European Journal of

Europ. J. appl. Physiol. 37, 111-121 (1977)

Applied

Physiology

and OccuDatlonal Physiology

~) by Springer-Verlag 1977

Signal Characteristics of EMG during Fatigue Jukka H. T. Viitasalo and Paavo V. Komi KinesiologyLaboratory, Department of Biologyof Physical Activity,Universityof Jyv/iskyl/i, SF-40100 Jyv/iskyl/i 10, Finland

Summary. Electromyographic (EMG) activity of m. rectus femoris muscle was registered from young male and female subjects during maintained isometric knee extension at 60% of maximal voluntary contraction. The following EMG parameters were analyzed for the entire fatigue time: integrated EMG (IEMG), averaged motor unit potential (AMUP) and power spectral density function (PSDF). The results indicated a slight but continuous rise of IEMG during the fatigue period. AMUP showed sensitivity to fatigue with increase in amplitude, rise time, and number of spikes counted. PSDF was also easily affected by fatigue so that the total power density curve was shifted towards lower frequencies with a high frequency decay. The mean power frequency decreased linearily as a function of fatigue time. The findings suggest that in addition to natural recruitment of new motor units the fatigue is characterized by marked reduction in the conduction velocities of action potential along the used muscle fibers. Key words: Muscular fatigue - Electromyography - Signal characteristics Spectral analysis - EMG amplitude.

Introduction The complexity of the neuromuscular system makes it difficult to evaluate all the mechanismic events which take place during fatigue. Therefore objective observation of neuromuscular fatigue would require simultaneous utilization of several physiologic and kinesiologic research tools. The various techniques of electromyography (EMG) can be used to assess the fatigue changes in one functional part of the neuromuscular system: activation pattern of the motor unit pools. In global type of EMG experiments integrated EMG (IEMG) has most often been employed in investigations of the fatigue phenomenon of both static and dynamic work. In recent reports (Viitasalo and Komi, 1975; Komi and Viitasalo, 1976) some other quantitative type of EMG analysis techniques have been explained. Utilizing these methods the present study was designed to investigate primarily the changes in the signal characteristics of EMG during sustained static contraction of submaximal level.

112

J.H.T. Viitasalo and P. V. Komi

Methods The subjects for the experiment were two girls and four boys with age ranges of 12-14 years.

Testing The subjects were first familiarized with the testing procedure on two occasions before the testing day. In testing situation the subject was seated at the edge of the dynamometer chair with the knee and hip angles of 240 and 130 degrees, respectively. The isometric knee extension was sensed with strain gauges installed on a steel wire connected to a cuff fastened around the subject's ankle. The steel wire was positioned so that the pull of the cuff was always at a 90 degree angle to the line joining the ankle and knee joints. The subjects were first asked to exert three maximum isometric contractions of the quadriceps muscle group. The rest periods between contractions were 3 rain. Then after 5 min additional recovery the relationship between EMG-parameters and muscle tension was determined. This was obtained with 4 submaximal contraction of 20%, 40%, 60% and 80% The recovery period between submaximal contractions was 2 min and each contraction was maintained at a steady level for 5 - 1 0 s. The fatigue contraction followed after 10 min rest from the last submaximal contraction. In this part of the experiment the subject was instructed to maintain the 60% contraction level as long as possible and to keep the measurement position constant throughout the contraction. The subject saw the required force level from a voltmeter placed in front of him. The testing was terminated after an abrupt decrease of force.

EMG Recording and its Processing EMG recording and its processing have been explained earlier (Viitasalo and Komi, 1975; Komi and Viitasalo, 1976), but for clarity it is also repeated here. EMG activity was picked up by two pairs of Beekmann miniature sized skin electrodes with a contact diameter of 4 ram. One pair was placed over the motor point area of the rectus femoris muscle and the second pair 25% in distance distally from the first pair along the line motor point-patella linearly with the muscle belly. The skin was carefully prepared for the electrode placement and the recordings were never started before the interelectrode resistance of below 5 k~2 was secured. Amplification of the EMG signals were performed with Tektronix RM 122 low level preamplifiers (60 dB; spectrum range 0.8 Hz--10 kHz). The amplified signals were subsequently stored together with the force signal in analog form on magnetic tape (Philips analog-7) using a recording speed of 15" per second. A 15/16" per second play speed of the tape recorder was used when the signals were fed into the automated processing system (Viitasalo and Komi, 1975). The digitization of the analog signals, two EMG's and one force signal, was performed with a final sampling frequency of 1600 per second. After this conversion the subsequent analysis was performed from three records of 625 ms duration. The records were timed to cover the early plateau phase of the force curve. The output of the EMG parameters from the processing phase were as follows: 1. Integral, IEMG (mV). 2. Power spectral density function, PSDF. 3. Averaged motor unit potential, AMUP. IEMG was timeaveraged for one second period. The methods given by Bendat and Piersol (1971) was used to compute PSDF. It started first with generation of autocorrelation function of the EMG signal for a given time period. The final PSDF was obtained through Fourier transformation. The frequency of the mean power (MPF) was calculated according to the formula given by Kwatny et al. (1970). In this computation the lower and higher cut off frequencies were set at 24 Hz and 800 Hz, respectively. PSDF were further treated so that the relative (per cent) portion of selected bandwidths (24--56 Hz, 64--96 Hz, 104--136 Hz and 144-800 Hz) were obtained from the total power density

Signal Characteristics of EMG during Fatigue

113

area. These bandwidths corresponded to those used in earlier reports (Kadefors et al., 1968; Viitasalo and Komi, 1975; Komi and Viitasalo, 1976). AMUP was computed by slightly modifying the method described by Lang et al. (1971). This computation requires a set threshold level, which determines whether an individual spike is included in the averaging. After finding the peak amplitude of a specific spike the computer counts 30 digitized points backwards and forwards from the peak. This gives a 37.5 ms length for time axis. The threshold level was set at 0.12mV. AMUP was further analyzed for its peak-to-peak amplitude and rise times. The latter was computed by finding the sign change of the second derivative of the peak rise. A tangent was calculated on this point and the rise time was obtained by taking the difference in x-axis between the amplitude peaks and the tangent.

Results

The results are summarized in Figures 1 - 1 0 which represent mean values of the six subjects. In describing the results an attempt has been made to show, in most cases, in the same picture a change in a particular EMG parameter during the course of fatigue and its relationships to muscle tension in a nonfatigued situation. In recording the fatigue time a sudden loss of tension was regarded as the point for the 100% fatigue time. The average maintainance of "steady" contraction was 86.7 + 34.6 s. The subjects were not able to hold the tension exactly at the required 60% level; the force dropped during the experiment from 6 2 - 5 5 % (Fig. 1). IEMG activity increased during the fatigue time both in the upper recording site and in the lower portion of m. rectus femoris (Fig. 1). In Figure 2 these changes are .3 iEMG

FORCE

(mV/

x (kg)

IEMG

UPPER

RF

IEMG

LOWER 9

RF

x

.2

9

9

9 40

.1

FORCE ~

K

Jl

K

I

M

lit 20

U 0 0

I

i

t

25

50

75

0 100

FATIGUE

TIME

(%)

Fig. 1. The values of IEMG of m. rectus femoris and the force of contraction of the quadriceps muscle group during the fatigue loading of constant isometric tension of 0.6 x P0

114

J.H.T. Viitasalo and P. V. Komi

shown together with the relationship between the IEMG and muscle tension measured before the fatigue situation. Figure 3 shows the AMUP's for the upper (A) and lower (B) recording site both in maximum contraction (100%) and before and after the fatigue loading with 60% contraction. The form of AMUP changed during the experiment so that it increased in duration and amplitude. The changes in selected parameters of AMUP are shown in detail in Figures 4 - 7 . For the upper recording site the number of spikes which exceeded the threshold of 0.12 mV and thus were included in AMUP, increased significantly. The number of spikes counted for AMUP fluctuated during the course of fatigue, but its end value was higher (p < 0.05) than in the beginning of contraction for the upper recording site (Fig. 4). Similarily, significant (p < 0.05) increase in AMUP amplitude was seen for the upper recording site only (Fig. 5).

[EMG

A

(mY)

,'

IEMG

IEMG

t

[mY)

B

IEMG

~mV)

(mV)

t // 3

/ F/

.2

3

/

2

/

t'

/

/

/

/ /

/

t

,1

0

20

40

215

I

I

50 F a t i g u e Time

60

~

75

10

I 0

.1

(oto)

100 Force

i 25

i 75

510

10~

F a t i g u e Time (~

4

100

(Olo)

Force

(%)

Fig. 2. Relationship between IEMG and the force of contraction for the EMG recordings taken from the upper portion (A) and lower portion (B) of the m. rectus femoris. Both figures indicate also how IEMG changes during the fatigue loading of constant isometric contraction of 0.6 x P0

mV

mV

A

,

!

9 -.~

Before

-_--

,,:,2

B

~

- ------

100 % 60% 60%

Before --=1After

a

9

L

--100%

",~7 -

2O m~

- -

j 2 0 ms

r

-.2

Fig. 3. Averaged motor unit potentials (AMUP) recorded from the upper portion (A) and lower portion (B) of m. rectus femoris in maximum voluntary contraction (100%) and before and after the fatigue loading with a maintained tension of 60% of the nonfatigued value

Signal Characteristics of EMG during Fatigue

40 t

A

No. s p i k eofs

F'- - ~ ~ /

115

No.of " 40 spikes

40[

30

30

20

20

40

Noof

,l

I

30

/ J

!

iI o

/

I"

. ~

I

30

iI //

/ {

No. of spikes

B

spikes

/ 20

t /

i

9

215

i i 50 75 F a t i g u e T i m e (olo)

//

100

/

i

i

0

25

/

1o

/

/

t

Fatigue

i

i

50

75

tOO

T i m e (~

/

z

//

/

I 20

I 40

I 60

8

I0

/

f

I 100

2O

q 40

i 60

Force(%)

~ 80

1010

F o r c e (~

Fig. 4. Relationship between the number of spikes in the AMUP potential (record time 1.9 s) and the force of contraction for the EMG recordings of the upper portion (A) and lower portion (B) of m. rectus femoris. Both figures show also the number of AMUP spikes during the fatigue loading of 0.6xP

0

Amplitude (mV)

A .'" /

/

Amplitude (mY)

/

.4

.4

Amp l i rude (mY)

B

/

/ /

.3

3

.2

2

I0~ 1

1

/

/

/

p

Ampli tude (mY)

/ / /

d

/

/

/

/

/ J

I 0

I 20

I 40

A 25

t 50

715

Fatigue

T i m e (~

I 80

I tOO

I 60 Force

0

i o

i 20

l 40

i 60

i 25

~ 50

Fatigue

Time

i 80

75

1001

{~

100

F o r c e (c/o)

(~

Fig. 5. Relationship between the amplitude of AMUP potential and the force of contraction for the EMG recordings of the upper portion (A) and lower portion (B) of m. rectus femoris. Effects of fatigue loading of 0.6 x P0 are also shown

Rise Time

//~

A

(ms)

Rise Time

Rise T i m e (ms)

3

,

0

I 20

l 40

9

,

,

25 50 75 F a t i g u e T i m e (~ I I I 60 80 100 Force (o/o)

1002

Rise Time (ms)

B

3

2

,

0

r 20

i 40

9

25 Fatigue I i 60 80 Force(eh)

,

,

50 75 Time(O/o) i 100

2 lOg

Fig. 6. Relationship between the rise time of AMUP potential and the force of contraction for the EMG recordings taken from the upper portion (A) and lower portion (B) of m. rectus femoris. Both figures indicate also how AMUP rise time changes during the fatigue loading of constant isometric contraction of 0.6 x P0

.12

A

A/RT

12

A/RT

A/RT

B

,,,*

/

t

08

/

/

/

.10 /

/

/

/

/

A/RT

/ .10

.08

// // 06

06

04

04

i

0 0

i 20

0

i 40

,O2 100

i

215 50 715 Fatigue Time(O/o)

I

60 Force

i

o

i

25

5;

;5

02

I00

Fatigue Time (o/o)

i

.J

i

i

80

100

40

60

(~

i

80 Force (o~)

t 10o

Fig. 7. Relationship between the amplitude/rise time ratio and the force of contraction for the EMG recordings fo the upper portion (A) and lower portion (B) of m. rectus femoris. Effects of fatigue loading of 0.6 x P0 are also shown 75 mV?H

z

A

I.

I

--100 ....

% Before

60%

.

.

.

.

..... 60 % After

!~ i az

0

0

200

400

75 mV/Hz

B

-" / ~'~

-100 % Before --60 % . . . . ..... 60 % After

/ I Is

x x

Hz 0

200

400

Fig, 8. Power spectrum density function (PSDF) recorded from the E M G signals of the upper portion (A) and lower portion (B) of m. rectus femoris in maximum voluntary contraction (100%) and before and after the fatigue loading with a maintained tension of 60% of the nonfatigued value

Signal Characteristics of EMG during Fatigue MPF

t17

A

(Hz) 'k

MPF C,Hz)

120

\

I1(]

B

T20

110

\

r" / ~ t

~

~

100

\

90

IO0 90

\

80

70

70 60

25 50 0

210

i 60

410

5'0

75

100

Fatilgue Tittle(~176 80 100

I

0

210

I

r

0 I

40

60

50

25

i

50

F a t i g u e Tinge 80 100

Force (~

(~

75

bO 100

Force (~

Fig. 9. Relationship between the mean power frequency (MPF) and the force of contraction for the EMG recordings taken fi'om the upper portion (A) and lower portion (B) of m. rectus femoris. Both figures indicate also how MPF changes during the fatigue loading of constant isometric contraction of 0.6 x P0

Relat ive Power (~

Relative Power

A 64 - 96 Hz

B

(~

_

64-96Hz ~ ~. . . . . .

. . . .

-

~ -

24 -56Hz

1 9.

%'~

0

4

~

104 - 136Hz

144 - 800 Hz I

i

i

i

20

40

60

80

100

i

i

i

20

40

60

Force (%) Relative Power

C

i

80 Force (%)

I00

D

(O/o)

60

6 4 - 96 Hz

20 ~ 2~

104-136Hz

2

..........

"~'~.~._.~

ioz~ - 136Hz

]44-800Hz" " - . . . . . . . . . . . . . . . . . . . .

~-'::: ::::~: Z'.'2":::2::~:~'~".................... o c

i

i

i

i

0

25

50

75

100

Fatigue Time (~

Ot O

I

/

25

50

I

75 Fatigue Time (o/o)

I

100

Fig. 10. Above: Relationship between per cent values of bandwidths of the total power spectrum and the force of contraction for the EMG recordings of the upper portion (A) and lower portion (B) of m. rectus femoris. Below: Changes in the relative power of different bandwidths during the maintalnance of constant isometric tension of 0.6 x P0. (A) upper portion; (B) lower portion of m. rectus femoris

118

J.H.T. Viitasalo and P. V. Komi

The rise time of AMUP increased during the fatigue contraction both for the upper (p < 0.01) and lower (p < 0.05) recording site (Fig. 6). The amplitude/rise time ratio (A/RT), however, stayed relatively the same during the entire fatigue time (Fig. 7). A slight, but not significant reduction of A/RT was observable at the very end of contraction in the lower EMG recording. Marked changes in power spectral density functions were noticed during the course of fatigue. Figure 8 shows how the total power density curve was shifted towards the lower frequencies with a marked high frequency decay. The peak power frequencies changed also from 80 Hz (before fatigue) to 48 Hz (end fatigue). The mean power frequency changes (Fig. 9) emphasized this shift to lower spectrums. The decrease of MPF value was almost linear as a function of time. Figure 10 shows the relative powers of the selected bandwidths first, as a function of load (upper part) and as a function of time during the maintainance of 60% isometric contraction (lower part). It can be seen that the fatigue caused a decrease (p < 0.05-0.01) of relative power in the bandwidth of 64-96 Hz, 104-186 Hz and 144--800 Hz. In 24-56 Hz bandwidth this change was, however, opposite showing a continuous increase (p < 0.01) of relative power.

Discussion

The continued muscular contraction at a certain submaximal tension causes physiological changes in the neuromuscular system which is manifested by an increased neural input to the muscle. The quantity of this neural energy is often measured with integrated EMG activity (IEMG). There are several reports in the literature to indicate an increase in IEMG in similar fatigue situations as used in the present study (e.g. Edwards and Lippold, 1956; Lippold, 1960; Courier, 1969; De Vries, 1968; Kuroda et al., 1970; Rau and Vredenbregt, 1970; Laurig, 1970). The increase in IEMG has been assumed to be due to failure in the muscle contractility, so that with maintained contraction the individual active fibers exert progressively less force. To compensate this effect new motor units are recruited and/or the active motor units will fire with an increased frequency (Lippold, 1970). In maximal contractions, on the other hand, all the motor units are assumed to be active and the natural effect of the continued contraction in this situation is a reduction of both muscle tension and IEMG (Komi and Rusko, 1974). Although the upper part of the rectus femoris seemed to show slightly greater changes in the measured parameters, the results are not at all conclusive enough to make judgements as to the possible functional differences between the two parts of the muscle. In general the changes in the EMG parameters during fatigue loading were unexpectedly low. For example, our results failed to produce an increase in IEMG in a similar magnitude as reported in most of the references cited. This was in spite of the fact that fatigue loading was performed with approximately 60% of maximal voluntary tension. At this tension level 2.5 time increase in IEMG has been reported for the biceps brachii muscle (Rau and Vredenbregt, 1970a; 1970b). Two reasons seem plausible for the small IEMG increase found in the present study: (1) The tension was not maintained exactly constant during the fatigue period, because it decreased from 62-55% of maximum voluntary contraction of the prefatigue

Signal Characteristics of EMG during Fatigue

119

level. (2) Migration of activity between the different muscles that take part in knee extension. Concerning the first possibility Rau and Vredenbregt (1970a) have shown with one subject (m. biceps brachii) that if the tension is changed during the fatigue loading the recorded IEMG will also change correspondingly. In our study, however, the greatest increase of IEMG was noticed for those subjects who were least able to hold the required tension all throughout the exhaustion time. Thus the possible inaccuracies in the loading procedure cannot explain the unexpected small increase of IEMG. A more likely explanation is therefore parttaking of other muscles in the executed tension. The quadriceps group alone includes four muscles - rectus femoris, vastus lateralis, vastus medialis and vastus intermedius - all of which can contribute to extension of the knee joint in the position used in the present study. Stephens and Taylor (1973) have correctly given their concern about relating EMG to forces, when this force is caused by more than one muscle. In fact, Lippold et al. (1960) have shown that when maintaining a constant force, IEMG of a muscle can decrease with time while that in the synergist muscle may increase. AMUP changes during fatigue were similar to those of IEMG. In upper recording site the AMUP amplitude increased significantly during the course of contraction time. Similarly, the number of spikes counted from AMUP was increased (Fig. 4). This can be seen attractive to explain so that the amplitude increase was brought about by recruitment of bigger (faster?) MU's. This is, however, indirect assumption, because the increase in the number of spikes for AMUP may come from either new motor units or increase in discharge frequency of already active MU's. A clear increase in duration of AMUP is in conformity with changes in individual MU spikes with fatigue (e.g. Lippold et al., 1960). Several authors (,e.g. Kadefors et al., 1968; Okada, 1971; Vredenbregt and Rau, 1973; O'Donnell et al., 1973; Chaffin, 1973) agree that power spectral changes during sustained submaximal isometric contractions are clearly demonstrable, so that there is increase in the low frequency components and decay in high frequencies. The present results are in good agreement with the previous reports (Fig. 8-- 10). Both the decay in high frequencies (64--96 Hz, 104-136 Hz, 144-800 Hz) and increase in low frequency band (24-56 Hz) occurred in a linear manner. Kadefors et al. (1968) have reported that the spectral changes during fatigue occur in two phases: low frequency augmentation takes place slowly and the high frequency decay shows a rapid initial stage. This could not be observed in our analysis. The uniformity in the spectrum change is seen also from Figure 9 which indicates the !inear lowering of the mean power frequency (MPF) in the course of fatigue. Relatively small changes observed in EMG parameters limit the possibilities for explaining the mechanismic events during fatigue. Increase of AMUP potential during fatigue is in line with changes in IEMG rather than in single MU potentials, which have been shown to decrease its amplitude with simultaneous increase of duration (Lippold et al., 1960). Person and Kudina (1972), however, found no changes in MU potentials but a clear decline in discharge frequency of individual spikes during maintained tension. The variations in spectral content (Fig. 10) also imply to the possibility of increase of MU potential duration. Similar explanation was given by Kadefors et al. (197'3) but for higher muscle forces than used in the present study.

120

J.H.T. Viitasalo and P. V. Komi

The increase in M U potential duration has been attributed to a decrease in the conduction velocity of the muscle fibers due to fatigue (Mortimer et al., 1970). The results are not at all conclusive for explaining the relative role of the fast ( F M U ) and slow m o t o r units (SMU) during the maintained tension. Earlier (Komi and Viitasalo, 1976) we have postulated that the A M U P analysis as to its amplitude and rise time components would reveal when F M U ' s give their strongest contribution to the increased muscle tension. G y d i k o v et al. (1974) have shown that for the biceps brachii muscle the threshold for the S M U ranges from 0 - 8 0 % and that for F M U from 3 0 - 8 0 % of m a x i m u m voluntary contraction. In fact our earlier results ( K o m i and Viitasalo, 1976) seemed to indicate that at approximately 60% P0 the A M U P changes were so m a r k e d that at this tension level S M U ' s had practically all attained their m a x i m u m discharge frequency and that the further rise in muscle tension were due to increase in recruitment and discharge frequency of F M U ' s . The fatigue conditions of the present study with a likely increase in individual M U potential durations and in the measured spike counts and amplitude of A M U P may, however, m a s k the actual events of the m o t o r unit recruitment during fatigue.

Acknowledgements. Supported in part by a grant No. 9729/79/73 from the Ministry of Education (Finland).

References Bendat, J. S., Piersol, A. G.: Random data: Analysis and measurement procedures. New York: Wiley & Sons 1971 Chaffin, D. B.: Localized muscle fatigue - definition and measurement. J. oceup. Med. 15, 346-354 (1973) Courier, D. P.: Measurement of muscle fatigue. Phys. Ther. 49, 724--730 (1969) De Vries, H. A.: Efficiency of electrical activity as a physiological measure of the functional state of muscle tissue. Amer. J. phys. Med. 47, 1, 10-22 (1968) Edwards, R. G., Lippold, O. C. J.: The relation between force and integrated electrical activity in fatigued muscle. J. Physiol. 132, 677--681 (1956) Kadefors, R., Kaiser, E., Peters~n, I.: Dynamic spectrum analysis of myopotentials and with special reference to muscle fatigue. Electromyography 9, 40-74 (1968) Kadefors, R.: Spectral analysis of events in the electromyogram. In: New developments in electromyography and clinical neurophysiology, Vol. I (J. E. Desmedt, ed.). Basel: Karger 1973 Komi, P. V., Viitasalo, J. H. T.: Signal characteristics of EMG at different levels of muscle tension. Acta physiol, scand. 96, 267-276 (1976) Komi, P. V., Rusko, H.: Quantitative evaluation of mechanical and electrical changes during fatigue loading of eccentric and concentric work. Scand. J. Rehab. Med., Suppl. 3, 121-126 (1974) Kuroda, E., Klissouras, V., Milsum, J. H.: Electrical and metabolic activities and fatigue in human isometric contraction. J. appl. Physiol. 29, 358--367 (1970) Kwatny, E., Thomas, D. H., Kwatny, H. G.: An application of signal processing techniques to the study of myoelectric signals. IEEE Trans. Bio-Med. Eng. 17, 303-312 (1970) Lang, A. H., Nurkkanen, P., Vaahtoranta, K. M.: Automatic sampling and averaging of electromyographic unit potentials. Electroenceph. clin. Neurophysiol. 31, 404--406 (1971) Laurig, W.: Electromyographie als arbeitswissenschaftliche Untersuchungsmethode zur Beurteilung von statischer Muskelarbeit. Schriftenreihe ,,Arbeitswissenschaft", Weinheim 1970 Lippold, O. C. J., Redfearn, J. W. T., Vuco, J.: The Electromyography of fatigue. Ergonomics 3, 121--131 (1960) O'Donnell, R. D., Rapp, J., Berkhout, J., Adey, W. R.: Autospectral and coherence patterns from two locations in the contracting biceps. Eiectromyogr. clin. Neurophysiol. 13, 259-269 (1973)

Signal Characteristics of EMG during Fatigue

121

Okada, M.: Some properties of the "global" electromyogram in man as revealed by frequency analysis. J. Fac. Sci. Univ. Tokyo, See. V. 4, 61-80 (1971) Person, R. S., Kudina, L P.: Discharge frequency and discharge pattern of human motor units during voluntary contraction of muscle. Electroenceph. clin. Neurophysiol. 32, 471-483 (1972) Rau, G., Vredenbregt, J.: The relationship between the EMG activity and the force during voluntary static contractions of' the human m. biceps. Instituut voor Perceptie Onderzoek, Insulindelaan 2, Eindhoven 1970a Ran, G., Vredenbregt, J.: Electromyographic activity during voluntary static muscle contractions. Instituut voor Perceptie Onderzoek, Insulindelaan 2, Eindhoven 1970b Stephens, J. A., Taylor, A.: The relationship between integrated electrical activity and force in normal and fatiguing human voluntary muscle contractions. In: New developments in electromyography and clinical neurophysiology, Vol. I (J. E. Desmedt, ed.). Basel: Karger 1973 Viitasalo, J. H. T., Komi, P. V.: Signal characteristics of EMG with special reference to reproducibility of measurements. Acta physiol, scand. 93, 531-539 (1975) Vredenbregt, J., Ran, G.: Surface electromyography in relation to force muscle length and endurance. In: New developments in electromyography and clinical neurophysiology, Vol. I (J. E. Desmedt, ed.). Basel: Karger 1973

Accepted February 25, .1977