klinini center ljubljana

University Medical Centre Ljubljana

SPS Nevroloka klinika

KO Intitut za klinino nevrofiziologijo

Slovenian-Italian Workshop on
Quantitative Needle and
High Resolution Surface EMG
University Medical Centre Ljubljana, Division of Neurology

Basic concepts and applications of

multichannel surface EMG
Roberto Merletti, Ph.D.
Lab. for Engineering of the Neuromuscular System, Politecnico di Torino, Italy

roberto.merletti@polito.it
www.lisin.polito.it

Prof. Lojze Vodovnik has

been one of my many
mentors and has strongly
influenced my career, my
approach to research and
my way of thinking about
problems.
He has also been a very
good friend.
I am in debt with him for
his teachings.
This lecture is dedicated
to his memory.
LISiN, Torino

WHAT SURFACE EMG IS AND IS NOT

1. Surface EMG is not a diagnostic technique and is not intended to
replace needle EMG.
2. Surface EMG is a monitoring technique suitable to study movement
and neuromuscular control and to assess muscle changes due to
aging, pathology, therapy, training, immobilization, lack of gravity,
occupational disorders, etc.
3. Standards are lacking: there have been successful EU efforts to
reach consensus and disseminate recommendations (SENIAM,
h.hermens@rrd.nl), and develop applications
(PROCID, roland.kadefors@niwl.se;
NEW, roberto.merletti@polito.it;
OASIS, paul.enck@uni-tuebingen.de;
CYBERMANS, alessandro.levizzari@crf.it)
4. There is a need for teaching and training in the medical schools.
LISiN, Torino

With respect to the needle technique, the

surface technique:
1. Is non invasive, non painful and without risks
2. Is global (provides global information)
3. Is simple and inexpensive
4. Is applicable by non medical personnel
5. Can be used over long times during work and sport
activities
6. Allows the measurement of quantities not measurable
with needles
7. Does not allow the measurement of quantities
measurable with needles
8. Is complementary (not a replacement) to the needle
technique

LISiN, Torino

At this time (a.d. 2006) the limitations of

surface EMG are:

1. Signals are dominated by the contributions of

superficial motor units
2. The thickness of skin and subcutaneous fat causes
strong blurring
3. Crosstalk from nearby muscles may be a serious
problem
4. Artifacts due to muscle movements (in dynamic
contractions) may be very strong

LISiN, Torino

Main applications of needle EMG

1. Diagnostics based on observations of single (or
very few) motor unit action potentials and of
their morphology and sound.
2. Fibrillation potentials in denervated fibers
3. Identification of MU territory (macro EMG)
4. All the above can be done or observed in either
superficial or deep muscles.

LISiN, Torino

Main applications of surface EMG

1. Biomechanics and movement analysis:
Identification of muscle activation intervals and levels, muscle coordination

2. Muscle fatigue and non invasive fiber typing:

Monitoring myoelectric manifestations of muscle fatigue, electrical and
mechanical responses of single motor units

3. Muscle physiopathology:
Monitoring muscle fiber conduction velocity, motor unit recruitment order

4. Occupational medicine:
Monitoring the Cinderellas, postural problems, muscle hyperactivity

5. Rehabilitation, space and sport medicine:

Assessment of effectiveness of treatments and training, monitoring
microgravity related changes and effectiveness of countermeasures

6. Pelvic floor analysis:

Detection of sphincter asymmetry, prevention of episiotomy related lesions.

7. Biofeedback:
Tension headache, muscle retraining, coordination retraining

Why EMG processing ?

To document differences between individuals and conditions (young-elderly,
before-after treatment or training, etc) by reporting EMG descriptors, that is
physical variables associated to the EMG signal. To observe central and
peripheral phenomena (such as myoelectric manifestations of muscle fatigue,
activation patterns, control strategies, etc), to assess effectiveness of treatments.

What descriptors ?
2 electrodes: Amplitude (ARV, RMS), frequency (MNF, MDF), amplitude envelope,
activation times during isometric or dynamic contractions.
4 electrodes: as above plus conduction velocity (CV), correlation coefficient (CC)
between the signals used for CV estimation.
5-16 or more electrodes in a linear array: descriptors of individual motor units
such as location of innervation zone, fiber length, highly accurate
CV estimates, firing rate, recruitment pattern, etc.

What conditions ?
Isometric conditions: these are special bench-tests to estimate values that are
much more difficult to estimate in dynamic conditions (rates of change of
descriptors, single Motor Unit features, etc).
Dynamic conditions: activation intervals (muscle on-off timing) during
movements, envelope detection, etc. These conditions are of greater physiological
interest but are affected by artifacts and may not produce reliable results.

Inputs from
other neurons
motoneuron

60

m
/s

The Motor Unit (MU)

(electrical activity)

Vm (mV)

Space or time
Action potential
(90-100 mVpp)

Axon

Schwann cells and

Ranvier nodes

- 70

Muscle fibers
1 ms o 4 mm

4 m/s = 4 mm/ms
One muscle: 10-1000 MU

4 m/s = 4 mm/ms

One MU: 50-1000 fibers of the same type (I o II)

V(t)

Potential distribution on the skin

x
CV

CV
Subcutaneous tissue

Innervation zone

Depolarized
Zone

V(x)

0 mV

- 70 mV

CV

Action potentials travelling towards the tendons

Skin

Muscle-tendon
junctions

CV
Single differential
amplifiers

Electrode array

CV

V2

V1

Subcutaneous
tissue

Differential
voltages

Monopolar
voltages in
space

V4

V3

Innervation
zone

V5

V7

V6

V1

V2
V3

V8

V4
V5

Skin

3 fiber motor unit

u r on
e
n
o
t
mo

0 mV

- 70 mV

Tendon
termination

Depolarized
zone

V(x)

Propagating
single fiber
action potentials

CV

Il passaggio delle zone depolarizzate

sotto una schiera di elettrodi genera
una sequenza di segnali scalati nel
tempo di un intervallo pari alla distanza
tra punti di prelievo divisa per la
velocit di propagazione (3-5 m/s).

bad

Array of equally
spaced electrodes

Propagating MUAP

good

small noisy
signals

bad

good
signals

good

small noisy
signals

good
signals

bad

small noisy
signals
Differential
amplifiers

Time

Two electrodes placed symmetrically over the I.Z. give unreliable information.

Biceps brachii muscle contracting at 70% MVC

Electrode
array

7
1 mV

15
10 mm
Depolarized
zones

50 ms

EMG signals detected with a linear array of

16 electrodes in SD mode. Innervation and
termination zones of single MUs are evident.

Global analysis and single MU analysis

Information may be extracted either from the interferential signal (global level)
or from the single MUAP (MU level).
1

1 mV

15

Electrode arrays
and amplifiers

Signals from right

and left trapezius
during typing
(project NEW)

Upper trap. activity

during typing
with forearms
on the desk.

Upper trap. activity

during typing
with forearms
off the desk.

Right Upper Trap

50 ms

0.4 mV

Left upper trap

50 ms

0.4 mv

Anal probe
16 Electrodes
Pressure sensor

Anal probes
(1 array and
3 arrays with
16 electrodes
each)

Urethral
probe

Stick-on array for

puborectalis muscle
(8 electrodes)

16 1

Anal Probe, MVC, depth 5cm, Electrodes 1,16 dorsal,

Example of ventral innervation (under electrode pairs 6-8)

D
R

Probe view
from outside

16

152 V

channels

16
1820

1840

1860

1880
1900
time (ms)

1920

1940

1960

Anal recording, max. voluntary contraction (observe asymmetry)

S01_02, OD, male
Probe location: near orifice
100 V
Contraction level: MVC

V = ventral
L = left
D = dorsal
R = right

mark

15 1
V

Probe view
from
outside

L
D
8

LEFT

RIGHT

15
0

2000

4000

6000

8000

10000

1150

1200

7350

7400
time (ms)

LEFT

RIGHT

15
2800

2850

2900

2950

3000

3050

3100

3150

3200
time (ms)

a) N = 34

[ch]
1

[ch]

b) N = 40
F1TQCA04.SIG

11

11

13

13

15

15

F1TWXA02.SIG

[ms] 0 10

20

30

40 50

60

70 80

[ms] 0 10 20

30

40

50 60 70 80

4
200 V

200 V

12

12

16

a)

16

D1TJPA2.sig 9.6875 - 9.7500 s

25 ms

Max. vol. contraction

14 mm

c)

depth 4-5 cm

d)

Anal
Orifice

Ch. 14

4 cm

Ch. 13
Ch. 12

10 mm

D1TWXA2.sig 1.7500 - 1.8125 s

25 ms
View from outside

Ch. 11
Fig. 1

b)
Max. vol. contraction

15 16 1 2
14
3
D
13
4
L
R
12
5
V
11
6
10

Ch 1
Ch 2
Ch 3
Ch 4

D
1

16
15
14
13

B
2

D
L

12
11

4
R

5
6

V
10 9

400 V

Innervation zone

A1

A2

Terminal zone

14 mm

Depth: 2-3 cm
12
2 cm

C1

D1

D2

10 mm
D1TISA10.sig
6.7500 - 6.8125 s

Anal Orifice

16

Max vol. contraction

Fig. 3

25 ms

C2

a
15 16 1 2
14
3
D
13
4
L
R
12
5
V
11
6
10 9 8 7

200 V

D
A
8

D1TJPA16.sig
2.1250 - 2.1875 s

Innervation zone

Max vol. contraction

Terminal zone

12

B1

B2

14 mm

depth 2-3 cm
2 cm

16
10 mm

Fig. 4

Anal
Orifice

25 ms

400 V

400 V

400 V

12

16
D1TQCA4.sig
0.3125 - 0.3750 s

D1TQCA6.sig
1.7500 - 1.8125 s

(4-5)
4-5

2-3

0-1

14 mm

Array depth (cm)

Fig. 8

(2-3)

25 ms

D1TQCA8.sig
9.1875 - 9.2500 s

Contraction
level : MVC

(0-1)
15

a = 1 cm

1 2
3

14
13

Anal
Orifice

16

12
11
10

V
9

5
7

b
16 1 2
15
3
D
14

13
L
R
5
12
6
V
11
7
10
8
9

C
D

A
B,D
C

400 V

Innervation zone
Terminal zone

12

14 mm

depth 4-5 cm
4 cm

16
10 mm
Anal Orifice
Fig. 5

25 ms

D1TQCA4.sig
9.3750 - 9.4375 s
Max vol. contraction

OASIS: possible sphincter damage due to episiotomy

V

Surgical
incision

Medium
risk

High
risk

Innervation
Low
risk

V = vaginal opening
P = perineal wall
A = anal opening

Very
high
risk

Myoelectric manifestations
of muscle fatigue

Segment b
spectrum 1

Segment c
spectrum 4

Sustained isometric
voluntary contraction

Spectral evolution of a
quasi stationary EMG
signal.

Normalized power

1.0

4
3

a)
2

0.5

1
0.0
0

Signal b, spectrum n. 1

mV

100
200
Frequency (Hz)

300

Signal b, spectrum n. 4

c)

b)

0.1

400

0.0
-0.1
0

0.25

t (s)

Beginning of the contraction

0.50

0.25

t (s)

End of the contraction

0.50

EMG power spectrum

The power of the EMG signal is distributed in the frequency
range 10-400 Hz
Power of the harmonics versus
their respective frequency.

Power

Harmonics

100

200

300

400 Hz

The spectrum of the EMG signal changes as a function of time during an

isometric constant force sustained contraction , because muscle fiber
conbduction velocity and motor unit action potential shape change in time.
These parameters recover quickly and there change may be small during
intermittent contractions.

Myoelectric manifestations of muscle fatigue

Signal at the beginning of an isometric
sustained contraction

Signal at the end of an isometric

sustained contraction

mV

0.1
0.0
-0.1

time (s)

0.25

0.50

Norm. power density

1.0

Power spectral
density during a
sustained
contraction.

0.5

0.0
0

100

200

300
400
frequency (Hz)

time (s)

0.25

0.50

During a sustained isometric

contraction the surface EMG signal
becomes slower, the power
spectral density is compressed
toward lower frequencies and
spectral variables (MNF, MDF)
decrease. The decrease of these
variables reflects a decrease of
muscle fiber conduction velocity
and changes of other variables
(such as active motor unit pool,
degree of synchronization, etc).

Mean and median spectral frequenclies of the

EMG signal (MNF and MDF)
MDF: splits the spectrum into two
parts of equal power
MNF: center of gravity line

fm =

200

f P(f)df

400 Hz

P (f)d

fmed

P(f)df =

fmed

P(f)df =

1
2 0

P(f)df

Myoelectric manifestations of muscle fatigue

Normalized EMG power spectrum

Centroid lines (Mean frequency)

Normalized power

100

80

Mean frequency
(MNF) pattern.

60
90s

One epoch

40
60s
20
30s
0

0s
0

100

200

frequency (Hz)

300

n
Co

n
io
t
c
tra

ra
u
d

n
tio

Example of power spectrum of the EMG of the biceps brachii during a sustained
isometric contraction at 60% MVC. The centroid value (MNF) progressively moves
towards the lower frequency values demonstrating myoelectric manifestations of
muscle fatigue. The rate of change can be taken as an index of fatigue.

(with respect to initial value)

normalized values

The Fatigue Plot

Root mean square value (RMS)
Average rectified value (ARV)
100

Force or torque
Conduction velocity (CV)
Mean spectral frequency (MNF)
Median spectral frequency (MDF)
0

contraction duration

The fatigue plot depicts the time course of some EMG signal variables
normalized with respect to their individual initial values.
It allows comparison of the patterns and rates of change of these
variables which reflect muscle properties.

Myoelectric and mechanical manifestations of muscle

fatigue during voluntary sustained isometric contractions
110
100

a)

+ - 5%

b)

+ - 5%

90

Torque
Torque

% of initial value

80

MDF

70
60

50% MVC

70% MVC

50

MDF

IMDF = 78 Hz

IMDF = 77 Hz

40
0

20

40
time (s)

60

80

20

40

60

80

100

120

time (s)

During strong contractions the pattern of MDF or MNF may be

exponential. Initial slope or time constant can be used as indexes of
myoelectric manifestations of muscle fatigue. These manifestations
begin at the beginning of the contraction and precede and predict
mechanical fatigue.

The Fatigue Plot during voluntary contractions

180

Normalized values w.r.t. initial values

ARVmean +/-SD
CVmean +/-SD
MNFmean +/-SD

120

160

Subject 1
Mean std. dev.
of 9 repetitions.
Small myoelectric
manifestations of
muscle fatigue.

TRQ(70%)MVC +/-SD

110

140
120

100

100
80
60
40

10

15

20

25

30

Time (sec)
ARVmean +/-SD
CVmean +/-SD
MNFmean +/-SD

180
160

Normalized values w.r.t. initial values

Normalized values w.r.t. initial values

Normalized values w.r.t. initial values

The Fatigue Plot is the graph of the time course of the EMG variables, normalized with
respect to their initial value, during a sustained voluntary or electrically evoked
contraction. It describes percent variations of different variables with respect to their
initial value. The graphs below show differences observable between two healthy
subjects during isometric 70% MVC contractions of the biceps brachii sustained for 30 s.

90
80

10

15

20

Time (sec)

25

30

Subject 8
Mean std. dev.
of 9 repetitions.
Large myoelectric
manifestations of
muscle fatigue.

120
110

140
120

TRQ(70%)MVC +/- SD

100

100
80
60
40

10

15

20

Time (sec)

25

30

90
80

10

15

20

25

30

Time (sec)

Rainoldi A., Galardi G., Maderna L., Comi G., Lo Conte L., Merletti R., Repeatability of surface EMG variables during
voluntary isometric contractions of the biceps brachii, J. Electrom Kinesiol., 9, 105-119, 1999.

The Fatigue Plot during voluntary contractions

Young subject (60%
MVC)

1.6

Elderly subject (60% MVC)

1.6
ARV

1.2
TRQ
1.0
0.8

CV

0.6

MNF

Normalized values

Normalized values

ARV

1.4

1.4

1.2
TRQ

1.0

CV
0.8
MNF

0.6
0.4

0.4
0

10

15

20

Time (s)

25

30

10

15

20

Time (s)

The number of type II (larger) muscle fibers decreases with age.

This is reflected by reduced MVC and myoelectric manifestations
of muscle ftigue.

25

30

M-wave changes during electrical stimulation of the

tibialis anterior muscle of two individuals for 30 s at 30 pps
6

4
Amplitude (mV)

D10C1

D3B1

0
0
-2

30

-2

30

-4

-4

-6

-6

-8
0

10

20

30

10

20

30

Time (ms)

Time (ms)

Subject a: limited myoelectric

manifestations of muscle fatigue

Subject a: marked myoelectric

manifestations of muscle fatigue

Fatigue plots obtained during electrical stimulation of the

tibialis anterior of 4 individuals for 30 s at 30 pps
Normalized values (%)

(individual differences are evident)

200
1
150

t5a3

ARV

ARV
RMS
RMS

100
CV
MNF

CV. MNF. MDF

50
0

Normalized values (%)

t703n1

10

20

30

MDF

10

20

30

200
3

t4a3

ARV

t401n1

ARV

150

RMS

RMS
CV

100

CV
MNF. MDF

MNF
MDF

50
0

10
Time (s)

20

30

10
Time (s)

20

30

Hopf, R.L.Herbort, M. Gnass, H. Gnther, K. Lowitzsch, Fast and slow contraction times associated
with fast and slow spike conduction of skeletal muscle fibers in normal subject and in spastic
hemiparesis, Z. Neurol, vol. 206,pp. 193-202,1974.

Conduction velocity (m/s)

5,5

5,0

Single twitches electrically

evoked from the biceps brachii
muscle.

4.65 m/s

4,5

4,0

r = -0,544
p < 0,005
y = 5,69 - 0,0296x

3,5

3.47 m/s

3,0
30

40

50

60

C ontraction tim e (m s)

70

80

Sadoyama T., T. Masuda, H. Miyata, and S. Katsuta , Fiber conduction velocity and fiber
composition in human vastus lateralis, Eur. J. Appl. Physiol. 57, 767-771, 1988.

5.4
Sprinters
Distance Runners

Conduction velocity (m/sec)

5.2
5.0
4.8
4.6

CVI(x=0) = 3.9 m/s

CVII(x=100) = 5.2 m/s

4.4
4.2

y = 0.013 x + 3.9
r = 0.84 ( p < 0.001 )

4.0
20

40

60

80

100

Relative area of FT fibers (%)

Komi P.V. and Tesch P., EMG frequency spectrum, muscle structure, and fatigue during
dynamic contractions in man. Eur J Appl Physiol Occup Physiol, 1979, 42(1):41-50.

Kupa, S.H. Roy, S.C. Kandarian, C.J. De Luca, Effects of muscle fiber type and size on EMG median
frequency and conduction velocity, J Appl Physiol, vol. 79(1), pp.23-32, 1995.

SO = Slow Oxidative fibers (I)

FOG = Fast Oxidative Glycolitic fibers (IIa)
FG = Fast Glycolytic fibers (IIb)

1.1

Fiber Type percentage by area

%SO (I)
%FOG (IIa) %FG (IIb)
SOL 87.2 4.3 12.8 4.4
0.00.0
38.02.3
DIA 28.2 2.5 33.9 3.0
66.62.6
EDL 1.5 0.4 31.9 2.5

Noemalized MDF

1.0
0.9
0.8
0.7

SOL

SOL: Soleus
DIA : Diaphragm
EDL: Ext. digitorum longus

0.6
0.5

DIA
EDL

Mean st. dev., N = 8 rats

0.4
0

10

Time (s)

15

20

Hvala Lepa !

roberto.merletti@polito.it
www.lisin.polito.it