Exp Brain Res (1984) 55:263-271
E,x_mental
Preparatory Process for Anticipatory Postural Adjustments:
Modulation of Leg Muscles Reflex Pathways
During Preparation for Arm Movements in Standing Man
M.H. Woollacott t, M. Bonnet, and K. Yabe 2
Department of Experimental Psychobiology, Institute of Neurophysiology and Psychophysiology, National Center for Scientific Research,
F-13402 Marseille C6dex 9, France
Summary. We have investigated, in 6 standing sub-
jects, the time course of amplitude changes in the
short latency (40-60 ms) and long latency (60-80 ms)
reflex response components of gastrocnemius
medialis (GM) and soleus (S) muscles during the
preparatory period (time between a warning signal
and a response signal) as a function of the precued
direction (pull or push) of arm movement. Subjects
maintained their standing posture by visual feedback
during the 1.5 s preparatory period of a reaction time
task. A warning signal gave advance information
concerning the voluntary response to be performed
which consisted of either a pull or push movement of
the right arm. The excitability of the reflex pathways
was evaluated by triggering a rapid rotation of the
right ankle joint (dorsiflexion) applied randomly at
100, 300 or 500 ms before the response signal for arm
movement. Statistical analysis of EMG amplitude
(ANOVA) showed that preparatory effects were
different for the two synergistic muscles with both
short (spinal) and longer latency components of GM
showing generalized facilitation and of S showing
generalized inhibition. In addition, the longer
latency, presumably supraspinal, component of the
two muscles was differentially modulated according
to directional advance information, showing relative
facilitation for pull as compared to push trials. These
reflex response modulations were emphasized in
faster reaction time (RT) performers at the end of
the preparatory period. It is concluded that postural
preparatory processes are reflected at the spinal level
1 Present address: Department of P.E. and Human Movement
Studies, University of Oregon, Eugene, Oregon, USA; 2 Insti-
tute for Developmental Research, Aichi Prepectural Colony,
Kasugai, Aichi, Japan
Offprint requests to: M.H. Woollacott (address see footnote)
BranResearch
9 Springer-Verlag 1984
in global effects and at the supraspinal level in
directionally specific effects.
Key words: Movement preparation - Postural adjust-
ments - Long latency reflex - Man - EMG
Introduction
Previous research by Requin et al. (1977), and
Bonnet and Requin (1982) has explored changes in
both spinal and transcortical reflex pathways during
preparation for a simple voluntary movement involv-
ing flexion or extension about a single joint. The
research has demonstrated a decrease in spinal reflex
excitability, and increase in transcortical reflex
excitability of the implicated muscle in the subject at
rest. The following experiments were designed to
determine if the postural regulatory system, which in
standing subjects is phasically activated in advance of
voluntary movement as a function of the direction
of movement (Belenkii et al. 1967; Elner 1973; Lee
1980; Bouisset and Zattara 1981), is also modulated
by preparatory processes, either in a global fashion,
or more specifically, in relation to the direction of the
prepared movement. Toward this goal, we have
studied the reflex excitability simultaneously in two
synergistic muscles which are differentiated by the
nature of their intervention during the maintenance
of standing posture. The soleus muscle (S) of the
triceps surae is continually active during quiet stance;
it furnishes the largest part of the stiffness of the
ankle. In contrast, the gastrocnemius medialis muscle
(GM) shows only low level, intermittent activity
during quiet stance, but proportionally greater con-
tribution during the postural adjustments which
accompany voluntary arm movements.
264
In this experiment we have elicited a brief stretch
of the triceps surae by a dorsiflexing rotation of the
support of the foot about the axis of the ankle joint.
This stimulation elicits in GM and S muscles a
response containing two successive components, R1
and R2. The first component to muscle stretch (at
35-45 ms latencies) has been shown to be of spinal
origin, while the second component (at 55-70 ms
latencies) has been postulated to be affected by a
supraspinal loop.
This technique was used to determine if either
spinal or purported long loop reflex pathways of the
GM and S muscles of the leg were modulated during
preparation for a push (elbow extension) or pull
(elbow flexion) arm movement.
It was predicted that: (1) The reflex responses of
the postural muscles would be modulated differently
according to directional advance information (push
versus pull), since the triceps surae has a different
function in the corresponding feedforward postural
adjustments; (2) the GM and S muscles would be
modulated differently during preparation due to their
different functions in quiet stance (S being tonically
active and GM intermittently active); (3) R2 would
be more sensitive to the postural preparatory process
than R1 because purely spinal reflex pathways have
been shown to be less sensitive to instructional set
than longer latency pathways (Bonnet et al. 1981;
Bonnet 1983).
Materials and Methods
Apparatus
In order to test for modulation of the postural regulatory system, a
strain gauge equipped platform consisting of separate foot plates
on which the subject stood was triggered to produce a rapid
dorsiflexing rotation in the right ankle joint axis and elicit triceps
surae stretch responses. Rotations were produced by a solenoid
driven hammer attached to the back of the right foot plate. The
impulse command duration was 60 ms and the amplitude of
rotation mechanically limited to 3 ~. This device generated repro-
ducible strain gauge signal changes from the platform foot plate
(see Fig. 1C).
Three light-emitting diodes (LED) were arranged vertically
on the oscilloscope screen in front of the standing subject. The
central LED served as a continuous warning signal, while the top
and bottom LED's (2 cm apart) served as directional preparatory
and response signals (RS = 200 ms duration).
In order to measure reaction time for the push-pull response
movement, two photocells placed 1 cm from the central (rest)
handle position were activated when the handle was displaced.
Procedure and Experimental Design
Successive events of a trial (Fig. 2) controlled by a PDP 12
computer were as follows: The subject triggered the trial when he
@
PULL PUSH
, r
I
"
w
i
,
i
o ! !
o TA I ~ . . . . . . _..a.
, i
! i
i i
i ,
RS
I !
2 0 0 ms
82
40 eo 12o ms ~ - - I
~ ankle rotation
Fig. la. Postural responses before and during voluntary activation
of biceps brachialis in a reaction time task (push or pull). Rectified
EMG's from one trial are illustrated, b The time course of reflex
responses to ankle rotation for gastroc, medialis and soleus,
showing the early (R1) and late (R2) components (averaged Raw
EMG from 10 trials. GM given is twice that of SOL). c Position of
the subject during experimental trials. RS: response signal; SOL:
soleus; GM: gastrocnemius medialis; TA: tibialis anterior; BB:
biceps brachialis
judged via platform force feedback on an oscilloscope screen that
his position was correct; after two seconds the illumination of the
central LED either by itself (control trial), or with the addition of
directional information (experimental trial), served as a warning
signal (WS). In the case of advance information, one (or both)
peripheral LEDs were illuminated for 200 ms, instructing (or not)
the subject in advance about the direction of the rapid arm
movement he would have to perform. This warning signal started a
preparatory period of 1.5 s during which the ankle joint rotation
was randomly distributed at three times: 100, 300 and 500 ms
before the response signal.
The preparatory period ended by illuminating again one of
the peripheral LED's, for 100 ms, as a response signal, triggering
the previously instructed movement in the corresponding direc-
tion. After 5 s, the subject was able to trigger a new trial.
In a series of 66 trials, there were 12 randomly distributed
control trials without the reaction time task which formed the
reference sample for stretch responses. The three advance infor-
mation conditions: "Push" - top LED, "Pull" - bottom LED or

self-triggered ........ w a r n i n g signal .............................................................
ITI advance information
min
PUSH,PULL,NO=.33
5 s
! PREPARATORY
//
waiting period
-2 ~f~/ 0 i
0.5
Fig. 2. Experimental design. ITI: Intertrial interval
"No advance information" - with simultaneous top and bottom
LED's, were equiprobable (0.33) and randomly distributed. For
each advance information condition, the probability that the
response signal would occur was 0.83. The remaining trials where
the response signal did not occur were randomly included in the
series as catch trials in order to prevent anticipatory responses
which may appear in a reaction time task with a preparatory period
of fixed duration. Thus, in a series of 66 trials, each combination
(advance information x delay) was tested 6 times. Each subject
performed the task during at least three sessions consisting of
4 separate experimental series. The last three series of the last
session were used for statistical analysis.
Subject's Instruction
Six normal subjects participated in the final experiment (5 males, 1
female aged 18-43). After a brief explanation about the apparatus
and general procedure, they were instructed as follows: (1) to
grasp the handle capable of anterior-posterior movements against
springs; (2) to precisely maintain the central position of the beam
of the oscilloscope that registered the two directional displacement
of the center of vertical forces of the two feet (measured via
platform strain gauges); (3) to trigger each trial when ready and
the posture was centered (a button was placed under the thumb on
the handle); (4) to maintain this position when the LED's were
illuminated as a warning signal; in some cases only the central
LED was illuminated (control trials); (5) to move the handle as
fast as possible in the precued direction when the corresponding
LED was illuminated again as a response signal; (6) to take
warning that in some cases the response signal would be omitted
(catch trials); (7) that on some trials, the warning signal would be
formed by the simultaneous lighting up of the LED's, thus making
movement direction unpredictable until the response signal; (8)
that the right foot position might be unexpectedly disturbed during
the task.
Recordings
The platform output (anterior-posterior force changes) and the
surface EMG from the gastrocnemius medialis (GM) and soleus
(S) muscles (using surface electrodes placed 3 cm apart) were
recorded. The experimenter continually monitored the subject's
265
voluntary
response signal ......... a r m .........
~ movement
P= 87
PERIOD
I I I
ttt
ankle rotation
Reaction Time
i
1 1.5 S
maintenance of a central postural position. Activity of the biceps
brachialis (BB) muscle was also recorded for 3 subjects in order to
verify that the task required advance postural adjustments before
activation of the prime mover (BB) during pull trials. EMG
responses to ankle rotations and platform signals were displayed
on a storage oscilloscope. EMG was filtered (10 Hz - 3000 Hz)
rectified, digitized on-line with an effective sampling rate of 1000
Hz and collected on magnetic tape for later off-line analysis. For
each session, averaged data obtained during 10 advance control
perturbations were used to adjust 2 computer integration windows
20 ms in width. In most cases, the windows were adjusted from 40
to 60 ms (for the RI component) and from 60 to 80 ms (for the R2
component) after rotation onset.
Statistical Analysis
In order to compare and to average EMG data from different
subjects, the method of standardized z score was used. For each
series, after calculation of the mean and standard deviation of the
12 control stretch responses, the experimental response values
were converted into z scores, expressing the deviation of these
values from the control mean in control standard deviation units.
EMG data collected during error trials were eliminated. Analysis
of variance (ANOVA) was used to determine significance of
changes in EMG response amplitude during the preparatory
period. The effect of the following experimental factors was
tested: nature of advance information, muscle tested, component
of reflex response (R1, R2) and time before the response signal. In
addition, the ANOVA furnished information concerning the
presence of interaction between these factors.
Results
It has b e e n v e r i f i e d t h a t t h e s t a n d i n g p o s t u r e d o e s
not systematically change during the preparatory
p e r i o d as a f u n c t i o n o f i n s t r u c t i o n . T h e f a c t t h a t t h e
subject continuously grasped a handle and visually
monitored the force feedback signal (stabilogram)
c l e a r l y a i d e d h i m in this t a s k . A s h a s b e e n p r e v i o u s l y
266

Table 1. Reaction Time (milliseconds)

Advance information No Push No Pull
Response signal Push Push Improvement Pull Pull Improvement
Faster subjects 1 307 265 42 310 250 60
Faster subjects2 353 291 62 370 295 75
Faster subjects 3 355 290 65 368 297 71

Mean 338 282 56 349 281 68

Slower subjects4 436 345 91 375 314 61

Slower subjects 5 439 392 47 365 317 48
Slower subjects6 447 364 83 421 347 74

Mean 441 367 74 387 326 61

Total mean 390 325 65 368 303 65

found for tasks involving rapid arm movements in the
context of quiet stance, during the course of the
reaction time period it is possible to observe varia-
tions in the activity of the postural muscles of the leg
which precede the activation of the muscles of the
ipsilateral arm responsible for carrying out the reac-
tion time task (Fig. tA). It is postulated that this
activation serves to preset postural muscle stiffness to
counteract the anticipated perturbing forces of the
rapid arm movement. In the present experimental
context involving the act of pushing or pulling on a
spring-loaded handle, GM and S were activated in
advance of BB (60 ms for the subject in Fig. 1) during
pull trials while they showed a contrasting decreasing
activity during push trials (there is some coactivation
of biceps and triceps brachialis for both pull and push
movements).
Timing o f E M G Activity Following a Sudden A n k l e
Rotation
Figure 1B shows the averaged reflex activity of GM
and S muscles in response to 10 successive dorsiflex-
ing ankle rotations (control conditions). The stretch
response, superimposed upon postural E M G activ-
ity, occurred with approximately a 40 ms delay
following the platform signal change, and continued
for at least 40 ms. On the basis of latency and shape
of EMG responses, two successive components of the
burst were identified. These components appeared
regularly in all subjects in both GM and S muscles of
the rotated leg (without a corresponding contralat-
eral effect). From 40 to 60 ms after the platform
displacement started, the first response (R1), corre-
sponding to the activation of the spinal monosynaptic
reflex pathway appeared. The second component of
the response (R2) occurred between 60 and 80 ms
after stimulus onset.
Reaction Time
Table 1 shows the RT's observed during the experi-
ments. In order to determine if there was a relation-
ship between the subjects' performance level and
amplitude of EMG response to probe stretch during
the preparatory period, subjects were split into two
groups according to the average reaction time
observed with directional advance information.
Average reaction time for the three faster performing
subjects was 312 ms, while average reaction time for
the three slower performing subjects was 380 ms, the
difference (68 ms) being significant (p < 0.05).
Concerning the direction of the voluntary move-
ment (Push or Pull), average reaction time for the
three faster subjects was similar for the two responses
when directional advance information was provided
(Push = 282 ms, Pull = 281 ms). In the slow subjects,
a tendency to elongate reaction time before Push
responses was observed. The diminution of reaction
time with directional advance information is both 1)
identical on average for Push and for Pull, and 2)
similar in the two performance groups: total mean
improvement was 65 ms, or 17%. This difference was
significant (p < 0.001).
Main Changes in the Reflex Responses During
the Task
Figure 3 shows the mean amplitude of the reflex
response according to the responding muscle (GM or
S). The corresponding A N O V A indicates that the
modulation of the reflex responses during the course
of the preparatory period differed according to the
muscle tested (p < 0.05); in particular, a marked
facilitation was observed in the responses of GM. In
addition, the time course of the modttlation at the
end of the preparatory period differed for the two
 267

MUSCLE
Z
muscles (Interaction Delay x Muscle; p < 0.01) with
~score
the responses of GM tending to increase with the
-40
approach of the response signal in contrast to a
decrease in S. The amplitude of the reflex responses
did not depend on the type of directional advance
-%
.20

information. Neither did it depend on the response

component considered (R1 or R2). These two factors
,.s :' "-..~ _,;o I
0
show an interaction. The amplitude of each compo-
-soo -~>o ---.i I -.,o nent varies differently as a function of directional
pp ~.
response advance information (p < 0.001). This differential
signal
modulation of R1 and R2 as a function of preparat-
Fig. 3. Time course during the preparatory period of the main ory information is significant for each muscle.
effect for average difference between reflex response amplitude of
solens and gastrocnemius medialis

Time Course of the Two Components of the Reflex

RI R2 Z
Response During the Preparatory Period According
40 - 60 ms 60 - 80 ms to the Tested Muscle GM or S and to Directional
9 60
.a p U LL
Advance Information, Push or Pull
o :o
.40 m
g Figure 4 shows the mean amplitude of the two
z
components (R1 and R2) of the reflex responses
9 20
according to the time when the stimulation occurred
( 0 O
during the preparatory period9 Data are presented
I
m
separately for GM and S according to directional
PREPARATORY PERIOD I advance information (Push or Pull). Table 2 presents
responsl 9 40 >
signal
"o the two corresponding ANOVA's. The only signifi-
~PULL .20
cant comparison in common to the two components
c
O
m is relative to the difference in amplitude of the reflex
I o
during the task, between GM and S (for R1,
p < 0.05, and for R2, p < 09 The preparatory
- .20
changes always consisted of a facilitation of the
responses of GM including the strictly spinal compo-
-500 -300 -100 0 ms nent (Fig. 4, R1, GM) which becomes progressively
Fig. 4. Changes in reflex response amplitude of two successive
larger toward the end of the preparatory period
components (R1 and R2) of the response recorded simultaneously without relation to directional advance information.
in the two synergistic muscles, gastrocnemius medialis and soleus The modulation of the amplitude of the late

Table 2. Analysis of variance for two components: R1 and R2 of the stretch response during the preparatory period preceding the execution
of directional-precued arm movement (Push or Pull)

R1 R2
Degrees of F ratio Level of F ratio Level of
freedom significance significance

Advance information I 1-5 2.99 NS 13.47 p<0.025

(Push or Pull)
Tested muscle M 1-5 6.68 p<0.05 11.92 p<0.025
(GM or S)
Mechanical stimulation delay D 2-10 1.35 NS 0.53 NS
(-500, -300 or -100 ms)
None of the interactions were significant.

Partial comparison
I for Gastroc. 1-5 0.01 NS 4.90 0.05<p<0.10
I for Soleus 1-5 3.14 NS 19.60 p<0.01
M for Push 1-5 1.15 NS 19.01 p<0.01
M for Pull 1-5 3.33 NS 5.60 0.05<p<0.10
268

SLOW FAST are inhibited. For the information Pull, the most
marked facilitation of GM is present for only the last
two delays tested during the preparatory period
GASTROC MEDIALIS
(0.05 < p < 0.10).
Supplementary statistical analysis has shown that
.70
the principal effects demonstrated above are more
.60
marked in the rapid subjects, and only significant in
.50
this group. As is demonstrated by the temporal
.40
evolution of the amplitude of the reflex during the
.30 course of the preparatory period (Figs. 3 and 4), the
.20 differential variations are more marked for the last
9 10 r- -I I delay tested (-100 ms), indicating an increasing effect
0 of preparatory processes as the response signal
- 910 approaches. Keeping in mind these two considera-
PUSH NO PULL
tions, we present in Fig. 5 the results from the group
of 3 slower subjects (left side) compared to the
SOLEUS
results from the group of the most rapid subjects
(right side), and using only the data obtained from
the last delay. The accentuation of the modulation
.10 which is observed when one takes into account both
r
0 the level of performance and the proximity of the
--10
--20
t response signal underlines the specific preparatory
character of the reflex modulation. Moreover, for the
!i '-
-.30 i: condition "no directional advance information"
-.40 ~]R1 (No), the modulations observed were of a value
-.50 intermediary between Push and Pull (Fig. 5). This
~R2 PUSH NO PULL fact confirms the directional specificity of the pre-
paratory processes which influence reflex excitabil-
ity. However, it should be noted that there is a total
Fig. 5. Amplitude of two successive stretch response components absence of a directional modulation for the R1
(R1 and R2) of the synergistic muscles gastrocnemius and soleus as
a function of directional advance information (push, no or pull).
component of GM, in contrast to the modulation of
Data were averaged for the three slower subjects (left side) and R2.
the three faster subjects (right side). Data were averaged for the
final delay (100 ms prior to the RS) only. Corresponding values in
percent (faster subjects, final delay) were: GM, Push, RI: 121%,
R2: 95%, Pull, RI: 125%, R2: 128%, S, Push, RI: 104%, R2:
87%, Pull, RI: 92%, R2: 94%
Discussion
The above experiments used a fixed foreperiod
reaction time task requiring advance postural stabili-
response, R2, during the course of the preparatory zation in order to study the effects of preparation on
period is dependent on the nature of directional the postural control system. Preparatory changes in
advance information (p < 0.025). This relation is two reflex response components, purported to be
more marked for S (p < 0.01) than for the GM associated with separate neuron subsystems, and in
(0.05 < p < 0.01). the synergistic muscles GM and S were examined.
In each muscle, the amplitude of the R2 compo- Results indicate that only the late reflex component
nent was consistently larger for Pull advance infor- was sensitive to directional advance information
mation trials (Fig. 4, right side). In spite of the during the time course of the preparatory period;
identical nature of this differential effect of advance however, both components of the two synergistic
information on the two muscles, the between muscle muscles were on average, regardless of directional
comparison of absolute variations introduced by each advance information, modulated in an opposite man-
type of advance information shows that these two ner. On the basis of the behavioral data and the re-
synergists are modulated in a contrasting fashion. For quirements of the reaction time procedure employed,
the information Push, the difference between GM it will be argued that changes in the activity of
and S is highly significant (p < 0.01); the GM supraspinal structures activating postural muscles are
responses are on average facilitated, while those of S triggered by preparatory advance information.
 269

(motor cortex)

primary synergy . . . . . . . . . . . . . . . . . . . . . . primary syner lyJ

- - selection . . . . . . . . . . . triggering

(association cortex
basal ganglia
cerebellum) Ji~ ~
voluntary
instructi~)na,
S e e c t on
[ of l Command T
I Movement I System ) automatic

Directional I /
Advance _~_ I / / (bu,bo-spi..,)
Information .... ] ] .... 1_. . . . ~.__
----~P ~Stur]a/c~Y~ r gY] . . . . . . . . . . . t ...... --L~-- P~stt!~/SlYnngrg YI

Warning
Signal
Response
Signal
l 1 postural adju
[ imovement
tment I executionl

PREPARATORY PERIOD REACTION TIME

Fig. 6. Model of possible psychologicalprocesses and neurophysiologicalmechanisms underlyingpreparation and initiation of postural
adjustments associated with voluntary movements. There are three main points: (1) the preparation is expressed by a choice between
alternative synergies,(2) the muscularreflex response modulationsobserved do not express themselvesthrough anticipatorychangesin the
postural activityitself, (3) it is hypothesizedthat there is an inhibition of voluntary instructionaloutput of the command system by the
automatic contextual output

Contrasting Changes of Soleus and Gastrocnemius element and the GM intervenes as a "stabilizing"
Medialis Stretch Response Modulation element (Gurfinkel et al. 1965). Correlated with
these functional differences are differences in motor
A general preparatory effect was noted, which on unit and muscle fiber make up, with S possessing a
average depended neither on the modality of ad- greater proportion of slow muscle fibers and GM
vance information nor on the muscle response com- possessing a greater proportion of fast muscle fibers
ponent (R1 or R2). This consisted of a reciprocal (Buchtal and Schmalbruch 1980).
modulation of reflex excitability in the two muscles This differential evolution of the GM and S
with GM facilitated and S inhibited. This global during the preparatory period constitutes an argu-
aspect of postural preparation is interesting because ment against the hypothesis according to which the
it demonstrates opposite effects on the GM and S modulations of the reflex responses result from
which are generally considered synergists and often preparatory variations of the standing posture itself.
studied together as the triceps surae. This finding In effect, the overt changes of posture in the sagittal
should serve as a cautionary note to the common plane would be expected to be accompanied by
assumption that it is possible to record the activity of parallel changes in the GM and S muscles.
one of the muscles and abstract the results to the
function of both. What functional differences under-
lie these differences in modulation of reflex excitabil- Differential Modulation of Reflex Response
ity? Though considered synergists, the GM and S Components
differ in the following ways: 1) GM is biarticular,
whereas S is monoarticular and gives the greatest The lack of R1 amplitude change related to direc-
contribution to the maintenance of stiffness. 2) GM is tional advance information at the end of the pre-
used predominantly in phasic movements and is paratory period confirms results previously collected
usually silent during quiet stance, whereas S is in experiments using the H-reflex (Requin et al.
tonically active during quiet stance. 3) During stance, 1977), or the spinal component in the reflex response
the S intervenes as a fundamental "supportive" to rapid stretch of upper limb flexors (Crago et al.
270
1976; Marsden et al. 1978; Bonnet and Requin 1982;
Bonnet 1983). This effect suggests that the lower
level of the nervous system is implicated only in
unspecific processes of preparation. If it is accepted
that the programming process occurs during the
preparatory period (as a function of the information
which the subject possesses concerning the task
which he must execute), it is probable that the
direction parameter of the postural synergy is supra-
spinally organized and stored, and has no spinal
consequences until it intervenes at the moment of
movement execution.
It was found that the R2 change consistently
depended upon the expected voluntary movement
direction and was related to performance level. In
each muscle, the amplitude of the R2 component was
larger for pull advance information trials. It is exactly
in this pull condition that one observes during the
reaction time period a burst of activity in the tested
muscles.
In experiments measuring multiple muscle
responses, Cordo and Nashner (1982) showed that
postural adjustments activated in advance of rapid
arm movements are essentially identical in organiza-
tional structure to previously described automatic
postural adjustments (Nashner and Woollacott 1979)
which are activated by feedback mechanisms in
response to body sway perturbations. Thus, it is
likely that the same postural control system is acti-
vated in the two cases, and simply accessed in the one
case by voluntary control structures and in the other
case by somatosensory (ankle joint) afferents. We
show here that, in goal-directed action organization,
these same centers which control postural synergies
are selectively modulated during the preparatory
period like previously described primary movement
synergies (Bonnet 1983). Thus, in addition to the
command system linkage of the voluntary and post-
ural output, it can be hypothesized that the preparat-
ory processes of the two responses are linked
together by a common preparatory system.
Possible Mechanisms and Model
for Postural Preparatory Process
The proposed model (Fig. 6) represents possible
psychological processes and neurophysiological
mechanisms underlying preparation and initiation of
postural adjustments associated with voluntary
movements. The preparation is represented by the
selection of the appropriate primary (arm move-
ment) and postural response synergies from a pre-
determined alternative (push or pull). It is
hypothesized that the central circuits involved in the
activation of the primary and postural response
synergies are progressively organized (dotted lines)
during the preparatory period after the selection
process has taken place. Direction specific changes
are hypothesized only to occur in central circuits,
since in the present experiment and in previous
experiments on preparation for arm movements
(Bonnet 1983) only supra-spinally modulated
response components were directionally changed.
At the time of the response signal a second
process is triggered, in which a command system
(involving association cortex, basal ganglia, and
cerebellum) activates the previously facilitated prim-
ary and postural response synergies. It is
hypothesized that the two execution processes are
separate, and that the postural synergy output ini-
tially inhibits the primary synergy output (as
hypothesized by Cordo and Nashner 1982) for the
following reasons: (1) we observed that occasionally
subjects responded to "catch trials" with a postural
output, but without subsequent activation of the
primary synergy. This suggests that the outputs are
not tightly linked. (2) It has been shown previously
(Mankovskii et al. 1980) that when aging subjects are
asked to perform a similar movement as rapidly as
possible, they occasionally activate the primary
synergy without previous activation of the postural
synergy, and subsequently lose balance. This also
argues for the separability of the two systems, and in
addition, suggests a possible inhibitory mechanism
(Fig. 6) operating in the normal young adult, which
delays the voluntary output until after the beginning
of the activation of the postural synergy.
In conclusion, our results support the concept
that the postural system is modulated by preparatory
processes. This modulation appears to have both
global and directionally related effects on the post-
ural response system. The global effect involves the
preparatory facilitation of GM and inhibition of S
spinal reflexes (without regard to directional advance
information). The directionally related effects con-
cern only the presumed supraspinal component and
consist of relative facilitation of both GM and S for
pull compared to push trials. The similarity of
supraspinal processes underlying the preparation of
the voluntary response system (Bonnet 1983) and the
postural responses seen here argues for a linking of
the two systems during the central selection process,
and could indicate that both systems are subunits of
the same motor program.
Acknowledgements. This work was supported by an INSERM
grant no. 826016. M.H. Woollacott was supported by the Fyssen
Foundation.

References
Belenkii VE, Gurfinkel VS, Paltsev RI (1967) On the elements of
voluntary movement control. Biofizika 12:135-141
Bonnet M (1983) Anticipatory changes of long latency stretch
responses during preparation for directional hand move-
ments. Brain Res (in press)
Bonnet M, Requin J (1982) Long-loop and spinal reflexes in man
during preparation for intended directional hand movements.
J Neurosci 2:90-96
Bonnet M, Requin J, Semjen A (1981) Human reflexology and
motor preparation. In: Miller D (ed) Exercise and sport
sciences reviews, vol 9. Franklin Institute Press, Philadelphia,
pp 119-157
Bouisset S, Zattara M (1981) A sequence of postural movements
precedes voluntary movement. Neurosci Lett 22:263-270
Buehtal F, Schmalbruch H (1980) Motor unit of mammalian
muscle. Physiol Rev 60:90-142
Cordo PJ, Nashner LM (1982) Properties of postural adjustments
associated with rapid arm movements. J Neurophysiol 47:
287-302
Crago PE, Houk JC, Hasan Z (1976) Regulatory actions of human
stretch reflex. J Neurophysiol 39:925-935
Elner AN (1973) Possibilities of correcting the urgent voluntary
271
movements and the associated postural activity of human
muscles. Biofizika 18:907-911
Gurfinkel VS, Kots YM, Shik ML (1965) Postural regulation in
man (in Russian), Moscow, Science Publishing House
Lee WA (1980) Anticipatory control of postural and task muscles
during rapid arm flexion. J Motor Behav 12:185-196
Mankovskii N, Mints YA, Lysenyuk UP (1980) Regulation of the
preparatory period for complex voluntary movement in old
and extreme old age. Human Physiol (Moscow) 6:46-50
Marsden CD, Merton PA, Morton HB, Adam JER, Hallett M
(1978) Automatic and voluntary responses to muscle stretch
in man. In: Desmedt JE (ed) Cerebral motor control in man:
Long-loop mechanisms. Karger, Basel, pp 167-177
Nashner LM, Woollacott M (1979) The organization of rapid
postural adjustments of standing humans: An experimental-
conceptual model. In: Talbot RE, Humphrey DR (eds)
Posture and movement. Raven Press, New York, pp 243-257
Requin J, Bonnet M, Semjen A (1977) Is there a specificity in the
supraspinal control of motor structures during preparation?
In: Dormic S (ed) Attention and performance IV Lawrence
Erlbaum Assoc, Hillsdale, pp 139-174
Received July 20, 1983