked to answer this question. This term
refers to some hypothetical variables that translate into required
movement patterns after being recalled. Most commonly, the nature
of variables forming motor programs (e.g., in what units they are
measured) is not explicitly defined.
 Bernstein (1935) introduced a different term, engram, to describe
brain representations of learned motor skills. In his theory, engrams
encoded topological features of actions, not their metric features.
Hence, they did not represent such variables as forces and
coordinates (which are metric in nature) but rather patterns of neural
variables encoding patterns of salient performance variables.
Can a learned pattern of neural variables be transferred to
different effectors? The answer seems to be yes based, for example,
on the demonstrated ability of humans to preserve their specific
features of handwriting while writing in different force fields and with
an implement held by such different effectors as the hand, the foot,
the elbow, and even the mouth (Bernstein 1967; Raibert 1977).
When humans learn an unusual modification of writing, such as
writing in the mirror with the dominant arm, the improvement can be
seen in the ability to write in this unusual way with the unpracticed
nondominant arm (Latash 1999; figure 34.5). Accepting the
affirmative answer has important implications for motor control; in
particular, it means that the hypothetical neural variables do not
prescribe patterns of neural commands to specific muscle groups,
muscle forces, and joint torques.

34.9 Learning Motor Synergies

All natural movements are performed in conditions of motor
redundancy, in which the number of elements is higher than the
number of task constraints (see chapter 22). According to the idea of
performance-stabilizing synergies (see section 22.7), such tasks are
associated with families of solutions that can all generate adequate
time profiles of important performance variables. Variance in the
space of elemental variables can be viewed as consisting of two
components, good variance and bad variance. By definition, only
bad variance has an effect on variance of the salient performance
variable, while good variance does not. It is natural to expect bad
variance to drop with practice; otherwise, accuracy of performance
would not improve. It is much less obvious, however, what could
happen with good variance.

Figure 34.5 Samples of writing by two subjects, who used (a) separate-letter
writing and (b) fused-cursive writing. Both subjects improved their timing index (TI)
and error index (EI) with practice. Note the improved handwriting by the
nondominant hand that had not practiced and also in the writing of a new,
nonpracticed phrase.
Reprinted by permission from M.L. Latash, “Mirror Writing: Learning, Transfer, and
Implications for Internal Inverse Models,” Journal of Motor Behavior 31, (1999): 107-112. ©
1999 HELDREF Publications. https://www.tandfonline.com/.

Let us consider two elements that try to produce an accurate

magnitude of their summed output. So, the task is E1 + E2 = C,
where E1 and E2 are the outputs of the elements, and C is a desired
constant value. Let us assume that prior to practice there was a
degree of covariation between E1 and E2 such that the good variance
(VGOOD) was larger than the bad variance (VBAD) and that with
practice, VBAD dropped. Figure 34.6 illustrates three scenarios of
possible changes in VGOOD: It can decrease proportionally to the
change in VBAD (unchanged synergy), it can remain unchanged or
even increase (improved synergy), and it can drop more than VBAD
(apparently synergy becomes worse).
 PROBLEM 34.6
By definition, “good variance” has no effect on important
performance variables. Why is it called “good” and not
“irrelevant”?

Figure 34.6 Prior to practice, there was covariation between E1 and E2

stabilizing the task variable (C). As a result, the good variance (VGOOD) was larger
than the bad variance (VBAD). With practice VBAD dropped. VGOOD can decrease
proportionally to the change in VBAD (left, unchanged synergy), it can remain
unchanged or even increase (bottom, stronger synergy), and it can drop more than
VBAD (right, weaker synergy).

Experiments have provided evidence for all three scenarios

illustrated in figure 34.6 (Domkin et al. 2002, 2005; Kang et al. 2004;
Wu et al. 2013; Wu and Latash 2014). Practicing a novel task is
associated with the creation and strengthening of synergies
stabilizing salient variables (VBAD drops more than VGOOD). After the
person reaches a floor in VBAD changes, further practice results in a
drop in VGOOD without much change in VBAD. These changes are
viewed as optimizing performance with respect to variables that are
not used in the task formulation—for example, searching for more
energy efficient or more comfortable solutions.

34.10 Stages in Motor Learning

In his classic books, Bernstein (1947/2020, 1996) suggested a three-
stage scheme of motor learning. He viewed the typical redundant
design of the motor apparatus as a computational problem that
required elimination of redundant degrees of freedom to make the
system controllable. So the first stage was associated with
elimination or freezing degrees of freedom. During the second stage,
the controller was assumed to start releasing degrees of freedom to
achieve stability of performance in the changing environment. The
third stage was assumed to lead to an optimal interaction with
external forces such that only those external forces that hurt
performance were counteracted, while those that help performance
were not.
This three-stage scheme has become classic (e.g., Vereijken et
al. 1992) despite the imprecise expressions freezing degrees of
freedom and releasing degrees of freedom. Indeed, if a variable is
involved in a task, one cannot simply cross it out. It is possible to
vary the range of changes in that variable, but this does not change
the total number of degrees of freedom. Note also that reducing the
range of motion in a joint of a multijoint limb during a fast movement
requires sophisticated control of muscles crossing that joint because
of the joint coupling (see chapter 24). So, trying to “freeze” a joint
does not automatically make the task simpler.
Viewing motor redundancy as abundance (see chapter 22)
questions the foundation of the three-stage scheme. Indeed, one has
to learn ways to stabilize salient performance variables with all the
available degrees of freedom to learn any task in natural conditions.
Further learning can go beyond performing the explicit task and
achieve such features as comfort, economy, and esthetics. This
corresponds to the two stages introduced at the end of the previous
section: stage 1 associated with developing and strengthening
relevant synergies, and stage 2 associated with selecting families of
solutions that satisfy additional explicit or self-imposed requirements.
Is there a third stage corresponding to Bernstein’s concept of
utilization of external forces? It seems that this concept is already
included under the concept of performance stabilization.

34.11 Effects of Practice on

Cortical Representations
As mentioned earlier (chapters 8-11), there are many motor maps
and sensory maps within the central nervous system that look like
distorted drawings of the whole body or of body parts. A number of
studies have shown that plastic changes happen within the central
nervous system of a healthy adult, leading to changes in the motor
and sensory maps following motor learning. Such changes have
been commonly documented using transcranial magnetic stimulation
and observing an increase in the cortical area from which a motor
response could be produced, in the threshold for inducing a
response, and in the size of the response. In particular, such
changes have been reported in people who learned how to read
Braille or to play a musical instrument (Pascual-Leone et al. 1995;
Sterr et al. 1998; Pascual-Leone 2001).
Several studies have shown that plastic changes within the central
nervous system can occur within a very short time in a single
experimental session limited to 1 to 1.5 h (Classen et al. 1998;
Latash et al. 2003). In one of the studies, the unexpected TMS-
induced responses were applied during the task performance, and
their effects were detrimental for performance since they produced
irregular, unexpected jerky movements. After about 1 h of practice in
such conditions, the amplitude of the response to the TMS
decreased. These results show that practice can lead not only to an
increase in the excitability of neural structures involved in the task
but also to a decrease in their excitability if this helps improve task
performance. More material on plastic changes in brain
representations with injury and rehabilitation will be presented in
chapter 41.

CHAPTER 34 IN A NUTSHELL
Practice leads to both transient,
easily reversible, changes in
performance (adaptation) and longer-
lasting changes (learning) based on
neuronal memory. The effects of memory
can be seen at different levels,
including habituation of reflexes,
classical or operant conditioning, and
skill acquisition. Pavlov’s theory
considers animal behavior to be a
combination of inborn and conditioned
reflexes. Bernstein’s physiology of
initiative emphasizes the principle of
activity and views motor skill
acquisition as an active search
process. Neuronal mechanisms of memory
are unknown. Long-term potentiation
and depression are frequently viewed
as the neuronal mechanism of memory
implying synapses as the storage site.
The classic theory of three stages in
learning a motor skill involves
freezing and releasing degrees of
freedom followed by optimized use of
external forces. Even short practice
can lead to plastic changes in the
brain motor and sensory maps. Within
the concept of abundance, there are
two stages: (1) developing and
strengthening relevant synergies and
(2) selecting families of solutions
that satisfy additional explicit or
self-imposed requirements.
Problems for Part VII
Self-Test Problems
1. A person practices a fast reaching movement to a target while
sitting in the center of a rotating centrifuge. Vision is removed
as soon as movement is initiated. What differences in
movement trajectories would you expect? What will happen
with the trajectories after prolonged practice? What will
happen after the practice when the same movements are
performed in a motionless centrifuge?
2. A person tries to keep maximal ankle plantarflexion force
against a stop for about 2 min. The force drops by about 50%.
A brief electrical stimulus is applied to the muscle nerve (n.
tibialis) at the end of this period on the background of
continuing flexion effort. The stimulus induces an M-response
and an H-reflex. What can you say about motor units that
contributed to the M- and H-responses? Would you expect the
responses (M-, H-, and force twitch) to be higher or lower than
those seen in response to the same stimulus in the same
muscle tested in nonfatigued, relaxed conditions? Why?
3. A toddler stands in a bus, and the bus suddenly starts to
move. What mechanisms would help the toddler not to fall
down? How do they differ from similar mechanisms in a
grown-up?
4. A single TMS stimulus applied over the primary motor cortex
of a person who generates a low level of hand force into
flexion produces a quick response in the wrist flexor (flexor
carpi radialis), followed by a silent period in the EMG. The
same person, after a fatiguing exercise, matches the
background hand force, and a TMS stimulus with the same
parameters is applied over the same spot. What changes
would you expect to see in characteristics of the muscle
response? Explain your answer.
5. A standing healthy elderly person is unexpectedly pushed
from behind by another person (not too strongly!). Describe all
 the differences in posture-stabilizing responses from those
seen in a young person.
6. Older persons frequently show increased tremor across a
variety of actions. Describe all possible mechanisms that
could bring about the increased tremor, assuming no
diagnosed neural pathology exists.

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question.
Write a short phrase explaining why you chose this answer.
1. Compared to a typical young person, a typical 80-year-old
woman is likely to show
a. lower grip force while drinking from a glass
b. larger motor units
c. earlier and stronger anticipatory postural adjustments
d. increased monosynaptic reflexes
e. none of the above
Why?
2. After a fatiguing exercise, the following changes are likely to
be seen in response to a quick muscle stretch:
a. an increase in the M1 response
b. an increase in the M3 response
c. a drop in the time delay of both responses
d. an increase in the percentage of fast fatigable motor units
responding to the stretch
e. none of the above
Why?
3. A young boy shows poor results in standard tests on
cognition, particularly spatial cognition, but talks a lot using
grammatically correct sentences. What is the likely cause?
a. impaired cerebellar development
b. impaired cortical development
 c. copper accumulation in the basal ganglia
d. autism
e. any of the above
Why?
4. A child with Down syndrome is asked to catch a quickly
thrown basketball. Which of the following would you expect to
see?
a. no anticipatory postural adjustments
b. tremor in the arms at 6 Hz, getting stronger as the ball
approaches
c. muscle cocontraction about the major leg joints
d. no preprogrammed corrections after the ball’s impact
e. all of the above
Why?
5. As compared to a typical young adult, during accurate finger
actions, an older person is expected to show
a. increased indices of finger independence
b. decreased indices of performance-stabilizing synergies
c. higher reliance on visual feedback
d. higher coactivation of agonist and antagonist muscles
e. all of the above
Why?
 Part VIII

Motor Disorders
 Chapter 35

Peripheral Muscular and

Neurological Disorders

KEY TERMS AND TOPICS

muscular dystrophies
myasthenia gravis
peripheral neuropathies
tetanus
stiff-man syndrome
diabetes

Clinical observations present a serious challenge to

neurophysiological, behavioral, or biomechanical analysis of motor
behavior. First, the language of clinical reports differs substantially
from the language of theoretical or experimental studies of
movements and requires some kind of translation to make the data
interpretable. For example, such commonly used clinical terms as
hypotonia and rigidity do not have clear neurophysiological
definitions and reflect only the impression that an expert (e.g., a
neurologist or a physical therapist) gets during clinical examination of
a patient. Second, movement pathologies are usually defined in
terms of signs and symptoms rather than in terms of
neurophysiological mechanisms, which is due, in part, to the lack of
knowledge about the underlying mechanisms. At best, some of the
commonly studied motor pathologies are associated with dysfunction
of a relatively well-localized anatomical formation within the central
nervous system—for example, with pathological changes in the
basal ganglia as in Parkinson’s disease and Huntington’s chorea
discussed in chapter 37.

Table 35.1 Sites of Damage in Peripheral Disorders

Site Disorder
Spinal root Radiculopathy (cervical or lumbar)
Peripheral axon Axonal neuropathy
Myelin sheath Guillain-Barré syndrome
Neuromuscular synapse Myasthenia gravis
Muscle Myopathy

35.1 Myopathies and

Neuropathies
We begin the review of common motor disorders with those that
originate from a dysfunction in the muscles, neuromuscular
synapses, peripheral axons, and bodies of α-motoneurons. Table
35.1 shows a brief list of sites where things can go wrong and the
associated movement disorders. The disorders can be classified into
neuropathies (a problem in the neural structures), and myopathies (a
problem in the muscles).

35.2 Muscular Dystrophies

Muscular dystrophies are a group of genetic diseases
characterized by progressive weakness and degeneration of the
skeletal muscles. Among males, Duchenne and Becker muscular
dystrophies together affect 1 in 3,500 to 5,000 births (Khurana and
Davies 2003; Tsao and Mendell 1999; Voisin and de la Porte 2004).
This results in about 500 boys born in the United States with these
conditions each year. Females are rarely affected by these forms of
muscular dystrophy.
Duchenne muscular dystrophy is the result of a mutation in the
gene that regulates dystrophin, a protein involved in maintaining the
integrity of muscle fibers (Hartigan-O’Connor and Chamberlain 1999;
Nowak and Davies 2004). It is the most common childhood form of
muscular dystrophy. Clinical symptoms begin between 2 and 6 yrs of
age, and the disease progresses rapidly. Children with Duchenne
dystrophy are often late in learning to walk; they show a waddling
and unsteady gait and can easily fall over. Since the disease affects
all muscles in the body, there is also difficulty in tasks like raising
arms. Most patients become unable to walk by the age of 12.
Breathing becomes affected during the later stages of Duchenne
dystrophy, leading to respiratory infections. By age 20, most patients
have to use a respirator to breathe.
Becker muscular dystrophy has a lot in common with Duchenne
dystrophy (Tsao and Mendell 1999; Nowak et al. 2005). In particular,
persons with Becker dystrophy also have a genetic abnormality
affecting dystrophin. Clinical symptoms of Becker dystrophy
generally appear later, during adolescence, and the disease
progresses more slowly than Duchenne dystrophy. The severity of
the disease varies, and boys and men with Becker dystrophy have a
longer life expectancy than those with Duchenne dystrophy.
Myotonic dystrophy is the most common adult form of muscular
dystrophy; it affects about 1 in 10,000 people worldwide (Amack and
Mahadevan 2004; Meola and Moxley 2004). The name of this
disorder underscores an unusual symptom found only in this form of
dystrophy—myotonia—which is a prolonged episode of muscle
activity, after its voluntary contraction. This activity consists of a
repetitive generation of action potentials on muscle fibers (figure
35.1) with an initially high frequency that gradually declines.
Myotonia is commonly seen in the fingers and facial muscles. Other
symptoms include a floppy-footed, high-stepping gait, cataracts,
cardiac abnormalities, and endocrine disturbances. Individuals with
myotonic muscular dystrophy tend to have long faces and drooping
eyelids.

PROBLEM 35.1
Suggest a physiological mechanism for myotonia.

Facioscapulohumeral muscular dystrophy appears in

adolescence and causes progressive weakness in facial muscles
and certain muscles in the arms and legs (Kissel 1999; Tawil 2004).
It progresses slowly and can vary in symptoms from mild to
disabling. This form of dystrophy leads to an impairment of walking,
chewing, swallowing, and speaking. About half of those with the
disorder retain the ability to walk throughout their lives.

Figure 35.1 An illustration of a myotonic discharge. Note the high frequency of

the discharge that gradually decreases; note also a decrease in the size of the
compound action potential.

35.3 Continuous Muscle Fiber

Activity Syndromes
One of the most widely known syndromes of continuous muscle fiber
activity is tetanus (Ernst et al. 1997). It typically arises from a skin
wound that becomes contaminated by a bacterium Clostridium
tetani, which is often found in soil. The toxin generated by this
bacterium, known as tetanospasmin, blocks postsynaptic inhibition at
the spinal level and, as a result, produces an uncontrolled long-
lasting muscle contraction. Approximately 10% of reported cases of
tetanus are fatal. In the United States, where 50 or fewer cases of
tetanus occur each year, deaths are more likely to occur in persons
age 60 and older.
A special form of tetanus, neonatal tetanus, occurs in newborns
who are delivered in unsanitary conditions. Prior to immunizations,
neonatal tetanus was much more common in the United States.
Now, routine immunizations for tetanus produce antibodies that
mothers pass to their babies before birth.
Bursts of muscle activity during tetanus can be stopped by a
neuromuscular or peripheral nerve block. Muscle activity also is
attenuated during sleep and under general or spinal anesthesia.
Another syndrome of excessive α-motoneuron excitation is the
stiff-man syndrome (Stayer and Meinck 1998; Thompson 2001;
Murinson 2004). It typically starts in adults aged 30 to 60. The onset
of symptoms is slow, commonly beginning with episodes of aching
and tightness in axial muscles (the neck, paraspinal, and abdominal
muscles). It can take a few months for the symptoms to become
constant, leading to a board-like rigidity of the trunk muscles. In most
patients, distal muscles of the extremities and facial muscles are
spared. But in about 25%, cranial musculature is involved, leading to
difficulties with swallowing, unusual facial expressions, and other
issues.
When a person with stiff-man syndrome is subjected to an
auditory stimulus and to an electrical finger stimulation there are
several bursts of muscle activity in muscles all over the body,
followed by their tonic contraction (Thompson 2001).
The stiff-man syndrome is associated with continuous motor unit
activity at rest; the symptoms disappear in sleep. The apparent
excessive motoneuron excitation in this condition is likely to be
associated with a disorder of presynaptic inhibition. This hypothesis
is supported by reports of strong therapeutic effects of baclofen
(Penn and Mangieri 1993; Silbert et al. 1995), an analog of a
neuromediator GABA that is likely to act in the spinal cord at a
presynaptic level (see chapter 36). Other signs of the stiff-man
syndrome include excessive coactivation of agonist–antagonist
muscle pairs. This state is somewhat similar to spasticity (discussed
in chapter 36); however, monosynaptic reflexes in the stiff-man
syndrome are typically unchanged, while they are exaggerated in
spastic muscles.

PROBLEM 35.2
How is it possible that monosynaptic reflexes are unchanged in
stiff-man syndrome if the problem is in reduced presynaptic
inhibition?

Neuromyotonia is a neuromuscular disorder characterized by

slow relaxation of muscles after voluntary contraction or electrical
stimulation (Vincent 2000). Individuals with this disorder may have
trouble releasing their grip on objects or may have difficulty rising
from a sitting position; they commonly show a stiff, awkward gait.
The disorder can affect all muscle groups. It may be acquired or
inherited. Neuromyotonia is associated with continuous activity of
single muscle fibers that can be seen at rest or on the background of
low muscle activity. Typical EMG recordings in this disorder show
small action potentials from individual muscle fibers on the
background of larger-amplitude compound action potentials of motor
units (Thompson 2001). The defect is probably in the motor axon
terminals that leak the neurotransmitter (acetylcholine) and lead to
action potential generation on the muscle fiber membrane.
Treatment involves anticonvulsant drugs as well as physical therapy
and other rehabilitative measures.

35.4 Myasthenia Gravis

One of the best-studied peripheral motor disorders is myasthenia
gravis—a disorder of transmission at the neuromuscular synapse
(Kothari 2004). This disorder affects about 60 persons per million,
and 3 to 4 new cases per million occur annually. Myasthenia gravis
can start at any age; it affects women twice as frequently as men. In
the first half of the 20th century, untreated myasthenia gravis carried
a mortality rate of 30% to 70%. In the modern era, these patients
have a nearly normal life expectancy.
Myasthenia gravis results from an autoimmune process with the
organism producing antibodies to acetylcholine receptors on the
muscle membrane. This leads to a decrease in the number of
acetylcholine-sensitive receptors and a corresponding reduction of
postsynaptic potentials on muscle fibers. This results in myopathic
features of motor unit potentials, in particular in an increase in the
duration of the potentials as compared to those in healthy persons
(Odabasi et al. 2000).
Clinical signs of myasthenia gravis include fatigue, exhaustion,
and muscle atrophy. Any muscle of the body can be affected, but
most frequently affected body parts include the eyes, face, lip,
tongue, throat, neck, and limbs. In cases when eye muscles are
affected, signs involve eyelid droop and inability to open one eye.
Facial involvement may lead to stiffness of the face and difficulties
with chewing, swallowing, laughing, and speech (dysarthria). If
breathing muscles are affected, disease progression may lead to
ventilatory insufficiency and to death.
There are two major approaches to treating myasthenia gravis
(Saperstein and Barohn 2004). One of them involves acting at the
neuromuscular junction, trying to counteract the drop in the number
of acetylcholine-sensitive receptors. Drugs such as neostigmine and
distigmine inhibit acetylcholinesterase, an enzyme that eliminates
acetylcholine from the synaptic cleft in a normally functioning
muscle. Using these drugs allows acetylcholine to stay longer in the
synaptic cleft and increases the chances of activation of the
postsynaptic muscle membrane. There is a danger of side effects
because these drugs can cause prolonged activation of healthy
muscles, resembling that seen in neuromyotonia.
An alternative method of treatment is to try to deal with the
autoimmune process itself. This can be done with
immunosuppressive drugs or thymectomy, but this radical method
may leave the immune system unable to protect the organism
against common infections. A less dramatic method is
plasmapheresis to remove autoimmune antibodies. This method
leaves the immune system relatively unchanged, but it has to be
repeated at regular time intervals. Thus, side effects seem to be
unavoidable with any treatment.

35.5 Mononeuropathies
Peripheral neuropathies are classified into mononeuropathies (when
a single nerve is affected), multiple mononeuropathies (when a
number of nerves are affected within a particular body area), and
polyneuropathies (these are systemic disorders affecting nerves
throughout the body). There are commonalities across these
disorders that include the slowing or complete interruption of
conduction along peripheral neural pathways as the main cause of
signs and symptoms.
Mononeuropathies typically are associated with slowed
conduction in a single nerve; the amplitude of motor or sensory
potentials is reduced. If a motor nerve is affected, signs of
denervation may be seen in the target muscle. Probably the best
known mononeuropathy is the carpal tunnel syndrome, which
represents entrapment of the median nerve in the carpal tunnel at
the wrist level. The carpal tunnel contains both the tendons of
extrinsic flexor muscles and the median nerve. There may be
different reasons causing increased pressure on the median nerve in
the carpal tunnel. These include thickening of the protective sheaths
that surround each of the tendons, resulting from repetitive overuse
of these muscles (for example, during work with power tools); the
tendon sheaths may also become swollen. There is also a genetic
predisposition for carpal tunnel syndrome that is related to the size of
the carpal tunnel.
Early symptoms of carpal tunnel syndrome commonly involve
painful tingling in one or both hands, particularly during the night.
These are felt in digits innervated by the median nerve up to the
medial half of the ring finger. As symptoms increase, tingling may be
felt during the day, commonly in the thumb, index, middle, and ring
fingers. There is also an impairment of the motor function leading to
poor ability to squeeze objects. In advanced cases, the thenar
muscle at the base of the thumb shows atrophy and strength loss.
Persons with this condition may start to appear clumsy and have
problems with simple tasks that involve finger coordination, such as
tying shoelaces or manipulating small objects. Figure 35.2 illustrates
schematically the delayed conduction of action potentials over the
carpal tunnel (cf. Nobuta et al. 2005); delays over 10 ms are
considered severe.

Figure 35.2 A schematic illustration of the delayed conduction of action

potentials along the axon in the median nerve to electrical stimulation of the
fingertip. The early large signal deviation is an artifact of the stimulus. (a) Normal
conduction delay is about 8 ms. (b) In a person with severe carpal tunnel
syndrome, the delay is over 10 ms.
 PROBLEM 35.3
What can you conclude from the fact that sensory impairments in
carpal tunnel syndrome are usually seen before motor
impairments?

Treatment of carpal tunnel syndrome includes nonsteroidal anti-

inflammatory drugs, such as aspirin, ibuprofen, and other
nonprescription pain relievers or locally applied antipain factors such
as ice packs and lidocaine injected directly into the wrist. Stretching
and strengthening exercises can be helpful. If symptoms last for
several months, surgical treatment may be used to reduce pressure
on the median nerve.
The second most frequently entrapped or injured nerve in the
human arm is the ulnar nerve. The ulnar nerve can be entrapped at
the elbow level (more frequently) and at the wrist level, leading to
ulnar palsy. This may cause denervation and paralysis of the
muscles supplied by that nerve: One of the most severe
consequences of the ulnar nerve entrapment is loss of intrinsic
muscle function in the hand. Numbness typically happens before any
motor signs can be detected.
Nerves in the lower extremities can also show cases of
entrapment. In particular, entrapment of the peroneal nerve is known
as peroneal pressure palsy, while entrapment of the tibial nerve
leads to tarsal tunnel syndrome, which is similar in its clinical
appearance to the carpal tunnel syndrome in the upper extremities.

35.6 Multiple Mononeuropathies

Multiple mononeuropathies may result from a number of disorders,
including diabetes mellitus and leprosy. It is relatively unusual to
consider diabetes mellitus (an impaired ability to metabolize glucose)
among motor disorders. Sensory-motor problems associated with
diabetes are secondary consequences of this long-standing disorder,
and they are not seen in all patients. However, the very large total
number of cases of diabetes makes these consequences rather
common. The total number of cases of diabetes mellitus in the
United States is about 27 million, with about 1.5 million new cases
diagnosed each year (Centers for Disease Control and Prevention
2020).
Long-term complications include, in particular, peripheral
neuropathy of both motor and sensory fibers, problems with
autonomic function, and atrophy of peripheral tissues. The loss of
sensory signals from peripheral receptors leads to problems with
motor coordination and balance (van Deursen et al. 1998; van
Deursen and Simoneau 1999). People who suffer from these
consequences of diabetes have to rely on visual information to be
able to stand; they may be completely unable to stand with their
eyes closed. This is accompanied by an increased probability of
falls.
In diabetes, distal areas of the lower extremities are usually more
affected than proximal areas. This results in a reorganization of
postural control mechanisms that relies more on information from the
relatively spared receptors in proximal leg muscles. One of the
consequences is weaker effects of muscle vibration on posture (see
chapter 25) in persons with diabetes when the vibration is applied to
the Achilles tendon and larger effects when the vibration is applied to
the patellar tendon (figure 35.3).
Figure 35.3 A scheme of reorganized postural control in diabetes. (a) In a
person with intact somatosensory signals, information from the lower leg plays a
major role in postural control. As a result, vibration (black dot) applied to the
Achilles tendon leads to large vibration-induced fallings (VIFs), while vibration of
the upper leg leads to smaller effects. (b) In a person with diabetes, unreliable
signals from the lower leg receptors make the postural control system rely more on
information from the upper leg receptors. As a result, larger VIFs are observed
during vibration of the upper leg than during vibration of the Achilles tendon.

PROBLEM 35.4
Imagine that a young adult person with diabetes stands on a
platform, and the platform suddenly starts to move at a moderate
 speed. Would you expect to see an ankle strategy or a hip
strategy in the early reaction to the perturbation?

35.7 Polyneuropathies
Polyneuropathies may be associated with a demyelinating process
or with toxic effects on peripheral axons. Demyelination may result
from a chronic inflammatory process or from an autoimmune
process, in which the body’s immune system attacks parts of the
peripheral nervous system. This latter condition is called Guillain-
Barré syndrome (Hughes and Cornblath 2005). The loss of myelin
may lead to blocked conduction along affected axons, which results
in a reduced ability to recruit motor units in muscles innervated by
the affected nerves. Monosynaptic reflexes such as knee jerks are
usually lost.
The Guillain-Barré syndrome has an annual incidence of 0.6 to
2.4 cases per 100,000. It is more frequent in males. About two-thirds
of patients have a history of gastrointestinal or respiratory infection 1
to 3 wk prior to the onset of the first clinical signs. The disease peaks
in the age ranges of 15 to 35 yrs and 50 to 75 yrs.
The first symptoms of this disorder include varying degrees of
weakness or tingling sensations in the legs. The onset of symptoms
is sudden and unexpected. In many instances, the weakness and
abnormal sensations spread to the arms and upper body. These
symptoms can increase in intensity until the patient is almost totally
paralyzed. Most people reach the stage of greatest weakness within
the first 2 wk after symptoms appear. A majority of patients, however,
recover from even the most severe cases of Guillain-Barré
syndrome, although some continue to have some degree of
weakness. The mortality rate ranges from 5% to 10%.
Guillain-Barré syndrome shares features of multiple sclerosis
(demyelination of neural pathways within the central nervous
system) and myasthenia gravis (an autoimmune disease). There are
several ways to treat the complications of the disease. Currently,
plasmapheresis and high-dose immunoglobulin therapy are used
most frequently.

35.8 Radiculopathies
The term radiculopathy refers to a group of disorders that originate
from a mechanical or inflammatory damage to a spinal root or a
group of spinal roots that have progressed enough to cause the
development of neurological symptoms in the areas supplied by the
affected root. These disorders can lead to sensory (tingling and pain)
or motor problems (weakness), depending on whether the dorsal or
ventral roots are affected. The separation of sensory and motor
symptoms makes radiculopathies different from peripheral
neuropathies that typically affect both sensory and motor neural
fibers that run together in common nerves.
One of the most frequent causes for radiculopathies is spinal disk
herniation. In this condition, the outer layer (annulus) of the disk
cracks, and the gel-like center (nucleus) breaks through. This causes
the disk to protrude, putting pressure on the nerve that exits the
spinal column at that point. Spinal disks can also suffer from age-
related changes, leading to a degenerative disk disease. As humans
age, the water content in the disks diminishes, causing the disk to
shrink. Without sufficient cushioning, the vertebrae may begin to
press against each other, pinching the nerve, or they can form bony
spurs. Finally, radiculopathies may result from spinal stenosis, a
condition in which the space in the center of the vertebrae narrows
and squeezes the spinal column and nerve roots.
Most common radiculopathies occur at the cervical and
lumbosacral levels. For example, in the United States, lumbosacral
radiculopathies occur in about 2% of the population. Of these cases,
10% to 25% develop persisting symptoms.

CHAPTER 35 IN A NUTSHELL
Peripheral motor disorders may
originate from a dysfunction in the
muscles, neuromuscular synapses,
peripheral axons, and bodies of α-
motoneurons. Muscular dystrophies are
a group of genetic diseases
characterized by progressive weakness
and degeneration of the skeletal
muscles. Duchenne and Becker
dystrophies have been linked to
defects in the same dystrophin gene.
Myotonic dystrophy is characterized by
myotonia, a prolonged episode of
muscle activity after its voluntary
contraction. Myasthenia gravis is a
disorder of transmission at the
neuromuscular synapse. It is likely to
be a consequence of an autoimmune
process. Peripheral neuropathies lead
to both sensory and motor consequences
in areas innervated by the affected
nerves. Guillain-Barré syndrome leads
to demyelination of peripheral axons
that may result from an autoimmune
process. Immunosuppressive treatment
and plasmapheresis are common
therapeutic strategies for both
myasthenia gravis and Guillain-Barré
syndrome. Advanced diabetes leads to
peripheral neuropathies that may be
associated with impaired coordination
and balance.
 Chapter 36

Spinal Cord Injury and

Spasticity

KEY TERMS AND TOPICS

spinal cord injury
signs and symptoms of spasticity
Babinski reflex
clonus
clasp-knife phenomenon
paresis
treatment
multiple sclerosis

Injury of the spinal cord is, unfortunately, a common consequence of

auto accidents (they account for 36% of all cases), falling accidents,
acts of violence, and some athletic activities. There are close to half
a million persons living in the United States who survived a spinal
cord injury. About 10,000 new cases happen every year, with over
80% involving young males. Spinal cord injuries may have a number
of very serious consequences. Only a fraction of these
consequences are directly related to the primary mechanical
damage to tissues, including neural structures, while the rest are
induced by secondary phenomena such as edema, loss of
presynaptic input to groups of neurons, and others.

36.1 Consequences of Spinal

Cord Injury
Two major groups of consequences of spinal cord injury may be
identified. The first group is related to the destruction of the
intraspinal neuronal apparatus, including both interneurons and
motoneurons. In particular, if all the α-motoneurons of a pool
controlling a certain skeletal muscle are destroyed, it is impossible to
restore normal voluntary control of the muscle. At the very best, it
may be possible to substitute for the lost muscle control with artificial
devices driven by voluntary effort generated by the subject—for
example, using EMG signals from healthy muscles (or from other
sources) to drive electrical stimulators applied to denervated
muscles. This approach is called functional electrical stimulation
(Stein et al. 2002; Kirschblum 2004).
The spinal cord also contains neural structures that control certain
components of voluntary actions. These include, in particular, central
pattern generators for locomotion (see chapter 26) as well as reflex
pathways whose normal functioning may be crucial for the
coordination of complex movements (chapters 24 and 27).
Destruction of interneurons within those structures may be expected
to lead to a major impairment of motor coordination.
The second group of consequences includes those related to an
injury to neural conduction pathways, both ascending and
descending. In particular, a total transection of the spinal cord leads
to an inability to transmit signals along both ascending and
descending neural tracts, leading to both complete paralysis and the
lack of sensations below the level of trauma. At present, all attempts
at restoring the function of the human spinal cord after its total
transection have been unsuccessful. As a result, therapies are often
directed not at restoring a lost function but at substituting for it—for
example, with the help of functional electrical stimulation or different
assisting devices, with the ultimate goal of making life more
comfortable for patients with such injuries.
The consequences of spinal cord injury are not limited to an
impairment in the control of movements and sensation. An equally or
even more important group of consequences are those related to the
functioning of internal organs, in particular, those of the bowels and
of the bladder, which are controlled by neurons in the spinal cord.
Another important consequence is chronic pain, which may
frequently be phantom. This means that a patient complains of
chronic pain in an area of the body that has no sensation with
respect to either potentially painful or nonpainful stimuli. Chronic pain
may persist even in cases of total transection of the spinal cord
above the level where the patient localizes the pain—that is, it may
be of a purely central origin.
The level of spinal cord injury is an important predictive factor for
the clinical picture (table 36.1). A supraspinal trauma typically leads
to clinical signs in both upper and lower extremities in the
contralateral to the trauma side of the body. A trauma at a cervical
level frequently leads to quadriparesis, or partial loss of voluntary
motor function in all four extremities, or even to quadriplegia (that is,
a total loss of motor function, combined with impaired sensation
below the level of trauma). A trauma at a thoracic level is commonly
associated with paraparesis or paraplegia (only the motor function of
the legs is impaired). Both cervical and thoracic traumas are
associated with a complex of signs and symptoms characteristic of
damaged descending spinal tracts that is commonly called
spasticity (described in the next section). A trauma at the lumbar
section of the spinal cord commonly leads to paraparesis of varying
degrees of severity that is not accompanied by spasticity and is
frequently associated with loss of muscle reflexes (flaccid
paraparesis).

Table 36.1 Consequences of Trauma at Different Levels

Level of trauma Likely consequences
Level of trauma Likely consequences
Supraspinal Hemiparesis or hemiplegia
(likely spastic)
Cervical Quadriparesis or quadriplegia
(likely spastic)
Thoracic Lower paraparesis or paraplegia
(likely spastic)
Upper lumbar Lower paraparesis or paraplegia
(likely spastic)
Lower lumbar Lower paraparesis or paraplegia
(likely flaccid)

Immediately following a spinal cord injury, there is typically a

period of spinal shock that lasts for a few minutes to a few hours.
During the spinal shock, all muscle reflexes are suppressed, and the
patient shows complete paralysis. Over the next days, weeks, or
months, typical consequences of the injury develop. There is
commonly partial recovery of voluntary movements and sensations
below the level of trauma accompanied, in some cases, by the
development of spasticity.

PROBLEM 36.1
Can you suggest why there is no spasticity in cases of lumbar
spinal cord injuries?

PROBLEM 36.2
Spinal cord injury at what level is likely to lead to the destruction of
the neuronal apparatus of locomotor central pattern generators?

36.2 Signs and Symptoms of

Spasticity
Spasticity is a common component of a variety of motor disorders
resulting from brain trauma (including cerebral palsy and stroke; see
chapters 39 and 40), spinal cord injury, and certain systemic
degenerative processes (including multiple sclerosis; see chapter
40). It is sometimes called an upper motor neuron disease, while
spastic symptoms may be called pyramidal symptoms, implying
damaged transmission of signals along the pyramidal tract from
cortical neurons to the spinal cord. These terms are jargon and may
be misleading. Indeed, most axons within the corticospinal tract
project on spinal interneurons, not motoneurons. Their activation
does not necessarily lead to muscle activation, as demonstrated in
studies of motor imagery, when the subjects only imagined
performing movements (Hétu et al. 2013; Mizuguchi and Kanosue
2017). So calling them “motor neurons” is unwarranted. Besides,
spasticity is a complex phenomenon reflecting interrupted
conduction along multiple descending pathways, not only along the
pyramidal tract.
Unfortunately, spasticity is a rather common disorder that in many
cases does not permit the patients to perform functionally significant
voluntary movements or even occupy certain postures. Uncontrolled
spastic spasms are frequently vigorous and painful. They may
prevent patients from having normal sleep. If a patient is sitting in a
wheelchair, the spasms may require strapping the limbs to the chair.
Since spasticity is frequently associated with a total loss of sensation
in the affected limbs, uncontrolled spastic movements may lead to a
trauma without the patient knowing about it. Spasticity is also
sometimes associated with chronic pain syndrome, which may lead
to constant excruciating pain.
Clinicians define spasticity as a disorder of spinal proprioceptive
reflexes manifested as profound changes in reflexes to muscle
stretch with a strong velocity-dependent component (Lance and
Burke 1974; Lance 1980), the emergence of pathological reflexes
and uncontrolled spasms, an increase in muscle tone, and an
impairment of voluntary motor function. This is a descriptive
definition in terms of signs and symptoms rather than in terms of
underlying mechanisms. Besides, it contains one of the worst
misnomers, “muscle tone.” Clinicians define “increased muscle tone”
as a “feeling of increased resistance when you try to move a joint.” In
such tests, the subject is asked to relax, and the clinician is moving
the joint over its range of motion. We will return to the notion of
muscle tone later in section 36.4.

PROBLEM 36.3
Suggest a physiological mechanism underlying increased muscle
tone. What about decreased muscle tone?

A great British neurologist, Hughlings Jackson, classified spastic

signs into positive and negative. Under positive signs, Jackson
implied phenomena that are seen in patients with spasticity but not in
healthy persons. Under negative signs, he implied phenomena that
are seen in healthy persons but lost in patients with spasticity.

Positive signs
Flexor/extensor spasms
Clonus
Clasp-knife phenomenon
“Babinski reflex”
Exaggerated cutaneous reflexes
Autonomic hyperreflexia
Tone abnormalities

Negative signs
Weakness (paresis or plegia)
Loss of dexterity
Fatigability
Contractures
Although it has been commonly assumed that spasticity is
associated with a deficit in spinal inhibitory mechanisms, including
both postsynaptic and presynaptic inhibition, there is no consensus
about what causes these deficits in the first place. To say that they
are due to the disruption of the normal functioning of certain
descending systems does not help much since these systems are
not well defined, and their role in voluntary motor control is unclear.
The relations between spasticity and muscle reflexes are also not as
unambiguous as implied by the definitions presented. Let us
consider typical changes in muscle reflexes associated with
spasticity.
• Spasticity can be associated with exaggerated, unchanged, and
even absent monosynaptic reflexes, including the H-reflex, although
an increase in monosynaptic reflexes is more typical.
• A common correlate of spasticity is spasm-like bursts of activity
in leg muscles in response to tactile stimulation of the sole of the foot
(figure 36.1). This response is sometimes imprecisely called the
Babinski reflex or a defensive reaction. It is quite variable across
patients and may involve bursts of activity in all major flexor muscles,
a sustained contraction of the flexor muscles, either with or without a
comparable activation of extensor muscles. However, this reflex can
be absent in certain patients with spasticity.
• Another typical sign of spasticity is clonus (figure 36.2), which
represents a series of alternating bursts of activity in the flexor and
extensor muscles of a joint at a frequency of about 6 to 8 Hz in
response to a single quick movement of the joint performed by the
experimenter (passively) or by the patient (if there is enough
voluntary motor control left). Clonus may last for only about a
second, or it may continue for tens of seconds or even minutes until
it is stopped mechanically—for example, by clamping the joint and
preventing it from moving. Clonus is likely to represent an auto-
oscillation in the hyperexcitable stretch reflex loop (Iansek 1984;
Latash et al. 1989; Hidler and Rymer 1999): When a muscle is
stretched, a monosynaptic stretch reflex leads to its phasic
contraction, leading to a reversal of the joint’s movement direction.
As a result, the antagonist muscle is stretched and demonstrates a
monosynaptic stretch reflex. And so on. Remember that joint
movements in healthy persons do not normally induce monosynaptic
reflexes. There is an alternative hypothesis claiming that clonus is a
result of the functioning of a central generator within the spinal cord
(Dimitrijevic et al. 1980; Beres-Jones et al. 2003).
Figure 36.1 An example of a spasm in leg muscles induced by tactile stimulation
of the sole of a foot in a spastic patient. This reaction is sometimes imprecisely
called the Babinski response. TA = tibialis anterior; SOL = soleus; QUAD =
quadriceps (rectus femoris); HAM = hamstrings.
Reprinted by permission from M.L. Latash, R.D. Penn, D.M. Corcos, and G.L. Gottlieb,
“Short-Term Effects of Intrathecal Baclofen in Spasticity,” Experimental Neurology, 103
(1989): 167. ©1989, with permission from Elsevier.

• If a joint of a person with spasticity is slowly moved into flexion,

joint extensors show an increase in activity and resist the motion
strongly. However, if enough external force is applied, this resistance
can be overcome. At some point, the resistance suddenly
disappears, and the joint collapses like a pocket knife. This behavior
is called the clasp-knife phenomenon. It is typically attributed to a
change in the balance of reflex effects from muscle spindles (which
resist the joint motion) and from Golgi tendon organs (which tend to
suppress this resistance). See chapters 17 and 18.
• A variety of changes in different components of muscle
reactions to stretch have been reported. No reproducible differences
have been found between gains of the long-latency reflexes to
muscle stretch (including the tonic stretch reflex) in spastic and
control subjects, and the increased resistance to muscle stretch has
been partially attributed to peripheral changes in muscle and tendon
stiffness (Dietz et al. 1981; Berger et al. 1984; Rack et al. 1984;
Thilmann et al. 1991). However, when spasticity was dramatically
reduced in a patient by intrathecal baclofen (we will discuss it
section 36.5), the gain of the tonic stretch reflex was also reduced
(Latash 1993).
Figure 36.2 An example of electrical muscle activity during ankle clonus induced
by a single rapid dorsiflexion of the foot. TA = tibialis anterior; SOL = soleus;
QUAD = quadriceps (rectus femoris); HAM = hamstrings.
 • Suppression of monosynaptic reflexes by muscle vibration,
which is presumably mediated by presynaptic inhibitory
mechanisms, has been suggested as a quantitative index for
spasticity (Ashby and Verrier 1976; Ongeboer de Visser et al. 1989).
This suppression is pronounced in healthy humans, leading to a 3- to
10-fold decrease in the peak-to-peak amplitude of the H-reflex in the
triceps surae muscle during vibration of the Achilles tendon. In
spastic patients, however, this effect is much lower (less than
twofold) or absent, or even reversed, representing an increase in the
amplitude of the H-reflex (figure 36.3). This index, however, does not
reflect the state of postsynaptic inhibition and, in some cases, poorly
correlates with clinical status. Also, monosynaptic reflexes are
sometimes absent in spasticity, making this method inapplicable.

Figure 36.3 (a) Changes in an H-reflex in the soleus muscle induced by vibration
of the muscle tendon in a healthy person. Note the suppression of the H-reflex. (b)
Changes in an H-reflex in a person with spasticity; no suppression of the H-reflex.

PROBLEM 36.4
Based on these two conflicting hypotheses, predict changes in the
clonus frequency in conditions when the limb is loaded inertially
and when the strength of contractions is reduced by a drug.
 Two clinical scales have been successfully used for quantitative
assessment of spasticity. The most frequently used is the Ashworth
scale, which reflects the degree of muscle resistance to passive limb
movements (table 36.2). The other is the spasm scale, which reflects
the frequency of spasms, their duration, and their general or local
character (table 36.3). Both scales are subjective and reflect a
physician’s general impression of the patient’s state, which is likely to
be important from the clinical view but is not very helpful for
understanding the mechanisms of the disorder.

Table 36.2 Ashworth Scale

Score Description of “muscle tone”
1 No increase in tone
2 Slight increase in tone, giving a catch when affected segment moved
into flexion or extension
3 More marked increase in tone, but the affected segment easily
flexed and extended
4 Considerable increase in tone; passive movement difficult
5 Affected segment rigid in flexion or extension

Table 36.3 Spasm Scale

Score Characteristics of spasms
0 No spasms
1 Mild spasms induced by stimulation
2 Infrequent spasms occurring less than once per hour
3 Spasms occurring more frequently than once per hour
4 Spasms occurring more frequently than 10 per hour

There is even more ambiguity in the relations between spasticity

and voluntary motor control. There has been controversy concerning
the relation between the positive and negative signs of spasticity
(according to Hughlings Jackson’s classification). Are increased
muscle reflexes and spasms (positive signs) interfering with impaired
voluntary motor control (negative signs)? Jackson thought that
eliminating the positive signs of spasticity should not be expected to
help voluntary motor function. This view has been challenged in a
number of studies, in which hyperactive reflexes and cocontraction
of antagonist muscles appeared to interfere with proper voluntary
motor function (Corcos et al. 1986; Latash and Penn 1996).
However, in order to solve this problem directly, an effective method
of treating spasticity was necessary.

36.3 Possible Mechanisms of

Spasticity
The following factors have been hypothesized to contribute to the
clinical picture of spasticity following spinal cord injury (reviewed in
Katz and Rymer 1989; Dietz 2000; Hultborn 2003):
Increased stiffness of peripheral structures
Lack of presynaptic inhibition of inputs to α-motoneurons
Lack of postsynaptic inhibition of α-motoneurons (in particular,
deficit in reciprocal inhibition, in Ib inhibition, and in Renshaw cell
action)
An increase in the gain of the tonic stretch reflex loop
A change in the threshold of the tonic stretch reflex
One of the influential hypotheses on the control of movements, the
equilibrium-point hypothesis (chapters 20 and 21) emphasizes the
importance of shifts of the threshold (λ) of the tonic stretch reflex
over the whole range for proper neural control of a muscle. The
equilibrium-point hypothesis has been developed to address
movements of persons with spasticity (Levin and Feldman 1994;
Jobin and Levin 2000).
In a healthy muscle, neural control is assumed to result in shifts of
the muscle force–length characteristic along the length axis (figure
36.4a). A healthy person can relax a muscle at any anatomically
accessible joint angle (muscle length). This means that the force–
length characteristic (and its threshold, λ) can be shifted outside the
anatomical muscle length range (the rightmost characteristic in figure
36.4a). On the other hand, a person can produce high muscle
forces, even if the muscle is close to its shortest possible length.
This implies a possibility of shifting λ beyond the anatomical range to
the left (the leftmost characteristic in figure 36.4a).

Figure 36.4 (a) In a healthy person, the brain can shift the control variable λ
beyond the anatomical range of muscle length values, from λMIN to λMAX. These
shifts can lead to large muscle force or muscle relaxation at any muscle length
(LMIN to LMAX). (b) In a person with spasticity, the controller loses its ability to shift
λ over the whole range, only from λ– to λ+. As a result, the muscle may not be able
to relax at high length values (shown as “spasms”) and to produce voluntary
activation at low length values (shown as “paralysis”).
 To account for spasticity, it has been assumed that the controller
loses the ability to shift λ over the whole range. Moreover, it has
been assumed that large λ shifts can be produced by sensory
feedback signals. Figure 36.4b illustrates that a contracted range of
voluntary λ shifts may be associated with unintentional muscle
activation when muscle length exceeds the range of λ shifts. The
same scheme suggests that no voluntary movement can be possible
when muscle length is shorter than the smallest value of λ within the
contracted range of its changes, leading to paralysis. A number of
recent studies linked the decreased range of λ changes in patients
with spasticity with their clinical state in support of the scheme in
figure 36.4 (Turpin et al. 2017; Subramaniam et al. 2018).
This hypothesis on the origins of spasticity leads to a few
predictions. Most importantly, it unites exaggerated muscle reflexes
and poor voluntary control into a single scheme. As such, it
eliminates the distinction between positive and negative signs of
spasticity and predicts that changes in one group of the signs may
be expected to be accompanied by changes in the other group. This
prediction received support in a number of studies that quantified the
effects of treatment on both muscle reflexes and voluntary
movements (Corcos et al. 1986; Latash et al. 1990; Latash and Penn
1996).

36.4 Defining Muscle Tone

Although the notion of muscle tone has been used broadly in clinical
practice and in the literature, it has remained poorly defined and
depends on the subjective impression during clinical examination.
Nikolai Bernstein defined muscle tone as a state of the muscle
reflecting its preparation for future action (Bernstein 1947). In
contrast to this definition, during typical clinical examination, the
patient is asked to relax, and the clinician moves effectors through
the range of motion (i.e., no preparation to action is involved). Note
that the instruction “to relax” is ambiguous and can be interpreted
differently by different persons.
 Figure 36.5 illustrates the dependence of muscle active force on
muscle length. If a person is asked to relax the muscle at a certain
length value (L0 in figure 36.5), this means that the threshold for
muscle activation, λ, has to correspond to muscle length greater than
L0. There is, however, a broad range of λ values that satisfy this
instruction. For example, the person can have λ just over L0, much
larger than L0, or beyond the range of anatomically possible muscle
length changes (λ1, λ2, and λ3 in figure 36.5). Imagine now that the
clinician moves the joint spanned by this muscle such that its length
becomes L1. There will be strong resistance in case of λ1, moderate
resistance in case of λ2, and no active resistance in case of λ3. Does
this mean that these three scenarios correspond to pathologically
increased muscle tone, normal tone, and reduced tone (hypotonia)?
Of course, not. A perfectly healthy person can follow the imprecise
instruction “to relax” in different ways, illustrated in figure 36.5.
Figure 36.5 allows for defining muscle tone as the inverse of the
distance between λ and the current muscle length. Note that if a
person prepares to perform a quick action that requires activation of
a muscle, keeping this distance short (apparently high tone) can
potentially reduce reaction time. Indeed, any physiological process in
the body proceeds at a limited rate. Even the fastest shifts of λ are
also limited in speed (available estimates are on the order of 800 °/s
for joint rotation; Latash et al. 1991; Latash 1993). So, keeping λ
close to the current muscle length value can potentially reduce the
time to the muscle activation.
Figure 36.5 If a person is asked to relax at a certain value of muscle length (L0),
the threshold of stretch reflex λ has to be larger than L0, but its value is undefined.
It can be λ1, λ2, or λ3. Depending on its value, slow muscle stretch to a new length
value (L1) can lead to different active forces (F1, F2, and 0, respectively), which
may be interpreted wrongly as different levels of “muscle tone.” LMIN and LMAX
show the anatomical range of muscle length.

This definition of muscle tone is compatible with Bernstein’s

definition because it links tone to preparation for future action. It is
not, however, easily compatible with the current practice of
measuring muscle tone, which is too imprecise (the instruction “to
relax”) and relies too much on the subjective perception of resistance
to motion by the examiner to be interpreted in terms of physiology.

36.5 Treatment of Spasticity

Treatment of spasticity, until 30 y ago, was rather unsuccessful.
Attempts at reducing spastic signs and symptoms involved drug
therapies, physical therapy, and in the worst cases, destructive
chemical procedures (such as damaging peripheral nerves with
phenol or alcohol, or weakening neuromuscular synapses with
botulinum toxin) or neurosurgical procedures (such as cutting the
dorsal roots of a number of segments, or damaging the area where
the dorsal roots enter the spinal cord). Drugs taken orally were rather
successful in a small percentage of cases, marginally successful in
some cases, and virtually ineffective in most cases. The major
problem was that taking a drug orally places it into the bloodstream
so that the same concentration of the drug is achieved in the spinal
cord (where it is needed) and in the brain (where it is not needed).
As a result, a therapeutic concentration of a drug at the spinal level
could lead to major side effects, including disorders of
consciousness and even coma.
A rather ingenious solution was pioneered by a Chicago team led
by the neurosurgeon Richard Penn (Penn and Kroin 1984, 1987;
also see Dralle et al. 1985; Zierski et al. 1988). It involves delivering
a drug directly into the spinal canal (intrathecally), avoiding the
systemic effects (figure 36.6). The drug is placed in a reservoir
connected to an electronically controlled pump. The pump and the
reservoir are placed under the skin and connected to the spinal
canal with a thin catheter. Intrathecal infusions of baclofen, an
agonist of gamma-aminobutyric acid (GABA), have been shown to
effectively reduce muscle spasms and exaggerated reflexes (Latash
et al. 1989; Penn et al. 1989). GABA is one of the major
neuromediators within the CNS. In particular, it is involved in the
presynaptic inhibition of α-motoneurons.
The clinical effectiveness of this new intrathecal method of drug
delivery combined with its very fast action led to dramatic effects in
many patients: virtually all the spastic signs were eliminated within 1
h after the beginning of baclofen delivery. Figure 36.7 illustrates the
ankle clonus in a patient with spinal cord injury prior to and after an
intrathecal injection of baclofen. The difference is striking. At the
same time, the scores for both Ashworth and spasm scales were
reduced and were accompanied by clear clinical gains. Intrathecal
baclofen also provides a rare opportunity in clinical motor control
studies to use a spastic patient as his or her own control in two
states, with and without spastic signs.
Figure 36.6 A scheme of intrathecal drug delivery with an implanted
programmable pump.
Figure 36.7 An example of ankle clonus in a spastic patient prior to the injection
of intrathecal baclofen and on intrathecal baclofen. Note that the stimulus (ankle
dorsiflexion) was the same during both tests (upper panel).
Adapted by permission from M.L. Latash and R.D. Penn, “Changes in Voluntary Motor
Control Induced by Intrathecal Baclofen in Patients with Spasticity of Different Etiology,”
Physiotherapy Research International 1 (1996): 229-246.
 Most of the aforementioned studies were carried out in patients
with multiple sclerosis and spinal cord injury who demonstrated very
severe spasticity resistant to all available destructive therapies,
including oral baclofen. The suppression of spastic signs, such as
spasms, in some patients with residual voluntary movements in
spastic limbs was accompanied by an improvement of movement
patterns, including both movement kinematics and EMG patterns
(Penn et al. 1989; Latash et al. 1990; Campbell et al. 1995). Figure
36.8 shows an improvement in elbow flexion movements in a patient
with spasticity after an injection of baclofen. Note that the
improvement in movement kinematics (faster movement, fewer
oscillations) is accompanied by a decrease in muscle cocontraction
and fewer bursts of muscle activity.
Some patients demonstrate spasticity in only one-half of the body,
left or right. These cases are called hemisyndromes and are rather
common consequences of brain traumas, strokes, and cerebral
palsy. A potentially very important and intriguing finding in the
patients with hemisyndromes was that they did not notice any
weakness in the unaffected limbs, despite the high intrathecal
baclofen doses. This apparent difference in the effects of baclofen
upon pathological muscle reflexes and voluntary motor control
suggests that intrathecal baclofen does not induce a nonspecific
widespread inhibition throughout all spinal structures.
The site of action of baclofen in the spinal cord is not clear. As a
GABA agonist, it binds to GABA receptor sites that occur widely in
the CNS. In particular, baclofen-sensitive GABA receptors were
found on primary afferent terminals. These findings suggest that
probably many of the effects of baclofen are mediated through an
increase in presynaptic inhibition. Thus, the observations in patients
with hemisyndromes can be explained by either different numbers of
baclofen-sensitive receptors on terminals of different descending
systems or their different ability to bind to intrathecal baclofen due to
anatomical and other factors (for example, diffusion of baclofen
throughout spinal cord structures). The lack of apparent changes in
intact muscles in the hemisyndrome cases suggests that one of the
long-term reactions of spinal structures to a spasticity-inducing
pathology can consist of an increase in the number of GABA-
sensitive receptors or sensitization of the existing receptors on
pathologically active reflex inputs.

PROBLEM 36.5
What conclusion can be drawn regarding the site of baclofen
action based on the observations of similar suppression of
monosynaptic reflexes on both sides of the body of a patient with
a hemisyndrome, effective suppression of spastic signs, and no
apparent changes in control of the intact side of the body?

Figure 36.9 shows schematically some of the possible

consequences of a neural trauma. Any trauma is likely to lead to the
lack of both descending inhibition and descending excitation to the
segmental levels. The lack of excitatory inputs may be expected to
lead to a decrease in the centrally induced levels of α-motoneuronal
activity, and consequently, to a decrease in voluntary muscle force
(weakness or paresis). At the present level of our knowledge, it is
impossible to adequately correct this deficiency. Functional electrical
stimulation (see chapter 41) is a way to induce stronger muscle
contractions, but it represents a substitution of the function rather
than its correction.
Figure 36.8 An example of kinematic and EMG patterns during elbow flexions by
a spastic patient prior to the action of intrathecal baclofen (thin traces) and on
intrathecal baclofen (thick traces). Note an increase in movement velocity and a
decrease in the level of muscle co-contraction. TrLon and TrLat = long and lateral
heads of triceps; BRR = brachioradialis; BIC = biceps; VEL = velocity; ACC =
acceleration.
Reprinted by permission from M.L. Latash, Control of Human Movement (Champaign, IL:
Human Kinetics, Inc., 1993), 270.
Figure 36.9 A scheme of possible consequences of a spinal cord injury and of
baclofen action.
Reprinted by permission from M.L. Latash, Control of Human Movement (Champaign, IL:
Human Kinetics, Inc., 1993), 273.

The lack of descending inhibition has multiple consequences,

including the characteristic features of spasticity, spasms,
exaggerated reflexes, and increased resistance to externally
imposed motion (we are trying to avoid using the term “muscle
tone”!). It may also lead to hypersensitivity to the lacking inhibitory
mediators, including GABA, below the level of trauma. The last
reaction may probably be considered compensatory since it
increases the effectiveness of remaining supplies of GABA and its
agonists, including baclofen. It looks as if the CNS “knew” that
intrathecal baclofen would be invented and made all the necessary
preparations to increase its effectiveness!

CHAPTER 36 IN A NUTSHELL
Spinal cord injury leads to sensory-
motor consequences reflecting the
disruption of transmission along
ascending and descending neural
pathways and the destruction of the
spinal neuronal apparatus. Spasticity
is a typical consequence of spinal
cord injury; it is characterized by a
partial or complete loss of voluntary
control over muscles, partial or
complete loss of sensation,
uncontrolled spasms and increased
reflexes, possibility of chronic pain,
and disruption of functions of
internal body organs below the level
of trauma. Treatment of spasticity
with intrathecal drug delivery, in
particular by intrathecal baclofen,
has been most successful. Suppression
of spasms and reflexes can be
accompanied by an unmasking of more
normal voluntary movements while
voluntary control of unaffected
muscles does not change. Adaptive
changes to the original trauma are
likely to play an important role in
the selective action of drugs.
Spasticity has been interpreted as a
problem in changing the threshold of a
stretch reflex over its whole normal
range. This interpretation links
typical spastic signs to problems with
voluntary muscle activation. It also
suggests a definition for muscle tone.
 Chapter 37

Disorders Involving the Basal

Ganglia

KEY TERMS AND TOPICS

cardinal signs of Parkinson’s disease
bradykinesia
tremor
rigidity
postural disorders
treatment strategies
Huntington’s chorea
hemiballismus
dystonia
tardive dyskinesia

The basal ganglia represent a constellation of paired (left and right)

nuclei that play an important role in many brain functions, including
the motor function (chapter 9). In this chapter, the focus will primarily
be on the role of the cortico-basal-thalamo-cortical loop in motor
function. A simplified scheme of this loop is presented in figure 37.1.
Actually, the loop contains two pathways, direct and indirect,
discussed in the earlier chapter on the basal ganglia. Despite its
simplicity and incompleteness (for example, the scheme does not
show the two parts of the substantia nigra, pars compacta and pars
reticulata, and does not mention projections of the basal ganglia to
the parapontine nuclei that are important for the initiation of
locomotor movements), this scheme is useful in the analysis of
disorders associated with dysfunction of particular components of
the basal ganglia. Based on animal studies and observations in
patients with disorders of the basal ganglia, this loop has been
assumed to play a central role in the coordination of voluntary
movements.

Figure 37.1 Direct and indirect pathways in the cortico-basal-thalamic-cortical

loop. Excitatory projections are shown with solid lines and open circles; inhibitory
projections are shown with dashed lines and filled circles. D and ID show the
effects of dopamine inputs on the direct and indirect loops, respectively. This is a
very much simplified scheme!

Disbalance of the projections involved in the cortico-basal-

thalamo-cortical loop may lead to a variety of motor disorders.
Commonly, these disorders are classified into hypokinesias and
hyperkinesias (figure 37.2). The former term refers to poverty of
movements, slowness, and problems with movement initiation. The
latter term implies excessive, poorly controlled movements. We will
start with the best-known disorder of the basal ganglia, Parkinson’s
disease, which belongs to hypokinesias.

Figure 37.2 Major hypokinetic and hyperkinetic disorders associated with

dysfunction of different structures in the basal ganglia. SN = substantia nigra; STN
= subthalamic nucleus; GP = globus pallidus.

37.1 Clinical Features of

Parkinson’s Disease
Parkinson’s disease is a complex disorder reflecting malfunctioning
of the basal ganglia associated with the death of dopamine-
producing neurons in the substantia nigra (pars compacta). Patients
with Parkinson’s disease typically demonstrate poverty of voluntary
movements, which is sometimes known as akinesia (synonymous
with bradykinesia and hypokinesia). Akinesia involves such typical
signs as slowness in movement initiation, long movement time,
difficulty in switching from one movement to another, masklike
expression of the face, stooped posture, shuffling gait, a lack of
associated arm movements during walking, and “frozen” postures.
Movements of these patients are slow and frequently ineffective.
During many everyday tasks, they have problems switching from an
apparently ineffective motor strategy to an alternative one. Hand
trembling (tremor) is another common manifestation of Parkinson’s
disease, making such activities as eating with a fork or a spoon and
drinking from a cup difficult.
Parkinson’s disease is a relatively common disorder that affects
about 1 out of 1,000 persons. This rate varies greatly depending on
a number of factors (reviewed in Marras and Tanner 2004). In
particular, genetic factors have been shown to contribute to
Parkinson’s disease: A person is much more likely to develop
Parkinson’s disease if one of his or her parents had this disease.
Parkinson’s disease typically affects older persons and, at some time
in the past, it was even considered accelerated aging; it is more
common in males. Somewhat surprisingly, such factors as smoking
and drinking alcohol have been shown to correlate with a lower
incidence of Parkinson’s disease.
Histological postmortem examination of the brains of patients with
Parkinson’s disease shows degeneration of neurons in the
substantia nigra; there is also a decrease in the dopamine content of
the striatum (more pronounced in the putamen) due to the
degeneration of the nigrostriatal connections. Degeneration may also
be seen in other areas of the brain. It is believed that Parkinson’s
disease occurs because of the striatal dopamine deficiency, and this
view is supported by the effectiveness of dopamine replacement (L-
DOPA) therapy. Based on the connections between the basal
ganglia and other brain structures illustrated in figure 37.1, one may
conclude that weakening of dopaminergic projections from the
substantia nigra to the striatum may lead to two types of effects:
weaker dopaminergic excitatory effects on the projection to the
internal pallidum, and weaker dopaminergic inhibitory effects on the
projection to the external pallidum (figure 37.3). As a result, both
direct and indirect pathways through the basal ganglia lead to a
decrease in the excitatory input to the brain cortex. This may explain
some of the features of Parkinson’s disease, such as poverty of
movements.
Figure 37.3 Parkinson’s disease is associated with weaker dopaminergic
projections from the substantia nigra, leading to changes in both direct and indirect
loops. These changes result in a decrease in excitatory projections from the
thalamus to the cortex. Excitatory projections are shown with solid lines and open
circles; inhibitory projections are shown with dashed lines and filled circles.
Projections that are weaker than in a healthy person are shown with thin lines,
projections that are stronger with thick lines.

There are four cardinal signs of Parkinson’s disease: tremor,

bradykinesia (akinesia), rigidity, and deficits in postural reflexes
(figure 37.4).
Tremor is characterized by a 5 to 6 Hz alternating activity of
antagonist muscles controlling a joint, leading to alternating joint
movements, that is primarily seen at rest (postural tremor) and is
commonly alleviated during voluntary movements. Tremor can
be seen in various muscle groups, and it is effector-specific. For
example, voluntary movement of an arm may lead to decreased
tremor in that arm, while tremor in the leg or in another part of
the body (e.g., jaw tremor) stays unchanged.
Bradykinesia usually refers to slowness of voluntary
movements and difficulty in their initiation, although deficits in
spontaneous or automated movements are also sometimes
 known as bradykinesia. It can affect any part of the body and
can be more or less generalized.
Rigidity is a sustained increase in the resistance to externally
imposed joint movements; this resistance does not show velocity
dependence, in contrast to the spasticity discussed in chapter
36.
Deficits in postural reflexes covers a range of deficits in
mechanisms involved in postural control; not all can be termed
reflexes (see chapter 15). In particular, patients with Parkinson’s
disease show poverty of anticipatory postural adjustments
(APAs) and an increase in preprogrammed corrections in the
activity of postural muscles associated with voluntary
movements or in response to an external perturbation (see
chapter 25).

Figure 37.4 Four major signs of Parkinson’s disease include tremor, rigidity,
bradykinesia, and deficit in postural reactions. The latter deficit has two
components: a deficit in APAs and a poorly controlled increase in later,
preprogrammed reactions.

PROBLEM 37.1
 There are obvious similarities between the symptoms of spasticity
and of Parkinson’s disease. How would you differentiate between
multiple sclerosis and Parkinson’s disease?

Since the primary cause of Parkinson’s disease is undoubtedly

supraspinal, it has been suggested that motor disorders in
Parkinson’s disease are due to changes in the descending motor
commands, while the spinal segmental apparatus is generally intact.
This view is supported by a number of observations, including
unchanged tendon jerk reflexes and generally normal short-latency
action of Ia muscle afferents. However, a number of changes in
segmental mechanisms have been reported. They include a deficit in
reciprocal inhibition (Hayashi et al. 1988; Lelli et al. 1991; Meunier et
al. 2002) that could, in particular, lead to considerable cocontraction
of antagonist muscles during voluntary movements, increased reflex
activity during tracking phases, in which the muscle is lengthening,
and a paradoxical shortening reaction or Westphal phenomenon
(Andrews et al. 1972; Angel and Lewitt 1978; Matthews et al. 1990).
The Westphal phenomenon represents an abrupt reflex excitation of
a muscle in response to an externally imposed muscle shortening. In
a sense, it is the inverse of the stretch reflex.

37.2 Voluntary Movements in

Parkinson’s Disease
Studies of voluntary movements in Parkinson’s disease have
revealed a number of differences from healthy people. While
performing a simple movement, when all the parameters of the
movement are known in advance, patients with Parkinson’s disease
initiate and perform the movement more slowly. In particular, they
demonstrate an increase in reaction time, which is larger for more
complex movements (Heilman et al. 1976; Evarts et al. 1981; Sanes
1985; Stelmach et al. 1986). An increase in movement time in
Parkinson’s disease is accompanied by considerable asymmetry of
the acceleration and deceleration phases and is mostly due to
prolongation of the deceleration phase (Inzelberg et al. 1990).
Movements of patients with Parkinson’s disease typically show
higher sensitivity to accuracy requirements when compared to
healthy persons (Rand et al. 2000).
Studies of multijoint movements revealed additional problems. In
particular, such movements show impaired interjoint coordination,
possibly associated with poor prediction of and compensation for
interaction torques (Dounskaia et al. 2005). Such movements are
less smooth, as reflected by their higher jerk indices (see chapters
22 and 24; Teulings et al. 1997).
Both temporal and spatial variability of limb movements to a target
have been shown to be higher in Parkinson’s disease (Sheridan et
al. 1987). In fact, increased motor variability is probably the most
common feature across virtually all motor disorders. It has been
suggested that bradykinesia in Parkinson’s disease may in part
result from the increased variability and the desire to preserve an
acceptable level of accuracy (i.e., it is a consequence of a
compensatory strategy adopted by the patient’s brain rather than a
primary deficit).
Figure 37.5 A typical example of muscle activation patterns during a fast
voluntary flexion movement in the elbow joint by a patient with Parkinson’s
disease. Note the increased cocontraction and typical repeated bursts of activity.
There was also an accompanying smaller wrist movement.
Reprinted by permission from M.L. Latash, A.S. Aruin, I. Neyman, J.J. Nicholas, and M.B.
Shapiro, “Feedforward Postural Adjustments in a Simple Two-Joint Synergy in Patients with
Parkinson’s Disease,” Electroencephalography and Clinical Neurophysiology, 97 (1995): 77-
89. © 1995, with permission from Elsevier.

Fast movements of Parkinson’s disease patients are typically

hypometric (that is, they undershoot the target), especially during
movements of large amplitude (Flowers 1976). The entire targeted
movement is frequently constructed of several discernible segments.
Correspondingly, the EMG patterns demonstrate a number of
repeated cycles of agonist–antagonist bursts (figure 37.5; Hallett and
Khoshbin 1980; Berardelli et al. 1986; Inzelberg et al. 1990). The
slow buildup of EMG during voluntary movements and the
considerable amount of cocontraction of antagonist muscles (Evarts
et al. 1979; Hayashi et al. 1988; Seidler et al. 2001) can also be
factors disrupting kinematic patterns during voluntary movements.
These findings allowed researchers to formulate a number of
hypotheses about the origins of the deficit in voluntary motor control
in Parkinson’s disease. However, all these formulations imply that
the EMGs or muscle forces are adequate measures of the voluntary
motor command, which, as we know, is not true (chapter 21).

PROBLEM 37.2
Why are muscle forces and EMGs not adequate reflections of
central commands during voluntary movements by healthy
persons?

In particular, it has been suggested that the mechanism controlling

EMG magnitude during rapid movements is impaired (Pullman et al.
1990). The failure to generate sufficient muscle forces has been
attributed to a basic failure to “sufficiently energize the muscles.” It
has also been concluded that, although the overall form of motor
programs (another poorly defined concept; see chapter 21) in
Parkinson’s disease is preserved, the details of the number and
frequency of activated motor units can be inaccurate (Dengler et al.
1990).
However, despite the slow buildup of EMGs during isometric
voluntary contractions, patients with Parkinson’s disease ultimately
achieve the correct final level of muscle activity. These patients are
able to produce accurate force levels, although the control of the rate
of force increase and decrease seems to be more affected
(Stelmach and Worringham 1988; Wing 1988). This group of
observations suggests that the problem is not in achieving absolute
levels of muscle activation but rather in the time pattern of muscle
activation, which is likely to depend upon the action of reflex
feedback loops and changes within the segmental spinal apparatus.
 The deficits in motor performance in Parkinson’s disease become
especially pronounced for sequential multijoint movements. In
particular, the intervals between components of sequential
movements are prolonged (Berardelli et al. 1986; Benecke et al.
1987). These patients also have troubles integrating several
components into one motor action (Benecke et al. 1986).

37.3 Vertical Posture and

Locomotion in Parkinson’s
Disease
As a reminder, postural control involves two types of corrective
actions whose function is to ensure postural stability in the presence
of perturbations (chapter 25). Some of these actions occur prior to a
perturbation and are called anticipatory postural adjustments. They
are generated by the central nervous system in a feedforward
manner, and they try to alleviate the effects of a predictable
perturbation. The second group involves reactions that are prepared
by the central nervous system in advance and are triggered by a
peripheral stimulus informing the central nervous system about a
postural perturbation (feedback based). These preprogrammed
corrections deal with actual perturbations that occur either because
of the suboptimal efficacy of APAs or because a perturbation comes
unexpectedly for the person.
APAs seen in leg and trunk muscles prior to a voluntary
movement are commonly smaller in amplitude in patients with
Parkinson’s disease (Bazalgette et al. 1986). These patients more
frequently demonstrate anticipatory cocontraction of antagonist
muscles acting at a postural joint (Viallet et al. 1987; Bouisset and
Zattara 1990), which apparently stiffens the joint and stabilizes it
against perturbations but is less efficient than the more commonly
observed pattern of alternating activity in postural agonist–antagonist
muscle pairs of healthy persons.
 Similar changes in feedforward motor adjustments in Parkinson’s
disease are seen during grip adjustments to manipulation of
handheld objects. These patients show larger absolute grip force
magnitudes and smaller modulation of the grip force during
movements when compared to age-matched healthy persons
(Gordon et al. 1997; Muratori et al. 2008; Jo et al. 2015). Taken
together, these observations suggest an overall impairment of
feedforward control over tasks and effectors.
Patients with Parkinson’s disease demonstrate profoundly
different preprogrammed postural adjustments, suggesting an
impairment in the hypothetical mechanism of preprogramming.
Stretching of a quiescent or voluntarily activated muscle of a patient
with Parkinson’s disease leads to long-latency muscle responses
(preprogrammed reactions or M2-M3; chapter 19) with an amplitude
that is considerably higher than in healthy persons (Rothwell et al.
1983; Cody et al. 1986; Hunter et al. 1988). Task-specific modulation
of preprogrammed reactions is seen in healthy persons; this ability is
impaired in Parkinson’s disease. So it is fair to say that these
patients demonstrate a poorly controlled increase in their feedback-
triggered corrective postural reactions.

PROBLEM 37.3
Suggest an experiment that would test the hypothesis that tremor
in Parkinson’s disease is an oscillation in a long-latency reflex
loop.

Patients with Parkinson’s disease show an ability to make much

more “normal-looking” steps, rather than their typical shuffling gait,
when they are asked to step over stripes painted on the floor (Giladi
2001; Morris et al. 2001). These observations suggest that the CNS
of these persons is able to generate a typical walking pattern but for
some reason does not do it during everyday activities. This
observation raises a practically important question: Should these
patients be trained and encouraged to produce more normal-looking
steps rather than their typical shuffling gait?
In chapter 41, we will return to a basic question of whether
apparently abnormal motor patterns should be corrected. Here we
would like to make a point that the shuffling gait in Parkinson’s
disease may represent an adaptive motor pattern to the cardinal sign
of impaired postural control. Shuffling gait is associated with smaller
transient forces when the foot touches the ground and, as such, it
may be expected to lead to smaller postural perturbations during
walking. In a friendly laboratory environment, where no unexpected
postural perturbations happen, the central nervous system of the
patient may lift the self-imposed restriction and facilitate more
normal-looking stepping. However, trying to reproduce this behavior
in the less predictable street environment may be dangerous,
potentially leading to falls.

37.4 Motor Synergies in

Parkinson’s Disease
Recent studies have documented major changes in movement
synergies (see chapter 22) in Parkinson’s disease, leading to poor
control of the stability of salient performance variables. Even at the
earliest stage of the disease, the patients show reduced indices of
synergies across a variety of tasks, from multifinger force production
to multi-muscle standing (Park et al. 2012; Falaki et al. 2016: Latash
and Huang 2015).
Figure 37.6 illustrates the multifinger synergy index in the task of
accurate force production followed by a rapid force pulse (starting at
time zero). Note the two major differences between the patients
(dashed lines with dark gray shades) and age-matched control
persons (solid lines with light gray shades). First, during steady-state
force production, the synergy index is smaller in the patients.
Second, there is a consistent drop in the synergy index in
preparation for the force pulse (anticipatory synergy adjustment,
ASA) seen in the control subjects but not in the patients. Note that
ASAs represent an important feedforward adjustment that facilitates
future action. To summarize, the patients have problems with stability
of the salient variable (force) and its agility.

Figure 37.6 The index of the multifinger synergy (ΔVZ) stabilizing total force
during the four-finger force production task by the right hand. The task involved
steady-state force production followed by a quick pulse into a target (starting at t0).
Note the higher steady-state value of the synergy index in the healthy control
subjects (dashed line, darker error shade) compared to patients with Parkinson’s
disease (PD, solid line, light error share). Note also the large ASA prior to the force
pulse initiation in the control group, but not in the PD patients.
Adapted by permission from J. Park, Y.-H. Wu, M.M. Lewis, X. Huang, and M.L. Latash,
“Changes in Multi-Finger Interaction and Coordination in Parkinson’s Disease,” Journal of
Neurophysiology 108 (2012): 915-924.

Similar changes in both stability and agility indices are seen also
in other tasks, including whole-body tasks such as standing. Both
indices are sensitive to dopamine replacement therapy (Park et al.
2014; Falaki et al. 2017). It is possible that the lack of ASAs
contributes to the impaired ability of patients with Parkinson’s
disease to initiate and modify actions, which can lead to episodes of
freezing, an inability to make a step in certain conditions. Freezing
can be a major contributor to falls and to the overall disability of
patients at later stages of the disease.
37.5 Treatment of Parkinson’s
Disease
Despite the recent progress in the understanding of the origins of
Parkinson’s disease and associated neurophysiological
mechanisms, treatment of this disorder is mostly symptomatic. The
most common strategy has been to take pills containing a precursor
of dopamine (L-DOPA). This drug does indeed induce a strong
therapeutic effect. However, it is also associated with a number of
side effects, including nausea, dyskinesias (to be discussed in
section 37.9), mood swings, hallucinations, delusions, and paranoid
psychosis. These problems are amplified by the fact that patients
develop tolerance to L-DOPA and require a continuous increase in
the dosage. In addition, high dosages of L-DOPA and large diurnal
variations in its concentration can accelerate the death of neurons in
the substantia nigra. Some of the aforementioned problems have
been partly avoided or minimized by using drugs with slow release of
L-DOPA. This prevents large daily variations in the drug’s
concentration.
Other pharmaceutical approaches to Parkinson’s disease include
using inhibitors of enzymes, COMT (catechol-0-methyltransferase)
and monoamineoxydase (MAO-B) involved in the L-DOPA
transformation in the brain. This helps keep the dopamine in the
brain for longer times. There is also a view that healthy brain
functioning depends on the balance between two types of neural
projections, those using dopamine and those using acetylcholine as
neuromediators. This view justifies using anticholinergic agents in
combination with dopamine replacement therapy.
Novel approaches to the treatment of Parkinson’s disease have
been developed lately. They involve, in particular, the implantation of
electrical brain stimulators into structures within the basal ganglia,
typically into the external part of the globus pallidus or into the
subthalamic nucleus, thus affecting the indirect loop within the basal
ganglia. Promising clinical results have also been reported in studies
with stimulation of thalamic structures (Benabid et al. 1998; Lozano
and Mahant 2004; Vaillancourt et al. 2004). This method, deep brain
stimulation (DBS), has so far led to variable results, from very strong
therapeutic effects to minor effects and even reports of an increase
in the incidence of falls in patients with DBS (Marconi et al. 2008;
Cossu and Pau 2017). Studies of the effects of DBS on motor
synergies have shown positive effects on indices of agility (longer
and larger ASAs) but no significant effects on indices of stability
(Falaki et al. 2018).

37.6 Huntington’s Chorea

Huntington’s disease is a relatively rare hereditary
neurodegenerative disorder with a prevalence of about 80 cases per
million of population (reviewed in Gusella and MacDonald 2004;
Marshall 2004). The disease typically starts in midlife, lasts for 15 to
20 y, and invariably leads to the death of the patient. The gene
responsible for Huntington’s disease has been identified, and so far
every person carrying this gene has developed the disease, while no
cases of Huntington’s disease have been reported among people
who do not have this gene.

PROBLEM 37.4
Should all newborns be tested for the gene responsible for
Huntington’s disease? Give your reasons pro and con.

Huntington’s disease is associated with atrophy of the caudate

nucleus and early loss of striatal GABAergic/enkephalinergic
neurons projecting to the external part of the globus pallidus and
neurons projecting to the substantia nigra (mediated by another
neurotransmitter, substance P). Glutamatergic activity is decreased
by as much as 85%, while acetylcholinergic activity may be
decreased by about 50%. These changes lead to an excessive
output from the thalamus to the cortex (figure 37.7), causing
excessive movements (hyperkinesia). As such, Huntington’s disease
may be seen as the pharmacological opposite of Parkinson’s
disease. However, at later stages, patients with Huntington’s disease
may develop typical signs of Parkinson’s disease. These
observations underscore the simplicity and inadequacy of the direct-
indirect loop scheme that we have been using until now.
Clinical features of Huntington’s disease include both motor and
nonmotor signs. The best-known motor feature of this disorder is
called chorea. Chorea represents generalized, irregular, restless,
often pseudopurposive movements that may include fidgeting hand
movements, dancelike gait (this is the origin of its name, chorea,
same root as in choreography), clumsiness, and slurred speech.
Typically, all parts of the body are involved. At early stages of the
disease, chorea can be suppressed voluntarily and looks like
restlessness or movements under emotional stress. At later stages,
choreic movements can be masked by rigidity and bradykinesia.

Figure 37.7 Atrophy of the caudate nucleus in Huntington’s disease leads to

changes in the indirect loop that result in a higher excitatory input from the
thalamus to the cortex. Excitatory projections are shown with solid lines and open
circles; inhibitory projections are shown with dashed lines and filled circles.
Projections that are weaker than in a healthy person are shown with thin lines,
projections that are stronger with thick lines. Compare to figure 37.1.
 Other motor abnormalities typical of Huntington’s disease include
nonchoreic gait abnormalities (slow, stiff, unsteady gait, with
episodes of freezing) and eye movement abnormalities (abnormal
saccades, problems with gaze fixation). At later stages, dysphagia
(inappropriate food selection and rate of eating, poor respiratory
control) and cachexia (weight loss, muscle wasting) can be seen.

PROBLEM 37.5
Some of the gait abnormalities in Huntington’s disease are very
similar to those in Parkinson’s disease (slow, stiff, unsteady gait
with episodes of freezing). However, one of these diseases is
classified as a hypokinesia, while the other is a hyperkinesia.
What can you conclude from these observations?

Nonmotor disorders of Huntington’s disease include depression,

irritability, loss of social skills and, at later stages, dementia.
Unfortunately, there is no treatment for this disorder. Attempts at
relieving the symptoms at different stages of the disease may
include cholinergic drugs or anticholinergic drugs, which at later
stages can be used in combination with L-DOPA.

37.7 Hemiballismus
Another hyperkinetic disorder associated with an injury to a structure
within the basal ganglia is ballism. Ballism is due to a lesion of the
subthalamic nucleus. Such a lesion may result from a
cerebrovascular accident (stroke) or from a malformation. Since the
left and right subthalamic nuclei are relatively far from each other,
cases of bilateral lesions of both subthalamic nuclei are extremely
unlikely. Most commonly, this disorder is associated with a lesion of
one of the subthalamic nuclei leading to clinical signs limited to the
contralateral side of the body. Therefore, the disorder is frequently
termed hemiballismus.
 Lesion of the subthalamic nucleus may be expected to lead to an
increase in the amount of excitation cortical structures get from the
thalamus (figure 37.8), leading to excessive movements on one side
of the body. Hemiballismus typically leads to uncontrollable rapid
movements of the contralateral limbs. These movements are
somewhat similar to those of chorea but typically faster and over
larger amplitudes. Treatment of hemiballismus is symptomatic,
similar to that of chorea.

37.8 Dystonia
Dystonia is defined as a syndrome of sustained muscle contractions
that produce twisting and repetitive movements and abnormal
postures. This rather vague definition unites cases of different
etiologies. However, a substantial number of patients with dystonia
show abnormalities in the basal ganglia, which justifies description of
this hyperkinetic disorder in this chapter.
Figure 37.8 Atrophy of the subthalamic nucleus leads to ballism (hemiballismus).
It results in changes in the indirect loop and higher excitatory input from the
thalamus to the cortex. Excitatory projections are shown with solid lines and open
circles; inhibitory projections are shown with dashed lines and filled circles.
Projections that are weaker than in a healthy person are shown with thin lines,
projections that are stronger with thick lines. Compare to figure 37.1.

Dystonias are classified into primary (or idiopathic) and

secondary. Primary dystonia may start in childhood (related to a
mutation in a DYT1 gene) or in midlife. There is, however,
epidemiological evidence suggesting a role for genetic factors in
adult-onset idiopathic dystonia. Secondary dystonias may be
associated with a variety of factors that include hereditary neurologic
syndromes (for example, Huntington’s disease), other neurological
disorders (such as multiple sclerosis and Parkinsonism), and
environmental causes (head trauma, drugs, toxins, encephalitis,
etc.). Secondary dystonia may also be psychogenic. Table 37.1
shows the highest percentage of psychogenic dystonia among
various psychogenic movement disorders.

Table 37.1 Incidence of Psychogenic Disorders

Disorder Frequency of occurrence
Dystonia 39%
Disorder Frequency of occurrence
Tremor 25%
Gait abnormality 5%
Myoclonus 11%
Parkinsonism 5%
Tic 0.7%
Stiff-man syndrome 0.3%

Another classification of dystonias refers to affected areas.

Dystonias are called focal when a single area is involved. Examples
of focal dystonias involve blepharospasm (the upper face is
affected, leading to intermittent or sustained bilateral eyelid closure
as a result of involuntary contractions), spasmodic dysphonia (the
vocal cords are affected, causing decreased smoothness of speech,
effortful voice, and reduction in loudness), torticollis (the neck is
affected, leading to twisting head movements and clumsy postures),
and writer’s cramp (an arm is affected, causing involuntary
contractions of limb muscles with twisting, repetitive movements and
abnormal postures). When two or more contiguous areas are
involved, dystonias are called segmental. For example, cranial
dystonias involve the face+jaw+tongue+vocal cords, while axial
dystonias involve the neck+trunk.
When two or more noncontiguous areas are involved, dystonias
are called multifocal. And finally, when the trunk, at least one leg,
and at least one more area are involved, dystonias are called
generalized. Focal dystonias are relatively common, with about 30
cases per 100,000 persons, while generalized dystonias are much
less common, with about 3 cases per 100,000.
Among focal dystonias, there is a group associated with overuse
of a limb during professional activities. These dystonias include
writer’s cramp (affecting writers, typists, and stenographers),
musician’s cramp (can be seen in guitar and trumpet players), and
athlete’s cramp (relatively common in golfers, snooker players, and
dart throwers). Interestingly, the affected limb can be used by these
persons in other activities that apparently involve contractions of the
same muscle groups—for example, writer’s cramp does not lead to
problems with using a fork and knife.

PROBLEM 37.6
What conclusions could you draw about the origins of writer’s
cramp based on the fact that the same muscles and joint may
show no dystonic signs in other activities?
Figure 37.9 Typical patterns of a voluntary wrist flexion movement in a person
with dystonia. EMG patterns are characterized by multiple, irregular bursts. The
trajectory is “bumpy” and may have hesitations and reversals.
Reprinted by permission from M.L. Latash and S.R. Gutman, “Abnormal Motor Patterns in
the Framework of the Equilibrium-Point Hypothesis: A Cause for Dystonic Movements?”
Biological Cybernetics 71 (1995): 87-94. © 1995 Springer-Verlag.

As mentioned earlier, increased movement variability is probably

the most general feature of all disordered voluntary movements. This
is particularly true for dystonic movements, whose distinctive
features include irregularity (figure 37.9). Oscillations, hesitations,
temporary reversals of the trajectory, and multiple EMG bursts (in
contrast to the commonly seen triphasic EMG pattern; see chapter
23) are common features typical of dystonic movements. Another
typical feature of dystonic movements that does not occur in all
patients is coactivation of antagonist muscles and frequently also of
distant muscles during attempts at unidirectional fast limb
movements.
Dystonia is apparently a problem of control, a problem of
disbalance among the descending signals that may not necessarily
correlate with any discernible pathology in the supraspinal or spinal
structures. A number of segmental abnormalities have been
described in dystonia, in particular a deficit in the Ia-mediated
inhibition and the Westphal phenomenon (Safronov and Kandel
1975; Chen et al. 1995; Berardelli et al. 1998). These, however, are
not seen in all patients, and they may represent secondary changes
in the system to a long-lasting primary disorder of motor control.
Indeed, motor patterns resembling those seen in dystonic
movements have been modeled within the framework of the
equilibrium-point hypothesis (Latash and Gutman 1994).
Treatment of dystonia is largely symptomatic. It can include such
different drugs as L-DOPA, antidopaminergic agents, and
anticholinergic therapy. In severe cases, botulinum toxin can be used
to prevent undesirable muscle contractions in the affected area.
Some patients discover particular maneuvers that help them control
dystonic movements. These may involve touching an affected part of
the body and sometimes even using inanimate objects (for example,
a purse) to control dystonic movements. Such maneuvers are
commonly called “gestes.”

37.9 Tardive Dyskinesia

Long-lasting pharmacological treatment of chronic neurological
disorders is sometimes associated with a particular motor disorder
called tardive dyskinesia (reviewed in Goetz and Horn 2004). This
term refers to late-appearing involuntary movements resulting from
the chronic administration of dopamine receptor blocking agents. It
can be induced by a variety of drugs, including antipsychotic drugs,
antidepressants, cough suppressants, and antihypertensive drugs.
Tardive dyskinesia may be associated with tardive dystonia and
tardive akathisia. The latter term refers to restlessness.
Clinical signs of tardive dyskinesia involve repetitive chewing
movements and tongue popping. Tardive dystonia may induce
oculogyric crisis—dystonia of extraocular muscles leading to
sustained ocular deviation, usually upwards, as well as involuntary
torsional movements of the neck, trunk, oral region, or face.

CHAPTER 37 IN A NUTSHELL
Disorders of the basal ganglia may
lead to movement poverty (hypokinesia)
or excessive movements (hyperkinesia).
Parkinson’s disease is a consequence
of a loss of dopamine-producing
neurons in the substantia nigra. Its
symptoms include poverty of movements,
tremor at about 5 to 6 Hz, rigidity,
bradykinesia (slowness), and deficits
in postural control. A deficit in
anticipatory postural adjustment and a
poorly controlled increase in
corrective postural reactions have
been reported as contributing to
postural deficits. The most common
treatment for Parkinson’s disease is a
precursor of dopamine (L-DOPA).
Promising clinical results have been
obtained with deep brain stimulation
of nuclei in the thalamus and basal
ganglia. Huntington’s disease is a
genetic disorder associated with a
dysfunction of the caudate nucleus. It
leads to choreic movements and
dementia. Hemiballismus is a
consequence of a lesion of one of the
subthalamic nuclei; it leads to poorly
controlled, fast, large-amplitude
movements in the contralateral side of
the body. Dystonia is a disorder that
in some cases is linked to a
dysfunction of the basal ganglia.
Voluntary movements in dystonia are
characterized by twisted, sustained
postures of the limb segments, limbs,
neck, or trunk. Long-lasting
pharmacological treatment of chronic
neurological disorders can lead to
tardive dyskinesia. Its clinical signs
involve repetitive chewing movements
and tongue popping.
 Chapter 38

Cerebellar Disorders

KEY TERMS AND TOPICS

abnormalities of stance and gait
cerebellar tremor
cerebellar rigidity
dysmetria
ataxia
Friedreich’s ataxia
cerebellar cognitive affective syndrome

The cerebellum is arguably the most enigmatic brain structure. On

the one hand, it contains more neurons than the rest of the brain
structures. On the other hand, an animal can survive a complete
removal of the cerebellum, recover, and function in a relatively
satisfactory way. Observations of behavioral changes after a
complete or partial injury to the cerebellum in animals resulted in a
view that the cerebellum mostly dealt with the motor function,
including balance. Only relatively recently has the cerebellum been
implicated in other functions, including cognition and emotions (see
chapter 10). In chapter 33, we discussed possible links of the
cerebellum to atypical behaviors, motor and nonmotor, in Down
syndrome, autism, Asperger’s disease, and the developmental
coordination disorder (DCD). These and other observations have led
to a shift in the understanding of the cerebellum not as simply a
“motor organ” of the brain but as a structure that is involved in many
brain functions crucial to a spectrum of subtle and sophisticated
behaviors.

38.1 Consequences of Cerebellar

Injuries in Animals
Cerebellar lesions produce rather different effects in different animals
(reviewed in Rothwell 1994; Bastian et al. 1999). In cats and dogs,
such lesions are reported to produce an increase in “muscle tone”
(this frequently used misnomer is discussed in chapter 36). The
increased tone is particularly strong in limb and neck extensor
muscles, leading to a typical picture of cerebellar rigidity or extensor
rigidity. These effects may be seen after ablation of only the anterior
cerebellar lobes.
The rigidity is not decreased after cutting the dorsal roots into the
corresponding spinal segments, which means that it does not
depend on stretch reflexes (or any other reflexes from
proprioceptors). Thus, it is assumed to be primarily due to an
increase in the excitability of α-motoneurons, and it is termed alpha-
rigidity. The mechanism of extensor rigidity is assumed to be based
on an increase in the activity of the vestibulospinal tract. This tract
originates in the Deiter’s nucleus (the lateral vestibular nucleus),
which normally receives inhibitory inputs from the anterior cerebellar
lobes. Cerebellar lesions in these animals also have effects on the
activity of the system of γ-motoneurons (fusimotor system); these
effects are inhibitory and lead to the depression of the stretch reflex.
The immediate effects of cerebellar lesions are followed by a
relatively quick recovery of a nearly normal extensor tone and stretch
sensitivity, which takes a few days.
In primates, there is smaller cerebellar input to the Deiter’s
nucleus. This might be the reason for the lack of direct strong effects
of cerebellar lesions on the excitability of α-motoneurons. So most of
the effects of cerebellar lesions on movements are assumed to be
mediated by an action on the fusimotor system. Such lesions lead to
depression of the “muscle tone” (hypotonia), which recovers very
slowly or may not recover at all (Gilman 1969). There are also
deficits in the rate and force of voluntary muscle contractions as well
as a low-frequency tremor, the cerebellar tremor (about 3 to 5 Hz).
The deficits are particularly pronounced during movements that
involve a postural component. One tentative, general conclusion
from these observations is that the cerebellum is responsible for
providing balance between the activation of the α- and γ-
motoneuronal systems (Thach et al. 1992b).

PROBLEM 38.1
What kinds of changes in descending effects on γ-motoneurons
may be expected to lead to decreased “muscle tone”?

In animal experiments, reversible local cooling is frequently used

to study the effects of local lesions of brain structures on
movements. Local cooling of the dentate and interpositus nuclei in
monkeys leads to motor deficits resembling those seen in patients
with cerebellar disorders (reviewed in Thach et al. 1992a,b). During
single-joint movements, typical changes include:
An increase in reaction time.
An increase in movement distance (target overshoot,
hypermetria) accompanied by a prolonged agonist EMG burst
and a delayed antagonist EMG burst.
Interruption of the rhythm of an oscillatory movement due to an
increase of time spent at the turning points.
Segmentation of slow movements, which become jerky and
demonstrate a 3 to 5 Hz tremor.
A perturbation induces an alternating EMG pattern in the flexor-
extensor muscles, leading to joint oscillation (figure 38.1).
 During tracking of a visual target, the movement becomes jerky,
and the animal seems to lose its ability to use velocity-related
information.
Note, however, that all these deficits may be relatively mild and
typically allow the animal to accomplish the task, albeit in a clumsy
manner.

Figure 38.1 After a cerebellar lesion, a perturbation of a joint may lead to an

alternating EMG pattern in the joint flexors and extensors, and a joint oscillation in
a healthy animal. Solid lines = before the lesion; dashed lines = after the lesion.

The deficits become much more pronounced when multijoint

movements are performed (reviewed in Bastian et al. 1999). Cooling
the fastigial nuclei in monkeys brings about the inability to sit, stand,
or walk, leading to frequent falls. Cooling the interpositus nuclei
leads to a severe 3 to 5 Hz tremor, which is particularly pronounced
during targeted movements (action tremor). Cooling the dentate
nuclei leads to a major overshooting of targets and loss of the ability
to use precision grip for food retrieval. So one may conclude that the
cerebellum may be much more involved in the coordination of many
muscles than in the control of simple, single-joint systems.
 PROBLEM 38.2
Earlier, an argument was made that single-joint movements do not
exist. How can this argument be reconciled with the
aforementioned difference in the effects of cerebellar lesions on
single-joint and multijoint actions?

Effects of cerebellar lesions in animals become much more

pronounced during the execution of novel tasks or tasks in novel
conditions. For example, a cat with a cerebellar lesion may be able
to walk downstairs successfully but only if all the stairs are of the
same height. The introduction of an unusually high or low step may
lead to stumbling, which is not seen in animals with an intact
cerebellum.

38.2 Consequences of Cerebellar

Disorders in Humans
Cerebellar disorders in humans may result from a number of causes
that lead to the death of neurons in the cerebellum or affect its input
or output pathways (reviewed in Cooper et al. 2004). Two major
causes of neuronal death are insufficient blood supply (ischemia)
and tumors. Neural pathways that connect the cerebellum to the rest
of the brain may suffer from demyelination or neurodegenerative
disorders of different etiologies. One of the most common disorders
within this group is olivopontocerebellar atrophy (OPCA), which
leads to atrophy of neural fibers that provide one of the two main
inputs into the cerebellum: the climbing fibers that originate in the
inferior olives (chapter 10).
The consequences of cerebellar injuries depend on the affected
area. In particular, the vermis receives input information mostly from
the periphery and the brainstem, with relatively little information
coming from the cortex. Its outputs are also projected mostly to the
brainstem and the spinal cord. An injury to the vermis or to its main
output cerebellar nucleus, the fastigial nucleus, commonly affects the
control of posture and balance and can induce gait and truncal
ataxia (a disorder characterized by discoordination, leading to
problems with movement initiation and its decomposition into
segments). At the same time, movements of the extremities may be
relatively spared.
The paravermal area receives inputs from the periphery,
brainstem, and cortex, including the motor cortex. Its outputs are
directed at the brainstem and the cortex. An injury to this area or to
its output nucleus, the interpositus nucleus, leads to changes in
motor cortex activity prior to movement initiation. Lesions restricted
to this area are very rare; therefore, it is hard to assign to them a
specific group of symptoms.
The lateral zone of the cerebellar cortex receives inputs mostly
from the cortex of the large hemispheres, and relatively little from the
periphery. Its outputs are also mostly targeting the cortex (and the
brainstem). An injury to this area or to the dentate nucleus leads to
problems with voluntary movements of extremities that show signs of
ataxia and tremor.

38.3 Abnormalities of Stance

and Gait
Disorders of balance and locomotion are the most common clinical
consequences of cerebellar injuries. These disorders may appear
alone (when the vermis area or the fastigial nucleus is affected) or in
combination with disorders of limb movements. Patients with
cerebellar disorders typically show increased postural sway (Diener
et al. 1984; Horak and Diener 1994). This increase in many patients
is seen mostly in the anterior–posterior direction. Closing the eyes
leads to more pronounced unsteadiness (Ohashi et al. 1993). These
patients spread their feet wider, apparently to increase postural
stability.
 Postural responses of patients with cerebellar disorders to
unexpected platform translations are increased (Horak and Diener
1994). Figure 38.2 shows that both muscles acting at the ankle joint
show an increase in their response to unexpected platform
translation backwards as compared to responses seen in a healthy
person. The responses in the patient show little modulation with the
speed and amplitude of the perturbation. Particularly dramatic
differences are seen in the EMG activity of the antagonist (tibialis
anterior).
Gait deficits in cerebellar disorders are characterized by series of
irregular steps with greatly variable foot placement in both anterior–
posterior and mediolateral directions (Diener and Dichgans 1992;
Thach and Bastian 2004). There is excessive lifting of the feet during
walking. These disorders become more pronounced when the
complexity of the task increases—for example, walking on tip-toes or
walking backwards.

PROBLEM 38.3
How would you distinguish the gait of a patient with a cerebellar
disorder from that of a person with Huntington’s chorea?

Postural and locomotor tasks can be accompanied by trunk or

head tremor that can be seen at a frequency of about 1 to 3 Hz. This
tremor is called titubation. It consists of rocking movements and can
be seen in association with any postural task, such as during sitting.
Figure 38.2 Postural responses of a person with a cerebellar disorder to platform
translation are much larger than in a healthy person. The person with the
cerebellar disorder shows substantial activation of the antagonist (tibialis), which is
not seen in the healthy person.
Reprinted by permission from F.B. Horak and H.C. Diener, “Cerebellar Control of Postural
Scaling and Central Set in Stance,” Journal of Neurophysiology 72 (1994): 479-493. With
permission from the American Physiological Society.

38.4 Voluntary Movements in

Cerebellar Disorders
Persons with cerebellar disorders can perform most everyday motor
tasks successfully; however, their movements look clumsy.
Clumsiness in cerebellar disorders gets contributions from the
following factors (figure 38.3). Voluntary movements of these
persons are initiated with longer delays. There are errors in the
magnitude of force, velocity, and timing leading to inaccurate
movement amplitudes (dysmetria) and problems with changes in
movement trajectory at appropriate times—for example, in tasks that
require tracking the trajectory of a particular shape (Beppu et al.
1984; Miall 1998; Topka et al. 1998). Dysmetria may include
overshoots (hypermetria) and undershoots (hypometria) of a target.
There have been attempts to link these problems to errors in
programming precise patterns of muscle activation and problems
with stopping the movement. However, as discussed in chapters 21
and 22, the brain cannot in principle program muscle activation
patterns because of the poorly predictable contributions from reflex
projection, which reflect the body’s interaction with changing external
force fields.
Impaired timing of movements in cerebellar disorders leads to two
phenomena called dysdiadochokinesis and dysrhythmokinesis. The
former term refers to an inability to maintain a constant rhythm
during alternating, fine repetitive movements such as, for example,
opposing each finger in a rapid succession against the thumb. The
latter terms refers to an inability to produce a required rhythm, such
as during tapping. Both types of tasks result in irregular, poorly
coordinated actions.
There is a minor decrease in muscle strength, but it does not
seem to be a limiting factor for everyday motor tasks. However,
persons with cerebellar disorders are commonly described as having
abnormal muscle tone. This abnormality is rather ambiguous
because it represents a combination of an apparently reduced (and
poorly defined; see chapter 36) “tone” and cerebellar rigidity. It has
been linked to a decreased level of activity of γ-motoneurons that
innervate intrafusal muscle fibers and control the sensitivity of
spindle sensory endings (Thach et al. 1992a).
Movements performed by persons with cerebellar disorders are
commonly decomposed into fragments, a phenomenon termed
movement segmentation. This phenomenon is particularly obvious
during more complex tasks that involve multijoint coordination
(Becker et al. 1991; Topka et al. 1998a). There is a marked disorder
of serial movements, possibly related to the aforementioned
problems with movement timing.
Figure 38.3 Schematic illustrations of typical features of disordered movements
after a cerebellar lesion. Left drawings are for a healthy person; right drawings are
for a person with a cerebellar disorder. (a) Irregular trajectory with multiple speed
peaks during reaching movement. (b) Inability to maintain rhythm. (c) Long time
delay and slow movement initiation.

During multijoint movements, persons with cerebellar disorders

are likely to show curved trajectories in contrast to the typically
straight trajectories of healthy persons (chapter 25). This may be due
to impaired coordination among the joints, which has been
hypothesized to result from an impaired ability of the neural
controller to predict and take into consideration joint interaction
torques (Bastian et al. 1996). Note, however, that this view accepts
as an axiom that joint coordination results from computation of
requisite torque profiles by the central nervous system, an axiom that
has received severe critique (e.g., chapter 21).

PROBLEM 38.4
What kinds of motor mistakes would you expect in a baseball
player with a mild cerebellar disorder who tries to hit or to catch a
baseball?

Cerebellar disorders are also seen in the activity of facial muscles

that lead to slurred speech and nystagmus (repetitive jumps of the
eyes). Other disturbances of extraocular movements include
impaired visual tracking and an impaired vestibulo-ocular reflex
(VOR).
Patients with cerebellar disorders show impaired adaptation to
novel conditions, which may include distorted sensory signals or
unusual force fields. Similar impairment can be seen in monkeys in
experiments with cooling of cerebellar structures. In particular, when
a healthy human or an animal is wearing prisms that distort visual
perception, the VOR (which keeps the axis of the eyeball in a
constant position during head movements) adapts quickly. Similarly
quick adaptation is seen in tasks where the person wearing prismatic
glasses is asked to perform an action toward a spatial target (e.g., to
throw a ball at the target). Such adaptation is lacking in cases of
cerebellar disorders (Martin et al. 1996).

38.5 Cerebellar Tremor

In addition to titubation, cerebellar disorders also lead to tremor (that
is, involuntary cyclic movements) of the limbs (reviewed in Gilman et
al. 1981). Cerebellar tremor has two components, static and kinetic;
its frequency is within the range from 3 to 5 Hz. Static or postural
tremor in patients with cerebellar disorders may look similar to that in
Parkinson’s disease. It is seen when a person holds a limb in a
certain position and tries to maintain this position over a few
seconds. The difference is that patients with Parkinson’s disease
show static tremor at complete rest—for example, in an arm resting
on a knee. Patients with cerebellar disorders show static tremor
when they actively keep a limb in a steady posture.
When a patient with a cerebellar disorder initiates a voluntary limb
movement, the movement may be associated with a kinetic tremor,
which typically involves alternating bursts of activation in proximal
muscle groups. The tremor may become stronger when the limb
approaches a target—for example, when a hand approaches a
glass. This tremor is called intentional, although this term has been
criticized as being misleading because it lumps together postural and
kinetic tremor (Gilman et al. 1981).
A number of mechanisms have been suggested as potential
contributors to cerebellar tremor (reviewed in Bastian et al. 1999).
These include activity in proprioceptive feedback loops from stretch-
sensitive receptors to the cortex, increased M2-M3 responses to
stretch (see chapter 19), and a general primary deficit in limb
stabilization during postural tasks.
Another atypical phenomenon sometimes leading to a rhythmic
motion in persons with cerebellar disorders is called check-and-
rebound. When a healthy person is asked to extend an arm and
close the eyes, and then another person taps the wrists lightly, there
is an accurate quick motion of the arm back to its original position.
Persons with cerebellar disorders show a large displacement of the
tapped limb, overshooting the original position. This response is
followed by a sequence of oscillatory motions that finally bring the
arm back.

PROBLEM 38.5
Suggest a physiological mechanism for the check-and-rebound
phenomenon.
38.6 Ataxias
In humans, cerebellar disorders are accompanied by ataxia of
voluntary movements, a decomposition of movement into a number
of jerky segments (reviewed in Gilman 2004). Ataxia has been
associated with problems with movement initiation, termination, and
velocity control. A particular case of ataxia, when movement of
speech articulators is affected, is called ataxic dysarthria.
There is a group of ataxic disorders that are inherited and
characterized by discoordination and problems with balance. These
are typically progressive neurodegenerative disorders that can lead
to significant impairment and death. Probably the most common of
inherited ataxias is Friedreich’s ataxia; it happens at a rate of 1
case per 30,000 to 50,000 among Caucasians. It is limited to certain
geographic areas such as Europe, Northern Africa, the Middle East,
and India. Clinical signs of Friedreich’s ataxia start before the age of
25; they include dysarthria, loss of reflexes, and axonal sensory
neuropathy, including loss of cutaneous and proprioceptive
sensation. The same genetic mutation has been shown to lead in
some people to a later-onset ataxia involving hyperreflexia and signs
of spasticity.
The second most common among inherited ataxias is ataxia
telangiectasia. This disorder starts early in childhood; the patients
are wheelchair bound by their second decade, and their life span
does not extend beyond early adulthood. Ataxia telangiectasia leads
to progressive ataxia, dysarthria, facial hypotonia, and oculomotor
abnormalities.
Other forms of inherited ataxias include spinocerebellar ataxias,
which are relatively rare and limited to particular subpopulations.
Ataxias rarely respond to pharmacological treatment. Supportive
care and rehabilitation remain the only options for helping these
patients and their families.

38.7 Changes in Motor Synergies

Cerebellar disorders have been described with such clinical terms as
asynergia and dyssynergia since the 19th century (Babinski 1899).
Those terms were used informally as synonyms of poor coordination.
The role of the cerebellum in organizing multiple effectors (e.g.,
muscles) into stable groups, addressed as synergies, has been
accepted by many researchers (Welsh and Llinas 1997; Berger et al.
2020). This view has been particularly influential within the
framework of internal models—hypothetical computational structures
in the brain that try to model interactions among body parts and
between the body and the environment (see chapter 22)—which
were assumed to use the available afferent and efferent signals for
predicting changes in the environment and to coordinate multiple
effectors to reach task-specific goals in this environment (Wolpert et
al. 1998; Imamizu and Kawato 2012; Herzfeld and Shadmehr 2014).
Within the definition of motor synergies as neural organizations of
elements with the purpose to stabilize performance (chapter 22),
studies of patients with cerebellar disorders have shown changes
similar to those in patients with Parkinson’s disease (Park et al.
2013). In particular, patients with OPCA show reduced indices of
multifinger synergies in accurate force production tasks as well as
smaller adjustments of those synergies (anticipatory synergy
adjustments [ASAs]; see chapter 37) in preparation for quick action.
As described in chapter 37, these signs can be viewed as problems
with action stability and agility.
Note, however, that in more recent literature, OPCA has been
viewed as a component of a more general disorder known as
multisystem atrophy (Peeraully 2014; Koeppen 2018; see chapter
40). It can involve atrophy not only of the cerebellum and associated
neural pathways but also of the circuitry within the basal ganglia.
Hence, more studies are needed to clarify the role of the cerebellum
in motor synergies that ensure the stability of actions.

38.8 Cerebellar Cognitive

Affective Syndrome
More attention has been paid to the nonmotor consequences of
cerebellar disorders. These consequences have been united under
the name cerebellar cognitive affective syndrome (Schmahmann and
Sherman 1998; Schmahmann 2004). They are particularly prominent
in patients with lesions of the posterior lobe and of the vermis.
Patients with cerebellar lesions demonstrate an impairment of
such general abilities as planning, abstract reasoning, working
memory, and spatial cognition. The latter feature is reflected in the
impaired ability of these patients to copy complex pictures and to
draw pictures of familiar objects that are characterized by regular
spatial features, such as the face of a clock. This feature can be
reflected in the performance of seemingly simple tasks, such as
dividing a straight line drawn on a piece of paper into two equal
parts.
Language deficits in persons with cerebellar disorders include the
improper use of grammar (agrammatism) and unusual voice
inflections (dysprosodia or dysprosody). Verbal fluency may be
lost. This condition may also lead to personality changes, resulting in
disinhibited and inappropriate behavior.
Drug treatment of cerebellar disorders has shown limited
effectiveness. In particular, such medications as 5-
hydroxytryptophan, clonazepam, and vitamin E have been reported
to have limited beneficial effects. In cases of infarctions,
hemorrhages, and neoplasms, treatment of the original cause of the
disorder (surgical or medical radiotherapy and chemotherapy) is
commonly followed by physical therapy.
To summarize, observation in animal studies and in persons with
cerebellar pathologies suggests that the cerebellum plays a major
role in the acquisition of “new things” rather than in the
implementation of what has already been learned. The “new things”
may apparently be related to different areas, including motor
coordination, language abilities, cognitive abilities, and social
interactions. All these phenomena seem to have an important
common feature: They all involve task-specific interactions of many
elements (muscles, joints, words, notions, and persons). It may be
that problems with the cerebellum lead to an impairment in the
formation of synergies that are reflected in all aspects of human
activities where stability of salient variables is crucial for success.

PROBLEM 38.6
Suggest an explanation for the suppressed tonic vibration reflex in
patients with cerebellar disorders.

CHAPTER 38 IN A NUTSHELL
In humans, cerebellar lesions are
associated with injuries, tumors, or
demyelinating processes affecting
neural tracts that carry information
to and from the cerebellum. Cerebellar
lesions lead to gross motor disorders
that can include rigidity, tremor,
hypotonia, dysmetria, ataxia, and
asynergia. Tremor in persons with
cerebellar disorders has a static,
postural component and a kinetic,
intentional component. Cerebellar
rigidity may be mediated by
abnormalities in inputs to α- or γ-
motoneurons. There are changes in
posture, which becomes unsteady, and
gait, which is characterized by
irregular steps and excessive lifting
of the swing foot. Voluntary movements
by persons with cerebellar disorders
are fragmented and show poor
interjoint coordination. They also
show major disorders of timing,
particularly in serial movements.
Motor learning and adaptation are
severely affected. Inherited cases of
ataxia, in particular Friedreich’s
ataxia, commonly represent progressive
neurological disorders leading to
serious impairment of motor function.
Cerebellar dysfunction leads to
impaired synergic control of movements
involving both reduced stability of
steady states and reduced agility when
variables have to be changed quickly.
Cerebellar cognitive affective
syndrome represents an impairment of
such general abilities as planning,
abstract reasoning, working memory,
and spatial cognition.
 Chapter 39

Cortical Disorders

KEY TERMS AND TOPICS

stroke
neglect
apraxia
myoclonus
tics
Tourette syndrome
Williams syndrome

The cortex of the large hemispheres contributes to a variety of

functions, including cognition, perception, and action (chapter 8).
Hence, a cortical injury can lead to a spectrum of consequences,
depending on the location and extent of the injury. Motor disorders
associated with malfunctioning of neural loops that involve cortical
neurons range from relatively mild (as, for example, in some forms of
tics) to disabling and even life-threatening (such as those that follow
a major stroke). First, we will consider the typical consequences of
lesions of different cortical lobes, assuming that the lesion is
moderate in size and is not associated with any neurodegenerative
disorder (for reviews, see Watts and Koller 2004). Such injuries may
result from malformations, small cerebrovascular accidents (CVAs or
strokes), or wounds.
39.1 Consequences of Lesions of
Different Cortical Lobes
Lesions of the frontal lobes typically lead to behavioral changes that
include irritation, perseveration (repetitive, apparently purposeless
actions), and general impairment of normal social conduct. These
consequences may be associated with the emergence of primitive
reflexes, such as the grasp reflex and the sucking reflex.
Common consequences of lesions of the parietal lobes involve a
defect of attention to events that take place in the contralateral visual
field (reviewed in Mesulam 1999; Vallar et al. 2003). In severe cases,
this can lead to neglect (a person does not recognize objects
presented in the contralateral visual field) or even denial (one’s own
contralateral limbs are viewed as belonging to someone else).
Interestingly, neglect may be associated with an impaired ability to
imagine acting in the contralateral field, while the ability to act in that
field may be relatively preserved (Danckert et al. 2002). For
example, if a person is asked to catch a ball and, after a few throws
in the unimpaired, ipsilateral (to the injury site) visual field, the ball is
thrown into the contralateral field, the person may be able to catch it.
However, later the person cannot account for this action and insists
that the ball got into the hand “by itself.” These observations suggest
that the brain fails to integrate the generally intact somatosensory
and visual sources of information into perception of the body.
Patients with injuries to the parietal lobes are also likely to show
apraxia, which is a disorder of simple motor skills. Apraxia is
particularly common in persons with damaged parietal areas of the
dominant hemisphere.
Lesions of the temporal lobes may lead to seizures, aggression,
and episodes of explosive rage. Aphasia (disordered spoken
language) may be seen in cases of lesions to the dominant temporal
lobe. Bilateral lesions of the tips of occipital lobes may lead to severe
neglect and virtual blindness.
 Cortical injuries frequently lead to disorders of language that
depend strongly on the affected area. There are different types of
aphasias. Nonfluent aphasia is characterized by slow speech and
perseveration in repeating words and word combinations. It is more
common in cases of injuries to the posterior frontal cortex, inferior
motor cortex, and adjacent insula. Fluent aphasia is a language
impairment resulting in nonsubstantial, fluent speech; it may result
from a lesion of the posterior superior temporal gyrus. Conduction
aphasia represents fluent speech with semantic confusions; it is
more common in cases of posterior lesions.

39.2 Stroke
A cerebrovascular accident (CVA), commonly known as a stroke,
involves an interruption of the normal blood supply to a brain area as
a result of a blood vessel rupture (hemorrhagic stroke) or blocked
blood flow (ischemic stroke). The clinical picture depends strongly on
the area affected by the stroke and the extent of the accident.
Strokes that affect the medulla are life-threatening. They can lead
to problems with breathing and cardiac function. Problems with
voluntary movements can also be seen, mostly caused by
interruption of transmission along descending pathways.
When a stroke affects the brainstem, there are consequences
related to dysfunction of the many nuclei that are located in the
brainstem. In particular, brainstem strokes commonly lead to
vestibular problems and associated problems with postural control.
Oculomotor problems may lead to nystagmus, while dysarthria
(disordered speech) may result from dysfunction of cranial nuclei
that innervate articulators. Discoordination of movements can be
seen, similar to that typical of cerebellar disorders produced by a
dysfunction of the climbing fiber input into the cerebellum. In cases
of an injury to the corticospinal pathways, patients can show signs of
spasticity typical of strokes that affect the large hemispheres.

PROBLEM 39.1
 What parts of the body are likely to show spastic signs following a
unilateral stroke affecting corticospinal pathways?

Strokes affecting the large hemispheres lead to both motor and

nonmotor consequences including, in particular, impaired cognition.
General motor consequences include hemiparesis, spasticity, and
discoordination, which affect posture and locomotion (Dietz and
Berger 1984; Diener et al. 1993; Levin 1996; Kautz et al. 2005).
Following a cortical stroke, patients frequently show narrow stance
(the feet placed too close to each other in the mediolateral direction)
and a scissor gait (with the leading foot placed such that it blocks the
future path of the trailing foot). Both features contribute to the
postural instability of patients after stroke and slow down their
recovery.
Clinical features of cortical strokes commonly show a proximal-to-
distal gradient of symptoms, with distal muscles more affected. In
particular, the hand function is likely to suffer more than arm
movements such as reaching that are controlled by more proximal
muscle groups (Shelton and Reding 2001; Mercier and Bourbonnais
2004; Michaelsen et al. 2004).
The consequences of lesions of the corticospinal tract are
sometimes called upper motoneuron syndromes in the clinical
literature. This term implies that there are two motoneurons involved
in the control of movements, an upper motoneuron in the cortex and
a lower motoneuron in the spinal cord (the α-motoneuron). The lack
of a one-to-one mapping between cortical areas and muscles (see
chapters 8 and 41) makes this term ambiguous and even
misleading. Following the terminology introduced by Hughlings
Jackson, clinical signs of such strokes are classified into positive and
negative. Note that these terms are not used as synonyms of “good”
and “bad,” but rather imply motor phenomena that are seen in
patients but not in healthy persons (positive signs, such as increased
muscle tone, increased reflexes, and spasms) and those that are lost
in patients as compared to healthy persons (negative signs, in
particular, weakness and discoordination).
 Limbs on the two sides of the body typically show dramatically
different degrees of impairment following cortical stroke. Sometimes,
one side of the body is even considered healthy. This is imprecise
since changes in the motor function of the relatively unimpaired
limbs have been documented (Colebatch and Gandevia 1989;
Cramer et al. 1997; Jones et al. 1989; Maenza et al. 2020). Besides,
stroke is followed by plastic changes of projections both within the
brain and between the brain structures and the spinal cord (reviewed
in Johansson 2000; Rossini and Pauri 2000; Hallett 2001). In
particular, the role of uncrossed fibers of descending tracts may
increase, leading to a shift in the balance of limb control from the
injured contralateral to the spared ipsilateral hemisphere.
Reaching movements by the contralateral arm after a unilateral
stroke are characterized by irregular trajectories that deviate from
the nearly straight trajectories typical of reaching in healthy persons
(top panels in figure 39.1). Reaching by the ipsilesional (same side
of the body as the lesion site) arm may show nearly normal
trajectories. When an object is placed further away from the body,
patients after a severe stroke are more likely to involve trunk motion
to help them to reach (the left bottom panel in figure 39.1). This
strategy may be viewed as adaptive to the loss of control of the arm
muscles.
Figure 39.1 Patients after stroke are likely to use trunk motion during reaching,
which may be viewed as an adaptive strategy. Note the lack of trunk motion in the
healthy persons (top panels) and mildly affected stroke patient (right bottom
panel). In contrast, the more affected patient (left bottom panel) showed
substantial trunk motion.
Reprinted by permission from M.F. Levin, “Interjoint Coordination During Pointing
Movements is Disrupted in Spastic Hemiparesis,” Brain 119 (1996): 281-293. ©1996 Oxford
University Press.
Figure 39.2 The index of the multifinger synergy (ΔVZ) stabilizing total force
during the four-finger force production task by the right hand. The task involved
steady-state force production followed by a quick pulse into a target (starting at t0).
Note the higher steady-state value of the synergy index in the healthy control
subjects (CS) compared to patients with Parkinson’s disease (PD), but no change
compared to stroke survivors (Stroke). Note also the large ASA prior to the force
pulse initiation in the control group, but not in the two patient groups.
Adapted by permission from M.L. Latash, Physics of Biological Action and Perception (New
York: Academic Press, 2019).

Following cortical stroke, patients commonly show major difficulty

in using the contralesional extremities, in particular the arm and hand
during such actions as reaching and object manipulation. In contrast,
indices of multijoint synergies stabilizing hand trajectory during
reaching and pointing and of multidigit synergies stabilizing total
force show only minor changes (Reisman and Scholz 2003; Jo et al.
2016). The contrast between stroke survivors and patients with
Parkinson’s disease is rather striking: figure 39.2 shows the synergy
index computed for the task of steady force production followed by a
quick force pulse. Note the very similar steady-state values for the
healthy control subjects and patients after stroke in contrast to the
much lower values for the group of patients with Parkinson’s
disease. These results suggest that the stability of salient
performance variables is more dependent on proper functioning of
subcortical loops (including those through the basal ganglia) than on
the state of cortical neurons. We should also note the much shorter
anticipatory changes in the synergy index (anticipatory synergy
adjustments, ASAs) in both patient groups shown in figure 39.2. This
finding reflects difficulties with movement initiation across all these
patients.

PROBLEM 39.2
What other brain circuits, in addition to those involving the basal
ganglia, contribute significantly to motor synergies?

Slow spontaneous recovery is typical after stroke. The degree of

recovery varies within a wide range, from nearly complete to
resulting in a severe disability. The process of recovery can be
helped by drugs and physical therapy. Pharmacological therapy after
stroke has several objectives. One of them is to prevent blood clot
formation that may result in secondary strokes. Antispastic
medications are used to alleviate the positive signs of spasticity and
to facilitate voluntary movements.
Since strokes typically lead to hemisyndromes with one side of the
body being much less affected than the other, patients tend to
perform most everyday tasks using the relatively spared limbs. For
example, they are more likely to stand with the center of pressure
shifted toward the relatively unimpaired foot and to use the relatively
unaffected hand during eating (even if this hand is nondominant).
These motor strategies help the patients perform many motor tasks,
but they also potentially slow down the recovery of the motor
function of the affected limbs. To encourage or even force the
patients to use their affected upper limbs, a suggested approach to
therapy has been termed constraint-induced therapy (reviewed in
Taub and Morris 2001; Mark and Taub 2004). This approach involves
limiting the ability of the relatively spared upper limb to participate in
everyday tasks (for example, by putting a mitten on the hand). As a
result, the patients are forced to use the impaired hand for a variety
of everyday activities. Such forced use represents a controversial
approach since it limits the range of tasks the patients can perform
without using the relatively spared hand. Besides, a number of
studies have shown that performing bilateral movements may be
beneficial for stroke rehabilitation (reviewed in Rose and Winstein
2004; Cauraugh and Summers 2005). Note that the main ideas of
bilateral movement therapy and constraint-induced therapy are
incompatible.
A strategy similar to constraint-induced therapy has also been
used to facilitate the recovery of postural control and stability during
walking. It involves placing a wedge-shaped insole into the shoe
worn on the relatively unimpaired foot (figure 39.3 wedge). This
leads to a shift of body weight toward the affected foot, forcing its
use in postural and locomotor tasks (Rodriguez and Aruin 2002).
More recently, the idea of inducing some degree of discomfort on the
less-affected body site has been developed with the goal of
increasing the use of more affected effectors and thus contributing to
their recovery (Aruin and Kanekar 2013; Oludare et al. 2017). Such
“discomfort-induced therapy” can include, in particular, using insoles
with sharp protrusions in the shoe on the less-affected foot (figure
39.3 discomfort), which leads to shifting the body weight toward the
more affected foot.
Figure 39.3 An illustration of using insoles in the shoe worn on the less-affected
(ipsilesional) foot after cortical stroke to encourage weight shift toward the more
affected foot. The dashed image shows a figure with symmetrical shoes with the
body weight shifted toward the less-affected side. The solid image shows
symmetrical weight distribution caused by wearing a wedge-shaped insole
(Wedge) or an insole with sharp protrusions (Discomfort).

39.3 Myoclonus
Myoclonus represents a brief muscle jerk caused by a neuronal
discharge (reviewed in Toro and Hallett 2004). It may be
accompanied by a single muscle burst or repetitive muscle bursts of
activity. Other movement disorders may accompany myoclonus,
such as dystonia (chapter 37) and essential tremor (chapter 40).
Myoclonus may occur in healthy people in special conditions.
Nocturnal leg myoclonus is the most common example: Jerky leg
movements occur rather frequently while falling asleep or in the
middle of the night. When myoclonus is the only or most important
neurological problem, it is called essential myoclonus. Most cases of
essential myoclonus are familial. Myoclonus may be part of
idiopathic epilepsy, particularly in children. Secondary myoclonus
may be associated with various causes including trauma, intoxication
(bismuth), renal failure, Huntington’s disease, olivopontocerebellar
atrophy, and corticobasal degeneration.
 Clinically, myoclonus is classified into spontaneous, reflex, and
action (reviewed in Obeso and Zamarbide 2004). Spontaneous
myoclonus may affect a single muscle or a small group of muscles
(focal), several muscle groups (multifocal), or nearly the whole body
(generalized). Spontaneous myoclonus may be associated with a
seizure, known as myoclonic epilepsy. Spontaneous myoclonic
bursts of muscle activity may be singular or rhythmic; in the latter
case, the frequency of the muscle EMG bursts is relatively low,
between 1 and 4 Hz. Myoclonic activity may even persist in sleep.
Reflex myoclonus is triggered by sensory stimuli that can be of
different modalities: visual, auditory, or somatosensory. The term
action myoclonus refers to myoclonic discharges associated with
attempts to perform a voluntary action. It can affect both postural
muscles and focal prime movers. Figure 39.4 illustrates
schematically action myoclonus that affects several limb muscles. In
particular, when a standing person performs an arm movement,
myoclonic muscle activity can be seen in both arm muscles and in
leg and trunk postural muscles. Action myoclonus may be negative;
this means that it leads not to a burst of muscle activity but rather to
a silence period in an ongoing steady-state muscle activation. An
example of negative myoclonus is shown by the arrow in the right
part of Figure 39.4. Note that rapid EMG bursts seen in a distal
muscle may be accompanied by transient gaps in steady-state
activation—negative myoclonus—of a proximal muscle (Obeso et al.
1983, 1985). Negative myoclonus in postural muscles may
compromise postural stability.
Figure 39.4 A schematic illustration of myoclonus in a distal limb muscle (lower
trace) and negative myoclonus (transient gap in steady-state muscle activation) in
a proximal muscle of the same limb.

Cortical myoclonus may result from a focal lesion, such as that

from a tumor, a trauma, or a stroke. It can also accompany focal
epilepsy. Myoclonic activity is believed to originate in the
sensorimotor cortex, with the signals from cortical neurons
propagating via the corticospinal tract. This hypothesis is supported
by different time delays of myoclonic EMG bursts in arm and leg
muscles to certain types of peripheral stimulation (C-reflex; Sutton
and Mayer 1974), 35 to 50 ms and 60 to 70 ms, respectively
(Shibasaki et al. 1978). Figure 39.5 illustrates schematically the
responses to electrical stimuli applied to an arm nerve (the median
nerve) and to a leg nerve (the tibial nerve). Stimuli produce unusually
large, or giant, somatosensory evoked potentials (SEPs) and EMG
responses with delays suggesting transmission of action potentials
via a transcortical loop.
Figure 39.5 Giant SEPs and C-reflexes in response to electrical stimulation of
(a) the median nerve and (b) the tibial nerve. Note the longer latencies of the
cortical response and reflex muscle response in b as compared to a.

The clinical features of cortical myoclonus involve spontaneous

localized muscle jerks, sometimes with secondary generalization;
that is, they are followed by muscle bursts of activity in neighboring
muscle groups. EMG bursts are brief, lasting for a few tens of
milliseconds. Evoked potentials recorded over the somatosensory
cortex in response to peripheral skin stimulation are typically
increased in amplitude.
A relatively rare disorder has received much publicity following
outbreaks of the mad-cow disease. This disorder, the Creutzfeldt-
Jakob disease, is a rare, fatal, transmissible encephalopathy that
leads to rapidly progressive dementia and myoclonus (reviewed in
Johnson and Gibbs 1998; Weihl and Roos 1999).
There are cases of myoclonus that are of noncortical origin. The
most common is reticular reflex myoclonus, which typically leads to
muscle contractions involving the whole body. Myoclonic activity may
also have a spinal origin, possibly induced by signals traveling along
propriospinal pathways.
The biggest problem in treating cortical myoclonus is that
myoclonic discharges use the same physiological pathways as those
involved in voluntary movements. As a result, drug therapy may
eliminate myoclonus but may also have significant side effects on
voluntary movements. Drugs that are believed to act by increasing
GABA activity in the cortex have been used, such as 5-
hydroxytryptophane (a serotonin precursor), clonazepam, and
piracetam.

PROBLEM 39.3
How can a myoclonic discharge be distinguished from a spastic
spasm?

39.4 Tics
A tic is defined as a brief, repetitive, and seemingly purposeless,
stereotyped action that may involve one muscle or muscle groups.
Typically, tics are seen in childhood, but they may persist for a
lifetime. In children, tics are commonly associated with Tourette
syndrome (next section).
Tics are classified as being either primary or secondary. Primary
tics may be transient (lasting from 1 mo to 1 yr) or chronic (lasting
over 1 yr). Secondary tics may follow an infection (encephalitis) or a
basal ganglia disorder, or they may be drug-induced. The range of
drugs reported to have induced tics is very broad and includes
stimulants, anticonvulsants, L-DOPA, antidepressants, and even
birth control pills.
Tics may be fast (clonic) or sustained (dystonic). Patients report
that motor tics and vocal tics are under voluntary control—they follow
an urge to produce an action or a sound. However, researchers and
physicians view tics as involuntary; unlike voluntary movements, tics
are not preceded by a readiness potential (chapter 8). Sometimes,
brief sensory experiences precede tics (these are called sensory
tics). Table 39.1 presents a summary of common motor, vocal, and
sensory tics, both simple and complex.
The neurophysiological origin of tics is unknown, They are
believed to result from abnormalities in corticostriatothalamic
pathways—consequences of a disorder involving signal transmission
through the basal ganglia. Therapies for tics reflect the potential
involvement of the basal ganglia in these hyperkinetic disorders:
dopamine blockers and acetylcholine-like drugs have been used to
treat tics.

PROBLEM 39.4
How would you distinguish a tic from a myoclonic discharge?

39.5 Tourette Syndrome

Tics in children may be associated with Tourette syndrome, which
was named after a French neurologist, Giles de la Tourette
(reviewed in State et al. 2001). The following five conditions have to
be met for a child to be diagnosed with Tourette syndrome:
1. Multiple motor and vocal tics must be present.
2. Tics should occur many times a day, nearly every day, or
intermittently for at least a year.
3. Tics should cause a marked distress or significant impairment
in social, occupational, and other areas.
4. The onset of tics should be before 18 yrs of age.
5. The tics should not be due to a direct action of a drug or a
general medical condition.
Tourette syndrome is traditionally viewed as a rare disorder,
although assessments of its prevalence range from 0.03% to 4% in
school-age children. It shows a 3:1 male predominance. A study
reported a genetic anomaly on chromosome 13 that may be linked to
Tourette syndrome, although in a small number of cases (Abelson et
al. 2005).
Children who suffer from Tourette syndrome show normal
intelligence but have difficulties with reading, writing, and arithmetic.
They are prone to self-injurious behavior, obsession, hostility, insults,
derogatory remarks, and destroying property. In general, these kids
may seem like typical misbehaving children. In about 50% of cases,
Tourette syndrome may be accompanied by attention deficit
hyperactivity disorder (ADHD). In about the same percentage of
cases, an obsessive-compulsive behavior may be observed.
Typically, Tourette syndrome shows a rostral-to-caudal
progression of symptoms with early signs occurring in the eye, face,
and head movements and progressing with age to movements of the
legs and trunk.
As with many neurological disorders, the neurophysiological
mechanisms underlying Tourette syndrome are unknown. Striatal
dopamine receptor hypersensitivity is suspected. This hypothesis is
supported by the effectiveness of dopamine receptor antagonist
drugs and the observations of worsening of symptoms under
dopaminergic drugs.
Tics in Tourette syndrome sometimes resemble other disorders. In
particular, they may resemble muscle jerks in myoclonus. However,
more complex motor tics and sustained dystonic tics are usually
present in Tourette syndrome. Another differentiating factor is that
myoclonus increases with intentional movement, while tics can be
suppressed intentionally, at least temporarily. Facial tics may
resemble blepharospasm (dystonia). However, tics are rarely limited
to eye movements. Besides, Tourette syndrome typically starts in
childhood, while idiopathic dystonia is typically adult-onset.

Table 39.1 Common Tics

Motor Vocal Sensory

Simple
Frequent blinking Sniffing Burning
Blepharospasm Grunting Tightness
Grimacing Throat clearing Muscle tension
Pouting Barking Tingling
Head jerking Coughing Itching
Jaw opening Growling Impulsion
Shoulder shrugging Moaning
Fist clenching Humming
Motor Vocal Sensory

Complex
Head twisting Panting Inner tension
Spitting Belching Pain syndrome
Hitting (self or others) Stuttering Phantom tics
Jumping Echolalia
Squatting Coprolalia
Pelvic/abdominal thrusting

In mild cases, most children with Tourette syndrome do not

require medications and adapt well. It is important to educate family
members, peers, and school personnel as well as to provide
supportive counseling. Pharmacotherapy is used only when tics are
functionally disabling.

39.6 Williams Syndrome

Williams syndrome is a rare disorder that affects 1 out of 50,000
children. It is associated with reduced cerebral volume and severely
depressed nonverbal IQ. In contrast, children with Williams
syndrome show high verbal and grammatical fluency (reviewed in
Mervis 2003). Their descriptions of objects and animals are colorful
and filled with detail. By contrast, they are unable to draw a simple,
coherent picture of that same object or animal. They also show
pronounced deficits in tasks that require copying a picture.

PROBLEM 39.5
Another rare disorder is associated with a basically intact cortex
and severely abnormal cerebellum. What striking differences
would you expect to see in persons with that disorder and in
children with Williams disease?

CHAPTER 39 IN A NUTSHELL
Cortical lesions lead to a variety of
motor, sensory, and cognitive
consequences, depending on their site
and extent. Strokes result from an
interruption of normal blood supply to
an area of the brain. They lead to a
mixture of positive and negative
signs, primarily in the contralateral
half of the body, including weakness,
discoordination, and spasticity.
Synergy indices show only minor
changes after stroke, but anticipatory
synergy adjustments are reduced in
patients, reflecting difficulties with
movement initiation. Exercise of the
affected body parts seems to be most
effective for recovery after stroke.
Myoclonus represents brief muscle
jerks that may be triggered by a
sensory stimulus (reflex myoclonus) or
by a voluntary action (action
myoclonus), or they may occur
spontaneously. Tics are brief,
repetitive, and seemingly purposeless
stereotyped actions. They can be
vocal, motor, or sensory. Tics are
sometimes associated with Tourette
syndrome, a condition in young
children that is typically transient.
Williams syndrome is a rare disorder
characterized by reduced cerebral
volume and severely depressed
nonverbal IQ. In contrast, children
with Williams syndrome show high
verbal and grammatical fluency.
 Chapter 40

Systemic Disorders

KEY TERMS AND TOPICS

amyotrophic lateral sclerosis
multiple sclerosis
multisystem atrophy
essential tremor
cerebral palsy
Wilson’s disease

In this chapter, we will consider a number of disorders that affect

multiple anatomical structures and pathways within the central
nervous system. The term systemic disorders is also used with
respect to disorders that are not limited to the central nervous
system. These may include endocrine and metabolic disorders (such
as diabetes mellitus, hypo- and hyperthyroidism, and liver failure),
infections (including encephalitis and Lyme disease), and
autoimmune disorders (Colcher and Hurtig 2004). These disorders
are not specifically targeting the neuromuscular structures involved
in movement production but lead to problems with movements.
Some of the disorders described in this chapter used to be linked to
specific groups of neurons. Amyotrophic lateral sclerosis for many
years had been considered a disorder that affected only
motoneurons. However, more recently, problems with other neuronal
formations have been reported. Other disorders, such as
olivopontocerebellar atrophy (OPCA), are associated with
degeneration of the system of climbing fibers and the cerebellum.
More recently, however, this disorder has been viewed as a member
of a larger family associated with atrophy of a number of neuronal
pathways, resulting in the term multisystem atrophy. Other disorders,
such as essential tremor, potentially unite under a single name a
number of underlying problems with different brain structures and
pathways that have a broad variety of clinical presentations.

40.1 Amyotrophic Lateral

Sclerosis
Amyotrophic lateral sclerosis (ALS; reviewed in Bruijn et al. 2004)
is also known as Lou Gehrig’s disease after the baseball great Lou
Gehrig, who died of ALS in 1941. This is a relatively rare disorder
that affects, on average, 1 out of 15,000 Americans, with 5,000 new
cases diagnosed in the United States each year. Men are affected by
ALS more frequently than women. The disease most commonly
strikes people between 40 and 60 yrs of age.
The most prominent consequence of this disease is progressive
death of α-motoneurons, resulting in muscle denervation and loss of
voluntary muscle control. Its earliest symptoms may include
twitching, cramping, or stiffness of muscles; muscle weakness
affecting an arm or a leg; slurred and nasal speech; or difficulty
chewing or swallowing. Patients have increasing problems with
voluntary movements, swallowing (dysphagia), and speaking
(dysarthria). About half of the people affected by ALS develop at
least mild difficulties with thinking and behavior, and most people
experience pain. Ultimately, the disease leads to paralysis that may
affect muscles involved in breathing, which is a common cause of
death, typically after 2 to 4 yrs following the diagnosis. In some
cases, disease progression stops for unknown reasons, and patients
can live for many years without significant worsening of the
symptoms. The best-known example is the great physicist Stephen
Hawking, who had been affected by ALS since his early 20s and
lived to age 76. Hawking preserved a clear mind over all those years
and wrote bestselling books on the history of the universe and black
holes.
More recent studies have provided evidence for neurons in other
parts of the central nervous system being affected by ALS (Silani et
al. 2017; Verde et al. 2017; Zhang et al. 2019). The affected parts
include brain areas involved in the control of movements (such as
the pyramidal tract) as well as other areas involved in the sensory
and cognitive functions, such as the frontal and parietal areas of the
cortex.
About 5% to 10% of all ALS cases are inherited, with about 20%
of all familial cases resulting from a specific genetic defect, a
mutation of the superoxide dismutase 1 (SOD1) gene (Andersen
2004). Most cases, however, have no clear cause and are classified
as sporadic. Unfortunately, there is no effective treatment for ALS.
Available drug therapies prolong the life expectancy only by a few
months. Physical, occupational and speech therapies are commonly
used to alleviate symptoms and make life more comfortable for these
patients.

40.2 Multiple Sclerosis

Multiple sclerosis is a disorder characterized by a loss of the
myelin sheath by the axons of certain neural tracts. Damage to the
myelin sheath along individual fibers varies in severity, resulting in
varying degrees of impairment. The conduction of action potentials
along affected tracts slows down (when the fastest-conducting fibers
lose their myelin sheath), loses its regularity, and may even be totally
interrupted. Saltatory transmission of action potentials (i.e., their
“jumps” from one Ranvier node to the next one) along the severely
affected axons becomes impossible because of the lack of the
myelin sheath. Note that the myelin sheath allows the local currents
to reach the threshold for action potential generation in the next
Ranvier node; this may become impossible if the sheath is damaged.
On the other hand, the number of ion channels is disproportionally
high in the Ranvier nodes and disproportionally low under the myelin
sheath, so action potential transmission typical of thinner axons
without the myelin sheath is also impossible when the myelin sheath
is destroyed.

PROBLEM 40.1
Within a neural tract, thicker axons are more likely to stop
conducting action potentials in multiple sclerosis. Suggest an
explanation.

Epidemiological data on multiple sclerosis suggest that its

prevalence ranges from 10 to 100 per 1,000,000 of population.
There are geographical areas that show relatively high rates of
incidence of this disease, especially in Northern Europe and North
America. Multiple sclerosis is a disorder of young adults; its
incidence peaks between ages 25 and 65 (Martinelli et al. 2004). It is
seen more frequently in females and shows a strong genetic
predisposition: the probability of having multiple sclerosis is 15 times
higher in a sibling or a parent of a patient diagnosed with this
disease.
The mechanism of demyelination involves an autoimmune
process, which either can be triggered by a viral infection or occur
spontaneously within the central nervous system. The autoimmune
disorder involves the patient’s macrophages and mononuclear cells
attacking myelin cells. Another important problem associated with
multiple sclerosis is disruption of the blood–brain barrier, which is
caused either from within the central nervous system or by an
external factor. As a result, macrophages and lymphocytes (in
particular, T cells) start to cross the blood–brain barrier and attack
myelin cells. Whether dysfunction of the blood–brain barrier is the
cause or the consequence of multiple sclerosis is disputed (Golan et
al. 2016; Lassman 2019) because activated T cells can cross a
healthy blood–brain barrier when they express adhesion proteins.
The attack on myelin starts inflammatory processes, which
activate other immune cells and lead to the release of cytokines
(small protein molecules important in cell signaling) and antibodies,
potentially leading to swelling. This may lead to further breakdown of
the blood–brain barrier.
There are four types of multiple sclerosis. The first type, known as
clinically isolated syndrome, corresponds to a clinical attack
suggestive of demyelination, which in a large percentage of cases (>
30%) leads to a diagnosis of multiple sclerosis. The relapsing-
remitting type typically follows the clinically isolated syndrome and,
as its name suggests, consists of episodes of remission and
relapses. Primary progressive multiple sclerosis affects 10% to 20%
of patients and does not show remission after the first symptoms.
The fourth type, most common (about 65% of patients), is secondary
progressive multiple sclerosis, which develops from the relapsing-
remitting type many years after the original diagnosis.
The clinical picture of multiple sclerosis is rather variable and is
dependent on which tracts are affected (Lublin 2004). If the optic
nerve is affected, the disease may start with a sudden onset of
blurred vision, a dull ache in the eye, impaired visual acuity, and
sometimes acute blindness. When olfactory and auditory cranial
nerves are affected, symptoms may include loss of the sense of
smell and unilateral deafness. Demyelination of brainstem pathways
leads to impairment of balance, intentional tremor, discoordination of
limbs, dysarthria, facial weakness and numbness, and unilateral
ophthalmoplegia (weakness of oculomotor muscles). Multiple
sclerosis may also lead to cognitive changes when pathways within
the large hemispheres are affected.
When multiple sclerosis affects major descending tracts that
participate in the control of voluntary movements, such as the
corticospinal tract, the signs and symptoms are similar to those seen
in patients with an incomplete spinal injury at a cervical or thoracic
level. These symptoms involve spasticity and paresis, which are
sometimes associated with changed somatosensory sensitivity in
parts of the body innervated by lower spinal segments. Another
prominent symptom of multiple sclerosis is the unusual sense of
tiredness, which is out of proportion to the degree of daily effort (see
chapter 31). This feeling is frequently called fatigue, although
patients with multiple sclerosis report that it is a different feeling from
ordinary fatigue. Unlike the typical muscle fatigue, fatigue in multiple
sclerosis has a strong central component, which is reflected in
increased responses of fatigued muscles to electrical stimulation
applied directly to the muscles (Latash et al. 1996). As described in
chapter 31, healthy persons show a drop in the muscle response to
stimulation under fatigue.
The prognosis in multiple sclerosis is very uncertain. An episode
of the disease may be followed by 20 yrs of no symptoms, and then
multiple sclerosis may strike again. The average duration of the
disease is 25 to 30 yrs, with the average rate of clinical relapse once
per 2 yrs. Spontaneous improvements are typical, but they may not
last long. Unfortunately, there is no cure for the primary cause—that
is, for demyelination. Therefore, treatment strategies include
addressing specific symptoms of patients, such as spasticity, and
using physical therapy and assisting devices. There have also been
attempts at treating this disease with hormone therapy (e.g.,
corticotropine and prednisolone), modification of diet
(polyunsuturated fatty acids), hyperbaric oxygenation, and
immunosuppressive treatment. Spasticity associated with multiple
sclerosis can be treated with intrathecal baclofen similar to treatment
of spasticity following a spinal cord injury (Penn et al. 1989;
Campbell et al. 1995).

PROBLEM 40.2
Patients with multiple sclerosis frequently feel better in a cold
room and worse in a hot room. Suggest an explanation.

40.3 Multisystem Atrophy

Multisystem atrophy (also known as multiple system atrophy) is a
relatively rare neurodegenerative disorder leading to a combination
of clinical signs and symptoms typical of patients with disorders of
the basal ganglia and of the cerebellum (described in chapters 37
and 38). It affects about 5 per 100,000 of the population and is
slightly more frequent in males (about 55% males, 45% females).
The disease is usually diagnosed in persons 50 to 60 yrs of age
(Shulman et al. 2004).
The disease can lead to the death of neurons in different brain
areas, including the cerebellum, the cortex of the large hemispheres,
the basal ganglia, and the inferior olives. Cases of multisystem
atrophy primarily involving the cerebellum and inferior olives were
known in the past as olivopontocerebellar atrophy (OPCA).
Disorders involving the basal ganglia used to be known as
striatonigral degeneration.
Patients with multisystem atrophy present a mixture of symptoms,
including those typical of Parkinson’s disease (rigidity, tremor,
bradykinesia, and problems with postural stability), cerebellar
disorders (ataxia), alterations of mood and personality, and
autonomic dysfunctions. The latter group involves orthostatic
hypotension, dry mouth, loss of sweating, and problems with the
urinary and sexual functions. In about 60% of patients, the disease
starts with symptoms resembling those of Parkinson’s disease,
including slowness in movement initiation (i.e., bradykinesia). These
symptoms may mask changes in the stretch reflex and other signs of
spasticity associated with damage of the pyramidal tract. Other
patients may initially show problems with balance resembling
cerebellar ataxia accompanied by visible degeneration of the pons
and cerebellum (seen during MRI examination).
Unlike patients with Parkinson’s disease, those with multisystem
atrophy show no consistent effect of dopamine-replacement
medications, such as L-DOPA and other dopamine agonists,
although parkinsonian signs may show improvements. Overall, there
is no effective treatment for multisystem atrophy. Treatment is
symptomatic and involves physical therapy, occupational therapy,
speech therapy, and other measures.
40.4 Essential Tremor
Essential tremor is rather common (reviewed in Cersosimo and
Koller 2004); it is seen in about 1% of young persons and in 2% to
5% of the elderly. The term “essential” means that the tremor is the
only significant problem in those patients. The movement frequency
of essential tremor can range broadly; typically it is between 4 and
12 Hz, suggesting that various mechanisms can lead to this disorder.
Tremor of higher frequency typically shows smaller peak-to-peak
amplitude. The frequency shows a tendency to drop with age, which
is associated with an increase in the tremor amplitude.
Essential tremor is most commonly seen in the hands, and it is
commonly asymmetrical or even unilateral. However, it may be
bilateral and may affect other muscle groups. It can be seen in
postural tasks and can show both an increase and a decrease with
voluntary movement. Chances of showing essential tremor are
higher in persons with another family member who has this disorder.
Short-term factors that lead to an increase in essential tremor
include high temperature, emotion, fatigue, and stimulants of the
central nervous system (such as coffee).
The origin of essential tremor is not known; it is believed to be
generated centrally rather than represent an oscillation in a neural
loop involving peripheral sensory endings. Brain structures
implicated in cases of essential tremor include the pyramidal tracts,
the cerebellum, and the inferior olives. The potential involvement of
the cerebellum is supported by common features of essential tremor
and intentional tremor in patients with cerebellar disorders.
Advanced essential tremor shows gait disturbances similar to those
seen in cerebellar patients.

PROBLEM 40.3
Suggest an experiment that would be able to test a hypothesis
that essential tremor represents an auto-oscillation in a loop
involving sensory receptors.
 Unless the tremor is disabling, it is not treated. In a large
percentage of patients, alcohol produces a dramatic temporary
suppression of the tremor. In cases where essential tremor
significantly interferes with daily activities, treatment strategies range
from beta-adrenergic blockers (e.g., propranolol), benzodiazepines
and other sedative drugs, peripheral agents inducing muscle paresis
(such as botulinum toxin), to anticonvulsant drugs, psychotherapy,
and hypnosis. In severe cases, thalamotomy and thalamic deep-
brain stimulation have been used.

40.5 Cerebral Palsy

Cerebral palsy represents an inborn, nonprogressive disorder seen
in young children. The prevalence of cerebral palsy is about 2 per
1,000 live births. The male-to-female ratio is 1.5 to 1 (reviewed in
Reddihough and Collins 2003). This condition may result from
genetic causes, such as congenital CNS malformations, as well as
complications of gestation, labor, and delivery. In particular, preterm
birth and very low birth weight correlate with an increased chance of
cerebral palsy. The lack of oxygen (hypoxia) during delivery is
another potentially contributing factor to this condition.
Disorders of the motor function in cerebral palsy include
discoordination, spasticity, dystonia, and dysarthria (reviewed in
Gorter et al. 2004). These may or may not be accompanied by
epilepsy, slow mental development, and disturbances of vision.
Spastic diplegia with or without ataxia can be seen in close to 50% of
children with cerebral palsy. Hemiplegias and tetraplegias account
for about 30% of cases.
Consequences of cerebral palsy are exacerbated by the fact that
this condition persists over the crucial periods of development. In
particular, spastic contractions of certain muscle groups can result in
awkward joint postures being occupied most of the time. As a result,
chronic pathological shortening of muscle fibers may occur, leading
to a permanent loss of the full joint range of motion. In such a
condition, even if the primary cause (such as spasticity) has been
successfully treated, joint movement may still be severely impaired.

PROBLEM 40.4
Children with cerebral palsy can sometimes show walking on tip-
toes. How would you distinguish this walk from Huntington’s
chorea?

Treatment of cerebral palsy is directed at both its primary

consequences (such as weakness, discoordination, and spasticity)
and delayed consequences, such as the aforementioned chronic
changes in muscle fiber length. Antispastic medications have shown
limited success in cerebral palsy, although cases of strong effects of
intrathecal baclofen have been described (Latash and Penn 1996).
Figure 40.1 illustrates the effects of intrathecal baclofen on ankle
clonus, while figure 40.2 shows patterns of voluntary foot
movements before (solid traces) and after (dashed traces)
intrathecal baclofen. Overall, baclofen led to a significant
suppression of the clonus, fewer uncontrolled bursts of muscle
activity during voluntary movements, and smoother movement
trajectories at higher speeds. The same can be said about the
effectiveness of physical therapy. Orthopedic procedures are used
relatively often to overcome the unwanted delayed consequences of
the atypical development of children with cerebral palsy.

40.6 Wilson’s Disease

Wilson’s disease is a rare condition characterized by copper
deposits in the brain (cortex and basal ganglia) and also in the liver
and maybe other organs (reviewed in Brewer 2005). It is likely
related to a genetic anomaly reflected, in particular, in changed
metabolism of copper. This disorder can be seen in young persons.
Its clinical signs include an unusual tremor, which is unlike other
involuntary cyclic actions we have considered until now. The tremor
is slow (1 to 2 Hz or even slower) and is commonly seen while a
person with Wilson’s disease tries to maintain a constant posture in
a limb. The tremor shows three features: wing-beat, crescendo, and
spreading. The first feature is related to the peculiar rotational
motions of the hand and arm that resemble a bird beating its wings.
The second refers to a gradual increase in the tremor’s amplitude:
the tremor may start as a relatively small trembling of distal limb
segments (e.g., fingers), and then increase in magnitude until it
leads to a large-amplitude motion of the whole limb. The spreading
feature refers to the tremor that commonly starts at one of the joints
of an arm (for example, the wrist) and then involves other joints of
the arm and maybe even the other arm.
Other clinical features of Wilson’s disease involve the masklike
facial expression, stooped posture, shuffling gait, and slurred
speech. All these features resemble signs of Parkinson’s disease
and suggest that the copper deposits affect signal transmission in
the cortico-basal-thalamo-cortical loop.

PROBLEM 40.5
What effects of cholinergic drugs would you expect in Wilson’s
disease?
Figure 40.1 Effects of intrathecal baclofen on ankle clonus produced by a quick
passive ankle dorsiflexion in a person with cerebral palsy (solid lines = before;
dashed lines = after). Note the smaller bursts of muscle activity on the drug. GM =
gastrocnemius medialis; GL = gastrocnemius lateralis; SOL = soleus; TA = tibialis
anterior.
Reprinted by permission from M.L. Latash and R.D. Penn, “Changes in Voluntary Motor
Control Induced by Intrathecal Baciofen,” Physiotherapy Research International 1 (1996):
229-246. ©1996 Whurr Publishers Ltd.
Figure 40.2 Effects of intrathecal baclofen on movement kinematics and EMG
patterns during voluntary ankle plantarflexion by a person with cerebral palsy (solid
lines = before; dashed lines = after). Note the smoother trajectory and higher
velocity on the drug; the EMG levels are generally reduced. GM = gastrocnemius
medialis; GL = gastrocnemius lateralis; SOL = soleus; TA = tibialis anterior.
Reprinted by permission from M.L. Latash and R.D. Penn, “Changes in Voluntary Motor
Control Induced by Intrathecal Baciofen,” Physiotherapy Research International 1 (1996):
229-246. ©1996 Whurr Publishers Ltd.

CHAPTER 40 IN A NUTSHELL
Systemic disorders affect multiple
anatomical structures and pathways
within the central nervous system.
Amyotrophic lateral sclerosis leads to
progressive death of α-motoneurons,
resulting in muscle denervation and
loss of voluntary muscle control. It
is associated with neuronal loss in
cortical areas involved in the motor,
sensory, and cognitive functions.
Multiple sclerosis is a disorder
characterized by a loss of the myelin
sheath by the axons of neural tracts
within the central nervous system,
leading to various clinical
consequences depending on which tracts
are affected. Its cause is an
autoimmune process that can be
triggered by a viral infection.
Multisystem atrophy leads to a
combination of clinical signs and
symptoms that are seen in patients
with basal ganglia and cerebellar
dysfunctions. Its initial signs
commonly resemble those seen in
Parkinson’s disease or in patients
with cerebellar ataxia. Essential
tremor is a common disorder involving
various brain structures, in
particular the cortex, the cerebellum,
and the inferior olives. Cerebral
palsy is an inborn nonprogressive
disorder caused by problems during
gestation and delivery. Disorders of
the motor function in cerebral palsy
include discoordination, spasticity,
dystonia, and dysarthria. Wilson’s
disease is caused by copper deposits
in the brain (cortex and basal
ganglia). It leads to a combination of
signs resembling those in Parkinson’s
disease and unusual low-frequency
tremor characterized by wing-beating,
crescendo, and spreading.
 Chapter 41

Motor Rehabilitation

KEY TERMS AND TOPICS

“normal movement”
neural plasticity
adaptive changes in motor patterns
functional electrical stimulation
constraint-induced therapy
brain–computer interface

Studies of populations whose ability to perform voluntary movements

is impaired due to natural reasons (for example, aging), genetic
anomaly (such as in Down syndrome), trauma (for example, spinal
cord injury), or illness (for example, Parkinson’s disease), frequently
result in two basic questions: Are observed motor patterns, which
may be rather different from those seen in the general population,
reflecting the inability of the central nervous system to demonstrate
normal patterns? and Should these patterns be corrected? These
questions are important not only for a deeper understanding of the
mechanisms of control of normal and disordered movements but
also to help physicians and researchers assess the effectiveness of
existing therapeutic approaches and to provide a focus within which
the development of new therapies can be considered.
 A common misconception is that any deviation from motor
patterns seen in the general unimpaired population is detrimental for
everyday motor functions. This misconception is revealed in the way
research findings are presented and interpreted, as well as in
prescriptions to correct the “wrong” motor patterns. Let us suggest
an alternative to this view and illustrate it with a few examples
(reviewed in Latash and Anson 1996, 2006).

41.1 Do “Normal Movements”

Exist?
In order to classify a movement as abnormal, one apparently has to
define what normal movement is. There are several reasons to claim
that such a definition does not exist and, moreover, cannot exist.
One of them is motor variability (reviewed in Newell and Corcos
1993; Davids et al. 2006). When a healthy person is asked to
perform the same task several times, motor patterns generated by
individual effectors vary across consecutive attempts. Bernstein
referred to this feature of human movements as “repetition without
repetition” (Bernstein 1947, 1967), implying that solving the same
task repeatedly does not involve repeated motor patterns. Natural
motor variability suggests that even for the simplest tasks performed
by a healthy person, there is no such a thing as a unique, optimal
“normal movement” that solves these tasks.
Another reason normal movement cannot be defined is due to the
interpersonal variability in motor patterns reflecting both the
genetically defined differences among individuals and the differences
in their lifetime experiences. A tall person can reach for a glass
placed on the top shelf of a kitchen cabinet, while a shorter person
may be forced to use a completely different strategy to solve this
motor problem, such as stepping up on a ladder or a stool. A
professional chef slices the onion using very different movements
from those used by a professional mathematician. Several studies
have shown that young, healthy persons, who are used as control
subjects in most movement studies, may or may not show certain
common features of motor performance, including those reflecting
motor synergies (Latash et al. 2001; Olafsdottir et al. 2005a). There
are consistent differences in how young, healthy persons solve tasks
involving large sets of elements that resemble personality traits (de
Freitas et al. 2019). There are also differences in motor patterns
across subgroups within the healthy population that reflect age and
gender.
These examples suggest that, for any everyday motor task, there
is a spectrum of movements that an external observer may or may
not view as corresponding to those most people commonly use.
These movements may range from highly refined and perfect for the
given task to rather clumsy. This classification is obviously going to
differ among observers. Nevertheless, there are common features
among motor patterns produced by most unimpaired persons.
Let us assume that there exists a set of rules used by the central
nervous system of most humans to select particular families of motor
patterns from the available infinite sets to solve everyday motor
problems. Such rules may reflect optimization principles (see chapter
22), such as the minimization of metabolic energy expenditure.
Indeed, energy may be viewed as a “universal currency,” and
extracting energy from the environment for personal goals of the
animal may be viewed as a central feature that differentiates living
beings from inanimate nature, a major factor in the process of
evolution (Schrödinger 1944; Yufik and Friston 2016; Ramstead et
al. 2018). Such rules can also reflect what the moving person
perceives as salient variables for functionally important behaviors
(for example, the trajectory of the index finger during pointing or the
trajectory of the whole body during walking) that have to reach
certain values and be stabilized at those values (reviewed in Latash
2021a,c).
Figure 41.1 The spectrum of typical movement patterns (Average person)
merges at one end with clumsiness and impaired movements and, at the other
end, with perfection and uniquely specified movements. Beyond the spectrum, in
the area that may be considered pathological, CNS priorities are changed,
potentially leading to atypical movement patterns.
Reprinted by permission from M.L. Latash and J.G. Anson, “What Are Normal Movements
in Atypical Populations,” Behavioral and Brain Sciences 19 (1996): 55-106. ©1996
Cambridge University Press.

We will address these unknown rules as priorities of the central

nervous system. These sets of rules are assumed to be common
across all people without highly specialized motor skills and without
any features that would disqualify them from serving as control
subjects in experiments run in typical laboratories that study human
movements.
Figure 41.1 illustrates this idea. Movement patterns produced by
an average person without specialized motor skills represent a
spectrum that merges at one end with clumsiness and impaired
movements and, at the other end, with perfection and uniquely
specialized movements. Priorities of the central nervous system are
assumed to be similar across all the people within the central part of
the spectrum. Clumsy children are at one end of the spectrum, and
at the other end are professional athletes, dancers, musicians,
watchmakers, jewelers, and other uniquely qualified persons across
many fields that require motor perfection.
When one moves beyond these limits into an area that may be
considered pathological or otherwise special, the priorities of the
central nervous system may change and lead to atypical motor
patterns. This may happen in the absence of any gross neurological
or motor pathology—for example, due to changes in understanding
the task and the environmental conditions (as in persons with Down
syndrome or schizophrenia). Further to the left, there is an area
associated with morphological, biochemical, or structural changes
within the central nervous system that may induce differences in
motor patterns by themselves and also by changing the priorities of
the central nervous system, as may occur with Parkinson’s disease
and spinal cord injury. Toward the left end of the axis, we have
placed peripheral changes, as in cases of amputation, which
certainly limit motor patterns by themselves, but may also lead to a
reorganization within the central nervous system and to resultant
changes in its priorities.
A few attempts have been made at deciphering priorities used by
the central nervous system. Most frequently, the researchers try to
guess the internal rules for solutions generated by the central
nervous system by investigating the consequences of optimization
of certain functions of performance (see chapter 22). Attempts at
minimizing or maximizing certain cost functions based on
movement kinematics (such as peak velocity, peak acceleration, or
jerk), movement dynamics (functions of joint torques), energy, or
functions related to such notions as fatigue, comfort, and effort have
not led to a breakthrough in understanding how natural movements
are actually controlled. Some of these approaches have
demonstrated an impressive correspondence to the actual
movement kinematics observed in experiments. This fit by itself does
not mean, however, that the intact central nervous system is
minimizing a function of jerk or joint torque, or calculating a “comfort
function,” or doing something else of the kind. It rather suggests that
the solutions preferred by the central nervous system do not violate
any of these principles too much. Thus, the actual priorities remain
unknown.

PROBLEM 41.1
Minimization of energy spending has been viewed as a general
principle underlying motor coordination. Present examples of
behaviors when this is likely to be true and when it is likely to be
false.

41.2 Changes in CNS Priorities

Let us consider situations of possible changes in priorities of the
central nervous system in more detail. As already mentioned, the
existence of choice at each level of the control hierarchy
(theoretically, at least) suggests that the existing priorities may be
modified in certain situations wherein the components of the system
for movement production are grossly changed, or the task is highly
unusual, or it is performed in unusual external conditions. A change
in the priorities may lead to a corresponding change in the externally
observed patterns of voluntary movements. For example, the
Fosbury flop is apparently not a coordinative pattern the central
nervous system prefers to use for jumping on an everyday basis. In
the artificial conditions of track-and-field competition, however, when
there are no unexpected changes in the external force field, no
hidden obstacles, and there is just one goal (to clear the bar at the
greatest height possible), the central nervous system may switch to
a new, quite unusual pattern of coordination. Thus, by changing the
external conditions of movement execution (the context of a motor
task in a broad sense) in combination with extensive practice, one
may be able to alter the priorities used by the healthy central
nervous system and observe movement patterns that are quite
different from those commonly seen in the general population;
compare, for example, the obviously different walking patterns in
ballet dancers with those in sumo wrestlers.
 Changes in the priorities of the central nervous system are likely
to occur during the early stages of human life. These developmental
jumps seem to follow the discovery by the child’s brain of the
biomechanics of its own effectors and the basic physical properties
of the external force fields (reviewed in Thelen 1995; Savelsbergh et
al. 2006). For example, crawling may be considered a temporary
solution for the problem of locomotion by babies whose system of
balance control is immature and does not allow adult bipedal
walking. Later, new solutions are discovered by the central nervous
system, and crawling is substituted with walking and running, with a
crucial role being played by environmental factors.
Consider the system of motor control of a chronically impaired or
otherwise atypical person. His or her lifetime experience is filled with
everyday voluntary movements in conditions of frequently changing
goals and external forces. If the differences between this person and
an average “healthy control subject” are large enough, there is a fair
chance that his or her central nervous system will reconsider its
priorities and elaborate, for everyday motor tasks, movement
patterns that will look different from those observed in the majority of
healthy persons. These atypical patterns can, however, be optimal
for the state of the body and tasks perceived as important.
This certainly does not mean that adaptive changes in the central
nervous system are the only important factor defining abnormal
motor patterns. An impaired system may well be genuinely unable to
display movement patterns seen in the general population, as for
example after a limb amputation or after a complete spinal cord
injury. However, let us focus on possible changes in movement
patterns that are not forced upon the system by a major chronic
impairment but instead result from a person’s reaction to a primary
impairment.
A car with a broken engine does not reconsider its priorities and
does not switch to alternative strategies. The ability of the human
body to adapt to pathological changes is a reflection of the basic
differences between the designs of the car and of the human body.
First, the design of the car does not involve redundancy, so there is
little room for such things as priorities, choice, and strategies.
Second, the car does not have a brain. The brain plus redundancy
make the design of the human body, including the system for the
production of voluntary movements, flexible and able to adapt not
only to changes in external conditions but also, at least to some
extent, to changes within the body itself.

41.3 Neural Plasticity

Adaptive reorganization of the central nervous system is likely to
result from the ability of neural circuits to display plasticity, including
changes in the efficacy of neural projections. Neural plasticity
represents one of the most remarkable neural phenomena and is
likely to contribute to the processes of motor development, motor
learning in healthy persons, and adaptation to trauma. Since the
classic works of Lashley in the 1930s (reviewed in Bruce 2001), it
has been proposed that an injury to the brain cortex may lead to a
dramatic topographic reorganization in adjacent cortical areas which
may, in particular, significantly contribute to recovery after stroke
(reviewed in Nudo 2003; Cauraugh 2004; Ward 2005). Changes in
peripheral afferent flow have been shown to induce changes in the
receptor field sizes and locations in the brain cortex of the cat.
Changes in somatosensory cortical representations in monkeys have
been shown after a specific training of one hand, and after digit
amputation or fusion (figure 41.2). Neural plasticity is not limited to
grossly changed pathological states and supraspinal structures. As
shown in classic studies by the group of Jonathan Wolpaw (chapter
34), changes in the H-reflex excitability can be produced by operant
conditioning in monkeys, suggesting that plastic changes can be
induced within the spinal cord neural machinery.
Figure 41.2 Changes in the somatosensory cortical representation in a monkey
after amputation of one of the digits (digit 3). Note that after the amputation,
representations of digits 2 and 4 expanded and occupied the area that previously
represented digit 3.
Reprinted by permission from M.M. Mersenich, et al., “Somatosensory Cortical Map
Changes Following Digit Amputation in Adult Monkeys,” The Journal of Comparative
Neurology 224 (1984):591-605. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of
John Wiley & Sons.

Plastic changes have been demonstrated in healthy persons after

long-lasting specialized training, such as reading Braille and playing
musical instruments (Pascual-Leone et al. 1995; Cohen et al. 1997;
Sterr et al. 1998; Pascual-Leone 2001). Such changes, however, can
also be seen following a relatively brief practice limited to 1 or 2 h
(Classen et al. 1998; Latash et al. 2003). This means that the human
central nervous system is always in the process of rewiring itself.
Effects of these processes can even be seen within a few seconds—
for example, as reflected in indices of finger interdependence
(enslaving; see chapter 27), which show an increase during steady
force production by a subset of fingers of a hand (Abolins et al. 2020;
Hirose et al. 2020).

PROBLEM 41.2
Plastic changes following training commonly lead to higher
excitability of involved cortical areas. Suggest a task when
practice can be expected to lead to decreased excitability of
involved cortical areas.

Many studies have demonstrated changes in neuronal projections

following a major injury to the central nervous system. For example,
a unilateral cortical stroke has been shown to lead to changes in
both interhemispheric projections and descending projections to the
spinal cord (Netz et al. 1997; Carmichael 2003; Celnik and Cohen
2004). These changes have been viewed as important factors that
promote adaptation within the injured organism and enable it to
perform better in a variety of everyday tasks.
Figure 41.3 A scheme of possible strategies to influence hand function after a
unilateral stroke. The “up” arrows show upregulation of neural signals and
processes in the affected hemisphere (arrow-1) and paretic hand (arrow-2). The
“down” arrows show downregulations of signals and processes in the unaffected
hemisphere (arrow-3) and unaffected hand (arrow-4).

In particular, stroke has been shown to lead to an increase in

interhemispheric inhibitory projections from the nonaffected
hemisphere to the affected hemisphere (Duque et al. 2005). These
changes may interfere with recovery of the residual function of the
body parts controlled by the affected hemisphere. Several strategies
have been suggested to deal with this problem, which is illustrated in
figure 41.3. They involve promoting plasticity in the affected
hemisphere with transcranial magnetic stimulation (TMS, the “up”
arrow-1 in figure 41.3; Butefisch et al. 2004), increasing the
somatosensory signals from the affected hand (the “up” arrow 2 in
figure 41.3; Conforto et al. 2002), downregulating activity in the
nonaffected motor cortex, also using TMS but with different
parameters (the “down” arrow-3 in figure 41.3; Schambra et al.
2003), and decreasing the somatosensory input from the
nonaffected hand (the “down” arrow-4 in figure 41.3; Werhahn et al.
2002).

41.4 Adaptive Changes in Motor

Patterns
Any major long-lasting difference between impaired and unimpaired
groups of human beings (for example, changes in many body
characteristics in Down syndrome, changed biomechanics and
afferent sources in amputees, changed reflexes and possibly other
mechanisms in spasticity, changed postural control in Parkinson’s
disease, and so on) makes the whole system of movement
generation different so that its priorities are likely to be reconsidered,
and control patterns that used to be optimal for an unimpaired
system are no longer optimal.
For any apparently abnormal motor pattern, the first question to
ask is: What does the brain perceive to be its primary goal during the
execution of this particular motor task? A straightforward answer,
such as to follow the exact instruction given by the experimenter,
may be true for motivated, healthy subjects, although, even in
healthy subjects, considerations such as minimizing discomfort and
pleasing the experimenter may be as important as optimizing the
performance. Other, frequently ignored components of a motor task
that may be considered important by the brain are, for example,
those related to maintaining gaze fixation, equilibrium of the head
and body, and posture of the limbs with respect to the trunk during
the required movements.
In impaired subjects, we may expect less obvious factors to play
an important role in making a choice of movement strategy. For
example, recall changes in the shuffling gait in patients with
Parkinson’s disease when they are asked to step over lines drawn
on the floor (Bernstein 1947; Morris et al. 2001). The observations of
the improved gait are very convincing. They were even viewed as
arguments that patients with Parkinson’s disease should be trained
in a physical therapy environment with lines drawn on the floor to
promote more normal-looking gait. However, as mentioned earlier,
two factors make one cautious with respect to such a
recommendation. First, the shuffling gait in Parkinson’s disease may
be a consequence of disordered postural control and the desire of
the patient to avoid large destabilizing contact forces during
stepping. Second, outside the laboratory, there are no lines drawn on
the pavement. A shuffling gait may not look “normal” (which is a
misnomer to begin with), but it may well be optimal for the group of
everyday tasks the patient faces and the state of his or her system
for the control of vertical posture.

PROBLEM 41.3
Suggest a method of helping patients with Parkinson’s disease to
overcome episodes of freezing during gait based on the “lines
drawn on the floor” idea but useful for everyday walking tasks.

Let us now consider in more detail one more example of changed

movements when some of the apparent abnormalities in motor
patterns are likely to be consequences of adaptive changes within
the central nervous system (also see the discussion of motor
disorders in spinal cord injury in chapter 36 and in Parkinson’s
disease in chapter 37).

41.5 Consequences of Amputation

With limb amputation, the primary cause of the apparent motor
disorders is unambiguously clear. Limb amputation leads to major
changes in the biomechanical and neurophysiological relations
developed during the person’s lifetime. There is evidence, however,
that the consequences of limb amputation may also involve a major
reorganization of both afferent (sensory) and efferent (motor)
projections that, by themselves, may contribute to the difference of
the motor patterns from those seen in unimpaired persons.
 Considerable changes in the biomechanics of walking occur after
a below-the-knee leg amputation. In healthy subjects, ankle plantar
flexors are the major energy generators. The role of hip extensors in
energy generation is relatively small. In below-the-knee amputees,
ankle plantar flexors are obviously unavailable, and hip extensors
become the main source of energy absorption and generation when
walking with a prosthesis (Czerniecki et al. 1991). This
rearrangement should be considered adaptive since it allows
amputees to walk with a prosthesis even though the gait may be less
energetically efficient.
The consequences of amputation also involve neural
reorganization at both the segmental and suprasegmental levels
(Cohen et al. 1991a,b; Fuhr et al. 1992). Obviously, the elimination of
a considerable number of proprioceptors residing in the amputated
portion of the leg leads to an abrupt change in the patterns of
afferent inflow and is likely to lead to changes in the relative weight
of the contribution of other, seemingly unaffected reflex projections.
Remember that reflex contribution is an important factor in natural
patterns of voluntary movements. Descending motor commands
should apparently take into account the existing state of reflex
connections. Besides that, proprioceptive inflow is used in the
process of generating short-latency, preprogrammed adjustments in
the activity of muscles that provide stability during voluntary
movements. Thus, amputation of a distal portion of a leg may be
expected to lead to a rearrangement of descending motor
commands and a shift of postural control from predominantly
proprioception-based to other modalities, such as visual and
vestibular signals.
Neurological reorganization of descending control signals after
below-the-knee amputation in humans was studied with TMS. In
those studies, stimuli at optimal positions of the coil recruited a larger
percentage of α-motoneurons controlling the muscles in the residual
leg (Fuhr et al. 1992). These muscles could also be activated from
larger areas of the scalp than the muscles at the intact side. Similar
results, also in human subjects, have been reported in a person with
congenital absence of the distal part of the left arm (Cohen et al.
1991b). Thus, descending corticospinal projections are likely to be
reorganized after an amputation.

41.6 Functional Electrical

Stimulation
Functional electrical stimulation (FES) is a method that uses the
physiological properties of unimpaired muscles to produce
movements in the absence of or in case of a major impairment in the
neural control of those muscles. The method involves direct
electrical stimulation applied to the muscles or to peripheral nerves
innervating those muscles. The strength and patterns of the
stimulation can be programmed by a computer, or the patient can be
allowed to trigger the stimulation by movement of unimpaired
effectors. Most common applications of FES have been in patients
after spinal cord injury or stroke, although applications to other
patient groups, including those with cerebral palsy and multiple
sclerosis, have also been developed.
For example, patients who cannot extend their fingers voluntarily
lose much of their hand function because grasping objects requires
opening the hand first. It is possible to trigger FES of the finger
extensor (extensor digitorum communis) by signals from the
relatively spared proximal muscles of the arm (Aoyagi and
Tsubahara 2004; Meilink et al. 2008). Another successful application
of FES is in patients with foot drop, which interferes with normal
locomotion (Burridge et al. 2008; Stein et al. 2010). FES can be
applied to the tibialis anterior muscle or its nerve and timed to the
swing phase during walking when ankle dorsiflexion happens in
healthy persons. Both examples target a local but vital component of
functionally important actions.
Attempts to use FES for the production of multijoint coordinated
movements in completely paralyzed persons have been marginally
successful. This should not be very surprising because, in healthy
persons, the central nervous system cannot prescribe muscle
activation patterns: they happen with important contributions from
reflex pathways that reflect the actual movement mechanics, which
are never 100% predictable (see chapters 21 and 24). The method
is, however, promising for the recovery of motor function in patients
with partially spared neural control of movements (Duffel and
Donaldson 2020; Springer and Khamis 2017).

41.7 Constraint-Induced and

Discomfort-Induced Therapies
As described in chapter 39, cortical stroke is a relatively common
disorder leading to major impairments in the sensory-motor function.
Many of the stroke survivors show spontaneous recovery within the
first weeks following the stroke, and then they show relatively minor
and slow improvements in functional motor activities years after the
stroke. Optimizing recovery over multiple years is a major goal of
physical and occupational post-stroke therapy.
Practically all studies of motor rehabilitation following stroke have
shown that the most important factor predicting recovery is the
amount of time spent by the patient performing prescribed exercise.
This factor seems to override the effects of different types of therapy.
There are several factors, both practical and financial, that limit the
amount of time a patient can spend as an inpatient in a rehabilitation
facility or as an outpatient receiving physical and occupational
therapy. Trying to maximize useful exercise of the affected effectors
during everyday activities led to the emergence of a number of
approaches to motor rehabilitation.
Constraint-induced therapy was designed to improve the recovery
of the affected hands and arms by maximizing involvement of the
contralesional (more impaired) arm in everyday movements (Fritz et
al. 2012; Kwakkel et al. 2015). This approach involves restricting the
contribution of the less affected arm and hand by placing a mitten on
that hand or splinting or bracing the arm. The mitten makes everyday
prehension movements difficult, and the patients are forced to use
their more affected hand for object manipulation, leading to an
increase of the exercise time, which promises quicker and larger
positive effects, including neural plasticity.
Somewhat similar approaches have been developed to optimize
recovery of the lower extremities. Stroke survivors tend to shift their
body weight to the less-affected leg, and their more-affected leg is
used as a kind of cane, which limits its involvement in locomotion
and potentially slows motor recovery. Placing a custom-made insole
into the shoe on the less-affected foot leads to a shift of the body
weight toward the more affected foot, which leads to larger
involvement of that lower extremity during locomotion that can
potentially accelerate recovery (Aruin et al. 2012; Aruin and Kanekar
2013). Three types of insoles have been used: lifting the less-
impaired foot, placing a wedge-shaped insole into the shoe for the
less-impaired foot, and putting an insole with sharp protrusions into
the shoe for the less-impaired foot. The last approach has been
known as discomfort-induced therapy (Aruin and Rao 2018). Placing
much of the body weight on the shoe with such an insole is
uncomfortable or even painful. As a result, the person shifts the body
weight to the side without the insole, leading to more exercise and
better recovery of the more affected leg.

41.8 Brain–Computer Interface

In a number of pathological conditions, the brain may be relatively
unaffected (although there may be secondary plastic changes!), but
its activity cannot lead to the production of movements. This can
happen in cases of spinal cord injury (chapter 36) and in some cases
of systemic disorders, such as amyotrophic lateral sclerosis (chapter
40). In the latter case, the patient may lose the ability to perform
movements altogether, leading to the “locked-in” syndrome
(Birbaumer 2006; Birbaumer and Cohen 2007; Chaudhary et al.
2016). The goal of research attempting to develop and improve
brain–computer interfaces is to decipher the neuronal activity in the
brain in order to understand the desires of the patient and to drive
robotic devices to perform desired actions.

PROBLEM 41.4
Suggest a method of communication with a “locked-in” person
who cannot perform any voluntary movements.

As described in chapter 8, neuronal populations in certain cortical

areas encode the mechanical characteristics of desired movements.
So, if a sufficient number of neurons are recorded, their activation
patterns can be used to decipher where the person wants to move.
Studies with chronically implanted arrays of microelectrodes in
monkeys have shown that this approach is feasible (Helms Tillery et
al. 2003; Wahnoun et al. 2006; Rajangam et al. 2016). The monkeys
were trained to use cortical neurons to control a variety of actions,
from using a robotic device to bring morsels of food to the mouth to
navigation in the environment.
Recent studies have confirmed that using chronically implanted
electrodes in the human brain is a feasible method to facilitate the
neural control of external robotic devices by humans with paralysis
(Lebedev and Nicolelis 2006; Wolpaw 2007; Chaudhury et al. 2016;
Bockbrader et al. 2018). This is an up-and-coming method that faces
many technical challenges and requires a better physiological
understanding and integration with theories of the neural control of
movement.

41.9 Practical Considerations

The purpose of this chapter has been to stress the potential role of
adaptive changes within the central nervous system. In particular,
practitioners should be warned against jumping to quick conclusions
about the inability of a patient’s central nervous system to produce
“correct movements” based on observations of atypical peripheral
motor patterns. The notion of normality with respect to peripheral
motor patterns should be treated very cautiously, possibly as another
commonly used misnomer.
Let us go through the most important practical conclusions based
on this and previous chapters:
Plastic changes within the central nervous system can play an
important role in shaping the patient’s behavior (motor patterns).
Therapy should be directed at optimizing functional behavior, not
movement patterns (this can be called a pragmatic approach).
Therapists should take advantage of the adaptive abilities of the
central nervous system; in other words, they should identify the
goals, provide the tools, and allow the central nervous system to
find a solution.
An important role in rehabilitation should be played by
procedures stimulating learning processes (and, therefore,
promoting plastic changes within the central nervous system).
For example, practicing a task in conditions of uncertainty may
be a promising method.
Correcting the primary cause of a disorder must certainly be the
first priority; unfortunately, this can rarely be done.
Physical and occupational therapists have important advantages
over the central nervous system of a patient. These include an ability
to predict long-term outcomes and an ability to understand that
exercise through discomfort or even pain may be necessary to
achieve a functional optimum. If the central nervous system is
allowed to generate adaptive patterns without any supervision, it
may have a tendency to settle down in a local optimum with respect
to a function because any exploratory activity leads away from this
optimum—that is, to a deterioration of the function. It might never
discover that there is a much more global optimum just behind a
nearby ridge. Pain is another factor that the central nervous system
does not like to experience. Thus, exercise through pain or through
temporary functional deterioration may never be discovered by the
central nervous system, but it may be prescribed by a therapist,
leading to an optimization of long-term functional goals.
CHAPTER 41 IN A NUTSHELL
Motor abundance during human voluntary
movements and plasticity within the
central nervous system allow for
adaptive changes in movement patterns
in persons with a variety of
disorders. Movement patterns in
atypical persons may differ from those
seen in healthy people, but may still
be optimal for the given state of the
system for movement production.
Adaptive changes within the central
nervous system may play an important
role, even after a limb amputation.
Functional electrical stimulation
attempts to produce muscle activations
timed to generate functionally useful
components of movements. Constraint-
induced and discomfort-induced
therapies try to maximize the
involvement of affected extremities in
everyday functional actions to promote
neural plasticity and recovery. Recent
studies have shown that using
chronically implanted electrodes in
the human brain (brain–computer
interface) is a feasible method to
facilitate the neural control of
external robotic devices by humans
with paralysis. Rehabilitation
strategies should be directed at
functional optimization rather than at
bringing movement patterns as close to
“normal” as possible.
Problems for Part VIII
Self-Test Problems
1. A person with Parkinson’s disease is picking up a heavy bag
with the right hand. Describe all the mechanisms that help the
person to maintain balance (their typical time delays, efficacy,
etc.). How will each of these mechanisms differ from those in
a healthy person in a similar situation?
2. A person after a relatively mild cortical stroke eats only half of
the portion placed on her plate. What part of the cortex is
likely affected? What movement disorders can be expected
during standing and walking?
3. A healthy person and a person with early-stage Parkinson’s
disease are asked to start walking as quickly as possible
following a sound signal. What differences would you expect
to observe in their sway (rambling and trembling) during quiet
standing before the signal, postural adjustments prior to the
first step (describe those), and gait characteristics?
4. A person with a unilateral atrophy of the olivo-ponto-cerebellar
pathway tries to draw a triangle as quickly as possible with the
affected hand. Describe all the differences of his movements
from those performed by a healthy age-matched person. Draw
the speed profile of the pencil trajectory.
5. A standing person with cerebellar ataxia is unexpectedly
pushed from behind by another person (not too strongly!).
Describe all the differences in posture-stabilizing responses
as compared to those seen in a healthy person.
6. A person with Parkinson’s disease has been taking dopamine-
replacement medications for many years. Now she complains
of sudden involuntary limb movements. What can be done to
alleviate the problem?

For Those Addicted to Multiple-Choice Tests

You have 20 minutes. Circle only one answer (statement) for each
question. Write a short phrase explaining why you chose this
answer.
1. A young boy complains about tingling sensations followed by
an urge to make a jerky head movement. What else would
you expect to see in this boy? Write the likely diagnosis.
a. clonus
b. increased monosynaptic reflexes in the leg muscles
c. ability to suppress the jerky movements by voluntary effort
d. dancelike gait
e. all of the above
Why?
2. A person shows brief, uncontrolled muscle jerks, while
sometimes postural muscle activity briefly disappears. What
would you recommend? Write the likely diagnosis.
a. stimulation of the external part of the globus pallidus
b. oral baclofen
c. botulinum toxin injection in the postural muscles
d. two shots of vodka
e. none of the above
Why?
3. A person shows uncontrolled clumsy head rotation to the left
leading to a sustained twisted posture. Which of the following
may also be observed? Write her diagnosis.
a. alleviation of the symptoms with sensory stimuli
b. sensory tics preceding motor tics
c. problems with movement adaptation to novel conditions
d. tremor at about 8 Hz and hypotonia
e. decreased anticipatory postural adjustments
Why?
4. A patient with an increased activity of one of the subthalamic
nuclei is likely to show
a. poverty of voluntary movements, particularly in
contralateral limbs
b. higher reflexes in the limbs on the same side of the body
 c. poorly controlled large-amplitude movements in one side of
the body
d. increased anticipatory postural adjustments
e. uncontrolled spasms preceded by tingling sensations
Why?
5. A 20-year-old person complains of involuntary shaking
movement of the hand that starts by itself while the person sits
quietly and then increases in magnitude and spreads to other
parts of the body. What is likely to be the problem?
a. myoclonus
b. Parkinson’s disease
c. copper accumulation in the basal ganglia
d. essential tremor
e. a cerebellar disorder
Why?
 Glossary
acetylcholine—A neurotransmitter and the mediator of
neuromuscular excitation.
actin—A molecule that is a major force-generating element within a
skeletal muscle.
action potential—A standard, brief pattern of changes in the
membrane potential that acts a unit of information transmission
within and among excitable tissues.
action tremor—A tremor appearing when a person tries to perform
a voluntary movement.
afferent fiber—An axon that transmits signals from a more
peripheral structure to a more central structure; commonly, a
sensory fiber, which is the peripheral axon branch of a
proprioceptor neuron.
agonist—A muscle whose activation leads to a required motor
effect.
akinesia—A movement disorder characterized by poverty of
movements.
all-or-none law—A law describing the generation of a standard,
nongraded response to a stimulus exceeding a certain threshold.
α-motoneuron—A neuron that innervates power-producing,
extrafusal muscle fibers.
α-γ coactivation—The simultaneous activation of α- and γ-
motoneurons during voluntary muscle activation.
amnesia—The loss of memories.
ampulla—A dilated area of a semicircular duct where vestibular
receptor cells (hair cells) are located; this area is innervated by
the ampullary nerve.
amyotrophic lateral sclerosis (ALS, Lou Gehrig’s disease)—A
disease leading to the progressive death of α-motoneurons and
resulting in muscle denervation and loss of voluntary muscle
control.
angiography—A method of visualizing the state of blood vessels by
injecting an X-ray opaque substance (a contrast) into the
circulatory system.
ankle strategy—A pattern of responses in postural muscles to an
external perturbation; this pattern shows the most pronounced
motion in the ankle joint.
antagonist—A muscle whose activation apparently counteracts a
required motor effect.
anticipatory postural adjustments (APAs)—Changes in the
activity of postural muscles seen before a self-generated postural
perturbation (e.g., before a fast voluntary movement).
antidromic conduction—Conduction of action potentials along the
axon to the soma.
aphasia—A disorder of spoken language.
apraxia—A disorder of simple motor skills.
articular receptors—Receptor endings located in and around the
joint capsule; these receptors are sensitive to joint angle (typically,
angles close to the anatomical limits of joint motion) and to
tension of the joint capsule.
ascending tracts—Neural tracts carrying information from the
peripheral receptors and from the spinal cord to the brain
(spinothalamic tract, spinocerebellar tracts, spinoreticular tract,
spinovestibular tract, and spinotectal tract).
Ashworth scale—The most commonly used clinical scale for
quantitatively assessing spasticity, the Ashworth scale reflects the
degree of muscle resistance to passive limb movements.
Asperger’s syndrome—A neurobiological disorder characterized by
normal intelligence and language development in combination
with behaviors resembling autism and marked deficiencies in
social and communication skills.
associative learning—One of the subtypes of nondeclarative
memories, associative learning involves creating a relationship
between two stimuli.
ataxia—Decomposition of movement into a number of jerky
segments.
ataxia telangiectasia—A genetic, neurodegenerative disorder
leading to progressive ataxia, dysarthria, facial hypotonia, and
oculomotor abnormalities.
autism—A spectrum disorder characterized by resistance to
change, distress for unclear reasons, difficulty in social interaction,
lack of responsiveness to words, and difficulty in verbal
expression.
axon—The output fiber of a neuron; commonly, the longest fiber
originating from the axon hillock.
axon hillock—An area where the axon exits the soma; this area is
characterized by the increased ability to generate action
potentials.
Babinski reflex—A response in leg muscles to a tactile stimulation
of the sole of the foot.
baclofen—A drug used to treat spasticity; an agonist of gamma-
aminobutyric acid (GABA).
ballism—A disorder of the basal ganglia characterized by fast,
large-amplitude, irregular, involuntary movements; associated with
an injury of the subthalamic nucleus.
basal ganglia—Several paired structures within the brain; these
structures play an important role in voluntary movement
generation and coordination.
Becker muscular dystrophy—A muscular dystrophy similar to
Duchenne muscular dystrophy but with clinical symptoms
appearing later, at adolescence, and with slower disease
progression.
Bernstein problem (redundancy problem)—An ill-posed problem
of choice; a problem of how the central nervous system chooses a
pattern of variables at a certain level of analysis based on a
required summed effect at a higher level of analysis.
blepharospasm—Commonly a sign of dystonia, blepharospasm
affects the upper face and leads to intermittent or sustained
 bilateral eyelid closure as a result of involuntary contraction.
blind spot—An area on the retina where the optic nerve exits the
eye; this area lacks photoreceptors.
blood–brain barrier (hematoencephalic barrier)—A membrane
between the cerebrospinal fluid and the blood, this barrier helps
maintain the homeostasis of the central nervous system.
bradykinesia (akinesia)—Slowness of movements that is typical of
Parkinson’s disease.
carpal tunnel syndrome—A mononeuropathy caused by
entrapment of the median nerve in the carpal tunnel.
catch phenomenon—A transient increase in muscle response to a
standard neural signal following a brief episode of a strong muscle
contraction.
caudate nucleus—A neural structure that is part of the basal
ganglia.
center of pressure (COP)—The point where the reactive force
acting from the support surface is applied to a standing person.
central pattern generator (CPG)—A hypothetical neural structure
that can generate a rhythmic neural activity that is later
transformed into a rhythmic muscle activity leading to a rhythmic
behavior such as locomotion.
cerebellar nuclei (dentate, fastigial, and interposed)—Brain
structures mediating the cerebellar output.
cerebellar peduncles—Six neural tracts connecting the cerebellum
with the rest of the central nervous system.
cerebellar tremor—A low-frequency tremor (3-5 Hz) seen in
patients with cerebellar disorders.
cerebellum—A large brain structure that lies just behind the medulla
and the pons.
cerebral palsy—An inborn, nonprogressive disorder seen in young
children.
cerebrospinal fluid—A fluid that plays an important role in the
diffusion-based exchange of substances of the central nervous
system; it fills the ventricles and the central canal.
cerebrospinal tract—One of the major descending tracts that
participate in controlling voluntary movements.
chorea (Huntington’s disease)—A disorder of the basal ganglia
characterized by excessive, irregular, involuntary movements and
a dancelike gait.
chronic fatigue syndrome—A sustained feeling of exhaustion and
inability to participate in any kind of activity involving even minimal
motor effort.
ciliary ganglion—A ganglion containing neurons that innervate the
smooth muscle of the pupillary sphincter.
clasp-knife phenomenon—A sudden drop in joint resistance to
passive motion; typical of spasticity.
clonus—A series of alternating bursts of activity in the flexor and
extensor muscles of a joint at a frequency of 6 to 8 Hz; clonus
may be induced by a quick joint movement.
coactivation—Simultaneous activation of agonist and antagonist
muscles acting at a joint.
colliculi (superior and inferior)—A structure in the midbrain that
helps process visual and auditory information.
complex spike—An unusual action potential generated by the
Purkinje cells in the cerebellum.
complex system—A system whose properties cannot be derived
from the properties of its elements.
computer tomography (CT)—A method of reconstructing three-
dimensional images of tissues from a series of two-dimensional
images.
concentric (contraction)—A muscle contraction that decreases
muscle length.
conditioned reflex—A reflex that develops to a new stimulus
because the new stimulus and an old stimulus associated with the
reflex were presented simultaneously.
cone—A photoreceptor; cones mediate color vision and provide
better spatial and temporal resolution than rods.
consolidation—Transferring information from the short-term
memory into the long-term memory.
convection—Movement of a solvent (e.g., water) and dissolved
particles due to differences in hydrostatic pressure.
convergence—When signals from neurons distributed over a
relatively wide cortical territory lead to responses in the same
muscle and move the same effector.
corpus callosum—A major neural tract connecting two large
cortical hemispheres.
corrective postural reactions—Preprogrammed reactions
occurring in postural muscles in response to external
perturbations.
corrective stumbling reaction—A reflex-like response to a
mechanical or electrical stimulation of a paw; this reaction
consists of a coordinated limb movement that leads to stepping
over the fictitious obstacle.
cortex (cerebral, cerebellar)—The external, thin layer of the brain;
this layer is densely packed with neuron bodies.
cost function—A rather arbitrary function introduced to solve a
problem of redundancy (Bernstein problem).
cranial nerves—Twelve paired nerves that originate from nuclei in
the supraspinal structures and play a major role in controlling the
head and neck muscles and autonomic functions.
Creutzfeldt-Jakob disease—A rare, fatal, transmissible
encephalopathy that leads to rapidly progressive dementia and
myoclonus.
cross-bridge—A molecular connection between an actin molecule
and a myosin molecule; the cross-bridge generates force during
muscle contractions.
crossed extensor reflex—A reflex contraction of major extensor
muscles in a contralateral limb in response to a stimulus that
induces a flexor reflex in the ipsilateral limb.
cutaneous receptors—Receptor endings that are sensitive to skin
displacement, pressure, temperature, and other stimuli; these
 receptors include Meissner corpuscles, Merkel disks, Ruffini
endings, and Pacinian corpuscles.
cytoarchitecture—The study of the structural arrangement of
neurons within the central nervous system.
degree of freedom—An independent variable describing a multi-
element system at a selected level.
dendrite—A relatively short fiber connected to the neuron body;
commonly serves as a site of input signals to the neuron.
denial—A state in which a person views their own limbs as
belonging to someone else; denial may be a consequence of
stroke.
depolarization—A decrease in the absolute value of the negative
membrane potential.
descending tracts—Neural tracts carrying information from the
brain to the spinal cord (corticospinal tract, corticobulbar tract,
pyramidal tract, rubrospinal tract, vestibulospinal tract,
reticulospinal tracts, and tectospinal tract).
developmental coordination disorder (DCD)—A childhood
disorder characterized by poor coordination and clumsiness.
diencephalon—A major brain structure that is almost completely
surrounded by the cerebral hemispheres.
diffusion—Movement of dissolved particles due to a difference in
particle concentrations.
divergence—When stimulating a single cortical neuron leads to
motor effects in many muscles of an extremity.
dopamine—A neurotransmitter whose deficit leads to Parkinson’s
disease.
dorsal root (spinal)—A set of neural fibers carrying peripheral
information into the spinal cord.
Down syndrome—A chromosomal disorder that affects both mental
and physical development and that is linked to trisomy of
chromosome 21.
dual-strategy hypothesis—A hypothesis that assumes that two
strategies exist during voluntary movements: with and without
 explicit or implicit control over movement time.
Duchenne muscular dystrophy—A muscular dystrophy resulting
from a mutation in the gene that regulates dystrophin, a protein
involved in maintaining the integrity of muscle fiber.
dynamic pattern generation (dynamic systems approach)—A
mathematical modeling approach that uses nonlinear differential
equations to describe the behavior of a complex system.
dysdiadochokinesia—An inability to maintain a constant rhythm
during alternating repetitive movements; typical of cerebellar
disorders.
dysmetria (hypometria, hypermetria)—An inability to achieve a
required final position, with undershooting or overshooting
predominating.
dysprosody—Unusual voice inflections.
dystonia—A complex neural disorder characterized by involuntary
movements with pronounced rotational components.
eccentric contraction—A muscle contraction that increases muscle
length (under the influence of external forces).
efferent fiber—An axon transmitting signals from a more central
structure to a more peripheral structure; commonly, an axon of a
motoneuron.
efferent copy (efference copy)—A hypothetical copy of motor
command signals participating in kinesthetic perception.
elastic element—A mechanical element that deforms under the
influence of an external force, generates force against the
deformation, and can store and release potential energy of the
deformation.
electroencephalography (EEG)—A method for registering waves of
brain activity by placing electrodes over the skull.
electrolyte—A fragment of a molecule with a nonzero total electric
charge.
electromyography (EMG)—A method for registering compound
action potentials generated by muscle fibers.
engram—A control function expressed in hypothetical variables that
can be scaled in time and magnitude and then projected onto
different effector systems.
enslaving—When intended force production by digits causes
unintended force production by other digits.
equifinality—An ability of a moving system to reach a planned final
position despite possible transient changes occurring in external
forces during the movement.
equilibrium-point hypothesis—A hypothesis of movement control
that assumes that the brain operates with control variables related
to spatial coordinates of muscle activation thresholds. Also, a
hypothesis of motor control that assumes that the central nervous
system manipulates equilibrium states of the system effector plus
load.
equilibrium potential (ion)—A potential that occurs when there is
no net passive movement of an ion across a membrane.
equilibrium potential (membrane)—A potential on a membrane
that is maintained in the absence of external stimuli.
equilibrium-trajectory hypothesis—A hypothesis that assumes
that the brain specifies a sequence of equilibrium points (an
equilibrium trajectory) for a working point, while its actual
trajectory is equally defined by external forces.
essential tremor—A tremor with a movement frequency between 4
and 12 Hz that is not accompanied by a diagnosed neurological
disorder.
evoked potential—A potential synchronized with an external event
(e.g., a stimulus).
excitatory postsynaptic potential (EPSP)—A brief, depolarizing
change in the potential of the postsynaptic membrane.
extensor mechanism—A network of passive elastic tissues that
produce an extensor action in the distal finger joints when intrinsic
finger flexors are activated.
exteroceptor—A receptor that transduces information from the
environment.
extrafusal fibers—Power-producing muscle fibers that are external
with respect to muscle spindles.
extrinsic muscles (hand)—Muscles that lie within the forearm,
outside of the hand.
facioscapulohumeral muscular dystrophy—A form of muscular
dystrophy that appears in adolescence and causes slow,
progressive weakness in facial muscles and certain muscles in
the arms and legs.
fatigue—A physiological state associated with a drop in maximal
force production and/or inability to continue performance at the
required level.
feedback control—When a controller changes command signals
based on their outcome.
feedforward control—When a controller generates command
variables independently of the outcome.
flexor reflex—A polysynaptic reflex seen in several major flexor
muscles in response to a mechanical or electrical stimulation of
flexor reflex afferents.
flexor reflex afferents—A group of peripheral receptors that
contribute to the flexor reflex.
force deficit—When a digit produces smaller maximal force when
acting together with other digits of the hand as compared to when
it acts independently.
fovea—The area of the retina with the highest density of
photoreceptors.
foveola—The central area of the fovea; provides for the best
perception of light stimuli.
Friedreich’s ataxia—A genetic, neurodegenerative disorder
characterized by dysarthria, loss of reflexes, and axonal sensory
neuropathy.
functional electrical stimulation—A method of substituting for a
lost motor function using EMG signals from healthy muscles (or
from other sources) to drive electrical stimulators applied to
paralyzed muscles.
functional magnetic resonance imaging (fMRI)—A method of
brain imaging that involves comparing MRI measurements taken
before and after a subject performs a task.
F-wave—A muscle response (induced by an electrical stimulation of
the muscle nerve) that does not involve central synaptic
transmission; the stimulus induces an antidromic volley in the
motor axons and leads to an orthodromic volley.
gamma-aminobutyric acid (GABA)—A common neurotransmitter
within the central nervous system.
γ-motoneurons—Small motoneurons that innervate intrafusal fibers
and that change the sensitivity of spindle endings to muscle
length (static γ-motoneurons) and to velocity (dynamic γ-
motoneurons).
ganglion—A group of neurons united by a common function.
gate control theory of pain—A theory suggesting that the
subjective feeling of pain is created by a disparity between signals
from proprioceptors and signals from nociceptors.
generalized motor program theory (schema theory)—A theory
that assumes that the brain stores patterns of variables directly
related to mechanical patterns associated with the production of
particular actions.
globus pallidus—A neural structure that is part of the basal ganglia.
Golgi cells—Inhibitory interneurons in the cerebellum that are
excited by the parallel fibers and that make inhibitory connections
with the dendrites of granular cells.
Golgi tendon organ—A receptor ending that is located at the
junction between muscle and tendon and is sensitive to changes
in muscle force.
gray matter—Neural tissue containing mostly neuron bodies.
Guillain-Barré syndrome—A polyneuropathy associated with
demyelination of peripheral neural fibers.
habituation—A decrease in a response to a stimulus seen with
repetitive presentations of the stimulus.
haptic receptors—Peripheral sensory endings participating in the
sense of touch.
hemiballismus—Ballism associated with a lesion of one of the
subthalamic nuclei.
hemisyndrome—An impairment of the sensorimotor function in the
left or right half of the body; commonly follows a supraspinal
injury.
Henneman principle (size principle)—The principle stating that
motor units are recruited from the smallest to the largest.
Hill equation—An equation describing the relationship between
muscle force and velocity of muscle shortening.
hip strategy—A pattern of responses in postural muscles to an
external perturbation; this pattern shows the most pronounced
motion in the hip joint.
hippocampus—An area deep in the forebrain that is suspected of
playing a major role in short-term memory and in consolidating
information from short-term to long-term memory.
homunculus—A fictitious little person that sits in the brain and
makes decisions with respect to appropriate actions to be taken.
H-reflex—A monosynaptic reflex induced by an electrical stimulation
of the muscle nerve.
Huntington’s disease—A hereditary neurodegenerative disorder
that is associated with atrophy of the caudate nucleus and leads
to poorly controlled movements, including dancelike gait (chorea).
hyperkinesia—A motor impairment characterized by excessive
movements.
hyperpolarization—An increase in the absolute value of the
negative membrane potential.
hypokinesia (akinesia)—A motor impairment characterized by
poverty of voluntary movements.
hypometria—A tendency to undershoot the target; typical of
Parkinson’s disease.
hypophysis—The pituitary gland, one of four major structures in the
diencephalon.
hypothalamus—A structure in the diencephalon playing an
important role in the autonomic and emotional functions; part of
the limbic circle.
hypotonia—A state of decreased muscle tone; commonly described
for Down syndrome. Also, a decrease in the resistance of a joint
to an external movement.
hysteresis—When a dependence of one variable on another
variable depends on the direction of change of the latter variable.
inactivation (of sodium channels)—A drop in the membrane
conductance for sodium ions that leads to an absolute refractory
period and prevents backfiring of action potentials.
inhibitory postsynaptic potential (IPSP)—A brief hyperpolarizing
change in the potential of the postsynaptic membrane.
insular cortex——The insular coortex lies folded deep within the
lateral sulcus of each hemisphere, hidden below parts of the
frontal, parietal, and temporal lobes. It plays a key role in emotion
management, perceptual awareness, social behavior, and
decision making.
interaction torques—Joint torque components related to the motion
of other joints.
internal models—An approach that assumes that the central
nervous system computes control signals that lead to the
production of adequate muscle force patterns.
interneurons—Neurons receiving information from and transmitting
information to other neurons. Ia interneurons mediate reciprocal
inhibition, Renshaw cells mediate recurrent inhibition, and Ib
interneurons mediate inhibitory effects from Golgi tendon organs.
interoceptor—A receptor that transduces information from within
the body.
intrafusal fibers—Muscle fibers that are found inside muscle
spindles and are innervated by a special system of fusimotor
neurons (γ-motoneurons).
intrinsic muscles (hand)—Muscles that lie within the hand.
inverse dynamics (problem of)—The problem of finding joint
torques that assure a particular trajectory of the endpoint of a
multijoint limb.
inverse kinematics (problem of)—The problem of finding a joint
configuration corresponding to a particular location of the endpoint
of a multijoint limb.
ischemia—A blockage of the blood flow into an area of the body;
disrupts the transmission of action potentials along nerves.
isometric conditions—Contracting the muscle while the length of
the muscle fibers does not change; typically unattainable in
experiments. Also, contracting the muscle when the length of the
muscle plus tendon system does not change.
isotonic conditions—Contracting the muscle while the apparent
external load on the muscle does not change; typically
unattainable in experiments.
kinesthesia—The awareness of the position of the body segments
in space and in relation to each other.
kinetic tremor—Tremor associated with the initiation of a voluntary
movement.
latency—The delay between a stimulus and a reaction.
lateral geniculate nucleus—An important subcortical region
participating in visual perception; projects to the primary visual
cortex.
limbic circle (limbic system)—Brain structures (the hypothalamus,
the fornix, the hippocampus, the amygdaloid nucleus, and the
cingulate gyrus of the cerebral cortex) that are involved in
generating emotional reactions.
lobes—Parts of the large brain hemispheres (frontal, parietal,
occipital, temporal, and insula).
locomotion—A motor action that changes the location of the whole
body in the environment.
locomotor area (mesencephalic area)—An area in the medulla
and in the midbrain whose electrical stimulation can induce
locomotion in a decerebrate animal.
locomotor strip—An area in the cervical spinal cord whose
electrical stimulation can induce locomotion in a decerebrate
animal.
long-term depression (LTD)—A long-lasting decrease in the
excitability of a neuron following a specific synaptic input.
long-term memory—A memory that lasts for life.
long-term potentiation (LTP)—A long-lasting increase in the
excitability of a neuron following a specific synaptic input.
magnetic resonance imaging (MRI)—An indirect method of
assessing neural activity in the brain.
magnetic misreaching—The reaching movements that are locked
to the object of gaze fixation even when the participant is
instructed to move to a different location in space. It happens
when patients have sustained damage to the posterior parietal
cortex.
magnetoencephalography (MEG)—A method of studying the
magnetic field produced by electric currents in the brain.
mass–spring models—Models that assume that voluntary
movement may be adequately represented as the motion of a
mass on a spring with modifiable parameters.
medulla—The part of the central nervous system that connects the
spinal cord and the brain; contains, among other vital structures,
the cardiac center, the respiratory center, and the vasomotor
center.
medium spiny neurons—The vast majority of neurons in the
striatum and interface between dopamine reward signals and the
cortico-basal ganglia circuits.
membrane—A biological, partially permeable structure separating
the inside structures of a cell from the environment.
membrane threshold—A membrane potential that leads to the
generation of an action potential.
mesencephalic locomotor region—An area in the medulla and in
the midbrain whose electrical stimulation can induce locomotion in
a decerebrate animal.
milestone (motor)—An ability of a baby to perform a certain motor
task, such as holding the head up, sitting independently, walking,
reaching for objects, and so on.
minimum jerk principle—An optimization principle that is based on
minimizing an integral measure of jerk (derivative of acceleration)
during voluntary movement.
mirror neurons—Cortical neurons that are active both during the
execution of hand actions and during the observation of the same
actions made by others.
motor primitive—A hypothetical building block for a variety of motor
actions; primitives are assumed to be scaled and combined to
produce actions.
motor program—A combination of hypothetical variables that are
stored in the memory and that translate into required motor
patterns after being recalled.
motor redundancy—The availability of more variables than
necessary to solve a motor task.
motor unit—An α-motoneuron and all the muscle fibers it
innervates; a unit of force production in skeletal muscles.
M-response—A direct muscle response (contraction) to an electrical
stimulation of the muscle nerve.
multiple sclerosis (MS)—A systemic disease leading to a loss of
the myelin sheath on tracts within the central nervous system.
muscle compartment—A subgroup of muscle fibers united by a
common function that is not shared by other fibers of the muscle.
muscle tone—A poorly defined notion of joint resistance to passive
motion reflecting the subjective perception of a neurologist or a
physical therapist.
muscular dystrophy—A genetic disease characterized by
progressive weakness and degeneration of the skeletal muscles.
myasthenia gravis—A disorder of transmission at the
neuromuscular synapse, with the body producing antibodies to
acetylcholine receptors; likely results from an autoimmune
process.
myelin—A substance made of glial cells that forms a protective
shield around axons, leading to an increase in the conduction
speed of action potentials.
myoclonus—A brief muscle jerk caused by a neuronal discharge;
may be physiological or a component of a neurological disorder.
myosin—A molecule that is a major force-generating element within
a skeletal muscle.
myotonia—A prolonged episode of muscle activity after its voluntary
contraction.
myotonic dystrophy—The most common adult form of muscular
dystrophies characterized by myotonia.
negative feedback—A feedback that decreases the magnitude of
an original stimulus.
negative signs (of spasticity)—Phenomena that are seen in
persons without spasticity but are lost in patients with spasticity,
particularly loss of strength and discoordination.
neglect—A state in which a person does not recognize objects
presented in the visual field; may be a consequence of stroke.
Nernst equation—An equation for calculating the equilibrium
potential of an ion in the presence of an electrical field and a
difference in ion concentrations.
neuromuscular synapse—A place where an action potential on the
presynaptic membrane of a motor axon excites the postsynaptic
muscle membrane.
neuromyotonia—A neuromuscular disorder characterized by slow
relaxation of muscles after voluntary contraction or electrical
stimulation.
neuron—An excitable cell that is a unit of the nervous system.
neurotransmitter (mediator)—A substance released through the
presynaptic membrane; this substance can depolarize or
hyperpolarize the postsynaptic membrane of a synapse.
nociceptors—Small sensory endings that generate action potentials
in response to potentially damaging stimuli, such as temperature,
 pressure, or certain chemicals; participate in creating the sense of
pain.
nondeclarative memory—Memory related to how to; involves motor
skills and habits, conditioned reflexes, and other phenomena.
nonelectrolyte—A molecule or a fragment of a molecule with no net
electric charge.
non-obligatory synapse—A synapse in which the arrival of a single
action potential on the presynaptic membrane does not produce
an action potential on the postsynaptic membrane.
norepinephrine—A biological amine. A major neuromediator within
the central nervous system.
nystagmus—Repetitive jumps of the eyes.
obligatory synapse—A synapse in which the arrival of an action
potential on the presynaptic membrane always produces an action
potential on the postsynaptic membrane.
olivopontocerebellar atrophy (OPCA)—A disorder leading to
atrophy of the climbing fiber input into the cerebellum.
operant conditioning—An experimental situation that occurs when
a relationship between an action by the animal and an external
stimulus is being learned.
optic chiasm—The place where two optic nerves join each other.
optimization—An approach to the problem of motor redundancy
that searches for a solution that achieves an optimal value of a
particular cost function.
optokinetic system—A system that keeps stable features of the
environment on the fovea during head movements.
orthodromic conduction—Conduction of action potentials along
the axon from the soma.
osmosis—Movement of water between two compartments caused
by different concentrations of water measured as the total
concentration of all particles.
otoliths—Crystals of calcium carbonate in the inner ear that
participate in detecting linear acceleration.
paresis—Partial loss of voluntary control over muscle within an area
of the body.
Parkinson’s disease—A complex disorder featuring a motor
component associated with an impairment of the functioning of the
basal ganglia.
Pavlov’s theory of conditioned reflexes—Theory that assumes
that behavior is a combination of unconditioned (inborn) and
conditioned reflexes.
perseveration—Repetitive, apparently purposeless actions.
persistent inward current (PIC)—A depolarizing current produced
by voltage-sensitive channels that do not show inactivation.
phasic stretch reflex—A monosynaptic reflex response to quick
muscle stretch (the same as the T-reflex).
photoreceptors (cones and rods)—Specialized neurons in the eye
that generate action potentials in response to visible light.
physiology of activity—A theory advanced by Bernstein suggesting
that voluntary movements are initiated by active processes within
the central nervous system.
plasticity (neural)—An ability to modify neural connections in
response to injury or to specific training.
pons—Part of the brain located just rostrally to the medulla.
positive feedback—Feedback that increases the magnitude of an
original stimulus.
positive signs (of spasticity)—Clinical signs seen in a patient with
spasticity that are not seen in persons without spasticity; involve
spasms and increased muscle tone.
positron emission tomography (PET)—A method of brain imaging
that traces a radioactive isotope injected into the circulatory
system.
postsynaptic inhibition—An inhibitory influence acting on the
postsynaptic membrane of a neuron.
postsynaptic membrane—An area of the membrane of an
excitable cell that receives excitatory or inhibitory stimuli through a
synapse.
posttetanic potentiation (PTP)—A brief increase in the twitch
contraction force following a brief tetanic stimulation.
postural sway—Spontaneous migration of the center of gravity and
of the center of pressure of the body during quiet standing.
posture–movement paradox—The question of how an active
movement can occur without triggering the resistance of posture-
stabilizing mechanisms.
prehension—Exploration of objects by touch and manipulation of
objects by applying adequate forces and moments of forces.
prehension synergies—Conjoint changes in the forces and
moments of forces produced by a set of digits on a handheld
object that stabilize the overall mechanical action of the hand on
the object.
premotor cortex (PM)—Part of the premotor area (area 6) of the
brain cortex.
preprogrammed reactions (long-latency reflexes, functional
stretch reflex, M2-M3, or triggered reactions)—Muscle reactions
to external signals (e.g., perturbations) prepared in advance by
the central nervous system and triggered by an appropriate
peripheral stimulus.
presynaptic inhibition—An inhibitory influence acting on the
presynaptic membrane of a synapse; selective with respect to the
involved synapse.
presynaptic membrane—An area of the membrane of a neural fiber
transmitting information through a synapse.
primary motor area—Area 4 of the precentral cortex; requires low
stimulation currents to induce visible movement.
primitive reflexes—Reflexes seen in newborns that typically
disappear as the child develops.
principle of abundance—A view that the neural controller does not
look for unique motor patterns but instead facilitates families of
solutions that are equally able to solve the task.
principle of reafference—A hypothesis that any voluntary action is
associated with readdressing afferent signals from proprioceptors
 to a new posture.
principle of superposition—A principle that assumes that the
output of a set of elements with several inputs equals the sum of
the outputs produced by each of the inputs applied separately.
proprioceptor—A receptor transducing information about the
relative configuration and state of the body segments.
propriospinal tracts—Neural tracts carrying information from one
segment of the spinal cord to another.
pupillary reflex—A reflex adaptation of pupil size to light mediated
by the pretectal area of the midbrain.
Purkinje cells—Large inhibitory cells in the cerebellum; provide the
only output of the cerebellum.
putamen—A neural structure that is part of the basal ganglia.
pyramidal cells—Large neurons of the cerebral cortex; the origin of
the pyramidal tract.
radiculopathy—A disorder that originates from mechanical or
inflammatory damage to a spinal root or to a group of spinal roots.
Ranvier nodes—Breaks in the myelin sheath that contain a high
concentration of sodium channels; places where action potentials
can be generated during transmission along myelinated fibers.
readiness potential (Bereitschafts potential)—A slow, negative
shift of the EEG seen as early as 1.5 s before a voluntary
movement.
receptor—A specialized neuron or a subcellular structure that
generates action potentials in response to specific sources of
energy.
reciprocal activation—A pattern of muscle activation in which one
muscle of an agonist–antagonist pair increases its activity while
the other muscle does not change or decreases its activity.
reciprocal inhibition—A system that uses inhibitory Ia interneurons
and suppresses the activity of a motoneuronal pool when an
antagonistic pool is being excited.
recurrent inhibition—A system that uses inhibitory interneurons,
Renshaw cells, activated by axonal branches of alpha-
 motoneurons and suppressing all alpha-motoneurons of the pool.
red nucleus—A structure within the midbrain; the source of the
rubro-spinal tract.
reductionism—An approach that tries to describe the properties of
a system by using the properties of its elements; it is, by definition,
inapplicable to the analysis of complex systems.
redundancy—Availability of more variables than absolutely
necessary to solve a task.
reflex—A misnomer implying a relatively stereotypical and
standardized reaction to an external stimulus; a monosynaptic
reflex involves only one central synapse, oligosynaptic reflexes
involve a few central synapses, and polysynaptic reflexes involve
many central synapses.
reflex arc—A loop typically involving a receptor, a central processing
unit (neural structures), and an output structure (a muscle).
refractory period (absolute)—A period when an excitable structure
cannot be excited even by a very strong external stimulus.
refractory period (relative)—A period when an excitable structure
needs a stronger-than-usual stimulus to generate a response.
reinnervation—A process of axonal sprouting that forms new
synapses.
Renshaw cell—An inhibitory interneuron excited by signals from
motoneurons of a pool and inhibiting the activity of the same
motoneuronal pool (recurrent inhibition).
reticular formation—A structure containing numerous small
neurons that occupy areas of the medulla and midbrain; the
source of the reticulospinal tract.
retina—A layer within the internal structures of the eye; contains
photoreceptors.
rheobase—The lowest stimulus amplitude that can lead to the
generation of an action potential for a stimulus of a very long
(infinite) duration.
rigidity—An increased resistance of a limb segment to attempts to
move it with an external force; typical of Parkinson’s disease.
rod—A photoreceptor that functions in the dim light and thus
provides night vision.
saccade—A very quick and accurate eye movement that is used to
shift the gaze from object to object.
sarcolemma—The membrane of a muscle cell.
sarcomere—A force-producing unit of a muscle filament.
sarcoplasmic reticulum—A system of cisternae containing Ca++
ions within a muscle fiber.
sensory ending—A part of a receptor cell able to generate action
potentials in response to specific influences (sources of energy).
serotonin—A biological amine. A major neuromediator within the
central nervous system.
servo—A feedback control system providing for the perfect
generation of a desired value of an output parameter.
servo hypothesis—A hypothesis of motor control that considers the
mechanism of the tonic stretch reflex as a perfect servo.
short-term memory—A memory that lasts for a few minutes or
hours.
sliding filament theory—A theory of muscle force production based
on molecular interactions, mostly between actin and myosin
molecules.
smooth pursuit—A relatively slow eye movement that keeps the
image of an object on the fovea.
sodium–potassium pump—An active mechanism that maintains
the difference of ion concentrations across biological membranes.
soma—The body of a neuron containing organelles; commonly, the
site of input signals.
somatosensory cortical areas—Areas 1, 2, 3a, and 3b in the
parietal cortex receiving inputs from the thalamus; contain sensory
maps that look like distorted images of the body.
spasmodic dysphonia—A focal dystonia that affects the vocal
cords, causing decreased smoothness of speech, effortful voice,
and a reduction in loudness.
spasticity—A complex of symptoms associated with the disruption
of transmission along descending spinal tracts; involves
uncontrolled spasms, increased muscle tone, and increased
muscle reflexes to stretch with a pronounced velocity-dependent
component.
spatial summation—An increase in the combined effect of two (or
more) stimuli when they simultaneously arrive at different sites
belonging to the same excitable structure (e.g., a neuron).
spinal shock—A state that occurs after a spinal cord injury, lasting
for a few minutes to a few hours, when muscle reflexes are
suppressed and the patient shows complete paralysis.
spindle (muscle)—A spindle-shaped structure located in parallel to
power-producing muscle fibers; contains primary and secondary
endings sensitive to changes in muscle length and velocity.
spindle endings—Primary receptors sensitive to changes in muscle
length and velocity, and secondary receptors sensitive to muscle
length but not to velocity.
sprouting—Growing new axonal terminals; commonly leads to
reinnervation.
stellate cells—Cortical neurons that are found in highest numbers in
layer IV of the cortex. They lack an apical dendrite and have a
restricted dendritic arbor that does not extend beyond the layer in
which the cell body resides.
stiff-man syndrome—A continuous muscle fiber activity syndrome
leading to a board-like rigidity of the trunk muscles.
stretch reflex—An increase in muscle activation and force caused
by stretching the muscle; it is mediated by the spinal cord and has
both phasic and tonic components.
stroke (cerebrovascular accident, CVA)—An interruption of the
normal blood supply to a brain area.
structural unit—A task-specific organization of elements within a
multi-element system.
substantia nigra—A structure in the midbrain; part of the basal
ganglia and a major source of dopamine in the brain.
subthalamic nucleus—A neural structure that is part of the basal
ganglia.
supplementary motor area (SMA)—A cortical area (area 6) whose
stimulation requires higher currents and induces more complex
movements than those seen during stimulation of the primary
motor area.
synapse—A place where signals are transmitted from one excitable
cell (a neuron) to another excitable cell (a neuron or a muscle
fiber).
synaptic cleft—A gap between the presynaptic membrane and the
postsynaptic membrane.
synergy (postural or movement)—A neural organization of signals
sent to elements (muscles, joints, effectors, etc.) of a multi-
element system that assures stable performance of a task.
tardive dyskinesia—Late-appearing involuntary movements
resulting from chronic administration of blocking agents for
dopamine receptors.
tarsal tunnel syndrome—A peripheral mononeuropathy caused by
entrapment of the tibial nerve; similar to carpal tunnel syndrome.
temporal summation—An increase in the effect of a stimulus when
it follows another stimulus after a brief delay.
tetanus—A sustained muscle contraction produced by a sequence
of action potentials in the motor axons. Also, a continuous muscle
fiber activity syndrome produced by the toxin tetanospasmin,
which blocks postsynaptic inhibition at the spinal level.
thalamus—A large structure in the diencephalon; plays an important
role in sensorimotor coordination.
tic—A brief, repetitive, and seemingly purposeless, stereotyped
action that may involve one muscle or muscle groups; a
component of Tourette syndrome.
titin—A macromolecule with elastic properties that plays a major
role in the mechanical behavior of skeletal muscles.
titubation—Trunk or head tremor at a frequency of 1 to 3 Hz; typical
of cerebellar disorders.
tone (muscle)—A misnomer implying a feeling of resistance
experienced by an examiner who tries to move a limb segment of
another person.
tonic stretch reflex—A polysynaptic reflex that increases the level
of muscle activation with slow muscle stretch.
tonic vibration reflex—A polysynaptic reflex that increases the level
of muscle activation induced by a low-amplitude, high-frequency
muscle or tendon vibration.
topographic organization—An organization of neural projections
that involves anatomically close neurons in one structure
projecting to anatomically close targets.
torticollis—Dystonia affecting the neck muscles.
Tourette syndrome—A complex disorder, more common in boys,
characterized by tics.
transcranial magnetic stimulation (TMS)—A noninvasive method
that uses a quickly changing magnetic field to stimulate brain
structures.
T-reflex—A monosynaptic reflex response to a quick muscle stretch
(e.g., to a tendon tap).
tremor—Alternating activity in agonist–antagonist muscle pairs
controlling a joint, leading to alternating joint movements; 3 to 5
Hz in cerebellar disorders; about 6 Hz in Parkinson’s disease; 8 to
12 Hz in persons without disorder (physiological tremor).
triphasic pattern—An EMG pattern typically accompanying
voluntary movements; consists of an agonist burst followed by an
antagonist burst and a second agonist burst.
tropomyosin—A long molecule that lies in parallel to an actin
molecule.
troponin—A molecule blocking a site for the formation of cross-
bridges.
T-tubule—An invagination of the sarcolemma where it comes close
to the cisternae within the sarcoplasmic reticulum.
twitch (contraction)—A brief muscle contraction in response to a
single presynaptic action potential or a single, synchronized volley
 of action potentials.
ulnar palsy—A peripheral mononeuropathy caused by entrapment
of the ulnar nerve, commonly at the elbow level.
uncontrolled manifold hypothesis—A hypothesis on the control of
multi-element systems; conforms to the principle of abundance.
unloading reflex—A decrease of the muscle activity when the load
is suddenly decreased.
ventral root (spinal)—A set of neural fibers carrying output (motor)
signals from motoneurons to their innervated structures.
ventricles (brain)—Hollow spaces within the brain filled with
cerebrospinal fluid.
vergence—An eye movement that fixes the gaze on targets at
different depths.
vestibular ganglion (Scarpa’s ganglion)—The ganglion
innervating vestibular receptors.
vestibular nuclei (lateral vestibular or Deiters’, medial
vestibular, superior vestibular, and inferior vestibular)—
Sources of the vestibulospinal tract; located in the medulla.
vestibulo-ocular reflex (VOR)—A reflex that coordinates eye and
head movements, helping to maintain a constant visual field.
vestibulospinal tract—A neural tract originating in the vestibular
nuclei and projecting on the spinal structures and cranial nerve
nuclei.
vibration-induced fallings (VIFs)—Postural disturbances
introduced by a low-amplitude, high-frequency vibration applied to
a postural muscle or to its tendon.
Westphal phenomenon—An abrupt reflex excitation of a muscle in
response to an externally imposed muscle shortening.
white matter—Neural tissue consisting mostly of conduction
pathways.
Williams syndrome—A rare inborn disorder associated with
reduced cerebral volume and severely depressed nonverbal IQ
but with high verbal and grammatical fluency.
Wilson’s disease—A rare condition characterized by copper
deposits in the brain (cortex and basal ganglia) and in other
organs; characterized by an unusual tremor.
wiping reflex—A reflex that is a coordinated movement of a spinal
animal; removes an irritating stimulus from the animal’s skin.
working point—The point whose trajectory is crucial for success
during a multijoint movement.
writer’s cramp—A task-specific dystonia causing involuntary
contractions of arm muscles with twisting, repetitive movements,
and abnormal postures.
References
RECOMMENDED BOOKS
Davids K, Bennett S, Newell KM (2005) Movement system variability. Champaign,
IL: Human Kinetics.
Enoka RM (2002) Neuromechanics of human movement. Third edition.
Champaign, IL: Human Kinetics.
Feldman AG (2015) Referent control of action and perception: Challenging
conventional theories in behavioral science. New York: Springer.
Kandel ER, Schwartz JH, Jessell TM (Eds.) (1999) Principles of neural science.
Fourth edition. New York: McGraw-Hill.
Kelso JAS (1995). Dynamic patterns: The self-organization of brain and behavior.
Cambridge, MA: MIT Press.
Kugler PN, Turvey MT (1987) Information, natural law, and the self-assembly of
rhythmic movement. Hillsdale, NJ: Erlbaum.
Latash ML (2019) Physics of biological action and perception. New York:
Academic Press.
Latash ML (Ed.) (2020) Bernstein’s construction of movement: Original text and
commentaries. Abingdon, UK: Routledge.
Latash ML, Zatsiorsky VM (Eds.) (2001) Classics in movement science.
Champaign, IL: Human Kinetics.
Latash ML, Zatsiorsky VM (2016) Biomechanics and motor control: Defining
central concepts. New York: Academic Press.
Newell KM, Corcos DM (Eds.) (1993) Variability in motor control. Champaign, IL:
Human Kinetics.
Orlovsky GN, Deliagina TG, Grillner S (1999) Neuronal control of locomotion.
From mollusk to man. Oxford: Oxford University Press.
Rothwell JC (1994) Control of human voluntary movement. Second edition.
London: Chapman & Hall.
Watts RL, Koller WC (Eds.) (2004) Movement disorders. Neurological principles
and practice. Second edition. New York: McGraw-Hill.
Windhorst U, Johansson H (Eds.) (1999) Modern techniques in neuroscience
research. Berlin: Springer-Verlag.
Winters JM, Woo SL-Y (Eds.) (1990) Multiple muscle systems. Biomechanics and
movement organization. New York: Springer-Verlag.
Zatsiorsky VM (1998) Kinematics of human motion. Champaign, IL: Human
Kinetics.
Zatsiorsky VM (2002) Kinetics of human motion. Champaign, IL: Human Kinetics.
Zatsiorsky VM, Prilutsky BI (2012) Biomechanics of skeletal muscles. Champaign,
IL: Human Kinetics.
Zigmond MJ, Bloom FE, Landis SC, Roberts JL, Squire LR (Eds.) (1999)
Fundamental neuroscience. San Diego: Academic Press.
REFERENCES
Abbs JH, Gracco VL (1984) Control of complex motor gestures: Orofacial muscle
responses to load perturbations of the lip during speech. J Neurophysiol 51:
705-723.
Abdusamatov RM, Feldman AG (1986) Description of the electromyograms with
the aid of a mathematical model for single joint movements. Biophysics 31: 549-
552.
Abelson JF, Kwan KY, O’Roak BJ, Baek DY, Stillman AA, Morgan TM, Mathews
CA, Pauls DL, Rasin MR, Gunel M, Davis NR, Ercan-Sencicek AG, Guez DH,
Spertus JA, Leckman JF, Dure LS 4th, Kurlan R, Singer HS, Gilbert DL, Farhi A,
Louvi A, Lifton RP, Sestan N, State MW (2005) Sequence variants in SLITRK1
are associated with Tourette’s syndrome. Science 310: 317-320.
Abolins V, Stremoukhov A, Walter C, Latash ML (2020) On the origin of finger
enslaving: Control with referent coordinates and effects of visual feedback. J
Neurophysiol 124: 1625-1636.
Agarwal GC, Gottlieb GL (1980) Effect of vibration on the ankle stretch reflex in
man. Electroencephalog Clin Neurophysiol 49: 81-92.
Akulin VM, Carlier F, Solnik S, Latash ML (2019) Sloppy, but acceptable, control of
biological movement: Algorithm-based stabilization of subspaces in abundant
spaces. J Human Kinet 67: 49-72.
Albus JS (1971) A theory of cerebellar function. Math Biosci 10: 25-61.
Alexander GE, Crutcher MD, DeLong MR (1990) Basal ganglia-thalamocortical
circuits: Parallel substrates for motor, coulomotor, ‘prefrontal’ and ‘limbic’
functions. Prog Brain Res 85: 119-146.
Alexandrov A, Frolov A, Massion J (1998) Axial synergies during human upper
trunk bending. Exp Brain Res 118: 210-220.
Allman BL, Rice CL (2002) Neuromuscular fatigue and aging: Central and
peripheral factors. Muscle Nerve 25: 785-796.
Allum JHJ (1975) Response to load disturbances in human shoulder muscles: The
hypothesis that one component is a pulse test information signal. Exp Brain Res
22: 307-326.
Allum JHJ (1983) Organization of stabilizing reflex responses in tibialis anterior
muscles following ankle flexion perturbations of standing man. Brain Res 264:
297-301.
Allum JHJ, Honneger F, Pfaltz CR (1989) The role of stretch and vestibulospinal
reflexes in the generation of human equilibrating reactions. Prog Brain Res 80:
399-409.
Amack JD, Mahadevan MS (2004) Myogenic defects in myotonic dystrophy. Dev
Biol 265: 294-301.
Ambike S, Mattos D, Zatsiorsky VM, Latash ML (2016) Synergies in the space of
control variables within the equilibrium-point hypothesis. Neuroscience 315:
150-161.
An CH, Atkeson CG, Hollerbach JM (1988) Model-based control of a robot
manipulator. Cambridge, MA: MIT Press.
Andersen PM (2004) The genetics of amyotrophic lateral sclerosis (ALS). Suppl
Clin Neurophysiol 57: 211-227.
Andersson G, Armstrong DM (1987) Complex spikes in Purkinje cells in the lateral
vermis (b zone) of the cat cerebellum during locomotion. J Physiol 385: 107-
134.
Andrews CJ, Burke D, Lance JW (1972) The response to muscle stretch and
shortening in Parkinsonian rigidity. Brain 95: 795-812.
Angel RW, Lewitt PA (1978) Unloading and shortening reactions in Parkinson’s
disease. J Neurol Neurosurg Psychiatry 41: 919-923.
Aoyagi Y, Tsubahara A (2004) Therapeutic orthosis and electrical stimulation for
upper extremity hemiplegia after stroke: A review of effectiveness based on
evidence. Top Stroke Rehabil 11: 9-15.
Arbib MA, Iberall T, Lyons D (1985) Coordinated control programs for movements
of the hand. In: Goodwin AW, Darian-Smith I (Eds.) Hand function and the
neocortex, pp. 111-129. Berlin: Springer-Verlag.
Arimoto S, Tahara K, Yamaguchi M, Nguyen PTA, Han HY (2001) Principles of
superposition for controlling pinch motions by means of robot fingers with soft
tips. Robotica 19: 21-28.
Aron, AR (2007) The neural basis of inhibition in cognitive control. Neuroscientist,
13(3), 214-228.
Aruin AS, Forrest WR, Latash ML (1998) Anticipatory postural adjsutments in
conditions of postural instability. Electroencephalog Clin Neurophysiol 109: 350-
359.
Aruin AS, Kanekar N (2013) Effect of a textured insole on balance and gait
symmetry. Exp Brain Res 231(2): 201-208.
Aruin AS, Latash ML (1995) The role of motor action in anticipatory postural
adjustments studied with self-induced and externally-triggered perturbations.
Exp Brain Res 106: 291-300.
Aruin AS, Latash ML (1996) Anticipatory postural adjustments during self-initiated
perturbations of different magnitude triggered by a standard motor action.
Electroencephalog Clin Neurophysiol 101: 497-503.
Aruin AS, Rao N (2018) The effect of a single textured insole in gait rehabilitation
of individuals with stroke. Int J Rehabil Res 41(3): 218-223.
Aruin AS, Rao N, Sharma A, Chaudhuri G (2012) Compelled body weight shift
approach in rehabilitation of individuals with chronic stroke. Top Stroke Rehabil
 19(6): 556-563.
Asaka T, Wang Y, Fukushima J, Latash ML (2008) Learning effects on muscle
modes and multi-mode synergies. Exp Brain Res 184: 323-338.
Asanuma H (1973) Cerebral cortical control of movement. Physiologist 16: 143-
166.
Asanuma H, Zarzecki P, Jankowska E, Hongo T, Marcus S (1979) Projection of
individual pyramidal tract neurons to lumbar motor nuclei of the monkey. Exp
Brain Res, 34(1), 73-89.
Asatryan DG, Feldman AG (1965) Functional tuning of the nervous system with
control of movements or maintenance of a steady posture. I. Mechanographic
analysis of the work of the limb on execution of a postural task. Biophysics 10:
925-935.
Ashby P, Verrier M (1976) Neurophysiological changes in hemiplegia, possible
explanation for initial disparity between muscle tone and tendon reflexes.
Neurology 26: 1145-1151.
Ashby FG, Turner BO, Horvitz JC (2010) Cortical and basal ganglia contributions
to habit learning and automaticity. Trends Cog Sci, 14(5), 208-215.
Asratyan EA (1953). I.P. Pavlov: His Life and Work. Moscow: Foreign Languages
Publishing House.
Atkeson CG (1989) Learning arm kinematics and dynamics. Ann Rev Neurosci 12:
157-183.
Atkeson CG, Hollerbach JM (1985) Kinematic features of unrestrained vertical arm
movements. J Neurosci 5: 2318-2320.
Atsuta Y, Garcia-Rill E, Skinner RD (1991) Control of locomotion in vitro: I.
Deafferentation. Somatosens Mot Res 8: 45-53.
Babinski F (1899) De l’asynergie cerebelleuse. Revue Neurologique 7: 806-816.
Bagesteiro LB, Sainburg RL (2002) Handedness: Dominant arm advantages in
control of limb dynamics. J Neurophysiol 88: 2408-2421.
Bagesteiro LB, Sainburg RL (2003) Nondominant arm advantages in load
compensation during rapid elbow joint movements. J Neurophysiol 90: 1503-
1513.
Baker R, Llinas R (1971) Electrotonic coupling between neurons in the rat
mesencephalic nucleus. J Physiol 212: 45-63.
Balestra C, Duchateau J, Hainaut K (1992) Effects of fatigue on the stretch reflex
in a human muscle. Electroencephalogr Clin Neurophysiol 85: 46-52.
Barash S, Melikyan A, Sivakov A, Zhang M, Glickstein M, Thier P (1999) Saccadic
dysmetria and adaptation after lesions of the cerebellar cortex. J Neurosci,
19(24), 10931-10939.
Barthel T, Mechau D, Wehr T, Schnittker R, Liesen H, Weiss M (2001) Readiness
potential in different states of physical activation and after ingestion of taurine
 and/or caffeine containing drinks. Amino Acids 20: 63-73.
Barto AG, Fagg AH, Sitkoff N, Houk JC (1999) A cerebellar model of timing and
prediction in the control of reaching. Neural Comput 11: 565-594.
Bastian AJ, Martin TA, Keating JG, Thach WT (1996) Cerebellar ataxia: Abnormal
control of interaction torques across multiple joints. J Neurophysiol 76: 492-509.
Bastian AJ, Mugnaini E, Thach WT (1999) Cerebellum. In: Zigmond MJ, Bloom
FE, Landis SC, Roberts JL, Squire LR (Eds.) (1999) Fundamental
neuroscience, pp. 973-992. San Diego: Academic Press.
Bawa P, McKenzie DC (1981) Contribution of joint and cutaneous afferents to
longer-latency reflexes in man. Brain Res 211: 185-189.
Bazalgette D, Zattara M, Bathien N, Bouisset S, Rondot P (1986) Postural
adjustments associated with rapid voluntary arm movements in patients with
Parkinson’s disease. Adv Neurol 45: 371-374.
Becker W, Fuchs AF (1985) Prediction in the oculomotor system: smooth pursuit
during transient disappearance of a visual target. Exp Brain Res 57(3): 562-575.
Becker WJ, Morrice BL, Clark AW, Lee RG (1991) Multi-joint reaching movements
and eye-hand tracking in cerebellar incoordination: Investigation of a patient
with complete loss of Purkinje cells. Can J Neurol Sci 18: 476-487.
Belanger AY, McComas AJ, Elder GBC (1983) Physiological properties of two
antagonistic human muscle groups. Eur J Appl Physiol 51: 381-393.
Belen’kii VY, Gurfinkel VS, Pal’tsev YI (1967) Elements of control of voluntary
movements. Biofizika 10: 135-141.
Bellugi U, Bihrle A, Jernigan T, Trauner D, Doherty S (1990) Neuropsychological,
neurological, and neuroanatomical profile of Williams syndrome. Amer J Med
Genet 6 (suppl.): 115-125.
Bemben MG (1998) Age-related alterations in muscular endurance. Sports Med
25: 259-269.
Benabid AL, Benazzouz A, Hoffmann D, Limousin P, Krack P, Pollak P (1998)
Long-term electrical inhibition of deep brain targets in movement disorders. Mov
Disord 13 Suppl 3:119-125.
Benecke R, Rothwell JC, Dick JPR, Day BL, Marsden CD (1986) Performance of
simultaneous movements in patients with Parkinson’s disease. Brain 109: 739-
757.
Benecke R, Rothwell JC, Dick JPR, Day BL, Marsden CD (1987) Disturbance of
sequential movements in patients with Parkinson’s disease. Brain 110: 361-379.
Bennett MVL, Crain SM, Grundfesst H (1959) Electrophysiology of supramedullary
neuron of Spheroides maculatus. I. Orthodromic and antidromic responses. J
Gen Physiol 43: 159-188.
Benwell NM, Byrnes ML, Mastaglia FL, Thickbroom GW (2005) Primary
sensorimotor cortex activation with task-performance after fatiguing hand
 exercise. Exp Brain Res 167: 160-164.
Beppu H, Suda M, Tanaka R (1984) Analysis of cerebellar motor disorders by
visually guided elbow tracking movement. Brain 107: 787-809.
Berardelli A, Dick JPR, Rothwell JC, Day BL, Marsden CD (1986) Scaling of the
size of the first agonist EMG burst during rapid wrist movements in patients with
Parkinson’s disease. J Neurol Neurosurg Psychiatr 49: 1273-1279.
Berardelli A, Rothwell JC, Hallett M, Thompson PD, Manfredi M, Marsden CD
(1998) The pathophysiology of primary dystonia. Brain 121: 1195-1212.
Beres-Jones JA, Johnson TD, Harkema SJ (2003) Clonus after human spinal cord
injury cannot be attributed solely to recurrent muscle-tendon stretch. Exp Brain
Res 149: 222-236.
Berger DJ, Masciullo M, Molinari M, Lacquaniti F, d’Avella A (2020) Does the
cerebellum shape the spatiotemporal organization of muscle patterns? Insights
from subjects with cerebellar ataxias. J Neurophysiol 123(5): 1691-1710.
Berger W, Horstmann D, Dietz V (1984) Tension development and muscle
activation in the leg during gait in spastic hemiparesis: Independence of muscle
hypertonia and exaggerated stretch reflexes. J Neurol Neurosurg Psychiat 47:
1029-1033.
Berkinblit MB, Feldman AG, Fukson OI (1986) Adaptability of innate motor
patterns and motor control mechanisms. Behav Brain Sci 9: 585-638.
Bernstein NA (1930) A new method of mirror cyclographie and its application
towards the study of labor movements during work on a workbench. Hygiene,
Safety and Pathology of Labor 5: 3-9 and 6: 3-11 (in Russian).
Bernstein NA (1935) The problem of interrelation between coordination and
localization. Arch Biol Sci 38: 1-35 (in Russian).
Bernstein NA (1947) On the construction of movements. Medgiz: Moscow (in
Russian). English translation is in: Latash ML (Ed.) (2020) Bernstein’s
construction of movements. Abingdon, UK: Routledge.
Bernstein NA (1966) Essays on the physiology of movements and physiology of
activity. Moscow: Meditsina (in Russian).
Bernstein NA (1967) The co-ordination and regulation of movements. Oxford:
Pergamon Press.
Bernstein NA (1996) On dexterity and its development. In: Latash ML, Turvey MT
(Eds.) Dexterity and its development, pp. 1-244. Mahwah, NJ: Erlbaum.
Bernstein NA (2003) Contemporary studies on the physiology of the neural
process. Moscow, Russia: Smysl.
Berntson GG, Torello MW (1982) The paleocerebellum and the integration of
behavioral function. Physiol Psychol 10: 2-12.
Bigland B, Lippold OCJ (1954) Motor unit activity in the voluntary contraction of
human muscle. J Physiol 125: 322-335.
Bigland-Ritchie B, Cafarelli E, Vollestad NK (1986) Fatigue of submaximal static
contractions. Acta Physiol Scand 128, Suppl. 556: 137-148.
Bigland-Ritchie B, Dawson NJ, Johansson RS, Lippold OCJ (1986) Reflex origin
for the slowing of motoneurone firing rates in fatigue of human voluntary
contraction. J Physiol 379: 451-459.
Bigland-Ritchie B, Johansson R, Lippold OCJ, Woods JJ (1983) Contractile speed
and EMG changes during fatigue of sustained maximal voluntary contractions. J
Neurophysiol 50: 313-324.
Bilodeau M, Henderson TK, Nolta BE, Pursley PJ, Sandfort GL (2001) Effect of
aging on fatigue characteristics of elbow flexor muscles during sustained
submaximal contraction. J Appl Physiol 91: 2654-2664.
Birbaumer N (2006) Breaking the silence: Brain-computer interfaces (BCI) for
communication and motor control. Psychophysiology 43(6): 517-532.
Birbaumer N, Cohen LG (2007) Brain-computer interfaces: Communication and
restoration of movement in paralysis. J Physiol 579: 621-636.
Bizzi E, Giszter SF, Loeb E, Mussa-Ivaldi FA, Saltiel P (1995) Modular organization
of motor behavior in the frog’s spinal cord. Trends Neurosci 18: 442-446.
Bizzi E, Hogan N, Mussa-Ivaldi FA, Giszter S (1992) Does the nervous system use
equilibrium-point control to guide single and multiple joint movements? Behav
Brain Sci 15: 603-613.
Bloedel JR (1992) Functional heterogeneity with structural homogeneity: How
does the cerebellum operate? Behav Brain Sci 15: 666-678.
Boatright JR, Kiebzak GM, O’Neil DM, Peindl RD (1997) Measurement of thumb
abduction strength: Normative data and a comparison with grip and pinch
strength. J Hand Surg (Amer) 22: 843-848.
Boehnke SE, Munoz DP (2008) On the importance of the transient visual response
in the superior colliculus. Curr Opin Neurobiol 18(6): 544-551.
Bockbrader MA, Francisco G, Lee R, Olson J, Solinsky R, Boninger ML (2018)
Brain computer interfaces in rehabilitation medicine. Phys Med Rehab 10(9
Suppl 2): S233-S243.
Bonasera SJ, Nichols TR (1996) Mechanical actions of heterogenic reflexes
among ankle stabilizers and their interactions with plantarflexors of the cat
hindlimb. J Neurophysiol 75: 2050-2070.
Boniface SJ (2001) Plasticity after acute ischaemic stroke studied by transcranial
magnetic stimulation. J Neurol Neurosurg Psychiatry 71: 713-715.
Bongaardt R (2001) How Bernstein conquered movement. In: Latash ML,
Zatsiorsky VM (Eds.) Classics in movement science, pp. 59-84. Champaign, IL:
Human Kinetics.
Bostan AC, Strick PL (2018) The basal ganglia and the cerebellum: nodes in an
integrated network. Nat Rev Neurosci, 19(6), 338-350. doi:10.1038/s41583-018-
 0002-7
Bottasso CL, Prilutsky BI, Croce A, Imberti E, Sartirana S (2006) A numerical
procedure for inferring from experimental data the optimization cost functions
using a multibody model of the neuro-musculoskeletal system. Multibody Syst
Dynam 16: 123-154.
Bouisset S, Zattara M (1987) Biomechanical study of the programming of
anticipatory postural adjustments associated with voluntary movement. J
Biomech 20: 735-742.
Bouisset S, Zattara M (1990) Segmental movement as a perturbation to balance?
Facts and concepts. In: Winters JM, Woo SL-Y (Eds.) Multiple muscle systems.
Biomechanics and movement organization, 498-506. New York: Springer-
Verlag.
Braitenberg V (1967) Is the cerebellar cortex a biological clock in the millisecond
range? Prog Brain Res 25: 2334-2346.
Brewer GJ (2005) Neurologically presenting Wilson’s disease: Epidemiology,
pathophysiology and treatment. CNS Drugs 19: 185-192.
Brooks VB, Thach WT (1981) Cerebellar control of posture and movement. In:
Brooks VB (Ed.) Handbook of physiology, the nervous system, sect. 1, vol. 2,
part 2, pp. 877-945. Baltimore: Williams & Wilkins.
Brown MC, Engberg I, Matthews PB (1967) The relative sensitivity to vibration of
muscle receptors of the cat. J Physiol 192: 773-800.
Brown P, Rothwell JC, Thompson PD, Britton TC, Day BL, Marsden CD(1991)
New observations on the normal auditory startle reflex in man. Brain 114(4):
1891-1902.
Bruce D (2001) Fifty years since Lashley’s In search of the Engram: Refutations
and conjectures. J Hist Neurosci 10: 308-318.
Bruijn LI, Miller TM, Cleveland DW (2004) Unraveling the mechanisms involved in
motor neuron degeneration in ALS. Annu Rev Neurosci 27: 723-749.
Bruton M, O’Dwyer N (2018) Synergies in coordination: A comprehensive overview
of neural, computational, and behavioral approaches. J Neurophysiol 120:
2761-2774.
Bruwer M, Cruse H (1990) A network model for the control of the movement of a
redundant manipulator. Biol Cybern 62: 549-555.
Buchman AS, Leurgans S, Gottlieb GL, Chen CH, Almeida GL, Corcos DM (2000)
Effect of age and gender in the control of elbow flexion movements. J Mot
Behav 32: 391-399.
Burke RE, Rudomin P, Zajac FE (1970) Catch property in single mammalian motor
units. Science 168: 122-124.
Burke RE, Rudomin P, Zajac FE (1976) The effect of activation history on tension
production by individual muscle units. Brain Res 109: 515-529.
Burnett RA, Laidlaw DH, Enoka RM (2000) Coactivation of the antagonist muscle
does not covary with steadiness in old adults. J Appl Physiol 89: 61-71.
Burrack A, Brugger P (2005) Individual differences in susceptibility to
experimentally induced phantom sensations. Body Image 2: 307-313.
Burridge JH, Haugland M, Larsen B, Svaneborg N, Iversen HK, Christensen PB,
Pickering RM, Sinkjaer T (2008). Patients’ perceptions of the benefits and
problems using the ActiGait implanted drop-foot stimulator. J Rehabil Med 40:
873-875.
Burstedt MK, Flanagan JR, Johansson RS (1999) Control of grasp stability in
humans under different frictional conditions during multidigit manipulation. J
Neurophysiol 82: 2393-2405.
Buschman TJ, Miller EK (2007) Top-down versus bottom-up control of attention in
the prefrontal and posterior parietal cortices. Science, 315(5820), 1860-1862.
Butefisch C, Khurana V, Kopylev L, Cohen LG (2004) Enhancing encoding of a
motor memory in the primary motor cortex by cortical stimulation. J
Neurophysiol 91: 2110-2116.
Butterworth G, Cicchetti D (1978) Visual calibration of posture in normal and motor
retarded Down’s syndrome infants. Perception 7: 513-525.
Cafarelli E, Kostka CE (1981) Effect of vibration on static force sensation in man.
Exp Neurol 74: 331-340.
Calancie B, Needham-Shropshire B, Jacobs P, Willer K, Zych G, Green BA (1994)
Involuntary stepping after chronic spinal cord injury. Evidence for a central
rhythm generator for locomotion in man. Brain 117: 1143-1159.
Campbell MJ, McComas AJ, Petito F (1973) Physiological changes in aging
muscles. J Neurol Neurosurg Psychiat 36: 174-182.
Campbell SK, Almeida GL, Penn RD, Corcos DM (1995) The effects of
intrathecally administered baclofen on function in patients with spasticity. Phys
Ther 75: 352-362.
Cannon SC, Zahalak GI (1982) The mechanical behavior of active human skeletal
muscle in small oscillations. J Biomech 15, 111-121.
Carmichael ST (2003) Plasticity of cortical projections after stroke. Neuroscientist
9: 64-75.
Casadio M, Morasso PG, Sanguineti V (2005) Direct measurement of ankle
stiffness during quiet standing: Implications for control modelling and clinical
application. Gait Posture 21: 410-424.
Cauraugh JH (2004) Coupled rehabilitation protocols and neural plasticity: Upper
extremity improvements in chronic hemiparesis. Restor Neurol Neurosci 22:
337-347.
Cauraugh JH, Summers JJ (2005) Neural plasticity and bilateral movements: A
rehabilitation approach for chronic stroke. Prog Neurobiol 75: 309-320.
Celnik PA, Cohen LG (2004) Modulation of motor function and cortical plasticity in
health and disease. Restor Neurol Neurosci 22: 261-268.
Centers for Disease Control and Prevention (2020) National Diabetes Statistics
Report, 2020. Washington, DC: U.S. Departmemnt of Health and Human
Services.
Cersosimo MG, Koller WC (2004) Essential tremor. In: Watts RL, Koller WC (Eds.)
Movement disorders. Neurological principles and practice, pp. 431-457. New
York: McGraw-Hill.
Chan CWY, Kearney RE (1982) Is the functional stretch reflex servo controlled or
preprogrammed? Electroencephalog Clin Neurophysiol 53: 310-324.
Chan CWY, Melvill Jones G, Kearney RE, Watt DGD (1979). The late
electromyographic response to limb displacement in man. I. Evidence for
supraspinal contribution. Electroencephalog Clin Neurophysiol 46: 173-181.
Chan KM, Raja AJ, Strohschein FJ, Lechelt K (2000) Age-related changes in
muscle fatigue resistance in humans. Can J Neurol Sci 27: 220-228.
Chaudhary U, Birbaumer N, Ramos-Murguialday A (2016) Brain-computer
interfaces for communication and rehabilitation. Nat Rev Neurol 12(9): 513-525.
Chen RS, Tsai CH, Lu CS (1995) Reciprocal inhibition in writer’s cramp. Mov
Disord 10: 556-561.
Cheney PD, Fetz EE (1984) Corticomotoneuronal cells contribute to long-latency
stretch reflexes in the rhesus monkey. J Physiol 349: 249-272.
Chiang H, Slobounov SM, Ray W (2004) Practice-related modulations of force
enslaving and cortical activity as revealed by EEG. Clin Neurophysiol 115:
1033-1043.
Christova P, Kossev A (1998) Motor unit activity during long-lasting intermittent
muscle contractions in humans. Eur J Appl Physiol Occup Physiol 77: 379-387.
Cioni M, Cocilovo A, Di Pasquale F, Araujo MB, Siqueira CR, Bianco M (1994)
Strength deficit of knee extensor muscles of individuals with Down syndrome
from childhood to adolescence. Amer J Ment Retard 99: 166-174.
Cirstea MC, Levin MF (2000) Compensatory strategies for reaching in stroke.
Brain 123: 940-953.
Classen J, Liepert J, Wise SP, Hallett M, Cohen LG (1998) Rapid plasticity of
human cortical movement representation induced by practice. J Neurophysiol
79: 1117-1123.
Cody FWJ, MacDermott N, Matthews PBC, Richardson HC (1986) Observations
on the genesis of the stretch reflex in Parkinson’s disease. Brain 109: 229-249.
Cohen LG, Bandinelli S, Findley TW, Hallett M (1991a) Motor reorganization after
upper limb amputation in humans: A study with focal magnetic stimulation. Brain
114: 615-627.
Cohen LG, Bandinelli S, Topka HR, Fuhr P, Roth BJ, Hallett M (1991b)
Topographic maps of human motor cortex in normal and pathological
conditions: Mirror movements, amputations and spinal cord injuries.
Electroencephalogr Clin Neurophysiol Suppl 43: 36-50.
Cohen LG, Celnik P, Pascual-Leone A, Corwell B, Falz L, Dambrosia J, Honda M,
Sadato N, Gerloff C, Catala MD, Hallett M (1997) Functional relevance of cross-
modal plasticity in blind humans. Nature 389: 180-183.
Colcher A, Hurtig HL (2004) Systemic illnesses that cause movement disorders.
In: Watts RL, Koller WC (Eds.) Movement disorders: Neurologic principles and
practice. Second edition, pp. 915-926. New York: McGraw-Hill.
Cole J, Paillard J. (1995) Living without touch and peripheral information about
body position and movement: Studies with deafferented subjects. In: Bermudes
JL, Marcel A, Eilan N (Eds.) The body and the self, pp. 245-266. Cambridge,
MA: MIT Press.
Cole KJ (1991) Grasp force control in older adults. J Mot Behav 23: 251-258.
Cole KJ, Abbs JH (1987) Kinematic and electromyographic responses to
perturbation of a rapid grasp. J Neurophysiol 57: 1498-1510.
Cole KJ, Rotella DL, Harper JG (1998) Tactile impairments cannot explain the
effect of age on a grasp and lift. Exp Brain Res 121: 263-269.
Cole KJ, Rotella DL, Harper JG (1999) Mechanisms for age-related changes of
fingertip forces during precision gripping and lifting in adults. J Neurosci 19:
3238-3247.
Colebatch JG, Gandevia SC (1989) The distribution of muscular weakness in
upper motor neuron lesions affecting the arm. Brain 112: 749-763.
Collins JJ, De Luca CJ (1993) Open-loop and closed-loop control of posture: A
random-walk analysis of center-of-pressure trajectories. Exp Brain Res 95: 308-
318.
Conforto AB, Kaelin-Lang A, Cohen LG (2002) Increase in hand muscle strength
of stroke patients after somatosensory stimulation. Ann Neurol 51: 122-125.
Connolly BH (2001) Aging in individuals with lifelong disabilities. Phys Occup Ther
Pediatr 21: 23-47.
Consalez GG, Goldowitz D, Casoni F, Hawkes R (2021) Origins, Development,
and Compartmentation of the Granule Cells of the Cerebellum. Frontiers in
Neural Circuits 14.
Contreras-Vidal JL, Grossberg S, Bullock D (1997) A neural model of cerebellar
learning for arm movement control: Cortico-spino-cerebellar dynamics. Learn
Mem 3: 475-502.
Contreras-Vidal JL, Teulings HL, Stelmach GE (1998) Elderly subjects are
impaired in spatial coordination in fine motor control. Acta Psychol (Amst) 100:
25-35.
Cooke DW, Thelen E (1987) Newborn stepping: A review of puzzling infant co-
ordination. Dev Med Child Neurol 29: 399-404.
Cooper SE, Johnson DS, Montgomery EB Jr (2004) Pathophysiology of cerebellar
disorders. In: Watts RL, Koller WC (Eds.) Movement disorders. Neurological
principles and practice, pp. 737-760. New York: McGraw-Hill.
Corcos DM, Gottlieb GL, Agarwal GC (1989) Organizing principles for single joint
movements. II. A speed-sensitive strategy. J Neurophysiol 62: 358-368.
Corcos DM, Gottlieb GL, Agarwal GC, Flaherty BP (1990) Organizing principles for
single joint movements. IV. Implications for isometric contractions. J
Neurophysiol 64: 1033-1042.
Corcos DM, Gottlieb GL, Penn RD, Myklebust B, Agarwal GC (1986) Movement
deficits caused by hyperexcitable stretch reflexes in spastic humans. Brain 109:
1043-1058.
Cordo PJ, Nashner LM (1982) Properties of postural adjustments associated with
rapid arm movements. J Neurophysiol 47: 1888-1905.
Cossu G, Pau M (2017) Subthalamic nucleus stimulation and gait in Parkinson’s
disease: A not always fruitful relationship. Gait Posture 52: 205-210.
Côté JN, Feldman AG, Mathieu PA, Levin MF (2008) Effects of fatigue on
intermuscular coordination during repetitive hammering. Motor Control 12: 79-
92.
Côté JN, Mathieu PA, Levin MF, Feldman AG (2002) Movement reorganization to
compensate for fatigue during sawing. Exp Brain Res 146: 394-398.
Courchesne E (1997) Brainstem, cerebellar and limbic neuroanatomical
abnormalities in autism. Curr Opin Neurobiol 7: 269-278.
Cramer SC, Nelles G, Benson RR, Kaplan JD, Parker RA, Kwong KK (1997) A
functional MRI study of subjects recovered from hemiparetic stroke. Stroke 28:
2518-2527.
Craske B (1977) Perception of impossible limb positions induced by tendon
vibration. Science 196: 71-73.
Crivello F, Mazoyer B (1999) Positron emission tomography of the human brain. In:
Windhorst U, Johansson H (Eds.) Modern techniques in neuroscience research,
pp. 1083-1098. Berlin: Springer-Verlag.
Crome LC, Stern J (1967) Pathology of mental retardation. London: Churchill.
Crowninshield RD, Brand RA (1981) A physiologically based criterion of muscle
force prediction in locomotion. J Biomech 14: 793-801.
Cruse H, Bruwer M (1987) The human arm as a redundant manipulator: The
control of path and joint angles. Biol Cybern 57: 137-144.
Crutcher MD, DeLong MR (1984) Single cell studies of the primate putamen, Parts
1 and 2. Exp Brain Res 53: 233-258.
Cuadra C, Corey J, Latash ML (2021) Distortions of the efferent copy during force
perception: A study of force drifts and effects of muscle vibration. Neuroscience
457: 139-154.
Cuadra C, Latash ML (2019) Exploring the concept of iso-perceptual manifold
(IPM): A study of finger force matching tasks. Neuroscience 401: 130-141.
Cuadra C, Wojnicz W, Kozinc Z, Latash ML (2020) Perceptual and motor effects of
muscle co-activation in a force production task. Neuroscience 437: 34-44.
Cullen, K (2018) Multisensory Integration and the Perception of Self-Motion.
Oxford Research Encyclopedia of Neuroscience.
Cullen KE (2019) Vestibular processing during natural self-motion: implications for
perception and action. Nat Rev Neurosci 20(6): 346–363.
Czerniecki JM, Gitter A, Munro C (1991) Joint moment and muscle power output
characteristics of below knee amputees during running: The influence of energy
storing prosthetic feet. J Biomech 24: 63-75.
Danckert J, Ferber S, Doherty T, Steinmetz H, Nicolle D, Goodale MA (2002)
Selective, non-lateralized impairment of motor imagery following right parietal
damage. Neurocase 8: 194-204.
D’Angelo E (2005) Synaptic plasticity at the cerebellum input stage: Mechanisms
and functional implications. Arch Ital Biol 143: 143-156.
Danion F, Latash ML, Li S (2003a) Finger interactions studied with transcranial
magnetic stimulation during multi-finger force production tasks. Clin
Neurophysiol 114: 1445-1455.
Danion F, Latash ML, Li Z-M, Zatsiorsky VM (2000) The effects of fatigue on multi-
finger coordination in force production tasks. J Physiol 523: 523-532.
Danion F, Latash ML, Li Z-M, Zatsiorsky VM (2001) The effect of a fatiguing
exercise by the index finger on single- and multi-finger force production tasks.
Exp Brain Res 138: 322-329.
Danion F, Schöner G, Latash ML, Li S, Scholz JP, Zatsiorsky VM (2003b) A force
mode hypothesis for finger interaction during multi-finger force production tasks.
Biol Cybern 88: 91-98.
Danna-Dos-Santos A, Slomka K, Zatsiorsky VM, Latash ML (2007) Muscle modes
and synergies during voluntary body sway. Exp Brain Res 179: 533-550.
d’Avella A, Saltiel P, Bizzi E (2003) Combinations of muscle synergies in the
construction of a natural motor behavior. Nat Neurosci 6: 300-308.
Davids K, Bennett S, Newell KM (2005) Movement system variability. Champaign,
IL: Human Kinetics.
Day BL, Riescher H, Struppler A, Rothwell JC, Marsden CD (1991) Changes in the
response to magnetic and electrical stimulation of the motor cortex following
muscle stretch in man. J Physiol 433: 41-57.
Deecke L, Scheid P, Kornhuber HH (1969) Distribution of readiness potential, pre-
motor positivity and motor potential of the human cerebral cortex preceding
voluntary finger movements. Exp Brain Res 7: 158-168.
De Freitas PB, Freitas SMSF, Lewis MM, Huang X, Latash ML (2019) Individual
preferences in motor coordination seen across the two hands: Relations to
movement stability and optimality. Exp Brain Res 237: 1-13.
DeLong MR (1999) The basal ganglia. In: Kandel ER, Schwartz JH, Jessell TM
(Eds.) Principles of neural science. Fourth edition, pp. 853-867. New York:
McGraw-Hill.
DeLong MR, Crutcher MD, Georgopoulos AP (1985) Primate globus pallidus and
subthalamic nucleus: Functional organisation. J Neurophysiol 53: 530-543.
DeLong MR, Georgopoulos AP (1979) Motor function of basal ganglia as revealed
by studies of single cell activity in the behaving primate. Adv Neurol 24: 131-
140.
DeLong MR, Georgopoulos AP (1981) Motor functions of the basal ganglia. In:
Brooks VB (Ed.) Handbook of physiology, sect. 1, vol. 2, part 2, pp. 1017-1061.
Baltimore: Williams & Wilkins.
De Luca CJ, Chang SS, Roy SH, Kline JC & Nawab SH (2015) Decomposition of
surface EMG signals from cyclic dynamic contractions. J Neurophysiol 113:
1941-1951.
Dengler R, Konstanzer A, Gillespie J, Argenta M, Wolf W, Struppler A (1990)
Behavior of motor units in Parkinsonism. Adv Neurol 53: 167-173.
De Schutter E, Maex R (1996) The cerebellum: Cortical processing and theory.
Curr Opin Neurobiol 6: 759-764.
Diedrichsen J, Criscimagna-Hemminger SE, Shadmehr R. (2007) Dissociating
timing and coordination as functions of the cerebellum. J Neurosci, 27(23),
6291-6301.
Diekmann V, Erne SN, Becker W (1999) Magnetoencephalography. In: Windhorst
U, Johansson H (Eds.) Modern techniques in neuroscience research, pp. 1025-
1054, Berlin: Springer-Verlag.
Diener HC, Bacher M, Guschlbauer B, Thomas C, Dichgans J (1993) The
coordination of posture and voluntary movement in patients with hemiparesis. J
Neurol 240: 161-167.
Diener HC, Dichgans J (1992) Pathophysiology of cerebellar ataxia. Mov Disord 7:
95-109.
Diener HC, Dichgans J, Bacher M, Gompf B (1984) Quantification of postural sway
in normals and patients with cerebellar diseases. Electroencephalogr Clin
Neurophysiol 57: 134-142.
Dietz V (2000) Spastic movement disorder. Spinal Cord 38: 389-393.
Dietz V, Berger W (1984) Interlimb coordination of posture in patients with spastic
paresis. Impaired function of spinal reflexes. Brain 107: 965-978.
Dietz V, Quintern J, Berger W (1981) Electrophysiological studies of gait in
spasticity and rigidity. Evidence that altered mechanical properties of muscle
contribute to hypertonia. Brain 104: 431-449.
Dietz V, Quintern J, Berger W (1984) Corrective reactions to stumbling in man:
Functional significance of spinal and transcortical reflexes. Neurosci Lett 44:
131-135.
Dijkstra TM, Schöner G, Giese MA, Gielen CCAM (1994) Frequency dependence
of the action-perception cycle for postural control in a moving visual
environment: Relative phase dynamics. Biol Cybern 71: 489-501.
Dimitrijevic MR, Nathan PW, Sherwood AM (1980) Clonus: The role of central
mechanisms. J Neurol Neurosurg Psychiatr 43: 321-332.
Dizio P, Lackner JR (1995) Motor adaptation to Coriolis force perturbations of
reaching movements: Endpoint but not trajectory adaptation transfers to the
nonexposed arm. J Neurophysiol 74: 1787-1792.
Doherty TJ, Brown WF (1997) Age-related changes in the twitch contractile
properties of human thenar motor units. J Appl Physiol 82: 93-101.
Dominici N, Ivanenko YP, Cappellini G, et al. (2011) Locomotor primitives in
newborn babies and their development. Science 334: 997-999.
Domkin D, Laczko J, Djupsjöbacka M, Jaric S, Latash ML (2005) Joint angle
variability in 3D bimanual pointing: Uncontrolled manifold analysis. Exp Brain
Res 163: 44-57.
Domkin D, Laczko J, Jaric S, Johansson H, Latash ML (2002) Structure of joint
variability in bimanual pointing tasks. Exp Brain Res 143: 11-23.
Dounskaia N, Ketcham CJ, Leis BC, Stelmach GE (2005) Disruptions in joint
control during drawing arm movements in Parkinson’s disease. Exp Brain Res
164: 311-322.
Dralle D, Muller H, Zierski J, Klug N (1985) Intrathecal baclofen for spasticity.
Lancet 2(8462): 1003.
Dreher JC, Grafman J (2002) The roles of the cerebellum and basal ganglia in
timing and error prediction. Eur J Neurosci 16: 1609-1619.
Dreissen YE, Tijssen MA (2012) The startle syndromes: Physiology and treatment.
Epilepsia 53: 3-11.
Duchateau J, Balestra C, Carpentier A, Hainaut K (2002) Reflex regulation during
sustained and intermittent submaximal contractions in humans. J Physiol 541:
959-967.
Duchateau J, Hainaut K. (1990) Effects of immobilization on contractile properties,
recruitment and firing rates of human motor units. J Physiol 422: 55-65.
Duffell LD, Donaldson NN (2020) A comparison of FES and SCS for neuroplastic
recovery after SCI: Historical perspectives and future directions. Front Neurol
11:607.
Dukelow SP, DeSouza JF, Culham JC, van den Berg AV, Menon RS, Vilis T (2001)
Distinguishing subregions of the human MT+ complex using visual fields and
pursuit eye movements. J Neurophysiol 86(4): 1991-2000.
Duque J, Hummel F, Celnik P, Murase N, Mazzocchio R, Cohen LG (2005)
Transcallosal inhibition in chronic subcortical stroke. Neuroimage 28: 940-946.
Duque J, Greenhouse I, Labruna L, Ivry RB (2017) Physiological markers of motor
inhibition during human behavior. Trends Neurosci, 40(4), 219-236.
Duysens J, Pearson KG (1976) The role of cutaneous afferents from the distal
hindlimb in the regulation of the stepcycle in thalamic cats. Exp Brain Res 24:
245-255.
Duysens J, Trippel M, Horstmann GA, Dietz V (1990) Gating and reversal of
reflexes in ankle muscles during human walking. Exp Brain Res 82: 351-358.
Edwards JM, Elliott D (1989) Asymmetries in intermanual transfer of training and
motor overflow in adults with Down’s syndrome and nonhandicapped children. J
Clin Exp Neuropsychol 11: 959-966.
Eisen A, Entezari-Taher M, Stewart H (1996) Cortical projections to spinal
motoneurons: Changes with aging and amyotrophic lateral sclerosis. Neurology
46: 1396-1404.
Eklund G (1969) Influence of muscle vibration on balance in man. A preliminary
report. Acta Soc Med Ups 74: 113-117.
Eklund G, Hagbarth KE (1966) Normal variability of tonic vibration reflexes in man.
Exp Neurol 16: 80-92.
Eklund G, Hagbarth KE (1967) Vibratory induced motor effects in normal man and
in patients with spastic paralysis. Electroencephalogr Clin Neurophysiol 23: 393.
Elble RJ, Moody C, Leffler K, Sinha R (1994) The initiation of normal walking. Mov
Disord 2: 139-146.
Ellaway PH, Davey NJ, Ljubisavljevic M (1999) Magnetic stimulation of the
nervous system. In: Windhorst U, Johansson H (Eds.) Modern techniques in
neuroscience research, pp. 869-892. Berlin: Springer-Verlag.
English AW (1984) An electromyographic analysis of compartments in cat lateral
gastrocnemius muscle during unrestrained locomotion. J Neurophysiol 52: 114-
125.
Enoka RM, Christou EA, Hunter SK, Kornatz KW, Semmler JG, Taylor AM, Tracy
BL (2003) Mechanisms that contribute to differences in motor performance
between young and old adults. J Electromyogr Kinesiol 13: 1-12.
Ernst ME, Klepser ME, Fouts M, Marangos MN (1997) Tetanus: Pathophysiology
and management. Ann Pharmacother 31: 1507-1513.
Evarts EV (1968) Relation of pyramidal tract activity to force exerted during
voluntary movement. J Neurophysiol 31: 14-27.
Evarts EV, Teravainen H, Beuchert D, Calne DB (1979) Pathophysiology of motor
performance in Parkinson’s disease. In: Fuxe K, Calne D (Eds) Ergot
Derivatives and Motor Function, pp. 45-59, London: Pergamon Press
Evarts EV, Teravainen H, Calne DB (1981) Reaction time in Parkinson’s disease.
Brain 104: 167-186.
Fadiga L, Craighero L, Olivier E (2005) Human motor cortex excitability during the
perception of others’ action. Curr Opin Neurobiol 15: 213-218.
Falaki A, Huang X, Lewis MM, Latash ML (2016) Impaired synergic control of
posture in Parkinson’s patients without postural instability. Gait Posture 44: 209-
215.
Falaki A, Huang X, Lewis MM, Latash ML (2017) Dopaminergic modulation of
multi-muscle synergies in postural tasks performed by patients with Parkinson’s
disease. J Electromyogr Kinesiol 33: 20-26.
Falaki A, Jo HJ, Lewis MM, O’Connell B, De Jesus S, McInerney J, Huang X,
Latash ML (2018) Systemic effects of deep brain stimulation on synergic control
in Parkinson’s disease. Clin Neurophysiol 129: 1320-1332.
Farina D, Holobar A, Merletti R, Enoka RM (2010) Decoding the neural drive to
muscles from the surface electromyogram. Clin Neurophysiol 121: 1616-1623.
Farina D, Merletti R, Enoka RM (2004) The extraction of neural signals from the
surface EMG. J Appl Physiol 96: 1486-1495.
Farina D, Merletti R, Enoka RM (2014) The extraction of neural strategies from the
surface EMG: An update. J Appl Physiol 117: 1215-1230.
Farivar, R. (2009). Dorsal–ventral integration in object recognition. Brain Res Rev
61(2): 144-153.
Feldman AG (1966) Functional tuning of the nervous system with control of
movement or maintenance of a steady posture. II. Controllable parameters of
the muscle. Biophysics 11: 565-578.
Feldman AG (1980) Superposition of motor programs. I. Rhythmic forearm
movements in man. Neuroscience 5: 81-90.
Feldman AG (1986) Once more on the equilibrium-point hypothesis (λ-model) for
motor control. J Mot Behav 18: 17-54.
Feldman AG (2009) New insights into action-perception coupling. Exp Brain Res
194: 39-58.
Feldman AG (2016) Active sensing without efference copy: Referent control of
perception. J Neurophysiol 116: 960-976.
Feldman AG (2019) Indirect, referent control of motor actions underlies directional
tuning of neurons. J Neurophysiol 121: 823-841.
Feldman AG, Latash ML (1982) Inversions of vibration-induced senso-motor
events caused by supraspinal influences in man. Neurosci Lett 31: 147-151.
Feldman AG, Latash ML (2005) Testing hypotheses and the advancement of
science: Recent attempts to falsify the equilibrium-point hypothesis. Exp Brain
Res 161: 91-103.
Feldman AG, Levin MF (1995) Positional frames of reference in motor control:
Their origin and use. Behav Brain Sci 18: 723-806.
Feldman AG, Orlovsky GN (1972) The influence of different descending systems
on the tonic stretch reflex in the cat. Exp Neurol 37: 481-494.
Findlater SE, Dukelow SP (2017) Upper extremity proprioception after stroke:
Bridging the gap between neuroscience and rehabilitation. J Mot Behav 49(1):
27-34.
Fink GR (2004) Functional MR imaging: From the BOLD effect to higher motor
cognition. Suppl Clin Neurophysiol 57: 458-468.
Fitts RH, Courtright JB, Kim DH, Witzmann FA (1982) Muscle fatigue with
prolonged exercise: Contractile and biochemical alterations. Am J Physiol 242:
C65-C73.
Flament D, Hore J (1986) Movement and electromyographic disorders associated
with cerebellar dysmetria. J Neurophysiol, 55(6), 1221-1233.
Flash T (1987) The control of hand equilibrium trajectories in multi-joint arm
movements. Biol Cybern 57: 257-274.
Flash T, Hogan N (1985) The coordination of arm movements: An experimentally
confirmed mathematical model. J Neurosci 5: 1688-1703.
Fleckenstein JL, Watumull D, Bertocci LA, Parkey RW, Peshock RM. (1992)
Finger-specific flexor recruitment in humans: Depiction by exercise-enhanced
MRI. J Appl Physiol 72: 1974-1977.
Flowers KA (1976) Visual « closed-loop » and « open-loop » characteristics of
voluntary movements in patients with Parkinsonism and intention tremor. Brain
99: 269-310.
Fooken, J, Kreyenmeier, P, Spering, M (2021). The role of eye movements in
manual interception: A mini-review. Vision Research, 183: 81-90.
Forssberg H, Grillner S, Rossignol S (1975) Phase dependent reflex reversal
during walking in chronic spinal cat. Brain Res 85: 103-107.
Forssberg H, Grillner S, Rossignol S (1977). Phasic gain control of reflexes from
the dorsum of the paw during spinal locomotion. Brain Res 132: 121-139.
Fortier PA, Kalaska JF, Smith AM (1989) Cerebellar neuronal activity related to
whole-arm reaching movements in the monkey. J Neurophysiol 62: 198-211.
Frahm J, Fransson P, Krüger G (1999) Magnetic resonance imaging of the human
brain function. In: Windhorst U, Johansson H (Eds.) Modern techniques in
neuroscience research, pp. 1055-1082. Berlin: Springer-Verlag.
Francis KL, Spirduso WW (2000) Age differences in the expression of manual
asymmetry. Exp Aging Res 26: 169-180.
Freal JE, Kraft GH, Coryell JK (1984) Symptomatic fatigue in multiple sclerosis.
Arch Phys Med Rehab 65: 135-138.
Freitas SMSF, Duarte M, Latash ML (2006) Two kinematic synergies in voluntary
whole-body movements during standing. J Neurophysiol 95: 636-645.
Fritz SL, Butts RJ, Wolf SL (2012) Constraint-induced movement therapy: From
history to plasticity. Expert Rev Neurother 12(2): 191-198.
Fuglevand AJ, Zackowski KM, Huey KA, Enoka RM (1993) Impairment of
neuromuscular propagation during human fatiguing contractions at submaximal
forces. J Physiol 460: 549-572.
Fuhr P, Cohen LG, Dang N, Findley TW, Haghighi S, Oro J, Hallett M (1992)
Physiological analysis of motor reorganization following lower limb amputation.
Electroencephalog Clin Neurophysiol 85: 53-60.
Fujita T, Nakamura S, Ohue M, Fujii Y, Miyauchi A, Takagi Y, Tsugeno H (2005)
Effect of age on body sway assessed by computerized posturography. J Bone
Miner Metab 23: 152-156.
Fukson OI, Berkinblit MB, Feldman AG (1980) The spinal frog takes into account
the scheme of its body during the wiping reflex. Science 209: 1261-1263.
Gale DJ, Areshenkoff CN, Honda C, Johnsrude IS, Flanagan JR, Gallivan JP
(2021) Motor planning modulates neural activity patterns in early human
auditory cortex. Cereb Cortex 31(6): 2952-2967.
Gale DJ, Flanagan JR, Gallivan JP (2021) Human somatosensory cortex is
modulated during motor planning. J Neurosci 41(27): 5909-5922.
Galganski ME, Fuglevand AJ, Enoka RM (1993) Reduced control of motor output
in a human hand muscle of elderly subjects during submaximal contractions. J
Neurophysiol 69: 2108-2115.
Gandevia SC (2001) Spinal and supraspinal factors in human muscle fatigue.
Physiol Rev 81: 1725-1789.
Gandevia SC, McCloskey DI, Potter EK (1980) Alterations in perceived heaviness
during digital anaesthesia. J Physiol 306: 365-375.
Garcia-Rill E, Skinner R (1987) The mesencephalic locomotor region. II.
Projections to reticulospinal neurons. Brain Res, 411(1), 13-20.
Garland SJ, Enoka RM, Serrano LP, Robinson GA (1994) Behavior of motor units
in human biceps brachii during a submaximal fatiguing contraction. J Appl
Physiol 76: 2411-2419.
Gelfand IM (1991) Two archetypes in the psychology of man. Nonlinear Sci Today
1: 11-16.
Gelfand IM, Latash ML (2002) On the problem of adequate language in biology. In:
Latash ML (Ed.) Progress in motor control. Vol. 2: Structure-function relations in
 voluntary movement, pp. 209-228. Champaign, IL: Human Kinetics.
Gelfand IM, Tsetlin ML (1966) On mathematical modeling of the mechanisms of
the central nervous system. In: Gelfand IM, Gurfinkel VS, Fomin SV, Tsetlin ML
(Eds.) Models of the structural-functional organization of certain biological
systems, pp. 9-26. Nauka: Moscow (in Russian; a translation is available in
1971 edition by MIT Press, Cambridge MA).
Georgopoulos AP (1986) On reaching. Ann Rev Neurosci 9: 147-170.
Georgopoulos AP, Kalaska JF, Caminiti R, Massey JT (1982) On the relations
between the direction of two-dimensional arm movements and cell discharge in
primate motor cortex. J Neurosci 2: 1527-1537.
Georgopoulos AP, Schwartz AB, Kettner RE (1986) Neural population coding of
movement direction. Science 233: 1416-1419.
Georgopoulos AP, Lurito JT, Petrides M, Schwartz AB, Massey JT (1989) Mental
rotation of the neuronal population vector. Science 243: 234-236.
Geuze RH (2005) Postural control in children with developmental coordination
disorder. Neural Plast 12: 183-196.
Ghez C, Gordon J (1987) Trajectory control in targeted force impulses. I. Role of
opposing muscles. Exp Brain Res 67: 225-240.
Ghez C, Shinoda Y (1978) Spinal mechanisms of the functional stretch reflex. Exp
Brain Res, 32: 55-68
Giampaoli S, Ferrucci L, Cecchi F, Lo Noce C, Poce A, Santaquilani A, Vescio MF,
Menotti A (1999) Hand-grip strength predicts insident disability in non-disabled
older men. Age Ageing 28: 283-288.
Gibson AR, Houk JC, Kohlerman NJ (1985) Relation between red nucleus
discharge and movement parameters in trained macaque monkeys. J Physiol
358: 551-570.
Gibson JJ (1979) The ecological approach to visual perception. Boston, MA:
Houghton Mifflin.
Gielen CCAM, van der Oosten K, Pull ter Gunne F (1985) Relations between EMG
activation patterns and kinematic properties of aimed movements. J Mot Behav
17: 421-442.
Giladi N (2001) Gait disturbances in advanced stages of Parkinson’s disease. Adv
Neurol 86: 273-278.
Gilbert CD, Wiesel TN (1979) Morphology and intracortical projections of
functionally characterised neurones in the cat visual cortex. Nature 280: 120-
125.
Gillberg C, Kadesjo B (2003) Why bother about clumsiness? The implications of
having developmental coordination disorder (DCD). Neural Plast 10: 59-68.
Gilles MA, Wing AM (2003) Age-related changes in grip force and dynamics of
hand movement. J Mot Behav 35: 79-85.
Gilman S (1969) The mechanism of cerebellar hypotonia: An experimental study in
the monkey. Brain 92: 621-638.
Gilman S (2004) Clinical features and treatment of cerebellar disorders. In: Watts
RL, Koller WC (Eds.) Movement disorders. Neurological principles and practice,
pp. 723-736. New York: McGraw-Hill.
Gilman S, Bloedel JR, Lechtenberg R (1981) Disorders of the cerebellum.
Philadelphia: FA Davis.
Giszter SF (2015) Motor primitives – new data and future questions. Curr Opin
Neurobiol 33: 156-165.
Giszter SF, Mussa-Ivaldi FA, Bizzi E (1993) Convergent force fields organized in
the frog’s spinal cord. J Neurosci 13: 467-491.
Glansdorf P, Prigogine I (1971) Thermodynamic theory of structures, stability and
fluctuations. New York: Wiley.
Goetz CC, Horn S (2004) Tardive dyskinesia. In: Watts RL, Koller WC (Eds.)
Movement disorders. Neurological principles and practice, pp. 629-638. New
York: McGraw-Hill.
Gogolla, N. (2017). The insular cortex. Curr Biol 27(12): R580-R586.
Golan D, Staun-Ram E, Miller A (2016) Shifting paradigms in multiple sclerosis.
Curr Opin Neurol 29: 354-361.
Goldberger ME (1977) Locomotor recovery after unilateral hindlimb deafferentation
in cats. Brain Res 123: 59-74.
Golenia L, Schoemaker MM, Otten E, Mouton LJ, Bongers RM (2018) Developing
of reaching during mid-childhood from a Developmental Systems perspective.
PLoS One 13(2): e0193463.
Goodale MA, Milner AD (1992) Separate visual pathways for perception and
action. Trends Neurosci 15: 20-25.
Goodale MA, Milner AD, Jakobson LS, Carey DP (1991). A neurological
dissociation between perceiving objects and grasping them. Nature 349: 154-
156.
Goodkin HP, Thach WT (2003) Cerebellar control of constrained and
unconstrained movements. II. EMG and nuclear activity. J Neurophysiol 89:
896-908.
Goodwin GM, McCloskey DI, Matthews PB (1972) The contribution of muscle
afferents to kinaesthesia shown by vibration induced illusions of movement and
by the effects of paralysing joint afferents. Brain 95: 705-748.
Gordon AM, Ingvarsson PE, Forssberg H (1997) Anticipatory control of
manipulative forces in Parkinson’s disease. Exp Neurol 145: 477-488
Gordon J, Ghez C (1987) Trajectory control in targeted force impulses. II. Pulse
height control. Exp Brain Res 67: 241-252.
Gorter JW, Rosenbaum PL, Hanna SE, Palisano RJ, Bartlett DJ, Russell DJ,
Walter SD, Raina P, Galuppi BE, Wood E (2004) Limb distribution, motor
impairment, and functional classification of cerebral palsy. Dev Med Child
Neurol 46: 461-467.
Gottlieb GL (1996) On the voluntary movement of compliant (inertial-viscoelastic)
loads by parcellated control mechanisms. J Neurophysiol 76: 3207-3229.
Gottlieb GL, Agarwal GC (1972) The role of the myotatic reflex in the voluntary
control of movements. Brain Res 40: 139-143.
Gottlieb GL, Agarwal GC (1980) Response to sudden torques about ankle in man.
II. Postmyotatic reactions. J Neurophysiol 43: 86-101.
Gottlieb GL, Agarwal GC (1988) Compliance of single joints: Elastic and plastic
characteristics. J Neurophysiol 59: 937-951.
Gottlieb GL, Corcos DM, Agarwal GC (1989a) Strategies for the control of
voluntary movements with one mechanical degree of freedom. Behav Brain Sci
12: 189-250.
Gottlieb GL, Corcos DM, Agarwal GC (1989b) Organizing principles for single joint
movements. I: A speed-insensitive strategy. J Neurophysiol 62: 342-357.
Gottlieb GL, Corcos DM, Agarwal GC, Latash ML (1990) Organizing principles for
single joint movements. III: Speed-insensitive strategy as a default. J
Neurophysiol 63: 625-636.
Graham Brown T (1914) On the nature of the fundamental activity of the nervous
centres; together with an analysis of the conditioning of the rhythmic activity in
progression and a theory of the evolution of function in the nervous system. J
Physiol 48: 18-46.
Grahn JA, Brett M (2007) Rhythm and beat perception in motor areas of the brain.
J Cognitive Neurosci 19(5): 893-906.
Grange RW, Houston ME (1991) Simultaneous potentiation and fatigue in
quadriceps after a 60-second maximal voluntary isometric contraction. J Appl
Physiol 70: 726-731.
Grant CM, Grayson A, Boucher J (2001) Using tests of false belief with children
with autism: How valid and reliable are they? Autism 5: 135-145.
Granzier HL, Labeit S (2004) The giant protein titin: A major player in myocardial
mechanics, signaling, and disease. Circ Res 94: 284-295.
Granzier HL, Labeit S. (2005) Titin and its associated proteins: The third
myofilament system of the sarcomere. Adv Protein Chem 71: 89-119.
Graybiel AM (1995) Building action repertoires: Memory and learning functions of
the basal ganglia. Curr Opin Neurobiol 5: 733-741.
Graybiel AM (1997) The basal ganglia and cognitive pattern generators. Schizophr
Bull 23: 459-469.
Graziano MS, Aflalo TN, Cooke DF (2005) Arm movements evoked by electrical
stimulation in the motor cortex of monkeys. J Neurophysiol 94: 4209-4223.
Graziano MSA, Taylor CSR, and Moore T (2002) Complex movements evoked by
microstimulation of precentral cortex. Neuron 34: 841-851.
Gribble PL, Ostry DJ (2000) Compensation for loads during arm movements using
equilibrium-point control. Exp Brain Res 135: 474-482.
Gribble PL, Ostry DJ, Sanguineti V, Laboissiere R (1998) Are complex control
signals required for human arm movements? J Neurophysiol 79: 1409-1424.
Grigg P, Finerman GA, Riley LH (1973) Joint-position sense after total hip
replacement. J Bone Joint Surg Am 55: 1016-1025.
Grillner S (1975) Locomotion in vertebrates: Central mechanisms and reflex
interaction. Physiol Rev 55: 247-304.
Grillner S, Wallen P (1985) Central pattern generators for locomotion, with special
reference to vertebrates. Ann Rev Neurosci 8: 233-261.
Grillner S, Zangger P (1975) How detailed is the central pattern generator for
locomotion? Brain Res 88: 367-371.
Grimby G, Danneskold-Samsoe B, Hvid K, Saltin B (1982) Morphology and
enzymatic capacity in arm and leg muscles in 78-81 year old men and women.
Acta Physiol Scand 115: 125-134.
Gurfinkel VS, Levik YS, Kazennikov OV, Selionov VA (1998) Locomotor-like
movements evoked by leg muscle vibration in humans. Eur J Neurosci 10:
1608-1612.
Gusella JF, MacDonald ME (2004) Genetics and molecular biology of Huntington’s
disease. In: Watts RL, Koller WC (Eds.) Movement disorders. Neurological
principles and practice, pp. 571-588. New York: McGraw-Hill.
Gutteling T P, Petridou N, Dumoulin SO, Harvey BM, Aarnoutse EJ, Kenemans JL,
Neggers SFW (2015) Action preparation shapes processing in early visual
cortex. J Neurosci 35(16): 6472-6480.
Gutman AM (1991) Bistability of dendrites. Int J Neural Syst 1: 291-304.
Hadjikhani N, Liu AK, Dale AM, Cavanagh P, Tootell RBH (1998) Retinotopy and
color sensitivity in human visual cortical area V8. Nat Neurosci 1(3): 235-241.
Hagbarth K-E, Bongiovanni LG, Nordin M (1995) Reduced servo-control of
fatigued human finger extensor and flexor muscles. J Physiol 485: 865-872.
Hagbarth K-E, Kunesch EJ, Nordin M, Schmidt R, Wallin EU (1986) Gamma loop
contributing to maximal voluntary contraction in man. J Physiol 380: 575-591.
Haken H (1977) Synergetics: An introduction. Nonequilibrium phase transitions
and self-organization in physics, chemistry, and biology. Berlin: Springer-Verlag.
Haken H, Kelso JAS, Bunz H (1985) A theoretical model of phase transitions in
human hand movements. Biol Cybern 51: 347-356.
Hakkinen K, Newton RU, Gordon SE, McCormick M, Volek JS, Nindl BC, Gotshalk
LA, Campbell WW, Evans WJ, Hakkinen A, Humphries BJ, Kraemer WJ (1998)
Changes in muscle morphology, electromyographic activity, and force
production characteristics during progressive strength training in young and
older men. J Gerontol Ser A, Biol Sci Med Sci 53: B415-B423.
Haley SM (1986) Postural reactions in infants with Down syndrome. Relationship
to motor milestone development and age. Phys Ther 66: 17-22.
Hallett M (2001) Plasticity of the human motor cortex and recovery from stroke.
Brain Res Rev 36: 169-174.
Hallett M, Khoshbin S (1980) A physiological mechanism of bradykinesia. Brain
103: 301-314.
Hansel C, Linden DJ (2000) Long-term depression of the cerebellar climbing fiber–
Purkinje neuron synapse. Neuron 26(2): 473-482.
Hartigan-O’Connor D, Chamberlain JS (1999) Progress toward gene therapy of
Duchenne muscular dystrophy. Semin Neurol 19: 323-332.
Hasan Z (1986) Optimized movement trajectories and joint stiffness in
unperturbed, inertially loaded movements. Biol Cybern 53: 373-382.
Hasanbarani F, Latash ML (2020) Performance-stabilizing synergies in a complex
motor task: Analysis based on the uncontrolled manifold hypothesis. Motor
Control 24: 238-252.
Hauber W (1998) Involvement of basal ganglia transmitter systems in movement
initiation. Prog Neurobiol 56: 507-540.
Hay L, Bard C, Fleury M, Teasdale N (1996) Availability of visual and
proprioceptive afferent messages and postural control in elderly adults. Exp
Brain Res 108: 129-139.
Hayashi A, Kagamihara Y, Nakajima Y, Narabayashi H, Okuma Y, Tanaka R (1988)
Disorder in reciprocal innervation upon initiation of voluntary movement in
patients with Parkinson’s disease. Exp Brain Res 70: 437-440.
Hayashi R, Miyake A, Jijiwa H, Watanabe S (1981) Postural readjustment to body
sway induced by vibration in man. Exp Brain Res 43: 217-225.
Heckman CJ, Gorassini MA, Bennett DJ (2005) Persistent inward currents in
motoneuron dendrites: Implications for motor output. Muscle Nerve 31: 135-
156.
Heckman CJ, Johnson M, Mottram C, Schuster J (2008) Persistent inward
currents in spinal motoneurons and their influence on human motoneuron firing
patterns. Neuroscientist 14: 264-275.
Heckman CJ, Lee RH, Brownstone RM (2003) Hyperexcitable dendrites in
motoneurons and their neuromodulatory control during motor behavior. Trends
Neurosci 26: 688-695.
Heilman KM, Bowers D, Watson RT, Greer M (1976) Reaction time in Parkinson’s
disease. Arch Neurol 33: 139-140.
Helms Tillery SI, Taylor DM, Schwartz AB (2003) Training in cortical control of
neuroprosthetic devices improves signal extraction from small neuronal
ensembles. Rev Neurosci 14(1-2): 107-119.
Henderson G, Tomlinson BE, Gibson PH (1980) Cell counts in human cerebral
cortex in normal adults throughout life using an image analysing computer. J
Neurol Sci 46: 113-136.
Herter TM, Takei T, Munoz DP, Scott SH (2015) Neurons in red nucleus and
primary motor cortex exhibit similar responses to mechanical perturbations
applied to the upper-limb during posture. Front Integrative Neurosci 9, 29.
Herzfeld DJ, Shadmehr R (2014) Cerebellum estimates the sensory state of the
body. Trends Cogn Sci 18: 66-67.
Hesse S, Schmidt H, Werner C, Bardeleben A (2003) Upper and lower extremity
robotic devices for rehabilitation and for studying motor control. Curr Opin
Neurol 16: 705-710.
Hétu S, Grégoire M, Saimpont A, Coll MP, Eugène F, Michon PE, Jackson PL
(2013) The neural network of motor imagery: An ALE meta-analysis. Neurosci
Biobehav Rev 37(5): 930-949.
Hidler JM, Rymer WZ (1999) A simulation study of reflex instability in spasticity:
Origins of clonus. IEEE Trans Rehabil Eng 7: 327-340.
Hill AV (1938) The heat of shortening and the dynamic constants of muscle. Proc
Roy Soc London Ser B 126: 136-195.
Hinder MR, Milner TE (2003) The case for an internal dynamics model versus
equilibrium point control in human movement. J Physiol 549: 953-963.
Hirano, T. (2018). Purkinje neurons: Development, morphology, and function.
Cerebellum 17: 699-700.
Hirose J, Cuadra C, Walter C, Latash ML (2020) Finger interdependence and
unintentional force drifts: Lessons from manipulations of visual feedback. Hum
Move Sci 74: 102714.
Hirsch JC, Fourment A, Marc ME (1983) Sleep-related variations of membrane
potential in the lateral geniculate body relay neurons of the cat. Brain Res 259:
308-312.
Hirschfeld H, Forssberg H (1991). Phase-dependent modulations of anticipatory
postural activity during human locomotion. J Neurophysiol 66: 12-19.
Hodgson M, Docherty D, Robbins D (2005) Post-activation potentiation:
Underlying physiology and implications for motor performance. Sports Med 35:
585-595.
Hogan N (1984) An organizational principle for a class of voluntary movements. J
Neurosci 4: 2745-2754.
Hogan N (1985) The mechanics of multi-joint posture and movement control. Biol
Cybern 52: 315-331.
Hogan N (1990) Mechanical impedance of single- and multi-articular systems. In:
Winters JM, Woo SL-Y (Eds.) Multiple muscle systems. Biomechanics and
movement organization, pp. 149-164. New York: Springer-Verlag.
Hollerbach JM, Atkeson CG (1987) Deducing planning variables from experimental
arm trajectories: Pitfalls and possibilities. Biol Cybern 56: 279-292.
Hollerbach JM, Flash T (1982) Dynamic interaction between limb segments during
planar arm movements. Biol Cybern 44: 67-77.
Horak FB, Diener HC (1994) Cerebellar control of postural scaling and central set
in stance. J Neurophysiol 72: 479-493.
Horak FB, Nashner LM (1986) Central programming of postural movements:
Adaptation to altered support-surface configurations. J Neurophysiol 55: 1369-
1381.
Horak FB, Shupert CL, Mirka A (1989) Components of postural dyscontrol in the
elderly: A review. Neurobiol Aging 10: 727-738.
Houk JC (1976) An assessment of stretch reflex function. Prog Brain Res 44: 303-
314.
Houk JC (1979) Regulation of stiffness by skeletomotor reflexes. Ann Rev Physiol
41: 99-114.
Houk JC, Wise SP (1995) Distributed modular architectures linking basal ganglia,
cerebellum, and cerebral cortex: their role in planning and controlling action.
Cereb Cortex 5(2): 95-110.
Houk JC, Buckingham JT, Barto AG (1996) Models of the cerebellum and motor
learning. Behav Brain Sci 19: 368-383.
Houk JC, Gibson AR (1987) Sensorimotor processing through the cerebellum. In:
King JS (Ed.) New concepts in cerebellar neurobiology, pp. 387-416. New York:
Liss.
Hubel DH, Wiesel TN (1979) Brain mechanisms of vision. Sci Amer 241: 150-162.
Hughlings Jackson J (1889, Aug. 17) On the comparative study of disease of the
nervous system. Brit Med J 355-362.
Hughes RA, Cornblath DR (2005) Guillain-Barre syndrome. Lancet 366: 1653-
1666.
Hughes S, Gibbs J, Dunlop D, Edelman P, Singer R, Chang RW (1997) Predictors
of decline in manual performance of older adults. J Am Geriatr Soc 45: 905-910.
Hulliger M, Nordh E, Vallbo AB (1982) The absence of position response in spindle
afferent units from human finger muscles during accurate position holding. J
Physiol 322: 167-179.
Hultborn H (2003) Changes in neuronal properties and spinal reflexes during
development of spasticity following spinal cord lesions and stroke: Studies in
 animal models and patients. J Rehabil Med 41 Suppl: 46-55.
Humphrey DR (1982) Separate cell systems in the motor cortex of the monkey for
the control of joint movement and of joint stiffness. In: Buser PA, Cobb VA,
Okuma (Eds.) Kyoto Symp, Electroencephalog Clin Neurophysiol Suppl 36:
393-408. Elsevier: Amsterdam.
Hunter JP, Ashby P, Lang AE (1988) Afferents contributing to the exaggerated long
latency reflex response to electrical stimulation in Parkinson’s disease. J Neurol
Neurosurg Psychiat 51: 1405-1410.
Iansek R (1984) The effects of reflex path length on clonus frequency in spastic
muscles. J Neurol Neurosurg Psychiat 47: 1122-1124.
Imamizu H, Kawato M (2012) Cerebellar internal models: Implications for the
dexterous use of tools. Cerebellum 11: 325-335.
Imamizu H, Kuroda T, Miyauchi S, Yoshioka T, Kawato M (2003) Modular
organization of internal models of tools in the human cerebellum. Proc Natl
Acad Sci USA 100: 5461-5466.
Imamizu H, Uno Y, Kawato M (1995) Internal representation of the motor
apparatus: Implications from generalization in visuomotor learning. J Exp
Psychol: Hum Percept Perform 21: 1174-1198.
Inglin B, Woollacott MH (1988) Anticipatory postural adjustments associated with
reaction time arm movements: A comparison between young and old. J
Gerontol 43: M105-M113.
Inzelberg R, Flash T, Korczyn AD (1990) Kinematic properties of upper-limb
trajectories in Parkinson’s disease and idiopathic torsion dystonia. Adv Neurol
53: 183-189.
Ito M (1989) Long-term depression. Ann Rev Neurosci 12: 85-102.
Ito M (2005) Bases and implications of learning in the cerebellum – adaptive
control and internal model mechanism. Prog Brain Res 148: 95-109.
Ivanenko YP, Dominici N, Cappellini G, Di Paolo A, Giannini C, Poppele RE,
Lacquaniti F (2013) Changes in the spinal segmental motor output for stepping
during development from infant to adult. J Neurosci 33: 3025-3036.
Ivanenko YP, Poppele RE, Lacquaniti F (2004) Five basic muscle activation
patterns account for muscle activity during human locomotion. J Physiol 556:
267-82.
Ivanenko YP, Poppele RE, Lacquaniti F (2006) Motor control programs and
walking. Neuroscientist 12: 339-348.
Ivanitsky AM, Nikolaev AR, Ivanitsky GA (1999) Electroencephalography. In:
Windhorst U, Johansson H (Eds.) Modern techniques in neuroscience research,
pp. 971-998. Berlin: Springer-Verlag.
Ivry RB (2003) Cerebellar involvement in clumsiness and other developmental
disorders. Neural Plast 10: 141-153.
Ivry RB, Keele SW, Diener HC (1988) Dissociation of the lateral and medial
cerebellum in movement iming and movement execution. Exp Brain Res 73:
167-180.
Ivry RB, Keele SW (1989) Timing functions of the cerebellum. J Cog Neurosci 1(2):
136-152.
Ivry RB, Spencer RM (2004) The neural representation of time. Curr Opin
Neurobiol 14: 225-232.
Izquierdo M, Hakkinen K, Ibanez J, Anton A, Garrues M, Ruesta M, Gorostiaga
EM (2003) Effects of strength training on submaximal and maximal endurance
performance capacity in middle-aged and older men. J Strength Cond Res 17:
129-139.
Jalinous R (1998) Guide to magnetic stimulation. The MagStim Company, Ltd.
Jankowska E (1979) New observations on neuronal organization of reflexes from
tendon organ afferents and their relation to reflexes evoked from muscle spindle
afferents. In: Granit R, Pompeiano O (Eds.) Reflex control of posture and
movement, pp. 29-36. Amsterdam: Elsevier.
Jankowska E, Lundberg A, Stuart D (1983) Propriospinal control of interneurons in
spinal reflex pathways from tendon organs in the cat. Brain Res 261: 317-320.
Jankowska E, McCrea DA (1983) Shared reflex pathways from Ib tendon organ
afferents and Ia muscle spindle afferents in the cat. J Physiol 338, 99-111.
Jansons H (1992) Bernstein: The microscopy of movement. In: Cappozzo A,
Marchetti M, Tosi V (Eds.) Biolocomotion: A century of research using moving
pictures, pp. 137-174. Rome, Italy: Promograph.
Jaric S, Latash ML (1999) Learning a pointing task with a kinematically redundant
limb: Emerging synergies and patterns of final position variability. Hum Move Sci
18: 819-838.
Jaric S, Milanovic S, Blezic S, Latash ML (1999) Changes in movement kinematics
during single-joint movements against expectedly and unexpectedly changed
inertial loads. Hum Move Sci 18: 49-66.
Jason LA, Corradi K, Torres-Harding S, Taylor RR, King C (2005) Chronic fatigue
syndrome: The need for subtypes. Neuropsychol Rev 15: 29-58.
Jeka JJ, Lackner JR (1994) Fingertip contact influences human postural control.
Exp Brain Res 100: 495-502.
Jeka JJ, Oie K, Schöner G, Dijkstra T, Henson E (1998) Position and velocity
coupling of postural sway to somatosensory drive. J Neurophysiol 79: 1661-
1674.
Jeneson JA, Taylor JS, Vigneron DB, Willard TS, Carvajal L, Nelson SJ, Murphy-
Boesch J, Brown TR (1990) 1H MR imaging of anatomical compartments within
the finger flexor muscles of the human forearm. Magn Reson Med 15: 491-496.
Jensen BR, Pilegaard M, Sjogaard G (2000) Motor unit recruitment and rate
coding in response to fatiguing shoulder abductions and subsequent recovery.
Eur J Appl Physiol 83: 190-199.
Jerde TE, Soechting JF, Flanders M (2003) Coarticulation in fluent fingerspelling. J
Neurosci 23: 2383-2393.
Jo HJ, Maenza C, Good DC, Huang X, Park J, Sainburg RL, Latash ML (2016)
Effects of unilateral stroke on multi-finger synergies and their feed-forward
adjustments. Neuroscience 319: 194-205.
Jo HJ, Park J, Lewis MM, Huang X, Latash ML (2015) Prehension synergies and
hand function in early-stage Parkinson’s disease. Exp Brain Res 233: 425-440.
Jobin A, Levin MF (2000) Regulation of stretch reflex threshold in elbow flexors in
children with cerebral palsy: A new measure of spasticity. Dev Med Child Neurol
42: 531-540.
Johansson BB (2000) Brain plasticity and stroke rehabilitation: The Willis lecture.
Stroke 31: 223-230.
Johansson RS, Westling G (1988) Coordinated isometric muscle commands
adequately and erroneously programmed for the weight during lifting task with
precision grip. Exp Brain Res 71: 59-71.
Johnson RT, Gibbs CJ Jr. (1998) Creutzfeldt-Jakob disease and related
transmissible spongiform encephalopathies. N Engl J Med 339: 1994-2004.
Johnson SK, DeLuca J, Natelson BH (1999) Chronic fatigue syndrome: Reviewing
the research findings. Ann Behav Med 21: 258-271.
Jones RD, Donaldson IM, Parkin PJ (1989) Impairment and recovery of ipsilateral
sensory-motor function following unilateral cerebral infarction. Brain 112: 113-
132.
Joyce, GC, Rack, PMH, Ross, HF (1974) The forces generated in the human
elbow joint in response to imposed sinusoidal movements of the forearm. J
Physiol 240: 351-374.
Joyce GC, Rack PMH, Westbury DR (1969) The mechanical properties of cat
soleus muscle during controlled lengthening and shortening movements J
Physiol 204: 461-474
Kadefors R, Kaiser E, Petersen I (1968) Dynamic spectrum analysis of myo-
potentials with special reference to muscle fatigue. Electromyography 8: 39-74.
Kaewmanee T, Liang H, Aruin AS (2020) Effect of predictability of the magnitude of
a perturbation on anticipatory and compensatory postural adjustments. Exp
Brain Res 238: 2207-2219.
Kakei S, Hoffman DS, Strick PL (1999) Muscle and movement representations in
the primary motor cortex. Science 285(5436): 2136-2139.
Kalezic I, Bugaychenko LA, Kostyukov AI, Pilyavskii AI, Ljubisavljevic M,
Windhorst U, Johansson H (2004) Fatigue-related depression of the feline
 monosynaptic gastrocnemius-soleus reflex. J Physiol 556: 283-296.
Kamen G, Sison SV, Du CC, Patten C (1995) Motor unit discharge behavior in
older adults during maximal-effort contractions. J Appl Physiol 79: 1908-1913.
Kang N, Roberts LM, Aziz C, Cauraugh JH (2019) Age-related deficits in bilateral
motor synergies and force coordination. BMC Geriatr 19(1): 287.
Kang N, Shinohara M, Zatsiorsky VM, Latash ML (2004) Learning multi-finger
synergies: An uncontrolled manifold analysis. Exp Brain Res 157: 336-350.
Karrer R, Wojtascek Z, Davis MG (1995) Event-related potentials and information
processing in infants with and without Down syndrome. Am J Ment Retard 100:
146-159.
Katz RT, Rymer WZ (1989) Spastic hypertonia: Mechanisms and measurement.
Arch Phys Med Rehabil 70: 144-155.
Kautz SA, Duncan PW, Perera S, Neptune RR, Studenski SA (2005) Coordination
of hemiparetic locomotion after stroke rehabilitation. Neurorehabil Neural Repair
19: 250-258.
Kawato M (1999) Internal models for motor control and trajectory planning. Curr
Opin Neurobiol 9: 718-727.
Kawato M, Gomi H (1992) The cerebellum and VOR/OKR learning models. Trends
Neurosci 15: 445-453.
Kawato M, Kuroda T, Imamizu H, Nakano E, Miyauchi S, Yoshioka T (2003)
Internal forward models in the cerebellum: fMRI study on grip force and load
force coupling. Prog Brain Res 142: 171-188.
Kazennikov O, Hyland B, Corboz M, Babalian A, Rouiller EM, Wiesendanger M
(1999) Neural activity of supplementary and primary motor areas in monkeys
and its relation to bimanual and unimanual movement sequences.
Neuroscience 89: 661-674.
Kazennikov OV, Shik ML (1988) Propagation of the activity along the “stepping
strip” of the spinal cord in the cat. Neirofiziologiia 20: 763-769 (in Russian).
Keele SW, Ivry R (1990) Does the cerebellum provide a common computation for
diverse tasks? A timing hypothesis. Ann New York Acad Sci 608: 179-211.
Keifer J, Houk JC (1994) Motor function of the cerebellorubrospinal system.
Physiol Rev 74: 509-542.
Keller E (1974) Participation of medial pontine reticular formation in eye movement
generation in monkey. J Neurophysiol 37(2): 316-332.
Kelso JAS (1984) Phase transitions and critical behavior in human bimanual
coordination. Amer J Physiol 246: R1000-R1004.
Kelso JAS (1995). Dynamic patterns: The self-organization of brain and behavior.
Cambridge MA: MIT Press.
Kelso JAS, Schöner G (1988) Self-organization of coordinative movement
patterns. Hum Mov Sci 7: 27-46.
Kelso JAS, Southard DL, Goodman D (1979) On the nature of human interlimb
coordination. Science 203: 1029-1031.
Kenzie JM, Semrau JA, Findlater SE, Herter TM, Hill MD, Scott SH, Dukelow SP
(2014) Anatomical correlates of proprioceptive impairments following acute
stroke: A case series. J Neurol Sci 342(1-2): 52-61.
Kernell D, Eerbeek O, Verhey BA (1983) Relation between isometric force and
stimulus rate in cat’s hindlimb motor units of different twitch contraction time.
Exp Brain Res 50: 220-227.
Ketcham CJ, Dounskaia NV, Stelmach GE (2004) Multijoint movement control: The
importance of interactive torques. Prog Brain Res 143: 207-218.
Ketcheson L, Pitchford EA, Kwon HJ, Ulrich DA (2017) Physical activity patterns in
infants with and without Down syndrome. Pediatr Phys Ther 29: 200-206.
Khurana TS, Davies KE (2003) Pharmacological strategies for muscular dystrophy.
Nat Rev Drug Discov 2: 379-390.
Kiemel T, Oie KS, Jeka JJ (2002) Multisensory fusion and the stochastic structure
of postural sway. Biol Cybern 87: 262-277.
Kilbreath SL, Gandevia SC (1994). Limited independent flexion of the thumb and
fingers in human subjects. J Physiol 479: 487-497.
Kinoshita H, Francis PR (1996) A comparison of prehension force control in young
and elderly individuals. Eur J Appl Physiol 74: 450-460.
Kinoshita H, Murase T, Bandou T (1996). Grip posture and forces during holding
cylindrical objects with circular grips. Ergonomics 39: 1163-1176.
Kirkendall DT. Garrett WE Jr (1998) The effects of aging and training on skeletal
muscle. Amer J Sports Med 26: 598-602.
Kirsch RF, Rymer WZ (1987) Neural compensation for muscular fatigue: Evidence
for significant force regulation in man. J Neurophysiol 57: 1893-1910.
Kirshblum S (2004) New rehabilitation interventions in spinal cord injury. J Spinal
Cord Med 27: 342-350.
Kissel JT (1999) Facioscapulohumeral dystrophy. Semin Neurol 19: 35-43.
Klein CS, Rice CL, Marsh GD (2001) Normalized force, activation, and
coactivation in the arm muscles of young and old men. J Appl Physiol 91: 1341-
1349.
Klous M, Mikulic P, Latash ML (2011) Two aspects of feed-forward postural control:
Anticipatory postural adjustments and anticipatory synergy adjustments. J
Neurophysiol 105: 2275-2288.
Koch M, Lingenhöhl K, Pilz P (1992) Loss of the acoustic startle response
following neurotoxic lesions of the caudal pontine reticular formation: possible
role of giant neurons. Neurosci 49(3): 617-625.
Koeppen AH (2018) The neuropathology of the adult cerebellum. Handb Clin
Neurol 154:129-149.
Korn H, Sotelo C, Crepel F (1973) Electrotonic coupling between neurons in the
rat lateral vestibular nucleus. Exp Brain Res 16: 225-275.
Koshland GF, Gerilovsky L, Hasan Z (1991) Activity of wrist muscles elicited during
imposed or voluntary movements about the elbow joint. J Mot Behav 23: 91-
100.
Kothari MJ (2004) Myasthenia gravis. J Am Osteopath Assoc 104: 377-384.
Krakauer J, Ghez C (2000) Voluntary movement. In: Kandel ER, Schwartz JH,
Jessell TM (Eds.) Principles of neural science. Fourth edition, pp. 756-780. New
York: McGraw-Hill.
Krauzlis RJ (2005) The control of voluntary eye movements: new perspectives.
Neuroscientist 11(2):124-137.
Kravitz DJ, Saleem KS, Baker CI, Ungerleider LG, Mishkin M. (2013) The ventral
visual pathway: an expanded neural framework for the processing of object
quality. Trends Cogn Sci Jan;17(1):26-49.
Krishnamoorthy V, Latash ML (2005) Reversals of anticipatory postural
adjustments during voluntary sway. J Physiol 565: 675-684.
Krishnamoorthy V, Latash ML, Scholz JP, Zatsiorsky VM (2003) Muscle synergies
during shifts of the center of pressure by standing persons. Exp Brain Res 152:
281-292.
Krishnamoorthy V, Latash ML, Scholz JP, Zatsiorsky VM (2004) Muscle modes
during shifts of the center of pressure by standing persons: Effects of instability
and additional support. Exp Brain Res 157: 18-31.
Krishnamoorthy V, Scholz JP, Latash ML (2007) The use of flexible arm muscle
synergies to perform an isometric stabilization task. Clin Neurophysiol 118: 525-
537.
Krishnan V, Aruin AS, Latash ML (2011) Two stages and three components of
postural preparation to action. Exp Brain Res 212: 47-63.
Krupp LB, Alvarez LA, LaRocca NG, Scheinberg LC (1988) Fatigue in multiple
sclerosis. Arch Neurol 45: 435-437.
Kuehni RG (2016) How many object colors can we distinguish? Color Res Appl
41(5): 439-444.
Kugelberg E, Lindegren B (1979) Transmission and contraction fatigue of rat motor
units in relation to succinate dehydrogenase activity of motor unit fibers. J
Physiol 288: 285-300.
Kugler PN, Turvey MT (1987) Information, natural law, and the self-assembly of
rhythmic movement. Hillsdale, NJ: Erlbaum.
Kwakkel G, Veerbeek JM, van Wegen EE, Wolf SL (2015) Constraint-induced
movement therapy after stroke. Lancet Neurol 14(2): 224-234.
Lackner JR, DiZio P (1994) Rapid adaptation to Coriolis force perturbations of arm
trajectory. J Neurophysiol 72: 1-15.
Lackner JR, Levine MS (1979) Changes in apparent body orientation and sensory
localization, induced by vibration of postural muscles; vibratory myesthetic
illusions. Aviat Space Environ Med 50: 346-354.
Lackner JR, Taublieb AB (1983) Reciprocal interactions between the position
sense representations of the two forearms. J Neurosci 3: 2280-2285.
Lacquaniti F, Maioli C (1989) The role of preparation in turning anticipatory and
reflex responses during catching. J Neurosci 9: 1134-1148.
Laidlaw DH, Bilodeau M, Enoka RM (2000) Steadiness is reduced and motor unit
discharge is more variable in old adults. Muscle Nerve 23: 600-612.
Lalonde R, Filali M, Bensoula AN, Lestienne F (1996) Sensorimotor learning in
three cerebellar mutant mice. Neurobiol Learn Mem 65:113-120.
Lamarre Y, Spidalieri G, Chapman CE (1983) A comparison of neuronal discharge
recorded in the sensory-motor cortex, parietal cortex, and dentate nucleus of
the monkey during arm movements triggered by light, sound or somesthetic
stimuli. Exp Brain Res Suppl 7: 140-156.
Lance JW (1980) The control of muscle tone, reflexes, and movement: Robert
Wartenberg lecture. Neurology 30: 1303-1313.
Lance JW, Burke D (1974) Mechanisms of spasticity. Arch Phys Med Rehabil 55:
332-337.
Lange DH, Inbar GF (1999) Modern techniques in ERP research. In: Windhorst U,
Johansson H (Eds.) Modern techniques in neuroscience research, pp. 997-
1024. Berlin: Springer-Verlag.
Larsson L-E (1975) On the relation between the EMG frequency spectrum and the
duration of symptoms in lesions of the peripheral motor neuron.
Electroencephalog Clin Neurophysiol 38: 69-78.
Lassman H (2019) The changing concepts in the neuropathology of acquired
demyelinating central nervous system disorders. Curr Opin Neurol 32: 313-319.
Latash LP (1979) Trace changes in the spinal cord and some basic problems of
the neurophysiology of memory. In: Oniani TN (ed.) Seventh Gagra talks: The
neurophysiological basis of memory, pp. 118-130. Tbilisi: Metsniereba.
Latash ML (1988) Spectral analysis of the electromyogram (EMG) in spinal cord
trauma patients. I. Different types of the EMG and corresponding spectra.
Electromyogr Clin Neurophysiol 28: 319-327.
Latash ML (1992) Virtual trajectories, joint stiffness, and changes in natural
frequency during single-joint oscillatory movements. Neuroscience 49: 209-220.
Latash ML (1993) Control of human movement. Champaign, IL: Human Kinetics.
Latash ML (1994) Reversals of the tonic vibration reflex in shoulder muscles. Hum
Physiol (Fiziologiya Cheloveka) 20(5): 56-60 (in Russian).
Latash ML (1996) How does our brain make its choices? In: Latash ML, Turvey MT
(Eds.) Dexterity and its development, pp. 277-304. Mahwah, NJ: Erlbaum.
Latash ML (1997) The answer may be 42. So, what is the question? Motor Control
1: 205-207.
Latash ML (1999) Mirror writing: Learning, transfer, and implications for internal
inverse models. J Mot Behav 31: 107-112.
Latash ML (2008) Synergy. New York: Oxford University Press.
Latash ML (2012) The bliss (not the problem) of motor abundance (not
redundancy). Exp Brain Res 217: 1-5.
Latash ML (2018a) Muscle co-activation: Definitions, mechanisms, and functions.
J Neurophysiol 120: 88-104.
Latash ML (2018b) Stability of kinesthetic perception in efferent-afferent spaces:
The concept of iso-perceptual manifold. Neuroscience 372: 97-113.
Latash ML (2019) Physics of biological action and perception. New York:
Academic Press.
Latash ML (2020) On primitives in motor control. Motor Control 24: 318-346.
Latash ML (2021a) Laws of nature that define biological action and perception.
Phys Life Rev 36: 47-67.
Latash ML (2021b) Efference copy in kinesthetic perception: A copy of what is it? J
Neurophysiol 125: 1079-1094.
Latash ML (2021c) One more time about motor (and non-motor) synergies. Exp
Brain Res 239: 2951-2967.
Latash ML, Anson JG (1996) What are normal movements in atypical populations?
Behav Brain Sci 19: 55-106.
Latash ML, Anson JG (2006) Synergies in health and disease: Relations to
adaptive changes in motor coordination. Phys Ther 86: 1151-1160.
Latash ML, Friedman J, Kim SW, Feldman AG, Zatsiorsky VM (2010) Prehension
synergies and control with referent hand configurations. Exp Brain Res 202:
213-229.
Latash ML, Gottlieb GL (1991) Reconstruction of elbow joint compliant
characteristics during fast and slow voluntary movements. Neuroscience 43:
697-712.
Latash ML, Gottlieb GL (1992) Virtual trajectories of single-joint movements
performed under two basic strategies. Neuroscience 47: 357-365.
Latash ML, Gutman SR (1994) Abnormal motor patterns in the framework of the
equilibrium-point hypothesis: A cause for dystonic movements? Biol Cybern 71:
87-94.
Latash ML, Gutman SR, Gottlieb GL (1991) Relativistic effects in single-joint
voluntary movements. Biol Cybern 65: 401-406.
Latash ML, Huang X (2015) Neural control of movement stability: Lessons from
studies of neurological patients. Neuroscience 301: 39-48.
Latash ML, Kalugina E, Nicholas JJ, Orpett C, Stefoski D, Davis F (1996)
Myogenic and central neurogenic factors in fatigue in multiple sclerosis. Multiple
Sclerosis 1: 236-241.
Latash ML, Kang N, Patterson D (2002a) Finger coordination in persons with
Down syndrome: Atypical patterns of coordination and the effects of practice.
Exp Brain Res 146: 345-355.
Latash ML, Li Z-M, Zatsiorsky VM (1998) A principle of error compensation studied
within a task of force production by a redundant set of fingers. Exp Brain Res
122: 131-138.
Latash ML, Penn RD (1996) Changes in voluntary motor control induced by
intrathecal baclofen. Physiother Res Intern 1: 229-246.
Latash ML, Penn RD, Corcos DM, Gottlieb GL (1989) Short-term effects of
intrathecal baclofen in spasticity. Exp Neurol 103: 165-172.
Latash ML, Penn RD, Corcos DM, Gottlieb GL (1990) Effects of intrathecal
baclofen on voluntary motor control in spastic paresis. J Neurosurg 72: 388-
392.
Latash ML, Scholz JF, Danion F, Schöner G (2001) Structure of motor variability in
marginally redundant multi-finger force production tasks. Exp Brain Res 141:
153-165
Latash ML, Scholz JP, Schöner G (2002b) Motor control strategies revealed in the
structure of motor variability. Exerc Sport Sci Rev 30: 26-31.
Latash ML, Scholz JP, Schöner G (2007) Toward a new theory of motor synergies.
Motor Control 11: 276-308.
Latash ML, Shim JK, Smilga AV, Zatsiorsky V (2005) A central back-coupling
hypothesis on the organization of motor synergies: A physical metaphor and a
neural model. Biol Cybern 92: 186-191
Latash ML, Yarrow K, Rothwell JC (2003) Changes in finger coordination and
responses to single pulse TMS of motor cortex during practice of a multi-finger
force production task. Exp Brain Res 151: 60-71.
Latash ML, Yee M, Orpett C, Slingo A, Nicholas JJ (1994) Combining electrical
muscle stimulation with voluntary contraction for studying muscle fatigue. Arch
Phys Med Rehab 75: 29-35.
Latash ML, Zatsiorsky VM (1993) Joint stiffness: Myth or reality? Hum Move Sci
12: 653-692.
Laufer Y (2005) Effect of age on characteristics of forward and backward gait at
preferred and accelerated walking speed. J Gerontol A Biol Sci Med Sci 60:
627-632.
Lebedev MA, Nicolelis MA (2006) Brain-machine interfaces: Past, present and
future. Trends Neurosci 29: 536-546.
Lee RG, Hendrie A (1977) Selective modification of human spinal and long-loop
reflexes by vibration. Electroencephalog Clin Neurophysiol 43: 606-610.
Lelli S, Panizza M, Hallett M (1991) Spinal cord inhibitory mechanisms in
Parkinson’s disease. Neurology 41: 553-556.
Lemon RN, Kirkwood PA, Maier MA, Nakajima K, Nathan P (2004) Direct and
indirect pathways for corticospinal control of upper limb motoneurons in the
primate. Prog Brain Res 143: 263-279.
Lepers R, Brenière Y (1995) The role of anticipatory postural adjustments and
gravity in gait initiation. Exp Brain Res 107: 118-124.
Levin MF (1996) Interjoint coordination during pointing movements is disrupted in
spastic hemiparesis. Brain 119: 281-293.
Levin MF, Feldman AG (1994) The role of stretch reflex threshold regulation in
normal and impaired motor control. Brain Res 657: 23-30.
Levin MF, Michaelsen SM, Cirstea CM, Roby-Brami A (2002) Use of the trunk for
reaching targets placed within and beyond the reach in adult hemiparesis. Exp
Brain Res 143: 171-180.
Li S, Latash ML, Yue GH, Siemionow V, Sahgal V (2003) The effects of stroke and
age on finger interaction in multi-finger force production tasks. Clin
Neurophysiol 114: 1646-1655.
Li Z-M, Latash ML, Zatsiorsky VM (1998) Force sharing among fingers as a model
of the redundancy problem. Exp Brain Res 119: 276-286.
Liddell EGT, Sherrington CS (1924) Reflexes in response to stretch (myotatic
reflexes). Proc Roy Soc London Ser B 96: 212-242.
Linden DJ (1996) Cerebellar long-term depression as investigated in a cell culture
preparation. Behav Brain Sci 19: 339-346.
Lindstrom L, Malmstrom J-E, Petersen I (1985) Clinical applications of spectral
analysis of EMH. In: Struppler A, Weindl A (Eds.) Electromyography and evoked
potentials. Theories and applications, pp. 108-113. Berlin: Springer-Verlag.
Lisin VV, Frankstein SI, Rechtman MB (1973) The influence of locomotion on
flexor reflex of the hind limb in cat and man. Exp Neurol 38: 180-183.
Liu JZ, Dai TH, Sahgal V, Brown RW, Yue GH (2002) Nonlinear cortical modulation
of muscle fatigue: A functional MRI study. Brain Res 957: 320-329.
Ljubisavljevic M, Milanovic S, Radovanovic S, Vukcevic I, Kostic V, Anastasijevic R
(1996) Central changes in muscle fatigue during sustained submaximal
isometric voluntary contraction as revealed by transcranial magnetic stimulation.
Electroencephalogr Clin Neurophysiol 101: 281-288.
Llinas R (1985) Functional significance of the basic cerebellar circuit in motor
coordination. In: Bloedel JR, Dichgans J, Precht W (Eds) Cerebellar functions.
Berlin: Springer-Verlag.
Loeb GE (2012) Optimal isn’t good enough. Biol Cybern 106: 757-765.
Longuet-Higgins HC, Willshaw DJ, Buneman OP (1970) Theories of associative
recall. Q Rev Biophys 3: 223-244.
Loram ID, Maganaris CN, Lakie M (2005) Active, non-spring-like muscle
movements in human postural sway: How might paradoxical changes in muscle
length be produced? J Physiol 564: 281-293.
Lozano AM, Mahant N (2004) Deep brain stimulation surgery for Parkinson’s
disease: Mechanisms and consequences. Parkinsonism Relat Disord 10 Suppl
1: S49-S57.
Lu MT, Preston JB, Strick PL (1994). Interconnections between the prefrontal
cortex and the premotor areas in the frontal lobe. J Comp Neurol 341(3): 375-
392.
Lublin FD (2004) Clinical features and diagnosis of multiple sclerosis. Neurol Clin
23: 1-15.
Luff AR (1998) Age-associated changes in the innervation of muscle fibers and
changes in the mechanical properties of motor units. Ann NY Acad Sci 854: 92-
101.
Lund S (1980) Postural effects of neck muscle vibration in man. Experientia 36:
1398-1340.
Lundberg A (1975) Control of spinal mechanisms from the brain. In: DB Tower
(Ed.) The nervous system, v. 2. New York: Raven Press.
Lundberg A (1979) Multisensory control of spinal reflex pathways. In: R Granit, O
Pompeiano (Eds.) Reflex control of posture and movement, pp. 11-28.
Amsterdam: Elsevier.
de Lussanet MHE, Osse JWM (2012) An ancestral axial twist explains the
contralateral forebrain and the optic chiasm in vertebrates. Animal Biology
62(2): 193-216.
Luu BL, Day BL, Cole JD, Fitzpatrick RC (2011) The fusimotor and reafferent origin
of the sense of force and weight. J Physiol 589: 3135-3147.
Macefield G, Hagbarth K-E, Gorman R, Gandevia S, Burke D (1991) Decline in
spindle support to α-motoneurones during sustained voluntary contractions. J
Physiol 440: 497-512.
MacKenzie CL, Iberall T (1994) The grasping hand. Amsterdam: North-Holland.
Madarshahian, S, Letizi, J, Latash, ML. (2021), Synergic control of a single
muscle: The example of flexor digitorum superficialis. J Physiol 599: 1261-1279.
Maenza C, Good DC, Winstein CJ, Wagstaff DA, Sainburg RL (2020) Functional
deficits in the less-impaired arm of stroke survivors depend on hemisphere of
damage and extent of paretic arm impairment. Neurorehabil Neural Repair
34:39-50.
Maki BE, Holliday PJ, Fernie GR (1990) Aging and postural control. A comparison
of spontaneous- and induced-sway balance tests. J Am Geriatr Soc 38: 1-9.
Marconi R, Landi A, Valzania F (2008) Subthalamic nucleus stimulation in
Parkinson’s disease. Neurol Sci 29(Suppl 5): S389-391.
Mark VW, Taub E (2004) Constraint-induced movement therapy for chronic stroke
hemiparesis and other disabilities. Restor Neurol Neurosci 22: 317-336.
Marr D (1969) A theory of cerebellar cortex. J Physiol 202: 437-470.
Marras C, Tanner CM (2004) Epidemiology of Parkinson’s disease. In: Watts RL,
Koller WC (Eds.) Movement disorders. Neurological principles and practice, pp.
177-196. New York: McGraw-Hill.
Marsden CD, Deecke L, Freund HJ, Hallett M, Passingham RE, Shibasaki H, Tanji
J, Wiesendanger M (1996) The functions of the supplementary motor area.
Summary of a workshop. Adv Neurol 70: 477-487.
Marsden CD, Merton RA, Morton HB (1976a) Stretch reflex and servo action in a
variety of human muscles. J Physiol 259: 531-560.
Marsden CD, Merton RA, Morton HB (1976b) Servo action in the human thumb. J
Physiol 257: 1-44.
Marsden CD, Rothwell JC, Traub M (1979) Long latency stretch reflex of the
human thumb can be reversed if the task is changed. J Physiol 293: 41P-42P.
Marshall FJ (2004) Clinical features and treatment of Huntington’s disease. In:
Watts RL, Koller WC (Eds.) Movement disorders. Neurological principles and
practice, pp. 589-602. New York: McGraw-Hill.
Martin TA, Keating JG, Goodkin HP, Bastian AJ, Thach WT (1996) Throwing while
looking through prisms. I. Focal olivocerebellar lesions impair adaptation. Brain
119: 1183-1198.
Martin N, Grafton S, Vinuela F, Dion J, Duckwiler G, Mazziotta J, Lufkin R, Becker
D (1992) Imaging techniques for cortical functional localization. Clin Neurosurg
38: 132-165.
Martin V, Scholz JP, Schöner G (2009) Redundancy, self-motion, and motor
control. Neural Comput 21:1371-1414.
Martín-Fernández J, Burunat I, Modroño C, González-Mora JL, Plata-Bello J
(2021) Music style not only modulates the auditory cortex, but also motor
related areas. Neurosci 457: 88-102.
Martinelli V, Rodegher M, Moiola L, Comi G (2004) Late onset multiple sclerosis:
Clinical characteristics, prognostic factors and differential diagnosis. Neurol Sci
25 Suppl 4: S350-S355.
Marzke MW (1992) Evolutionary development of the human thumb. Hand Clinics
8: 1-8.
Massion J (1992) Movement, posture and equilibrium: Interaction and
coordination. Prog Neurobiol 38: 35-56.
Matthews PBC (1959) The dependence of tension upon extension in the stretch
reflex of the soleus of the decerebrate cat. J Physiol 47: 521-546.
Matthews PBC (1970) The origin and functional significance of the stretch reflex.
In: Andersen P, Jansen JKS (Eds.) Excitatory synaptic mechanisms, pp. 301-
315. Oslo, Norway: Universitets forlaget.
Matthews PBC (1972) Mammalian muscle receptors and their central actions.
Baltimore: Williams & Wilkins.
Matthews PBC, Cody FW, Richardson HC, MacDermott N. (1990) Observations on
the reflex effects seen in Parkinson’s disease on terminating a period of tendon
vibration. J Neurol Neurosurg Psychiatry 53: 215-219.
Matthews PB, Stein RB (1969) The sensitivity of muscle spindle afferents to small
sinusoidal changes of length. J Physiol 200: 723-743.
Mathewson RC, Nava PB (1985) Effects of age on Meissner corpuscles: A study
of silver-impregnated neurites in mouse digital pads. J Comp Neurol 231: 250-
259.
Mattay VS, Fera F, Tessitore A, Hariri AR, Das S, Callicott JH, Weinberger DR
(2002) Neurophysiological correlates of age-related changes in human motor
function. Neurology 58: 630-635.
Mattos D, Latash ML, Park E, Kuhl J, Scholz JP (2011) Unpredictable elbow joint
perturbation during reaching results in multijoint motor equivalence. J
Neurophysiol 106: 1424-1436.
Maurer C, Mergner T, Peterka RJ (2006) Multisensory control of human upright
stance. Exp Brain Res 171: 231-250.
McAlonan GM, Cheung V, Cheung C, Suckling J, Lam GY, Tai KS, Yip L, Murphy
DG, Chua SE (2005) Mapping the brain in autism. A voxel-based MRI study of
volumetric differences and intercorrelations in autism. Brain 128: 268-276.
McComas AJ (1977) Neuromuscular function and disorders. Boston, MA:
Butterworths.
McCormick DA, Thompson RF (1984) Cerebellum: Essential involvement in the
classically conditioned eyelid response. Science 223: 296-299.
McDonagh MJ, White MJ, Davies CT (1984) Differential effects of ageing on the
mechanical properties of human arm and leg muscles. Gerontology 30: 49-54.
McKinley PA, Smith, JL, Gregor RJ (1983) Responses of elbow extensors to
landing forces during jump downs in cats. Exp Brain Res 49: 218-228.
McNamee D, Wolpert DM (2019) Internal models in biological control. Ann Rev
Contr Robot Autonom Syst 2:339-364.
Mechsner F, Kerzel D, Knoblich G, Prinz W (2001) Perceptual basis of bimanual
coordination. Nature 414(6859): 69-73.
Meijer O (2002) Bernstein versus Pavlovianism: An interpretation. In: Latash ML
(Ed.) Progress in motor control. Vol. 2: Structure-function relations in voluntary
movement, p. 229-250. Champaign, IL: Human Kinetics.
Meilink A, Hemmen B, Seelen HA, Kwakkel G (2008) Impact of EMG-triggered
neuromuscular stimulation of the wrist and finger extensors of the paretic hand
after stroke: A systematic review of the literature. Clin Rehabil 22: 291-305.
Meissirel C, Wikler KC, Chalupa LM, Rakic P. (1997). Early divergence of
magnocellular and parvocellular functional subsystems in the embryonic
primate visual system. Proc Natl Acad Sci USA, 94(11): 5900–5905.
Melzack R, Wall PD (1965) Pain mechanisms: A new theory. Science 150: 971-
979.
Melzer I, Benjuya N, Kaplanski J (2004) Postural stability in the elderly: A
comparison between fallers and non-fallers. Age Ageing 33: 602-607.
Meola G, Moxley RT (2004) Myotonic dystrophy type 2 and related myotonic
disorders. J Neurol 251: 1173-1182.
Merchant H, Grahn J, Trainor L, Rohrmeier M, Fitch WT (2015) Finding the beat: a
neural perspective across humans and non-human primates. Philos T R Soc B
370(1664): 20140093.
Mercier C, Bourbonnais D (2004) Relative shoulder flexor and handgrip strength is
related to upper limb function after stroke. Clin Rehabil 18: 215-221.
Mergner T (2007) Modeling sensorimotor control of human upright stance. Prog
Brain Res 165: 283-297.
Merton PA (1953) Speculations on the servo-control of movements. In: Malcolm
JL, Gray JAB, Wolstenholm GEW (Eds.) The spinal cord, pp. 183-198. Boston:
Little, Brown.
Mervis CB (2003) Williams syndrome: 15 years of psychological research. Dev
Neuropsychol 23: 1-12.
Merzenich MM, Nelson RJ, Stryker MS, Cynader MS, Schoppman A, Zook JM
(1984) Somatosensory cortical map changes following digit amputation in adult
monkeys. J Comp Neurol 224: 591-605.
Mesulam MM (1999) Spatial attention and neglect: Parietal, frontal and cingulate
contributions to the mental representation and attentional targeting of salient
extrapersonal events. Philos Trans R Soc Lond B Biol Sci 354: 1325-1346.
Meunier S, Katz R, Simonetta-Moreau M (2002) Central nervous system lesions
and segmental activity. Adv Exp Med Biol 508: 309-313.
Meyer CH, Lasker AG, Robinson DA (1985) The upper limit of human smooth
pursuit velocity. Vision Res 25(4): 561-563.
Miall RC (1998) The cerebellum, predictive control and motor coordination.
Novartis Found Symp 218: 272-284.
Michaelsen SM, Jacobs S, Roby-Brami A, Levin MF (2004) Compensation for
distal impairments of grasping in adults with hemiparesis. Exp Brain Res 157:
162-173.
Middleton FA, Strick PL (2000) Basal ganglia and cerebellar loops: Motor and
cognitive circuits. Brain Res Rev 31: 236-250.
Miller AD, Brooks VB (1981) Late muscular responses to arm perturbations persist
during supraspinal dysfunctions in monkeys. Exp Brain Res 41: 146-158.
Miller LE, Houk JC (1995) Motor co-ordinates in primate red nucleus: Preferential
relation to muscle activation versus kinematic variables. J Physiol 488: 533-548.
Milner TE, Cloutier C (1998) Damping of the wrist joint during voluntary movement.
Exp Brain Res 122: 309-317.
Mizuguchi N, Kanosue K (2017) Changes in brain activity during action
observation and motor imagery: Their relationship with motor learning. Prog
Brain Res 234: 189-204.
Molnar GE (1978) Analysis of motor disorder in retarded infants and young
children. Amer J Ment Defic 83: 213-222.
Monks J (1989) Experiencing symptoms in chronic illness: Fatigue in multiple
sclerosis. Int Disabil Stud 11: 78-83.
Morasso P (1981) Spatial control of arm movements. Exp Brain Res 42: 223-227.
Morasso P (1983) Coordination aspects of arm trajectory formation. Hum Mov Sci
2: 197-210.
Morasso PG, Sanguineti V (2002) Ankle muscle stiffness alone cannot stabilize
balance during quiet standing. J Neurophysiol 88: 2157-2162.
Morris ME, Huxham FE, McGinley J, Iansek R (2001) Gait disorders and gait
rehabilitation in Parkinson’s disease. Adv Neurol 87: 347-361.
Morris AF, Vaughan SE, Vaccaro P (1982) Measurements of neuromuscular tone
and strength in Down’s syndrome children. J Ment Defic Res 26: 41-46.
Muratori LM, McIsaac TL, Gordon AM, Santello M (2008) Impaired anticipatory
control of force sharing patterns during whole-hand grasping in Parkinson’s
disease. Exp Brain Res 185: 41-52.
Murinson BB (2004) Stiff-person syndrome. Neurologist 10: 131-137.
Muris P, Steerneman P, Merckelbach H (1998) Difficulties in the understanding of
false belief: Specific to autism and other pervasive developmental disorders?
Psychol Rep 82: 51-57.
Mushiake H, Inase M, Tanji J (1991) Neuronal activity in the primate premotor,
supplementary, and precentral motor cortex during visually guided and internally
determined sequential movements. J Neurophysiol 66: 705-718.
Mussa-Ivaldi FA, Morasso P, Zaccaria R (1989) Kinematic networks. A distributed
model for representing and regularizing motor redundancy. Biol Cybern 60: 1-
16.
Mutha PK, Haaland KY, Sainburg RL (2012) The effects of brain lateralization on
motor control and adaptation. J Mot Behav 44: 455-469.
Myklebust BM (1990) A review of myotatic reflexes and the development of motor
control and gait in infants and children: A special communication. Phys Ther 70:
188-203.
Myklebust BM, Gottlieb GL (1993) Development of the stretch reflex in the
newborn: Reciprocal excitation and reflex irradiation. Child Dev 64: 1036-1045.
Nachev P, Kennard C, Husain M (2008) Functional role of the supplementary and
pre-supplementary motor areas. Nat Rev Neurosci 9(11): 856-869.
Nakao M, Inoue Y, Murkami H (1989) Aging process of leg muscle endurance in
males and females. Eur J Appl Physiol 59: 209-214.
Nann M, Cohen LG, Deecke L, Soekadar SR (2019) To jump or not to jump - The
Bereitschaftspotential required to jump into 192-meter abyss. Sci Rep 9(1):
2243.
Napier JR (1956) The prehensile movements of the human hand. J Bone Joint
Surg 38B: 902-913.
Nardone A, Schieppati M (1988) Postural adjustments associated with voluntary
contractions of leg muscles in standing man. Exp Brain Res 69: 469-480.
Narici MV, Bordini M, Cerretelli P (1991) Effect of aging on human adductor pollicis
muscle function. J Appl Physiol 71: 1277-1281.
Nashner LM (1976) Adapting reflexes controlling human posture. Exp Brain Res
26: 59-72.
Nashner LM (1980) Balance adjustments of humans perturbed while walking. J
Neurophysiol 44: 650-664.
Nashner LM (1982) Adaptation of human movement to altered environments.
Trends Neurosci 5: 358-361.
Nashner LM, Cordo PJ (1981) Relation of automatic postural responses and
reaction-time voluntary movements of human leg muscles. Exp Brain Res 43:
395-405.
Nashner LM, McCollum G (1985) The organization of human postural movements:
A formal basis and experimental synthesis. Brain Behav 8: 135-172.
Nashner LM, Woollacott M, Tuma G (1979) Organization of rapid responses to
postural and locomotor-like perturbations of standing man. Exp Brain Res 36:
463-479.
Nelson W (1983) Physical principles for economies of skilled movements. Biol
Cybern 46: 135-147.
Netz J, Lammers T, Homberg V (1997) Reorganization of motor output in the non-
affected hemisphere after stroke. Brain 120: 1579-1586.
Newell KM, Corcos DM (1993) Variability in motor control. Champaign, IL: Human
Kinetics.
Nichols TR (1989) The organization of heterogenic reflexes among muscles
crossing the ankle joint in the decerebrate cat. J Physiol 410: 463-477.
Nichols TR (2002) Musculoskeletal mechanics: A foundation of motor physiology.
Adv Exp Med Biol 508: 473-479.
Nichols TR (2018). Distributed force feedback in the spinal cord and the regulation
of limb mechanics. J Neurophysiol 119: 1186-1200.
Nichols TR, Houk JC (1976) Improvement in linearity and regulation of stiffness
that results from actions of stretch reflex. J Neurophysiol 39: 119-142.
Nichols TR, Lawrence JH 3rd, Bonasera SJ (1993) Control of torque direction by
spinal pathways at the cat ankle joint. Exp Brain Res 97: 366-371.
Nicol C, Kuitunen S, Kyrolainen H, Avela J, Komi PV (2003) Effects of long- and
short-term fatiguing stretch-shortening cycle exercises on reflex EMG and force
of the tendon-muscle complex. Eur J Appl Physiol 90: 470-479.
Nicoll RA, Schmitz D (2005) Synaptic plasticity at hippocampal mossy fibre
synapses. Nat Rev Neurosci 6: 863-876.
Nieuwenhuijzen PHJA, Schillings AM, Galen GPV, Duysens J (2000) Modulation of
the startle response during human gait. J Neurophysiol 84(1): 65-74.
Nobuta S, Sato K, Komatsu T, Miyasaka Y, Hatori M (2005) Clinical results in
severe carpal tunnel syndrome and motor nerve conduction studies. J Orthop
Sci 10: 22-26.
Nouillot P, Bouisset S, Do MC (1992) Do fast voluntary movements necessitate
anticipatory postural adjustments even if equilibrium is unstable? Neurosci Lett
147: 1-4.
Nowak KJ, Davies KE (2004) Duchenne muscular dystrophy and dystrophin:
Pathogenesis and opportunities for treatment. EMBO Rep 5: 872-876.
Nowak K, McCullagh K, Poon E, Davies KE (2005) Muscular dystrophies related to
the cytoskeleton/nuclear envelope. Novartis Found Symp 264: 98-111.
Nudo RJ (2003) Adaptive plasticity in motor cortex: Implications for rehabilitation
after brain injury. J Rehabil Med 41 Suppl: 7-10.
Nudo RJ, Masterton RB (1989) Descending pathways to the spinal cord: II.
Quantitative study of the tectospinal tract in 23 mammals. J Comp Neurol
286(1): 96-119.
Obeso JA, Rothwell JC, Lang AE, Marsden CD (1983) Myoclonic dystonia.
Neurology 33: 825-830.
Obeso JA, Rothwell JC, Marsden CD (1985) The spectrum of cortical myoclonus:
From focal reflex jerks to spontaneous motor epilepsy. Brain 108: 193-224.
Obeso JA, Zamarbide I (2004) Classification, clinical features, and treatment of
myoclonus. Watts RL, Koller WC (Eds.) Movement disorders. Neurological
principles and practice, pp. 659-669. New York: McGraw-Hill.
Odabasi Z, Kuruoglu R, Oh SJ (2000) Turns-amplitude analysis and motor unit
potential analysis in myasthenia gravis. Acta Neurol Scand 101: 315-320.
O’Hare A, Khalid S (2002) The association of abnormal cerebellar function in
children with developmental coordination disorder and reading difficulties.
Dyslexia 8: 234-248.
Ohashi N, Nakagawa H, Asai M (1993) Contribution of vision to the stabilization of
body sway in patients with spinocerebellar degeneration. Acta Otolaryngol
Suppl 504: 117-119.
Ohtsuki T (1981). Inhibition of individual fingers during grip strength exertion.
Ergonomics 24: 21-36.
Ojakangas CL, Ebner TJ (1992) Purkinje cell complex and simple spike changes
during a voluntary arm movement learning task in the monkey. J Neurophysiol
68: 2222-2236.
Olafsdottir H, Yoshida N, Zatsiorsky VM, Latash ML (2005a) Anticipatory
covariation of finger forces during self-paced and reaction time force production.
Neurosci Lett 381: 92-96.
Olafsdottir H, Zatsiorsky VM, Latash ML (2005b) Is the thumb a fifth finger? A
study of digit interaction during force production tasks. Exp Brain Res 160: 203-
213.
Olafsdottir H, Zhang W, Zatsiorsky VM, Latash ML (2007) Age related changes in
multi-finger synergies in accurate moment of force production tasks. J Appl
Physiol 102: 1490-1501.
Olafsdottir H, Kim SW, Zatsiorsky VM, Latash ML (2008). Anticipatory synergy
adjustments in preparation to self-triggered perturbations in elderly individuals. J
Appl Biomech 24: 175-179.
Olafsdottir HB, Zatsiorsky VM, Latash ML (2008) The effects of strength training
on finger strength and hand dexterity in healthy elderly individuals. J Appl
Physiol 105: 1166-1178.
Oludare SO, Ma CC, Aruin AS (2017) Unilateral discomfort increases the use of
contralateral side during sit-to-stand transfer. Rehabil Res Pract 2017: 4853840.
Ongeboer de Visser BW, Bour LJ, Koelman JHTM, Speelman JD (1989)
Cumulative vibratory indices and the H/M ratio of the soleus H-reflex: A
quantitative study in control and spastic patients. Electroencephalog Clin
Neurophysiol 73: 162-166.
Orlovsky GN, Deliagina TG, Grillner S (1999) Neuronal control of locomotion.
From mollusk to man. Oxford: Oxford University Press.
Ostry DJ, Gribble PL (2016) Sensory plasticity in human motor learning. Trends
Neurosci 39(2): 114-123.
Oullier O, Marin L, Stoffregen TA, Bootsma RJ, Bardy BG (2006) Variability in
postural coordination dynamics. In: Davids K, Bennett S, Newell KM (Eds.)
Movement system variability, pp. 25-47. Champaign, IL: Human Kinetics.
Overduin SA, d’Avella A, Roh J, Carmena JM, Bizzi E (2015) Representation of
muscle synergies in the primate brain. J Neurosci 35: 12615-12624.
Owings TM. Grabiner MD (1998) Normally aging older adults demonstrate the
bilateral deficit during ramp and hold contractions. J Gerontol Ser A, Biol Sci &
Med Sci 53: B425-B429.
Oztop E, Bradley NS, Arbib MA (2004) Infant grasp learning: A computational
model. Exp Brain Res 158(4): 480-503.
Packard MG, Knowlton BJ (2002) Learning and memory functions of the basal
ganglia. Ann Rev Neurosci 25: 563-593.
Palmen SJ, van Engeland H, Hof PR, Schmitz C (2004) Neuropathological findings
in autism. Brain 127: 2572-2583.
Panzer VP, Bandinelli S, Hallett M (1995) Biomechanical assessment of quiet
standing and changes associated with aging. Arch Phys Med Rehabil 76: 151-
157.
Park J, Lewis MM, Huang X, Latash ML (2013) Effects of olivo-ponto-cerebellar
atrophy (OPCA) on finger interaction and coordination. Clin Neurophysiol 124:
991-998.
Park J, Lewis MM, Huang X, Latash ML (2014) Dopaminergic modulation of motor
coordination in Parkinson’s disease. Parkinsonism Rel Disord 20: 64-68.
Park J, Wu Y-H, Lewis MM, Huang X, Latash ML (2012) Changes in multi-finger
interaction and coordination in Parkinson’s disease. J Neurophysiol 108: 915-
924.
Partridge LD (1965) Modifications of neural output signals by muscles: A
frequency response study. J Appl Physiol 20: 150-156.
Pascual-Leone A (2001) The brain that plays music and is changed by it. Ann New
York Acad Sci 930: 315-329.
Pascual-Leone A, Dang N, Cohen LG, Brasil-Neto JP, Cammarota A, Hallett M
(1995) Modulation of muscle responses evoked by transcranial magnetic
stimulation during the acquisition of new fine motor skills. J Neurophysiol 74:
1037-1045.
Peeraully T (2014) Multiple system atrophy. Semin Neurol 34(2): 174-81.
Penfield W, Rasmussen T (1950) The cerebral cortex of man. A clinical study of
localization of function. New York: MacMillan.
Penn RD, Kroin JS (1984) Intrathecal baclofen alleviates spinal cord spasticity.
Lancet 1(8385): 1078.
Penn RD, Kroin JS (1987) Long-term intrathecal baclofen infusion for treatment of
spasticity. J Neurosurg 66: 181-185.
Penn RD, Mangieri EA (1993) Stiff-man syndrome treated with intrathecal
baclofen. Neurology 43: 2412.
Penn RD, Savoy SM, Corcos D, Latash M, Gottlieb G, Parke B, Kroin JS (1989)
Intrathecal baclofen for severe spinal spasticity. N Engl J Med 320: 1517-1521.
Pereira HS, Landgren M, Gillberg C, Forssberg H (2001) Parametric control of
fingertip forces during precision grip lifts in children with DCD (developmental
coordination disorder) and DAMP (deficits in attention motor control and
perception). Neuropsychologia 39: 478-488.
Person AL (2019) Corollary discharge signals in the cerebellum. Biological
Psychiatry: Cognitive Neuroscience and Neuroimaging, 4(9), 813-819.
Pessoa L, Medina L, Hof PR, Desfilis E (2019) Neural architecture of the
vertebrate brain: implications for the interaction between emotion and cognition.
Neurosci Biobehav Rev 107: 296-312.
Petersen NT, Pyndt HS, Nielsen JB (2003) Investigating human motor control by
transcranial magnetic stimulation. Exp Brain Res 152: 1-16.
Philipp R, Hoffmann KP (2014) Arm movements induced by electrical
microstimulation in the superior colliculus of the macaque monkey. J Neurosci
34(9): 3350-3363.
Phillips CG (1969) Motor apparatus of the baboon’s hand. Proc Roy Soc London
Ser B 173: 141-174.
Pilon JF, De Serres SJ, Feldman AG (2007) Threshold position control of arm
movement with anticipatory increase in grip force. Exp Brain Res 181: 49-67.
Pilowsky T, Yirmiya N, Arbelle S, Mozes T (2000) Theory of mind abilities of
children with schizophrenia, children with autism, and normally developing
children. Schizophr Res 42: 145-155.
Poggio T (1973) On holographic models of memory. Kybernetik 12: 237-238.
Pratt J, Chasteen AL, Abrams RA (1994) Rapid aimed limb movements: Age
differences and practice effects in component submovements. Psychol Aging 9:
325-334.
Prilutsky BI (2000) Coordination of two- and one-joint muscles: Functional
consequences and implications for motor control. Motor Control 4: 1-44.
Prilutsky BI, Zatsiorsky VM (2002) Optimization-based models of muscle
coordination. Exerc Sport Sci Rev 30: 32-38.
Prochazka A, Clarac F, Loeb GE, Rothwell JC, Wolpaw JR (2000) What do reflex
and voluntary mean? Modern views on an ancient debate. Exp Brain Res 130:
417-432.
Prochazka A, Gritsenko V, Yakovenko S (2002) Sensory control of locomotion:
Reflexes versus higher-level control. Adv Exp Med Biol 508: 357-367.
Proske U, Gandevia SC (2012) The proprioceptive senses: Their roles in signaling
body shape, body position and movement, and muscle force. Physiol Rev 92:
1651-1697.
Pruszynski JA, Johansson RS, Flanagan JR (2016) A rapid tactile-motor reflex
automatically guides reaching toward handheld objects. Curr Biol 26: 788-792.
Pruszynski JA, Kurtzer I, Nashed JY, Omrani M, Brouwer B, Scott SH (2011)
Primary motor cortex underlies multi-joint integration for fast feedback control.
Nature 478(7369): 387-390.
Pullman SL, Watts RL, Juncos JL, Sanes JN (1990) Movement amplitude choice
reaction time performance in Parkinson’s disease may be independent of
dopaminergic status. J Neurol Neurosurg Psychiat 53: 279-283.
Purves D, Augustine GJ, Fitzpatrick D, Hall WC, LaMantia AS, Mooney RD, …
White LE (2017) Neuroscience. Sixth edition. Oxford: Sinauer Associates.
Rack PMH, Ross HF, Thilmann AF (1984) The ankle stretch reflexes in normal and
spastic subjects: The response to sinusoidal movement. Brain 107: 637-654.
Raibert MH (1977) Motor control and learning by the state space model. Doctoral
dissertation, Massachusetts Institute of Technology, Cambridge, MA.
Rainor AJ (2001) Strength, power, and coactivation in children with developmental
coordination disorder. Dev Med Child Neurol 43: 676-684.
Rajangam S, Tseng PH, Yin A, Lehew G, Schwarz D, Lebedev MA, Nicolelis MA
(2016) Wireless cortical brain-machine interface for whole-body navigation in
primates. Sci Rep 6: 22170.
Ramstead MJD, Badcock PB, Friston KJ (2018) Answering Schrödinger’s
question: A free-energy formulation. Phys Life Rev 24: 1-16.
Rand MK, Stelmach GE, Bloedel JR (2000) Movement accuracy constraints in
Parkinson’s disease patients. Neuropsychologia 38: 203-212.
Ranganathan VK, Siemionow V, Sahgal V, Yue GH (2001) Effects of aging on
hand function. J Am Geriatr Soc 49: 1478-1484.
Rankin LL, Enoka RM, Volz KA, Stuart DG (1988) Coexistence of twitch
potentiation and tetanic force decline in rat hindlimb muscle. J Appl Physiol 65:
2687-2695.
Rantanen T, Guralnik JM, Foley D, Masaki K, Leveille S, Curb JD, White L (1999)
Midlife hand grip strength as a predictor of old age disability. JAMA 281: 558-
560.
Rarick GL, Dobbins DA, Broadhead GG (1976) The motor domain and its
correlates in educated handicapped children. Englewood Cliffs, NJ: Prentice
Hall.
Ravaioli E, Oie KS, Kiemel T, Chiari L, Jeka JJ (2005) Nonlinear postural control in
response to visual translation. Exp Brain Res 160: 450-459.
Recanzone GH, Merzenich MM, Jenkins WM, Grajski KA, Dinse HR (1992)
Topographic reorganization of the hand representation in cortical area 3b of owl
monkeys trained in a frequency-discrimination task. J Neurophysiol 67: 1031-
1056.
Reddihough DS, Collins KJ (2003) The epidemiology and causes of cerebral palsy.
Aust J Physiother 49: 7-12.
Redfern MS, Jennings JR, Martin C, Furman JM (2001) Attention influences
sensory integration for postural control in older adults. Gait Posture 14: 211-216.
Reisman D, Scholz JP (2003). Aspects of joint coordination are preserved during
pointing in persons with post-stroke hemiparesis. Brain 126: 2510-2527.
Reschechtko S, Cuadra C, Latash ML (2018) Force illusions and drifts observed
during muscle vibration. J Neurophysiol 119: 326-336.
Reschechtko S, Latash ML (2017) Stability of hand force production: I. Hand level
control variables and multi-finger synergies. J Neurophysiol 118: 3152-3164.
Reschechtko S, Pruszynski JA (2020) Voluntary modification of rapid tactile-motor
responses during reaching differs from its visuomotor counterpart. J
Neurophysiol 124: 284-294.
Riccio GE (1993) Information in movement variability about the qualitative
dynamics of posture and orientation. In: Newell KM, Corcos DM (Eds.)
Variability and motor control, pp. 317-358. Champaign, IL: Human Kinetics.
Riley MA, Wong S, Mitra S, Turvey MT (1997) Common effects of touch and vision
on postural parameters. Exp Brain Res 117: 165-170.
Rizzolatti G, Gentilucci M (1988) Motor and visual-motor functions of the premotor
cortex. In P. Rakic & W. Singer (Eds.), Neurobiology of Neocortex (42, pp. 269-
284): John Wiley & Sons.
Rizzolatti G, Craighero L (2004) The mirror-neuron system. Ann Rev Neurosci 27:
169-192.
Rizzolatti G, Fogassi L, Gallese V (2001) Neurophysiological mechanisms
underlying the understanding and imitation of action. Nat Rev Neurosci 2: 661-
670.
Robinson DA (1976) Adaptive gain control of the vestibuloocular reflex by the
cerebellum. J Neurophysiol 39: 954-969.
Rodriguez GM, Aruin AS (2002) The effect of shoe wedges and lifts on symmetry
of stance and weight bearing in hemiparetic individuals. Arch Phys Med Rehabil
83: 478-482.
Rogers MW, Kukulka CG, Soderberg GL (1992) Age-related changes in postural
responses preceding rapid self-paced and reaction time arm movements. J
Gerontol 47: M159-M165.
Roll JP, Vedel JP (1982) Kinaesthetic role of muscle afferents in man, studied by
tendon vibration and microneurography. Exp Brain Res 47: 177-190.
Roll JP, Vedel JP, Roll R (1989) Eye, head and skeletal muscle spindle feedback in
the elaboration of body references. Prog Brain Res 80: 113-123.
Roos MR, Rice CL, Vandervoort AA (1997) Age-related changes in motor unit
function. Muscle Nerve 20: 679-690.
Rose DK, Winstein CJ (2004) Bimanual training after stroke: Are two hands better
than one? Top Stroke Rehabil 11: 20-30.
Roseberry Thomas K, Lee AM, Lalive Arnaud L, Wilbrecht L, Bonci A, Kreitzer
Anatol C (2016) Cell-type-specific control of brainstem locomotor circuits by
basal ganglia. Cell 164(3): 526-537.
Rosenbaum DA, Engelbrecht SE, Busje MM, Loukopoulos LD (1993) Knowledge
model for selecting and producing reaching movements. J Mot Behav 25: 217-
227.
Rossi-Pool R, Zainos A, Alvarez M, Diaz-deLeon G, Romo R (2021) A continuum
of invariant sensory and behavioral-context perceptual coding in secondary
somatosensory cortex. Nat Commun 12(1): 1-13.Rossini PM, Barker AT,
Berardelli A, Caramia MT, Caruso G, Gracco RQ, Dimitrijevic MR, Hallett M,
Katayama Y, Lucking CH, Maertens-de Noordhout AL, Marsden CD, Murray
NMF, Rothwell JC, Swash M, Tomberg C (1994) Non-invasive electrical and
magnetic stimulation of the brain, spinal cord and roots; basic principles and
procedures for routine clinical application. Electroencephalog Clin Neurophysiol
91: 79-92.
Rossini PM, Pauri F (2000) Neuromagnetic integrated methods tracking human
brain mechanisms of sensorimotor areas ‘plastic’ reorganisation. Brain Res Rev
33: 131-154.
Rothwell JC (1994) Control of human voluntary movement. Second edition.
London: Chapman & Hall.
Rothwell JC, Obeso JA, Traub MM, Marsden CD (1983) The behaviour of the long-
latency stretch reflex in patients with Parkinson’s disease. J Neurol Neurosurg
Psychiatry 46: 35-44.
Rothwell JC, Traub MM, Day BL, Obeso JA, Thomas PK, Marsden CD (1982a)
Manual motor performance in a deafferented man. Brain 105: 515-542.
Rothwell JC, Traub MM, Marsden CD (1982b) Automatic and “voluntary”
responses compensating for disturbances of human thumb movements. Brain
Res 248: 33-41.
Safronov VA, Kandel EI (1975) The reflex on shortening (the Westphal
phenomenon) in deforming muscular (torsion) dystrophy. Zh Nevropatol
Psikhiatr Im S S Korsakova 75: 1495-1500 (in Russian).
Sainburg RL (2005) Handedness: Differential specializations for control of
trajectory and position. Exerc Sport Sci Rev 33: 206-213.
Sainburg RL, Ghez C, Kalakanis D (1999) Intersegmental dynamics are controlled
by sequential anticipatory, error correction, and postural mechanisms. J
Neurophysiol 81: 1045-1056.
Sainburg RL, Ghilardi MF, Poizner H, Ghez C (1995) Control of limb dynamics in
normal subjects and patients without proprioception. J Neurophysiol 73: 820-
835.
Sainburg RL, Poizner H, Ghez C (1993) Loss of proprioception produces deficits in
interjoint coordination. J Neurophysiol 70: 2136-2147.
Saltiel P, Wyler-Duda K, D’Avella A, Tresch MC, Bizzi E (2001) Muscle synergies
encoded within the spinal cord: Evidence from focal intraspinal NMDA
iontophoresis in the frog. J Neurophysiol 85: 605-619.
Samson MM, Crowe A, de Vreede PL, Dessens JA, Duursma SA, Verhaar HJ
(2001) Differences in gait parameters at a preferred walking speed in healthy
subjects due to age, height and body weight. Aging 13: 16-21.
Sanes JN (1985) Information processing deficits in Parkinson’s disease during
movement. Neuropsychologia 23: 381-392.
Sanes JN, Mauritz KH, Dalakas MC, Evarts EV (1985) Motor control in humans
with large-fiber sensory neuropathy. Hum Neurobiol 4: 101-114.
Sanes JN, Shadmehr R (1995) Sense of muscular effort and somesthetic afferent
information in humans. Can J Physiol Pharmacol 73: 223-233.
Santello M, Flanders M, Soechting JF (2002) Patterns of hand motion during
grasping and the influence of sensory guidance. J Neurosci 22: 1426-1435.
Santello M, Soechting JF (2000) Force synergies for multifingered grasping. Exp
Brain Res 133: 457-467.
Santos MJ, Kanekar N, Aruin AS (2010) The role of anticipatory postural
adjustments in compensatory control of posture: 1. Electromyographic analysis.
J Electromyogr Kinesiol 20: 388-397.
Saperstein DS, Barohn RJ (2004) Management of myasthenia gravis. Semin
Neurol 24: 41-48.
Sarlegna FR, Malfait N, Bringoux L, Bourdin C, Vercher J-L (2010) Force-field
adaptation without proprioception: Can vision be used to model limb dynamics?
Neuropsychologia 48(1): 60-67.
Sato H (1982) Functional characteristics of human skeletal muscle revealed by
spectral analysis of the surface electromyogram. Electromyogr Clin
Neurophysiol 22: 459-516.
Savelsbergh GJP, van der Kamp J, Rosengren KS (2006) Functional variability in
perceptual motor development. In: Davids K, Bennett S, Newell KM (Eds.)
Movement system variability, pp. 185-198. Champaign, IL: Human Kinetics.
Schambra HM, Sawaki L, Cohen LG (2003) Modulation of excitability of human
motor cortex (M1) by 1 Hz transcranial magnetic stimulation of the contralateral
M1. Clin Neurophysiol 114: 130-133.
Scheidt RA, Ghez C (2007) Separate adaptive mechanisms for controlling
trajectory and final position in reaching. J Neurophysiol 98: 3600-3613.
Schepens B, Drew T (2004) Independent and convergent signals from the
pontomedullary reticular formation contribute to the control of posture and
movement during reaching in the cat. J Neurophysiol 92(4): 2217-2238.
Schieber MH (1999) Voluntary descending control. In: Zigmond MJ, Bloom FE,
Landis SC, Roberts JL, Squire LR (Eds.) Fundamental neuroscience, pp. 931-
 949. San Diego: Academic Press.
Schieber MH (2001) Constraints on somatotopic organization in the primary motor
cortex. J Neurophysiol 86: 2125-2143.
Schieber MH, Santello M (2004) Hand function: Peripheral and central constraints
on performance. J Appl Physiol 96: 2293-2300.
Schlesinger G (1919) Der Mechanische Aufbau der Kunstlichen Glieder. In:
Borchardt M (Ed.) Ersatzglieder und Arbeitshilfen fur Kriegsbeschadigte und
Unfallverletzte, pp. 321-699. Berlin: Springer.
Schmahmann JD (2004) Disorders of the cerebellum: Ataxia, dysmetria of thought,
and the cerebellar cognitive affective syndrome. J Neuropsychiatry Clin
Neurosci 16: 367-378.
Schmahmann JD, Sherman JC (1998) The cerebellar cognitive affective
syndrome. Brain 121: 561-579.
Schmidt RA (1975). A schema theory of discrete motor skill learning. Psychol Rev
82: 225-260.
Schmidt RA (1980) Past and future issues in motor programming. Res Quart Exer
Sport 51: 122-140.
Schmidt RA, McGown C (1980) Terminal accuracy of unexpected loaded rapid
movements: Evidence for a mass-spring mechanism in programming. J Mot
Behav 12: 149-161.
Schmidt RA, Wrisberg CA (2007) Motor learning and performance: A situation-
based learning approach. Fourth edition. Champaign, IL: Human Kinetics.
Schmidt RC, Carello C, Turvey MT (1990) Phase transitions and critical
fluctuations in the visual coordination of rhythmic movements between people. J
Exp Psychol: Human Percept Perform 16: 227-247.
Schmidt RC, Turvey MT (1994) Phase-entrainment dynamics of visually coupled
rhythmic movements. Biol Cybern 170: 369-376.
Scholz JP, Danion F, Latash ML, Schöner G (2002) Understanding finger
coordination through analysis of the structure of force variability. Biol Cybern 86:
29-39.
Scholz JP, Schöner G (1999) The uncontrolled manifold concept: Identifying
control variables for a functional task. Exp Brain Res 126: 289-306.
Scholz JP, Schöner G, Latash ML (2000) Identifying the control structure of
multijoint coordination during pistol shooting. Exp Brain Res 135: 382-404.
Schöner G (1995) Recent developments and problems in human movement
science and their conceptual implications. Ecol Psychol 8: 291-314.
Schöner G, Kelso JAS (1988) Dynamic pattern generation in behavioral and neural
systems. Science 239: 1513-1520.
Schrödinger E (1944, 2012) What is life? Canto Classics. Cambridge, UK:
Cambridge University Press.
Schwindt P, Crill WE (1977) A persistent negative resistance in cat lumbar
motoneurons. Brain Res 120: 173-178.
Schwindt P, Crill WE (1981) Negative slope conductance at large depolarizations
in cat spinal motoneurons. Brain Res 207: 471-475.
Seidler RD, Alberts JL, Stelmach GE (2001) Multijoint movement control in
Parkinson’s disease. Exp Brain Res 140: 335-344.
Seidler-Dobrin RD, He J, Stelmach GE (1998) Coactivation to reduce variability in
the elderly. Motor Control 2: 314-330.
Seif-Naraghi AH, Winters JM (1990) Optimized strategies for scaling goal-directed
dynamic limb movements. In: Winters JM, Woo SL-Y (Eds.) Multiple muscle
systems. Biomechanics and movement organization, pp. 312-334. New York:
Springer-Verlag.
Seitz RJ, Butefisch CM, Kleiser R, Homberg V (2004) Reorganisation of cerebral
circuits in human ischemic brain disease. Restor Neurol Neurosci 22: 207-229.
Selionov VA, Shik ML (1984) Medullary locomotor strip and column in the cat.
Neuroscience 13: 1267-1278.
Senkowski D, Schneider TR, Foxe JJ, Engel AK (2008) Crossmodal binding
through neural coherence: implications for multisensory processing. Trends
Neurosci 31(8): 401-409.
Serlin DM, Schieber MH (1993) Morphologic regions of the multitendoned extrinsic
finger muscles in the monkey forearm. Acta Anat 146: 255-266.
Shadmehr R, Mussa-Ivaldi FA (1994) Adaptive representation of dynamics during
learning of a motor task. J Neurosci 14: 3208-3224.
Shadmehr R, Mussa-Ivaldi FA, Bizzi E (1993) Postural force fields of the human
arm and their role in generating multijoint movements. J Neurosci 13: 45-62.
Shadmehr R, Wise SP (2005) The computational neurobiology of reaching and
pointing. Cambridge, MA: MIT Press.
Shaklai S, Mimouni-Bloch A, Levin M, Friedman J (2017) Development of finger
force coordination in children. Exp Brain Res 235: 3709-3720.
Shapiro DC, Walter CB (1986) An examination of rapid positioning movements
with spatiotemporal constraints. J Mot Behav 18: 373-395.
Shapkov YuT, Shapkova EYu, Mushkin AYu (1995) Spinal generators of human
locomotor movements. In 4th IBRO World Congr Neurosci Abstracts, p. 349.
Kyoto, Japan.
Shapkova EYu (2004) Spinal locomotor capability revealed by electrical stimulation
of the lumbar enlargement in paraplegic patients. In: Latash ML, Levin MF
(Eds.) Progress in motor control-3, pp. 253-290. Champaign, IL: Human
Kinetics.
Shelton FNAP, Reding MJ (2001) Effect of lesion location on upper limb motor
recovery after stroke. Stroke 32: 107-112.
Sheridan MR, Flowers KA, Hurrell J (1987) Programming and execution of
movement in Parkinson’s disease. Brain 110: 1247-1271.
Sherrington CS (1910) Flexion reflex of the limb, crossed extension reflex, and
reflex stepping and standing. J Physiol 40: 28-121.
Shibasaki H, Yamashita Y, Kuroiwa Y (1978) Electroencephalographic studies of
myoclonus: Myoclonus-relate cortical spikes in progressive myoclonic epilepsy.
Brain 108: 225-240.
Shik ML, Orlovskii GN, Severin FV (1966) Organization of locomotor synergism.
Biofizika 11: 879-886 (in Russian).
Shik ML, Orlovsky GN (1976) Neurophysiology of locomotor automatism. Physiol
Rev 56: 465-501.
Shik ML, Severin FV, Orlovskii GN (1967) Structures of the brain stem responsible
for evoked locomotion. Fiziol Zh SSSR Im I M Sechenova 53:1125-1132 (in
Russian).
Shim JK, Latash ML, Zatsiorsky VM (2003) Prehension synergies: Trial-to-trial
variability and hierarchical organization of stable performance. Exp Brain Res
152: 173-184.
Shim JK, Latash ML, Zatsiorsky VM (2005) Prehension synergies in three
dimensions. J Neurophysiol 93: 766-776.
Shim JK, Lay B, Zatsiorsky VM, Latash ML (2004) Age-related changes in finger
coordination in static prehension tasks. J Appl Physiol 97: 213-224.
Shinohara M, Latash ML, Zatsiorsky VM (2003a) Age effects on force production
by the intrinsic and extrinsic hand muscles and finger interaction during maximal
contraction tasks. J Appl Physiol 95: 1361-1369.
Shinohara M, Li S, Kang N, Zatsiorsky VM, Latash ML (2003b) Effects of age and
gender on finger coordination in maximal contractions and submaximal force
matching tasks. J Appl Physiol 94: 259-270.
Shinohara M, Scholz JP, Zatsiorsky VM, Latash ML (2004) Finger interaction
during accurate multi-finger force production tasks in young and elderly
persons. Exp Brain Res 156: 282-292.
Shulman LM, Minagar A, Weiner WJ (2004) Multiple-system atrophy. In: Watts RL,
Koller WC (Eds.) Movement disorders: Neurologic principles and practice.
Second edition, pp. 359-369. New York: McGraw-Hill.
Shumway-Cook A, Woollacott MH (1985) Dynamics of postural control in the child
with Down syndrome. Phys Ther 65: 1315-1322.
Siebner HR, Rothwell J (2003) Transcranial magnetic stimulation: New insights
into representational cortical plasticity. Exp Brain Res 148: 1-16.
Siegel M, Buschman TJ, Miller EK (2015) Cortical information flow during flexible
sensorimotor decisions. Science 348(6241): 1352-1355.
Silani V, Ludolph A, Fornai F (2017) The emerging picture of ALS: A multisystem,
not only a “motor neuron disease. Arch Ital Biol 155(4): 99-109.
Silbert PL, Matsumoto JY, McManis PG, Stolp-Smith KA, Elliott BA, McEvoy KM
(1995) Intrathecal baclofen therapy in stiff-man syndrome: A double-blind,
placebo-controlled trial. Neurology 45: 1893-1897.
Singh T, Latash ML (2011) Effects of muscle fatigue on multi-muscle synergies.
Exp Brain Res 214: 335-350.
Singh T, SKM V, Zatsiorsky VM, Latash ML (2010) Fatigue and motor redundancy:
Adaptive increase in force variance in multi-finger tasks. J Neurophysiol 103:
2990-3000.
Singh T, Zatsiorsky VM, Latash ML (2012) Effects of fatigue on synergies in a
hierarchical system. Hum Mov Sci 31: 1379-1398.
Singh T, Zatsiorsky VM, Latash ML (2013) Contrasting effects of fatigue on multi-
finger coordination in young and older adults. J Appl Physiol 115: 456-467.
Sipila S, Suominen H (1995) Effects of strength and endurance training on thigh
and leg muscle mass and composition in elderly women. J Appl Physiol 78:
334-340.
Sittig AC, Denier van der Gon JJ, Gielen CC (1985) Separate control of arm
position and velocity demonstrated by vibration of muscle tendon in man. Exp
Brain Res 60: 445-453.
Sjogaard G, Kiens B, Jorgensen K, Saltin B (1986) Intramuscular pressure, EMG
and blood flow during low-level prolonged static contraction in man. Acta
Physiol Scand 128: 475-484.
Sjogaard G, Savard G, Juel C (1988) Muscle blood flow during isometric activity
and its relation to muscle fatigue. Eur J Appl Physiol Occup Physiol 57: 327-
335.
Skinner HB, Barrack RL, Cook SD, Haddad RJ Jr (1984) Joint position sense in
total knee arthroplasty. J Orthop Res 1: 276-283.
Slobounov S, Chiang H, Johnston J, Ray W (2002) Modulated cortical control of
individual fingers in experienced musicians: An EEG study.
Electroencephalographic study. Clin Neurophysiol 113: 2013-2024.
Smeets JB, Brenner E (1999) A new view on grasping. Motor Control 3: 237-271.
Smits-Engelsman BC, Wilson PH, Westenberg Y, Duysens J (2003) Fine motor
deficiencies in children with developmental coordination disorder and learning
disabilities: An underlying open-loop control deficit. Hum Mov Sci 22: 495-513.
Solopova IA, Selionov VA, Zhvansky DS, Gurfinkel VS, Ivanenko Y (2016) Human
cervical spinal cord circuitry activated by tonic input can generate rhythmic arm
movements. J Neurophysiol 115: 1018-1030.
Spencer RM, Zelaznik HN, Diedrichsen J, Ivry RB (2003) Disrupted timing of
discontinuous but not continuous movements by cerebellar lesions. Science
 300(5624): 1437-1439.
Spering M, Schütz AC, Braun DI, Gegenfurtner KR (2011) Keep your eyes on the
ball: smooth pursuit eye movements enhance prediction of visual motion. J
Neurophysiol 105(4): 1756-1767.
Sperry RW (1950) Neural basis of the spontaneous optokinetic response produced
by visual inversion. J Comp Physiol Psychol 43: 482-489.
Spidalieri HJ, Busby L, Lamarre Y (1983) Fast ballistic arm movements triggered
by visual, auditory, and somesthetic stimuli in the monkey. II. Effects of
unilateral dentate lesion on discharge of precentral cortical neurons and
reaction. J Neurophysiol 50: 1359-1379.
Spiegel KM, Stratton J, Burke JR, Glendinning DS, Enoka RM (1996) The
influence of age on the assessment of motor unit activation in a human hand
muscle. Exp Physiol 81: 805-819.
Spira ME, Yarom Y, Parnas I (1976) Modulation of spike frequency by regions of
special axonal geometry and by synaptic inputs. J Neurophysiol 39: 882-899.
Springer S, Khamis S (2017) Effects of functional electrical stimulation on gait in
people with multiple sclerosis - A systematic review. Mult Scler Relat Disord 13:
4-12.
State MW, Pauls DL, Leckman JF (2001) Tourette’s syndrome and related
disorders. Child Adolesc Psychiatr Clin N Am 10: 317-331.
Stayer C, Meinck HM (1998) Stiff-man syndrome: An overview. Neurologia 13: 83-
88.
Stein RB, Chong SL, James KB, Kido A, Bell GJ, Tubman LA, Belanger M (2002)
Electrical stimulation for therapy and mobility after spinal cord injury. Prog Brain
Res 137: 27-34.
Stein RB, Everaert DG, Thompson AK, Chong SL, Whittaker M, Robertson J,
Kuether G (2010) Long-term therapeutic and orthotic effects of a foot drop
stimulator on walking performance in progressive and nonprogressive
neurological disorders. Neurorehab Neural Repair 24: 152-167.
Stelmach GE, Goggin NL, Amrheim PC (1988) Aging and the restructuring of
precued movements. Psychol Aging 3: 151-157.
Stelmach GE, Goggin NL, Garcia-Colera A (1987) Movement specification time
with age. Exp Aging Res 13: 39-46.
Stelmach GE, Worringham CJ (1988) The preparation and production of isometric
force in Parkinson’s disease. Neuropsychologia 26: 93-103.
Stelmach GE, Worringham CJ, Strand EA (1986) Movement preparation in
Parkinson’s disease: The use of advance information. Brain 109: 1179-1194.
Sternad D (2002) Wachholder and Altenberger 1927: Foundational experiments for
current hypotheses on equilibrium-point control in voluntary movements. Motor
Control 6: 299-302.
Sterr A, Muller MM, Elbert T, Rockstroh B, Pantev C, Taub E (1998) Perceptual
correlates of changes in cortical representation of fingers in blind multifinger
Braille readers. J Neurosci 18: 4417-4423.
Strick PL (1983) The influence of motor preparation on the response of cerebellar
neurons to limb displacements. J Neurosci 3: 2007-2020.
Stuart DG, Pierce PA, Callister RJ, Brichta AM, McDonagh JC (2001) Sir Charles
S. Sherrington : Humanist, mentor, and movement neuroscientist. In: Latash
ML, Zatsiorsky VM (Eds.) Classics in movement science, pp. 317-374.
Champaign, IL: Human Kinetics.
Stuphorn V, Hoffmann KP, Miller LE (1999) Correlation of primate superior
colliculus and reticular formation discharge with proximal limb muscle activity. J
Neurophysiol 81(4): 1978-1982.
Subramanian SK, Feldman AG, Levin MF (2018) Spasticity may obscure motor
learning ability after stroke. J Neurophysiol 119: 5-20.
Sutherland DH, Olshen R, Biden EN, Wyatt MP (1988) The development of mature
walking. London: MacKeith Press.
Sutton GG, Mayer RF (1974) Focal reflex myoclonus. J Neurol Neurosurg Psychiat
37: 207-217.
Sylos-Labini F, La Scaleia V, Cappellini G et al. (2020) Distinct locomotor
precursors in newborn babies. Proc Natl Acad Sci USA 117: 9604-9612.
Takakusaki K (2017) Functional neuroanatomy for posture and gait control. J
Movement Disord 10(1): 1-17.
Tang PF, Woollacott MH (1998) Inefficient postural responses to unexpected slips
during walking in older adults. J Gerontol A Biol Sci Med Sci 53: M471-M480.
Tatton WG, Bawa P, Bruce IC, Lee RG (1978) Long loop reflexes in monkeys: An
interpretive base for human reflexes. Prog Clin Neurophysiol 4: 229-245.
Taub E, Morris DM (2001) Constraint-induced movement therapy to enhance
recovery after stroke. Curr Atheroscler Rep 3: 279-286.
Tawil R (2004) Facioscapulohumeral muscular dystrophy. Curr Neurol Neurosci
Rep 4: 51-54.
Taylor JL, Allen GM, Butler JE, Gandevia SC (2000) Supraspinal fatigue during
intermittent maximal voluntary contractions of the human elbow flexors. J Appl
Physiol 89: 305-313.
Taylor JL, Gandevia SC (2001) Transcranial magnetic stimulation and human
muscle fatigue. Muscle Nerve 24: 18-29.
Teasdale N, Simoneau M (2001) Attentional demands for postural control: The
effects of aging and sensory reintegration. Gait Posture 14: 203-210.
Teitelbaum O, Benton T, Shah PK, Prince A, Kelly JL, Teitelbaum P (2004) Eshkol-
Wachman movement notation in diagnosis: The early detection of Asperger’s
syndrome. Proc Natl Acad Sci USA 101: 11909-11914.
Teitelbaum P, Teitelbaum O, Nye J, Fryman J, Maurer RG (1998) Movement
analysis in infancy may be useful for early diagnosis of autism. Proc Natl Acad
Sci USA 95: 13982-13987.
Terekhov AV, Pesin YB, Niu X, Latash ML, Zatsiorsky VM (2010) An analytical
approach to the problem of inverse optimization: An application to human
prehension. J Math Biol 61: 423-453.
Teulings HL, Contreras-Vidal JL, Stelmach GE, Adler CH (1997) Parkinsonism
reduces coordination of fingers, wrist, and arm in fine motor control. Exp Neurol
146: 159-170.
Thach WT (1978) Correlation of neural discharge with pattern and force of
muscular activity, joint position, and direction of intended next movement in
motor cortex and cerebellum. J Neurophysiol 41: 654-676.
Thach WT (1998) A role for the cerebellum in learning movement coordination.
Neurobiol Learn Mem 70: 177-188.
Thach WT, Bastian AJ (2004) Role of the cerebellum in the control and adaptation
of gait in health and disease. Prog Brain Res 143: 353-366.
Thach WT, Goodkin HG, Keating JG (1992a) Cerebellum and the adaptive
coordination of movement. Ann Rev Neurosci 15: 403-442.
Thach WT, Kane SA, Mink JW, Goodkin HP (1992b) Cerebellar output: Multiple
maps and motor modes in movement coordination. In: Llinas R, Sotelo C (Eds.)
The cerebellum revisited, pp. 283-300. New York: Springer-Verlag.
Thelen E (1986) Development of coordinated movement: Implications for early
human development. In: Whiting HTA, Wade MG (Eds.) Motor development in
children: Aspects of coordination and control, pp. 107-124. Dordrecht, The
Netherlands: Martinus Nijhoff.
Thelen E (1995) Motor development. A new synthesis. Amer Psychol 50: 79-95.
Thelen E, Cooke DW (1987) Relationship between newborn stepping and later
walking: A new interpretation. Dev Med Child Neurol 29: 380-393.
Thelen E, Corbetta D, Kamm K, Spencer JP, Schneider K, Zernicke RF (1993) The
transition to reaching: Mapping intention and intrinsic dynamics. Child Dev 64:
1058-1098.
Thelen E, Corbetta D, Spencer JP (1996) Development of reaching during the first
year: Role of movement speed. J Exp Psychol Hum Percept Perform 22: 1059-
1076.
Thilmann AF, Fellows SJ, Ross HF (1991b) Biomechanical changes at the ankle
joint after stroke. J Neurol Neurosurg Psychiat 54: 134-139.
Thompson AK, Wolpaw JR (2015) Restoring walking after spinal cord injury:
Operant conditioning of spinal reflexes can help. Neuroscientist 21: 203-215.
Thompson PD (2001) The stiff-man syndrome and related disorders. Parkinsonism
Relat Disord 8: 147-153.
Thorpe SJ, Fabre-Thorpe M. (2001) Neuroscience. Seeking categories in the
brain. Science. 291(5502): 260-263.
Timmann D, Watts S, Hore J (1999) Failure of cerebellar patients to time finger
opening precisely causes ball high-low inaccuracy in overarm throws. J
Neurophysiol 82: 103-114.
Ting LH, Macpherson JM (2005) A limited set of muscle synergies for force control
during a postural task. J Neurophysiol 93: 609-613.
Ting LH, McKay JL (2007) Neuromechanics of muscle synergies for posture and
movement. Curr Opin Neurobiol 17: 622-628.
Topka H, Konczak J, Dichgans J (1998a) Coordination of multi-joint arm
movements in cerebellar ataxia: Analysis of hand and angular kinematics. Exp
Brain Res 119: 483-492.
Topka H, Konczak J, Schneider K, Boose A, Dichgans J (1998b) Multijoint arm
movements in cerebellar ataxia: Abnormal control of movement dynamics. Exp
Brain Res 119: 493-503.
Toro C, Hallett M (2004) Pathophysiology of myoclonic disorders. In: Watts RL,
Koller WC (Eds.) Movement disorders. Neurological principles and practice, pp.
671-681. New York: McGraw-Hill.
Traub MM, Rothwell JC, Marsden CD (1980) A grab reflex in the human hand.
Brain 103: 869-884,
Tresch MC, Jarc A (2009) The case for and against muscle synergies. Curr Opin
Neurobiol 19: 601-607.
Tsao CY, Mendell JR (1999) The childhood muscular dystrophies: Making order
out of chaos. Semin Neurol 19: 9-23.
Tseng Y, Scholz, JP, Schöner G, Hotchkiss L (2003). Effect of accuracy constraint
on the underlying joint coordination of pointing movements. Exp Brain Res 149:
276-288.
Ts’o DY, Frostig RD, Lieke EE, Grinvald A (1990) Functional organization of
primate visual cortex revealed by high resolution optical imaging. Science 249:
417-420.
Tsutsumi T, Don BM, Zaichkowsky LD, Delizonna LL (1997) Physical fitness and
psychological benefits of strength training in community dwelling older adults.
Appl Human Sci 16: 257-266.
Tuite DJ, Renstrom PA, O’Brien M (1997) The aging tendon. Scand J Med Sci Sp
7: 72-77.
Turner R (2000) fMRI: Methodology – sensorimotor function mapping. Adv Neurol
83: 213-220.
Turpin NA, Feldman AG, Levin MF (2017) Stretch-reflex threshold modulation
during active elbow movements in post-stroke survivors with spasticity. Clin
Neurophysiol 128: 1891-1897.
Turvey MT (1990) Coordination. Amer Psychol 45: 938-953.
Turvey MT (2007) Action and perception at the level of synergies. Hum Move Sci
26: 657-697.
Turvey MT, Carello C (1996) Dynamics of Bernstein’s level of synergies. In: Latash
ML, Turvey MT (Eds.) Dexterity and its development, pp. 339-376. Mahwah, NJ:
Erlbaum.
Tyson SF, Hanley M, Chillala J, Selley AB, Tallis RC (2008) Sensory loss in
hospital-admitted people with stroke: Characteristics, associated factors, and
relationship with function. Neurorehabil Neural Repair 22(2): 166-172.
Uchiyama T, Johansson H, Windhorst U (2003) Static and dynamic input-output
relations of the feline medial gastrocnemius motoneuron-muscle system
subjected to recurrent inhibition: A model study. Biol Cybern 89: 264-273.
Vaillancourt DE, Larsson L, Newell KM (2003) Effects of aging on force variability,
single motor unit discharge patterns, and the structure of 10, 20, and 40 Hz
EMG activity. Neurobiol Aging 24: 25-35.
Vaillancourt DE, Newell KM (2003) Aging and the time and frequency structure of
force output variability. J Appl Physiol 94: 903-912.
Vaillancourt DE, Prodoehl J, Verhagen Metman L, Bakay RA, Corcos DM (2004)
Effects of deep brain stimulation and medication on bradykinesia and muscle
activation in Parkinson’s disease. Brain 127: 491-504.
Vallar G, Bottini G, Paulesu E (2003) Neglect syndromes: The role of the parietal
cortex. Adv Neurol 93: 293-319.
Vallbo AB (1971) Muscle spindle response at the onset of isometric voluntary
contractions. Time difference between fusimotor and skeletomotor effects. J
Physiol 218: 405-431.
Vallbo AB (1981) Basic patterns of muscle spindle discharge in man. In: Taylor A,
Prochazka A (Eds.) Muscle receptors and movement, pp. 263-275. London:
MacMillan Press.
van Asten WN, Gielen CC, Denier van der Gon JJ (1988) Postural adjustments
induced by simulated motion of differently structured environments. Exp Brain
Res 73: 371-383.
Vandekerckhove I, De Beukelaer N, Van den Hauwe M, et al. (2020) Muscle
weakness has a limited effect on motor control of gait in Duchenne muscular
dystrophy. PLoS One 15(9): e0238445.
Vandervoort AA, Hayes KC (1989) Plantarflexor muscle function in young and
elderly women. Eur J Appl Physiol Occup Physiol 58: 389-394.
Vandervoort AA, Quinlan J, McComas AJ (1983) Twitch potentiation after voluntary
contraction. Exp Neurol 81: 141-152.
van Deursen RW, Sanchez MM, Ulbrecht JS, Cavanagh PR (1998) The role of
muscle spindles in ankle movement perception in human subjects with diabetic
 neuropathy. Exp Brain Res 120: 1-8.
van Deursen RW, Simoneau GG (1999) Foot and ankle sensory neuropathy,
proprioception, and postural stability. J Orthop Sports Phys Ther 29: 718-726.
Van Doren CL (1995) Pinch force matching errors predicted by an equilibrium-
point model. Exp Brain Res 106: 488-492.
Van Doren CL (1998) Differential effects of load stiffness on matching pinch force,
finger span, and effort. Exp Brain Res 120: 487-495.Van Groeningen CJ, Nijhof
EJ, Vermeule FM, Erkelens CJ (1999) Relation between torque history, firing
frequency, decruitment levels and force balance in two flexors of the elbow. Exp
Brain Res 129: 592-604.
Van Heijst JJ, Vos JE, Bullock D (1998) Development in a biologically inspired
spinal neural network for movement control. Neural Networks 11:1305-1316.
Van Kan PL, Houk JC. Gibson AR (1993) Output organization of intermediate
cerebellum of the monkey. J Neurophysiol 69: 57-73.
Verde F, Del Tredici K, Braak H, Ludolph A (2017) The multisystem degeneration
amyotrophic lateral sclerosis – neuropathological staging and clinical
translation. Arch Ital Biol 155(4): 118-130.
Vereijken B, van Emmerick REA, Whiting HTA, Newell KM (1992) Free(z)ing
degrees of freedom in skill acquisition. J Motor Behav 24: 133-142.
Verrel J, Lövden M, Lindenberger U (2012) Normal aging reduces motor synergies
in manual pointing. Neurobiol Aging 33: 200.e1-10.
Viallet F, Massion J, Massarino R, Khalil R (1987) Performance of a bimanual
load-lifting task by Parkinsonian patients. J Neurol Neurosurg Psychiat 50:
1274-1283.
Viitasalo JJ, Era P, Leskinen AL, Heikkinen E (1985) Muscle strength profiles and
anthropometry in random samples of men aged 31-35, 51-55, and 71-75 years.
Ergonomics 28: 1563-1574.
Vilensky JA, Moore AM, Eidelberg E, Walden JG (1992) Recovery of locomotion in
monkeys with spinal cord lesions. J Mot Behav 24: 288-296.
Vincent A (2000) Understanding neuromyotonia. Muscle Nerve 23: 655-657.
Viviani P, Terzuolo C (1980) Space-time invariance in learned motor skills. In:
Stelmach GE, Requin J (Eds.) Tutorials in motor behavior, pp. 525-533.
Amsterdam: North-Holland.
Voisin V, de la Porte S (2004) Therapeutic strategies for Duchenne and Becker
dystrophies. Int Rev Cytol 240: 1-30.
Von Holst E, Mittelstaedt H (1950/1973) Daz reafferezprincip. Wechselwirkungen
zwischen Zentralnerven-system und Peripherie, Naturwiss. 37: 467-476, 1950.
The reafference principle. In: Martin R (translator), The behavioral physiology of
animals and man. The collected papers of Erich von Holst, pp. 139-173. Coral
Gables, FL: University of Miami Press.
Wada M, Kawahara H, Shimada S, Miyazaki T, Baba H (2002) Joint proprioception
before and after total knee arthroplasty. Clin Orthop Relat Res 403: 161-167.
Wade MG, Van Emmerik RV, Kernozek TW (2000) Atypical dynamics of motor
behavior in Down syndrome. In: Weeks DJ, Chua R, Elliott D (Eds.) Perceptual-
motor behavior in Down syndrome, pp. 277-304. Champaign, IL: Human
Kinetics.
Wahnoun R, He J, Helms Tillery SI (2006) Selection and parameterization of
cortical neurons for neuroprosthetic control. J Neural Eng 3(2): 162-171.
Wall PD (1978) The gate control theory of pain mechanisms. A re-examination and
re-statement. Brain 101: 1-18.
Wang Y, Watanabe K, Asaka T (2017) Aging effect on muscle synergies in
stepping forth during a forward perturbation. Eur J Appl Physiol 117: 201-211.
Wang Y, Zatsiorsky VM, Latash ML (2005) Muscle synergies involved in shifting
center of pressure during making a first step. Exp Brain Res 167: 196-210.
Warabi T, Noda H, Kato T (1986) Effect of aging on sensorimotor functions of eye
and hand movements. Exp Neurol 92: 686-697.
Ward NS (2005) Neural plasticity and recovery of function. Prog Brain Res 150:
527-535.
Ward NS, Cohen LG (2004) Mechanisms underlying recovery of motor function
after stroke. Arch Neurol 61: 1844-1848.
Watson JD, Colebatch JG, McCloskey D (1984) Effects of externally imposed
elastic loads on the ability to estimate position and force. Behav Brain Res 13:
267-271.
Weihl CC, Roos RP (1999) Creutzfeldt-Jakob disease, new variant Creutzfeldt-
Jakob disease, and bovine spongiform encephalopathy. Neurol Clin 17: 835-
859.
Welford AT (1984) Psychomotor performance. Ann Rev Gerontol Geriatr 4: 237-
273.
Welsh JP, Llinás R (1997) Some organizing principles for the control of movement
based on olivocerebellar physiology. Prog Brain Res 114: 449-461.
Welsh T, Elliott D (2000) Preparation and control of goal-directed limb movements
in persons with Down syndrome. In: Weeks DJ, Chua R, Elliott D (Eds.)
Perceptual-motor behavior in Down syndrome, pp. 49-70. Champaign, IL:
Human Kinetics.
Werhahn KJ, Mortensen J, Van Boven RW, Zeuner KE, Cohen LG (2002)
Enhanced tactile spatial acuity and cortical processing during acute hand
deafferentation. Nat Neurosci 5: 936-938.
Werner W (1993) Neurons in the primate superior colliculus are active before and
during arm movements to visual targets. Eur J Neurosci 5(4): 335-340.
Wetts R, Kalaska JF, Smith AM (1985) Cerebellar nuclear cell activity during
antagonist cocontraction and reciprocal inhibition of forearm muscle. J
Neurophysiol 54: 231-244.
Windhorst U, Christakos CN, Koehler W, Hamm TM, Enoka RM, Stuart DG (1986)
Amplitude reduction of motor unit twitches during repetitive activation is
accompanied by relative increase of hyperpolarizing membrane potential
trajectories in homonymous α-motoneurons Brain Res 398: 181-184.
Winegard KJ, Hicks AL, Vandervoort AA (1997) An evaluation of the length-tension
relationship in elderly human plantarflexor muscles. J Gerontol Ser A, Biol Sci
Med Sci 52: B337-B343.
Wing AM (1988) A comparison of the rate of pinch grip force increases and
decreases in Parkinsonian bradykinesia. Neuropsychologia 26: 479-482.
Winter DA, Patla AE, Prince F, Ishac M, Gielo-Perczak K (1998) Stiffness control
of balance in quiet standing. J Neurophysiol 80: 1211-1221.
Winter DA, Prince F, Frank JS, Powell C, Zabjek KF (1996) Unified theory
regarding A/P and M/L balance in quiet stance. J Neurophysiol 75: 2334-2343.
Wolpaw JR (1987) Operant conditioning of primate spinal reflexes: The H-reflex. J
Neurophysiol 57: 443-459.
Wolpaw JR (2007) Brain-computer interfaces as new brain output pathways. J
Physiol 579: 613-619.
Wolpaw JR, Carp JS (1993) Adaptive plasticity in spinal cord. Adv Neurol 59: 163-
174.
Wolpert DM, Miall RC, Kawato M (1998) Internal models in the cerebellum. Trends
Cogn Sci 2: 338-347.
Woods JJ, Furbush F, Bigland-Ritchie B (1987) Evidence for a fatigue-induced
reflex inhibition of motoneuron firing rates. J Neurophysiol 58: 125-137.
Woollacott M, Inglin B, Manchester D (1988) Response preparation and posture
control. Neuromuscular changes in the older adult. Ann New York Acad Sci 515:
42-53.
Woollacott MH, Shumway-Cook A (1990) Changes in posture control across the
life span – a systems approach. Phys Ther 70: 799-807.
Wu Y-H, Pazin N, Zatsiorsky VM, Latash ML (2013) Improving finger coordination
in young and elderly persons. Exp Brain Res 226: 273-283.
Wu Y-H, Latash ML (2014) The effects of practice on coordination. Exerc Sport Sci
Rev 42: 37-42.
Yamagata M, Falaki A, Latash ML (2019) Effects of voluntary agonist-antagonist
co-activation on stability of vertical posture. Motor Control 23: 304-326.
Yamagata M, Gruben K, Falaki A, Ochs WL, Latash ML (2021) Biomechanics of
vertical posture and control with referent joint configurations. J Mot Behav 53:
72-83.
Yang J-F and Scholz JP (2005) Learning a throwing task is associated with
differential changes in the use of motor abundance. Exp Brain Res 163: 137-
158.
Yeo CH, Hardiman MJ, Glickstein M (1984) Discrete lesions of the cerebellar
cortex abolish the classically conditioned nictitating membrane response of the
rabbit. Behav Brain Res 13: 261-266.
Yufik YM, Friston K (2016) Life and understanding: Origins of the understanding
capacity in self-organizing nervous systems. Front Syst Neurosci 10: 98.
Zajac FE, Gordon ME (1989) Determining muscle’s force and action in multi-
articular movements. Exer Sport Sci Rev 17: 187-230
Zatsiorsky VM (1998) Kinematics of human motion. Human Kinetics: Champaign,
IL.
Zatsiorsky VM (2002) Kinetics of human motion. Human Kinetics: Champaign, IL.
Zatsiorsky VM, Duarte M (1999) Instant equilibrium point and its migration in
standing tasks: Rambling and trembling components of the stabilogram. Motor
Control 3, 28-38.
Zatsiorsky VM, Duarte M (2000) Rambling and trembling in quiet standing. Motor
Control 4: 185-200.
Zatsiorsky VM, Kraemer WJ, Fry AC (2021) Science and practice of strength
training. Third edition. Champaign, IL: Human Kinetics.
Zatsiorsky VM, Latash ML (2004) Prehension synergies. Exer Sport Sci Rev 32:
75-80.
Zatsiorsky VM, Li Z-M, Latash ML (1998) Coordinated force production in multi-
finger tasks. Finger interaction and neural network modeling. Biol Cybern 79:
139-150.
Zatsiorsky VM, Li Z-M, Latash ML (2000) Enslaving effects in multi-finger force
production. Exp Brain Res 131: 187-195.
Zhang L, Feldman AG, Levin MF (2018) Vestibular and corticospinal control of
human body orientation in the gravitational field. J Neurophysiol 120: 3026-
3041.
Zhang LQ, Rymer WZ (2001) Reflex and intrinsic changes induced by fatigue of
human elbow extensor muscles. J Neurophysiol 86: 1086-1094.
Zhang Y, Qiu T, Yuan X, Zhang J, Wang Y, Zhang N, Zhou C, Luo C, Zhang J
(2019) Abnormal topological organization of structural covariance networks in
amyotrophic lateral sclerosis. Neuroimage Clin 21:101619.
Zhou T, Zhang L, Latash ML (2015) Intentional and unintentional multi-joint
movements: Their nature and structure of variance. Neuroscience 289: 181-
193.
Zierski J, Muller H, Dralle D, Wurdinger T (1988) Implanted pump systems for
treatment of spasticity. Acta Neurochir Suppl (Wien) 43: 94-99.
Zimmerman SD, McCormick RJ, Vadlamudi RK, Thomas DP (1993) Age and
training alter collagen characteristics in fast- and slow- twitch rat limb muscle. J
Appl Physiol 75: 1670-1674.
Zoia S, Castiello U, Blason L, Scabar A (2005) Reaching in children with and
without developmental coordination disorder under normal and perturbed vision.
Dev Neuropsychol 27: 257-273.
 Index
Note: The italicized f and t following page numbers refer to figures
and tables, respectively.
A
abundance
principle of abundance, 205-206, 206f
accuracy, 213, 275-276, 289, 291-292
action potential
all-or-none law, 22
conduction
antidromic, 48, 155-158, 156f, 160, 160f
local currents, 28-29, 28f, 29f
orthodromic, 48, 155-158, 156f, 160, 160f
saltatory, 29
speed
along myelinated fibers, 29-30, 29f
along nonmyelinated fibers, 29
frequency
modulation, 32
generation, 22-25
transmission, 19
action–perception coupling, 265, 273-274
adaptation, 325
afferents
cutaneous, 166
exafference, 196, 269
flexor reflex afferents (FRA), 166, 166f, 170f, 225
Ia, 49, 151, 152f, 156-158, 169f
Ib, 164,164f, 169, 169f
II, 49
reafference, 196, 269, 269f
affordances, 6
aging
coactivation, 310-311, 311f
effects of training, 314
gait, 311-312
handwriting, 313
locomotion
speed, 307, 308f, 312, 312f
 muscle
apparent stiffness, 308, 311
contraction speed, 310
denervation, 308-309, 309f
distal, 310
fatigability, 309
motor units, 308-310, 308f, 309f
proximal, 310
reinnervation, 308-309, 309f
posture
anticipatory postural adjustments, 311
falling, 311
hip strategy, 312
prehension
grip force, 312, 313
synergy, 313
reaching movements, 307
reflexes
H-reflex, 310
tendon tap (T-reflex), 310
sarcopenia, 308
sensory function
peripheral neuropathy, 310
slowness, 307, 308f, 312, 312f
synergies
anticipatory synergy adjustments, 313
variability, 308-309, 309f
weakness, 312
agonists, 4, 98, 98f, 195f, 214-221, 214f, 220f, 221f
Altenburger, H., 5
amputation
locomotion, 394
plastic changes, 255, 392, 392f
amyotrophic lateral sclerosis (Lou Gehrig’s disease)
death of α-motoneurons, 381
dysarthria, 381
dysphagia, 381
epidemiology, 381
antagonists, 4, 98, 98f, 151-152, 152f, 159, 159f, 163-164, 163f, 195f, 214-221,
214f, 220f, 221f
Aristotle, 3
arm
dominant, 225
 nondominant, 225
Asperger, H., 321
Astruc, J., 4, 142
ATP, 20, 21f, 41f, 59
autism
cerebellum, 322
false belief, 322
prevalence, 321
stereotypy, 321
autoimmune process
Guillain-Barré syndrome, 344
multiple sclerosis, 304, 382-383
myasthenia gravis, 341-342
Avogadro number, 15
axon
myelinated, 29, 29f
nonmyelinated
B
basal ganglia
direct pathway, 86-87, 87f
disorders
ballism (hemiballismus), 362, 362f
dystonia, 362-364, 363f
Huntington’s disease
chorea, 361
Parkinson’s disease, 90, 356-361, 357f, 358f, 360f
globus pallidus
external, 84, 86, 88f
internal, 84, 86, 88f
indirect pathway, 86-87, 87f
medium spiny neurons, 84-86, 85f, 87, 87f
putamen, 83, 84-86, 84f
striatum, 83-85, 84f, 85f, 87f, 88f
substantia nigra
dopamine, 85, 85f
pars compacta, 84, 85f, 86f
pars reticulata, 84, 85f, 86f
subthalamic nucleus, 84, 86, 87f, 88f, 362, 362f
behaviors
catching, 293
locomotion, 176-177, 177f, 243-251
playing music, 254
 prehension, 253-260
reaching, 72, 74, 192f, 195f, 224-225, 224f, 230f, 276, 276f, 277-279, 278f, 280,
280f, 288-289, 290-292, 292f
speech, 125
standing, 176, 205f, 234-238, 235f, 237f, 238f, 240-241, 271, 271f
throwing
basketball, 231
Frisbee, 231
typing, 194
writing, 6, 193, 194f, 331f
Bernstein, N.A.
levels of movement construction, 201, 203-204
physiology of activity, 146, 327
Bizzi, E., 229
Borelli, G.A., 4
brain–computer interface (BCI)
brain–machine interface (BMI), 81-82, 82f
locked-in syndrome, 395
neuronal populations, 395
brainstem
midbrain
cerebral peduncles, 94, 94f, 95-96, 97
inferior colliculus, 85, 106, 125
superior colliculus, 86f, 102f, 103-104, 283, 289
tegmentum, 102f
medulla oblongata, 101, 101f
pons, 101, 101f
red nucleus, 5, 99f, 101, 104, 227, 228f
reticular formation, 102-103, 102f
Brodmann, K., 73, 73f
C
cells
amacrine, 128-129, 128f
Betz, 75, 76
bipolar, 128-129, 128f
horizontal, 128-129
magnocellular, 286
parvocellular, 286
pyramidal, 73, 76
retinal ganglion, 128-129, 128f
stellate, 73, 96, 96f
central pattern generators (CPG)
 in humans, 246-247
cerebellar disorders
asynergia (dyssynergia), 370
ataxia
ataxia telangiectasia, 370
ataxic dysarthria, 370
Friedreich’s ataxia, 370
cerebellar cognitive affective syndrome
agrammatism, 371
dysprosodia, 371
spatial cognition, 371
dysdiadochokinesis, 368
dysmetria, 368
dysrhythmokinesis, 368
hypotonia, 318, 366
motor adaptation, 369
motor learning, 369
movement segmentation, 368-369
reaction time, 366, 369, 369f
rigidity, 365
tremor
action, 366
intentional, 370
kinetic, 369-370
postural, 369-370
cerebellum
cerebellar cortex
cerebrocerebellum, 93-95, 97
spinocerebellum, 93, 95, 97, 99
vestibulocerebellum, 93-94, 96, 98, 99, 107
cerebellar nuclei
dentate nuclei, 95, 97, 100
fastigial nuclei, 95, 97, 107
interposed nuclei, 95, 97
cerebellar peduncles
inferior cerebellar, 94, 95-96
middle cerebellar, 94, 95, 97
superior cerebellar, 94, 95, 97
climbing fibers, 95, 96-97, 96f, 97f, 99
granule cells, 95, 96, 96f
hemispheres, 93
inferior olives, 95, 96, 96f, 99, 99f, 102
cerebellar layers
 granule cell layer, 96, 97f
molecular layer, 95, 96, 96f, 97f
Purkinje cell layer, 96, 97f
mossy fibers, 95, 96, 96f
nuclei, 93
Purkinje cells
complex spikes, 25
spinocerebellar tracts, 95, 97, 115
vermis, 93, 95, 97-98
cerebral palsy, 385f, 386f
diplegia, 384
epidemiology, 384
hemiplegia, 384
spasticity, 346, 384
treatment, 384-385
coactivation
command, 197, 220-222, 220f
increased, 310-311
zone, 231
cochlea, 125
conduction
antidromic, 155, 156f
local currents, 28-29, 29f
orthodromic, 155, 156f
saltatory, 29
contraction
concentric, 400
eccentric, 41
isometric, 212, 217-218, 217f, 218f, 219f
isotonic, 44
tetanus
sawtooth, 42
smooth, 42, 42f
twitch, 41-42, 42f, 58, 58f, 301, 301f
control theory
comparator, 184, 185f, 186
control variables, 184-185, 184f, 185f, 228, 230f
feedback control, 184-186, 185f, 289, 292f
feedforward control, 184-185, 184f
servo, 186-188, 186f
corpus callosum, 71-72
cortex
auditory cortex, 125-126
 canonical neurons, 74, 290
convergence, 95
divergence, 95
frontal eye fields, 289
medial superior temporal, 131f, 132, 132f, 135f
middle temporal, 131f, 132, 132f, 135f
mirror neurons, 74, 258
parietal area, 71, 72f, 77, 116
parietoinsular vestibular cortex, 122, 123f
plasticity, 392
prefrontal, 77-79
premotor area, 71, 72, 74-75
presupplementary motor area, 74, 75f
primary motor area, 71, 75-76
primary visual cortex (V1), 125, 130f, 131, 132
representations
plasticity, 392
somatosensory areas, 115, 115f
supplementary motor area, 71, 74-75, 75f, 79
cranial nerves, 104f, 105-106, 105f
currents
local, 28, 28f
persistent, 25, 31, 32f, 153-154
D
Darwin, C., 5
deafferentation, 173, 244
degree of freedom (DOF)
definition, 212
freezing, 332
problems of DOF, 202
releasing, 332
Democritus, 3
Descartes, R., 4
descending pathways
pyramidal tract, 106
rubrospinal tract, 5
development
atypical
Asperger’s disorder, 321-322
autism, 321-322
developmental coordination disorder, 322
Down syndrome, 317-321
 Duchenne muscular dystrophy, 340
of synergies
locomotion, 317
reaching, 317
typical
dynamic systems
ecological psychology, 316
emergence of patterns, 316
exploration, 316
milestones, 316, 316f
myelinization, 315
primitive reflexes, 316
diabetes mellitus
peripheral neuropathy, 343
postural control, 238, 343, 343f
disinhibition, 152
Down, J.H.L., 317
Down syndrome
cerebellum, 319
clumsiness, 318
coactivation, 318-319
effects of training
on EMG patterns, 319f
on movement speed, 320
on synergies, 320
genetics, 317
hypotonia, 318
nonmotor problems, 317
posture, 317-318
preprogrammed reactions, 318
prevalence, 317
slow motor development, 317, 318
slowness of movements, 318
submovements, 318
synergies, 320
dual-strategy hypothesis
excitation pulse, 219, 219f
isometric contractions, 218f, 219f, 220
isotonic movements, 220
speed-insensitive strategy, 218-219, 219f, 221f
speed-sensitive strategy, 219, 219f, 221-222, 221f
DuBois-Reymond, E., 4
dynamic system, 198-199, 243-244, 248-249, 316
dystonia
focal
athlete’s cramp, 363
blepharospasm, 362-363
musician’s cramp, 363
spasmodic dysphonia, 363
torticollis, 363
writer’s cramp, 363
generalized, 363
geste, 364
idiopathic, 362
secondary, 362
segmental, 363
variability, 363
E
ecological psychology
direct perception, 265
motor development, 316
perception–action coupling, 265, 273-274, 316
efference copy
classical definition, 196, 269
corollary discharge, 269
referent coordinate, 230-231, 269
electroencephalography (EEG)
evoked potential, 79, 377
readiness potential, 79
electromyography (EMG)
envelope, 64
filtering, 63-65, 65f
integration, 64
intramuscular, 61-63, 62f
normalization, 64, 302f
rectification, 63-65, 64f, 65f
spectrum, 304, 304f
surface, 61-63, 62f, 63f, 65f
triphasic pattern, 5, 174, 214, 214f, 215f, 216f, 217, 217f, 218f, 221
encephalization, 13
engineering, 6-7, 81, 191-192
engram, 193-194, 330
equifinality
violations of equifinality, 197
equilibrium-point hypothesis
 c-command, 197, 220-222, 220f
control variable, 220-221
EMG patterns, 221f, 222
equilibrium point, 189-190, 189f, 196f
equilibrium trajectory, 229-230, 230f
history, 195-196
invariant characteristic, 189f, 190
N-shaped control trajectory, 221
r-command, 197, 220-222, 220f, 221f
referent coordinate, 198
equilibrium potential
ion equilibrium potential, 17, 20-21, 20f
essential tremor
effects of alcohol, 384
frequency, 384
evolution of movements, 5-6
excitation
neuromediators, 32, 150
postsynaptic potential (EPSP), 33, 33f, 34f, 35f, 150, 150f
eye movements
nuclei
abducens, 123, 289
oculomotor, 123, 124f, 135, 289
trochlear, 289
reflexive
optokinetic reflex, 135
vestibulo-ocular reflex, 98, 105, 123-124, 124f, 135
voluntary
saccades, 78, 98, 102-103, 134-135, 135f, 289
smooth pursuits, 98, 102-103, 134, 135f, 290, 293
vergence, 134
F
fatigue
adaptation to fatigue, 303-304
central, 300, 304
chronic fatigue syndrome, 304
effects on
preprogrammed reactions (M2-M3), 302
reflexes, 299, 301-303
synergies, 304
multiple sclerosis, 304, 383
muscle activation
 conduction velocity, 300
relation to force, 299-301, 303-304
spectrum, 302
synchronization, 303-304
peripheral, 300, 300f
presynaptic inhibition, 302
spinal mechanisms, 301-303
supraspinal mechanisms, 303
Fechner, G.T., 266
Feldman, A.G., 229
equilibrium-point hypothesis, 77, 189-190, 195-198, 220-222
feedback
negative, 23, 23f, 151-152, 184-185, 185f
positive, 23, 23f, 184-185, 185f
finger
cortical representations, 254-255
enslaving, 255-257, 257f
error compensation, 257, 321f
extensors
extensor digitorum communis (EDC), 394
flexors
intrinsic, 254
extrinsic
flexor digitorum profundis (FDP), 254, 254f
flexor digitorum superficialis (FDS), 63, 254, 254f
force deficit, 256, 257f
minimization of secondary moments, 257
principle of superposition, 259-260
sharing, 257
synergies, 256-260
thumb, 256, 259f
virtual finger, 259, 259f
Flash, T., 229
Foster, M., 4
functional electrical stimulation (FES)
foot drop, 394
grasping, 394
G
Galen, 4
Galvani, L., 4
ganglia
basal, 6, 77, 77f, 83-90, 100, 289, 355-364
Gehrig, L., 381
Gelfand, I.M., 206, 228
Gibson, J., 6, 306
Goldman-Hodgkin-Katz equation, 21
Golgi, C., 4
Golgi tendon organs
reflex effects, 164
Graham Brown, T., 4, 244, 246
Grillner, S., 246
grip
aperture, 72, 258, 259f, 272
control with referent coordinates, 230-231
force, 174, 258, 258f, 312
power, 258
precision, 76, 258, 290
referent aperture, 258, 259f
safety margin, 312, 322
H
hand
anatomy, 253-254, 254f
dominant, 193, 194f, 223, 331f
joints, 253-254, 254f
muscles
extrinsic
compartments, 254
intrinsic, 254
nondominant, 193, 194f, 331f
Hawking, S., 382
hemiballismus
subthalamic nucleus, 362, 362f
Hill, A.V., 197
history of movement studies, 3-5
Hodgkin, A., 21
Hogan, N., 229
homeostasis, 151, 327
homunculus
motor, 75, 76f
sensory, 75, 115, 116f
Houk, J.C., 227
Hughlings Jackson, J., 75, 347, 349, 374
Humphrey, D., 77, 229
Huntington’s disease
 cachexia, 362
caudate nucleus, 361, 361f
chorea, 361
dysphagia, 361
epidemiology, 361
genetics, 361
Huxley, A., 21
I
ill-posed problems, 192, 201
illusions
Ebbinghaus, 285-286
flash-lag effect, 279
kinesthetic, 270-271, 271f
Müller-Lyer, 291, 291f
inhibition
neuromediators, 150
postsynaptic
potential (IPSP), 33, 33f, 150, 150f
presynaptic, 150, 152-153, 153f
reciprocal, 5, 151-152
recurrent
Renshaw cells, 151-152, 152f
internal model
direct, 194, 195f
inverse, 194, 195f
predictor, 194
interneurons
Ia, 151-152, 152f, 159, 159f, 163, 163f, 169, 169f
Ib, 164, 164f, 169, 169f
Renshaw cells, 151-152, 152f
inverse problems
inverse dynamics, 192, 201, 229
inverse kinematics, 192, 201, 229
ions
channels
density, 28, 29, 31
inactivation, 24-25, 24f, 28, 28f, 153
calcium, 20, 38, 40-41, 41f
chlorine, 20-21
conductance, 17, 21, 23-25, 23f, 25f
potassium, 17, 20-21, 23f, 24-25, 25f
sodium
 sodium–potassium pump, 20-21, 21f
isometric
EMG patterns, 217-219, 217f, 218f, 219f
step contraction, 217
pulse contraction, 217-218, 219f
isotonic
EMG patterns, 214-215, 215f, 216f
J
Jackson, J.H., 75, 347, 349, 374
Jankowska, E., 169
K
Kanner, L., 321
Kelso, J.A.S., 248, 249
kinematics
inverse kinematics, 192, 201, 229
kinesthetic perception
afferent components, 269-270, 269f
efference copy, 269-270, 269f, 271-272
effort, 272-273
illusions
Pinocchio effect, 271
vibration-induced, 270-271, 271f
iso-perceptual manifold, 273
perceptual equivalence, 273
preconceptions, 271
kinetics
external forces, 6
gravity, 8-9, 8f, 199, 199f, 233-234, 234f
interaction forces, 6
Kugler, P., 248
L
large-fiber peripheral neuropathy, 238
Lashley, K., 328
laws of behavior
Stevens law, 266
Weber-Fechner law, 48, 266
laws of nature
gravity law, 8, 199
Hooke’s law, 7, 7f, 43
second Newton’s law, 7, 7f
Levin, M.F., 229
Liddell, E., 166
linear system, 43, 191
lobes
frontal, 71, 72f, 84
occipital, 72f, 131, 374
parietal, 72f, 132, 373
temporal, 71, 72f, 132, 374
local field potential, 78
locomotion
central pattern generator (CPG)
half-center model, 244, 244f
in humans, 246-247
in lumbar enlargement, 246-247, 317
corrective stumbling reaction, 176-177, 250-251, 250f
in decerebrate preparations, 102, 245
dynamic systems, 198-199, 243-244, 248-249, 316
gait
gait transitions, 247
gallop, 246, 247, 247f
trot, 246, 247, 247f
walk, 246, 247, 247f
locomotor strip, 245, 245f
mesencephalic locomotor region, 102, 245f
motor programming, 243, 248, 248f
in spinal preparations, 246, 247
step initiation, 249-250
treadmill walking, 246, 247
long-term depression (LTD), 99, 329
long-term potentiation (LTP), 99, 329
Lorente de No, R., 328
Lundberg, A., 169
M
Marey, E.-J., 4
mass-spring model, 197
matrix factorization
independent component analysis (ICA), 205
non-negative matrix factorization (NNMF), 205
principal component analysis (PCA), 205
Matteucci, C., 4
membrane
afterpotential, 25
depolarization, 22-25, 22f, 23f, 29f, 33, 33f, 34, 39f, 150f, 152-153
hyperpolarization, 22-25, 22f, 33, 33f, 150f
 potential, 19-25, 20f, 22f, 31-32, 32f
refractoriness
absolute, 24, 24f
relative, 24, 24f
threshold, 22
memory
amnesia
anterograde, 328
retrograde, 328
conditioning
classical, 327
operant, 327
consolidation, 328-329
declarative (explicit), 326
habituation, 326-327
hysteresis, 326, 326f
long-term, 328-329, 328f
muscle memory, 326
nondeclarative (implicit), 326, 327, 330
reverberating circuits, 328
short-term, 328-329, 328f
Merton, R.A., 187, 188
mitochondria, 21f, 30, 40, 59
Mittelstaedt, H., 196, 269
motoneurons
alpha
size principle, 59-60, 301
gamma
dynamic, 50-51, 50f, 51f
static, 50-51, 50f
α-γ coactivation, 188, 188f, 267, 273
motor control
control theory, 6, 8, 8f
definition, 6-7
engineering approach, 6-7
physical approach, 7-9
motor learning
adaptation
to force field, 329-330
to visual rotation, 292-293
cortical representations
plasticity, 255
learning synergies, 330-332
 stages, 332
motor program
engram, 193-194, 330
generalized motor program, 193-194
schema theory, 194
motor skill
development, 6, 330
repetition without repetition, 6, 389
variability, 6, 141, 142, 205-206, 211, 213, 308
motor unit
changes with aging, 308-309, 308f, 309f
derecruitment, 59f, 60, 301, 304
fatigability, 301, 309
innervation ratio, 58, 309
recruitment, 60-61, 304
size principle, 59-60, 59f, 301
types
fast fatigable, 59, 59t
fast fatigue resistant, 59, 59t
slow, 59, 59t
Mountcastle, V., 292
movement of particles
convection, 14-15, 14f
diffusion, 15, 15f
movement of ions, 15, 16-17, 19-21, 19f
osmosis, 15-16
multijoint movement
equilibrium trajectory, 229-230, 230f
interaction torques, 193, 224-225
multijoint synergy, 231, 231f
optimization, 230-231
multisystem atrophy
ataxia, 383
bradykinesia, 383
epidemiology, 383
olivopontocerebellar atrophy (OPCA), 383
postural instability, 383
rigidity, 383
striatonigral degeneration, 383
tremor, 383
multiple sclerosis
autoimmune disease, 382
effects on
 coordination, 383
hearing, 383
muscle force, 304
sense of smell, 383
vision, 383
epidemiology, 382
fatigue, 383
myelin sheath
demyelination, 382
progression, 383
spasticity, 383
therapy, 383
tremor, 383
types, 382-383
multisensory integration
Bayesian integration, 277
multisensory neurons, 275
multisensory nuclei, 283
sensory reference frames, 279-280, 279f
sensory weighting, 276
muscle
biarticular, 212, 224, 226, 226f
compartment, 37, 61, 254
contraction
tetanic, 42, 42f, 58, 58f
twitch, 41-42, 42f, 58, 58f, 301, 301f
elasticity, 166-167, 240, 311
excitation–contraction coupling, 41
force
length dependence, 166f
velocity dependence, 167
models
contractile element, 42, 44
damping element, 42, 42f
elastic element
parallel, 42-43, 42f, 44
serial, 42-43, 42f, 44
Hill’s model, 44, 197
muscle–tendon complex, 37, 44f, 212, 268
skeletal, 37-45
tone (tonus)
decreased (hypotonia), 318, 366
increased (hypertonia), 365
muscle fiber
cross-bridge, 39, 41, 44
myofilament
A-band, 39, 39f
actin, 38-39, 39f, 41, 41f
I-band, 39, 39f
myosin, 38-39, 41, 41f
titin, 39
tropomyosin, 39, 39f, 41
troponin, 39, 39f, 41, 41f
Z-line, 39, 39f
sarcolemma, 38, 38f, 40, 41f
sarcomere, 39, 44
sarcoplasma
sarcoplasmic reticulum, 38, 38f, 40, 41, 41f
sliding filament mechanism, 41, 41f
T-tubule, 38, 38f, 40, 41f
muscular dystrophy
Becker, 340
Duchenne, 317, 340
dystrophin, 340
epidemiology, 340
myotonia, 340
myotonic dystrophy, 340
Mussa-Ivaldi, F.A., 229
Muybridge, E.J., 4
myasthenia gravis
acetylcholine receptors, 341
antibodies, 341-342
autoimmune process, 341-342
symptoms, 341-342
treatment, 342
myelin
Schwann cells, 29
sheath, 29, 29f, 52, 382
Ranvier nodes, 29, 29f, 31f, 382
myoclonus
action, 377
C-reflex, 377, 377f
Creutzfeldt-Jacob disease, 378
essential, 376
giant somatosensory evoked potentials, 377, 377f
negative, 377, 377f
 nocturnal, 376
reflex, 377, 378
reticular, 378
spina, 378
spontaneous, 377
N
Nernst equation, 16-17
neural fibers
afferent
Ia, 49, 151, 152f, 155, 156-158, 157f, 160, 163f, 165, 168
Ib, 51f, 164, 164f, 169
II, 49, 166, 166f
conduction speed, 7, 28
diameter, 29, 30, 30t
efferent, 30
myelinated, 29, 29f, 59
nonmyelinated, 29
neural tracts
ascending, 101
corticobulbar, 76, 106
colliculospinal, 106, 107f
corticospinal, 76-77, 77f, 78, 79, 95, 97, 106, 254
descending, 106-107
internal capsule, 114
medial lemniscus, 114-115, 114f, 270
optic nerve, 105, 105f, 128-129, 128f, 130-131, 133
optic radiations, 130-131, 130f, 133
optic tract, 130-131, 130f, 133
reticulospinal, 102, 106, 107f, 124
rubrospinal, 97, 104, 106, 107, 107f
spinocerebellar, 95, 97, 115
tectospinal, 106-107, 107f
vestibulospinal, 98, 104, 104f, 105, 106-107, 107f, 124, 236-237, 366
neuromyotonia
axon terminals, 341
neuron
axon
axonal hillock, 31, 31f, 150f, 157f
terminal branching, 31
dendrites, 31-32, 31f, 32f, 73, 96
soma, 30-31, 31f, 95
neurotransmitters
 amino acids
gamma aminobutyric acid (GABA), 34, 85, 85f, 152, 341, 351
glutamic acid, 34
leucine, 34
biogenic amines
acetylcholine, 34, 39f, 40, 85, 85f, 341-342, 361
dopamine, 34, 85, 87-88, 355f, 356, 361
norepinephrine, 34
serotonin, 34
neuropeptides
endorphins, 34
enkephalins, 34
neurotensin, 34
Nichols, T.R. 167, 169, 226
normal movement, 389-391
O
olivopontocerebellar atrophy, 367, 370-371, 381, 383
operant conditioning
changes in reflexes, 159, 326, 327, 328
optimization
cost functions
metabolic energy, 390
minimal effort, 202-203
minimal fatigue, 203
minimal jerk, 202
minimal movement time, 202, 202f
minimal torque-change, 203
inverse optimization, 203
Orlovsky, G.N., 245, 247
otolith organs
saccule, 119-121, 120f, 236
utricle, 119-121, 120f, 236
P
parameters
parametric control, 8-9, 199
Parkinson’s disease
cardinal signs
bradykinesia (akinesia), 90, 356-357, 356f, 357f, 358
postural instability, 356f, 357, 357f, 359-360
rigidity, 356f, 357, 357f
tremor, 356-357, 356f, 357f
epidemiology, 356
 locomotion
associated arm movements, 356
freezing of gait, 360
shuffling gait, 356, 360, 393-394
neurophysiology
direct loop, 357f
indirect loop, 357f, 361
substantia nigra, 356, 356f, 357f, 360
postural control
anticipatory postural adjustments (APAs), 357, 357f, 359
preprogrammed reactions, 357f
stooped posture, 90, 359
treatment
deep brain stimulation, 361
dopamine replacement, 356, 360-361
voluntary movement
interjoint coordination, 358
movement time, 356, 358
reaction time, 358
synergies, 360-361, 360f
variability, 358
Westphal phenomenon, 357
Pavlov, I.P.
theory of conditioned reflexes, 145-146, 327
Penfield, W., 75, 115, 255
Penn, R.D., 351
perception
action–perception coupling, 265, 273-274
auditory, 119, 123, 277
conscious, 48, 116, 270
direct, 265
exteroception, 47
haptic, 47, 53
interoception, 47
percept, 266, 272, 272f, 273
proprioception, 47, 114-115, 237-238, 276, 276f
smell, 47-48, 106t, 383
somatosensory, 113-116
visual, 6, 72, 240, 276, 276f
peripheral neuropathy
mononeuropathy
carpal tunnel syndrome
treatment, 343
 nerve conduction, 342
peroneal pressure palsy, 343
tarsal tunnel syndrome, 343
ulnar palsy, 343
polyneuropathy
Guillain-Barré syndrome, 344
permeability
ion permeability, 21, 22, 40
modulation during action potential, 22-25
peripheral large-fiber neuropathy, 54, 116
persistent ion currents, 25
Pflüger, E.F.W., 4
phase
in-phase, 198, 249, 258
out-of-phase, 198, 249, 258
transitions, 247
plasticity
cortical
maps, 255
H-reflex, 392
spinal, 392-393
topographic reorganization, 392
Plato, 3
posttetanic potentiation, 301
posture
anticipatory postural adjustments (APA), 311, 357, 357f, 359
center of mass (COM), 234-235, 234f, 235f, 238, 238f, 241
center of pressure (COP), 103, 234-235, 235f, 238-240, 239f, 241-242, 249,
250f
disorders
aging, 311-312
cerebellar disorders, 367, 368f
Down syndrome, 317-318
Parkinson’s disease, 90, 356-357, 357f, 359-360
tabes dorsalis, 238
effects of muscle vibration
vibration-induced fallings (VIFs), 238, 271, 271f, 343, 343f
inverted pendulum model, 234, 234f
posture-movement paradox, 196-197
preflexes, 240
preprogrammed reactions
ankle strategy, 240-241, 312
hip strategy, 240-241, 312
 role of vestibular system
galvanic vestibular stimulation, 237
role of vision, 237
stability, 233-234, 234f
sway
characteristics, 234
rambling-trembling decomposition, 235, 235f
synergies, 241-242
potentials
chemical, 16
equilibrium, 17, 20-21, 20f, 24-25
membrane, 19-25
miniature, 34, 40
plateau, 25
postsynaptic
excitatory (EPSP), 33, 33f, 34f, 35f, 150, 150f
inhibitory (IPSP), 33, 33f, 150, 150f
psychophysics, 266
prehension
grasping, 258
grip
aperture, 72, 258, 259f, 272
force, 312, 313
power, 258
precision, 76, 258, 290
safety margin, 312, 322
synergies
force-stabilizing, 260, 313f
moment-stabilizing, 257, 260
preprogrammed reactions
changes in
Down syndrome, 318
the elderly, 311, 313
Parkinson’s disease, 357f
corrective postural reactions
ankle strategy, 240-241, 312
hip strategy, 240-241, 312
corrective stumbling reaction, 176-177, 250-251, 250f
effects of fatigue, 302
effects of instruction, 171-175, 172f
latency, 172, 172f
M1, M2, and M3, 171-172
primitives, 227
proprioceptors
articular receptors, 52, 52f, 267-268
sensitivity to joint capsule tension, 52
sensitivity to joint position, 52
cutaneous and subcutaneous receptors
Meissner corpuscles, 53, 53f
Merkel disks, 53, 53f
Pacinian corpuscles, 52, 53, 53f
Ruffini corpuscles, 53, 53f
Golgi tendon organs, 51-52, 51f, 164, 267-268, 268f, 269, 270f
muscle spindles
bag fibers, 49, 49f, 50, 50f
chain fibers, 49, 49f, 50, 50f
fusimotor system, 50-51
primary endings, 49, 49f, 50f
secondary endings, 49-50, 49f, 50f
Pythagoras, 3
R
radiculopathy
disk herniation, 344
dorsal, 344
ventral, 344
Ramon Y Cajal, S., 4
receptors
articular, 52, 52f, 267-268
cutaneous, 52-53, 53f, 165, 166f
Golgi tendon organs, 51-52, 51f, 164, 267-268, 268f, 269, 270f
pain (nociceptors), 47, 52, 166
retinal photoreceptors
cones, 128-130, 129f
rods, 128-130, 129f
spindle endings
primary, 49, 49f, 50f
secondary, 49-50, 49f, 50f
subcutaneous, 53f, 165, 167, 190
vestibular, 122, 135, 173, 236, 240
reciprocal inhibition, 5, 151-152
red nucleus
magnocellular red nucleus, 227
rubrospinal tract, 5, 97, 104, 228f
redundancy
optimization, 202-203
 problem (Bernstein problem), 201-202, 225
referent coordinate (RC)
hierarchical control, 230-231
theory of movement control with RC, 230-231, 249, 258
reflex
arc
attenuation, 144, 144f
autogenic, 165
conditioned
Pavlov theory, 145-146, 146f, 327
definition, 141-143
F-wave, 160, 160f
gain, 144
grasp, 174
habituation, 144, 144f
inborn, 145-146, 327
interjoint and interlimb, 169-170, 224, 225-226
Jendrassik maneuver, 160
latency, 135, 143, 143f, 156, 160, 164, 166, 172
long-loop, 171-176
M-response, 155-158, 156f, 157f, 158f
monosynaptic
H-reflex
effects of voluntary activation, 159-160
suppression by vibration, 348-349, 349f
tendon tap (T-reflex), 158-159, 159f, 165, 310
oligosynaptic, 144-145, 145f, 163-164
phasic, 144, 145f, 164-165
polysynaptic
crossed extensor reflex, 170, 170f, 244
flexor reflex
flexor reflex afferents, 166, 166f, 170f, 225
tonic vibration reflex (TVR), 167-169, 167f, 168f
withdrawal, 166
startle response, 103, 103f, 107, 126
tonic
tonic stretch reflex
characteristic, 166f, 167
threshold, 166f, 167
tonic vibration reflex (TVR), 167-169, 167f, 168f
transcortical, 172
vestibulocervical, 107, 123
vestibulospinal, 105, 107, 123-124
 wiping, 226-227, 226f
Rexed, B., 148
S
Schmidt, R., 194
Schöner, G., 248
semicircular canals, 98, 119-123, 120f, 121f, 124f
sensation
Stevens’ law, 266
Weber-Fechner law, 48, 266
servo hypothesis, 187-189, 187f, 188f
Severin, F., 245, 247
Sherrington, C., 4, 75, 141, 166, 170, 244, 246
Shik, M.L., 245, 247
single-joint movements
dual-strategy hypothesis, 218-220
EMG patterns
within the EP hypothesis, 221
kinematics, 213-214, 213f, 220f
spasticity
Babinski reflex, 347, 348f
clasp-knife phenomenon, 347, 348, 348f, 352f
clonus, 347-348
epidemiology, 347
H-reflex
suppression by vibration, 348-349, 349f
muscle tone
Ashworth scale, 349, 349t
positive signs, 347
postsynaptic inhibition, 347, 348-349
presynaptic inhibition, 347, 349, 351, 352
effects of baclofen, 348, 351-352, 352f, 353f
GABA, 351, 352, 354
spasms
extensor, 347
flexor, 347
spasm scale, 349, 349t
stretch reflex threshold, 349
treatment
destructive chemical, 351
intrathecal baclofen, 348, 351-352, 352f, 353f, 354
oral drugs, 351
physical therapy, 351
 surgical, 351
voluntary movement, 347, 350, 352, 353f
spatial field, 6, 83
spinal cord
cauda equina, 149, 149f
dorsal column, 54, 113-114, 114f, 270, 270f
gray matter, 54, 147
horns
dorsal, 113, 148
ventral, 54, 113-114, 148
injury
chronic pain, 346
clinical consequences, 345-346, 346t
epidemiology, 345
flaccid syndromes, 346
functional electrical stimulations, 345, 346
paraparesis, 346
paraplegia, 346
spasticity, 346, 349
spinal shock, 346
laminar structure, 147, 147f
Rexed laminae, 147f, 148
roots
dorsal, 48f, 54f, 114-115, 148-149, 148f
ventral, 57f, 148-149, 148f
segments
cervical, 107, 148, 346t
lumbar, 107, 148, 246, 346t
sacral, 148
thoracic, 148, 346t
spinal preparation, 143, 246
white matter, 107, 147
spine
intervertebral disks, 148, 344
spinal processes, 148f
vertebrae, 148-149, 148f, 149f
stability
dynamical, 204-205, 207, 208, 227
Stevens, S.S., 266
stiff-man syndrome
effects of baclofen, 341
presynaptic inhibition, 341
stiffness
 apparent, 166f, 167, 229, 272, 273f, 308, 311, 313, 318
definition, 43
joint, 77, 229, 311, 313
muscle, 90, 167, 235, 267-268, 308, 348
tendon, 267-268, 348
stimulation
subthreshold, 79
suprathreshold, 79
stroke
aphasia
conduction, 374
fluent, 374
nonfluent, 374
apraxia, 373
denial, 373
hemorrhagic, 374
ischemic, 374
neglect, 373
plasticity, 374
proximal-to-distal gradient, 374
rehabilitation
constraint-induced therapy, 376
discomfort-induced therapy, 376
scissor gait, 374
spasticity, 374
synergy changes, 375-376, 375f
summation
spatial, 34-35, 35f
temporal, 34-35, 34f
synapse
exocytosis, 33, 40
neuromuscular
acetylcholine, 39f, 40
acetylcholinesterase, 39f, 40
active zone, 40
miniature endplate potential (MEPP), 40, 40f
re-uptake, 39f
neuro-neural, 33, 144, 153, 158
neurotransmitter (mediator), 33-34, 33f
non-obligatory, 33
obligatory, 33
presynaptic membrane, 32-34, 33f, 34f, 35f, 39f, 40, 150, 150f, 152-153, 153f
postsynaptic membrane, 32-35, 33f, 40,40f, 150, 150f, 152-153, 153f
 synaptic cleft, 32-33, 33f, 40, 152-153
synaptic transmission, 32-33, 158
synaptic vesicles, 33f, 39f, 40, 158
transmission, 32-33, 158
synergy
anticipatory synergy adjustments, 313, 360, 370, 376
asynergia (dyssynergia), 370
definition, 201, 312
changes with
aging, 313
atypical development, 317
neurological disorder, 360
practice, 204, 320-321, 330-332
error compensation, 208-209, 321f
grouping of elements
factors, 204
modes, 204-205
modules, 204
primitives, 204
level of synergies, 5, 83, 203-204
stability, 5, 204, 206-208
structural units, 206-207
T
tardive dyskinesia
oculogyric crisis, 364
tardive akathisia, 364
tardive dystonia, 364
tetanus
toxin, 341
thalamus
lateral geniculate nucleus, 130-132, 130f, 131f, 132f, 133f, 283, 286
ventral posterior nucleus, 114, 270
tics
blepharospasm, 362-363
clonic, 378
complex, 379t
dystonic, 378
motor, 378, 379t
sensory, 378, 379t
simple, 379t
Tourette syndrome, 378-379
treatment, 378, 379
 vocal, 378, 379t
voluntary suppression, 379
topology, 193, 330
Tourette syndrome
epidemiology, 378
rostral-to-caudal progression, 379
tics, 378-379
transcranial magnetic stimulation (TMS)
muscle responses, 79
physics, 254
Tsetlin, M.L., 206, 228
Turvey, M.T., 248
U
uncontrolled manifold (UCM)
hypothesis, 208-209
variance along the UCM, 231, 231f
variance orthogonal to the UCM, 208, 231, 231f
V
Vallbo, A., 188
variability
“bad”, 208-209, 209f
“good”, 208-209, 209f
role in skill acquisition, 224, 327
variables
control, 184-185, 185f, 190, 228-229, 230f
essential, 228
kinematic, 213, 214, 233, 330
mechanical, 113, 194, 205, 209, 330
state, 184-185, 186, 199
velocity
bell-shaped, 202, 213, 224f, 230, 318
patterns, 44, 44f
vestibular nuclei, 104-105, 104f, 107, 122, 123, 123f, 236, 236f
vibration
effects on monosynaptic reflexes
sensitivity of spindle endings, 49-50, 50f, 167
tonic vibration reflex (TVR), 167-169, 167f, 168f
vibration-induced falling, 238, 271, 343f
vibration-induced illusions, 270-271, 271f
vibration-induced locomotion, 168
voluntary suppression, 168
visual deficits
 anopsias
heteronymous hemianopsia, 133
homonymous hemianopsia, 133
homonymous quadrantanopsia, 133
cerebral achromatopsia, 133, 286
cerebral akinetopsia, 133, 286
color blindness, 130, 286
glaucoma, 133
macular degeneration, 133
optic ataxia, 286, 292
prosopagnosia, 286
spatial neglect, 71, 286
visual streams
dorsal, 72, 72f, 131f, 132-133, 132f, 133f, 286-290
dorsolateral, 72, 72f
dorsomedial, 72, 72f
ventral, 72, 72f, 131f, 132-133, 132f, 133f, 286-290, 288f
Von Holst, E.W., 196, 269
W
Wachholder, K., 5, 195-196
Weber, E.F., 4
Weber, E.H., 4, 266
Weber, W.E., 4
Williams syndrome, 379
Willis, T., 4
Wilson’s disease
basal ganglia, 385
copper, 385
shuffling gait, 385
stooped posture, 385
tremor
crescendo, 385
spread, 385
wing-beat, 385
wiping reflex
in spinal frog, 226-227, 226f
Wolpaw, J., 328, 392
Woodworth, R.S., 288-289
working point, 223-224, 224f, 225f
 About the Authors

Mark L. Latash, PhD, is a distinguished professor of kinesiology and

the director of the Motor Control Laboratory at Penn State University.
His research interests are the control and coordination of human
voluntary movements, movement disorders in neurological patients,
and effects of rehabilitation.
Latash has authored five books, edited 10 books, and had more
than 400 peer-reviewed articles published. He was the founding
editor of the journal Motor Control and is a former president of the
International Society of Motor Control (ISMC). He is a fellow of the
National Academy of Kinesiology (NAK) and member of the Society
for Neuroscience (SfN). He has served as director of the annual
Motor Control Summer School series and is a recipient of the ISMC’s
Bernstein Prize in motor control and Penn State’s Pauline Schmitt
Russell Distinguished Research Career Award.
Tarkeshwar Singh, PhD, is an assistant professor of kinesiology at
Penn State University, where he serves as the director of the
Sensorimotor Neuroscience and Learning Laboratory. Singh’s
research interests are in the areas of multisensory integration, motor
control, eye movements, and movement disorders. He has published
over 30 peer-reviewed papers and three book chapters. He is a
member of the Society for Neuroscience (SfN), the International
Society of Motor Control (ISMC), the Society for the Neural Control
of Movement (NCM), and the American Physiological Society (APS).