Some Factors Affecting the Power Spectra of Surface Electromyograms in Isometric Contractions

Transcription

1 Some Factors Affecting the Power Spectra of Surface Electromyograms in Isometric Contractions Haruhiko SATO Department of Ergonomics, Kyushu Institute of Design Abstract In order to investigate the variations in the EMG power spectra, the power spectra of the bipolar surface EMG of the biceps brachii were obtained for the frequency range from 6 to 192 Hz by the autocorrelation and its Fourier transformation technique. The EMG power spectra of the biceps brachii showed relatively definite pattern with some intra- and inter-individual variations. The difference in the EMG power spectra between the right and left biceps brachii was not larger than the inter-individual variation. No systematic variation was observed in the EMG power spectra either with the contraction level or with the muscle length. It was concluded that the power spectra of the bipolar surface EMG of a muscle show relatively definite pattern irrespective of subjects, right- or left-side, contraction level and muscle length at least for normal biceps brachii if constant electrode condition is maintained. INTRODUCTION Electromyographic (EMG) signals led by surface electrodes are highly complex in nature. Therefore frequency analysis of the EMG has been considered to be an effective tool for obtaining information about motor functions. Most of the investigators used the band pass filter technique for frequency analysis of the EMG led by surface electrodes or subcutaneous electrodes (WALTON, 1952; FEX & KRA- KAU, 1957; NIGHTINGALE, 1959 ; HAYES, 1960; KOGI & HAKAMADA, 1962a, 1962b; KAISER & PETERS] N, 1963; SATO, 1964, 1965; SATO et al, 1965; KOPEd & HAUSMAN- PETRUSEWICZ, 1966; SATO & TSURUMA, 1967; HAYAMI et al, 1967; GERSTEN et al, 1967; KoNDO et al, 1968; KIKUCHI, 1968; KADEFORS et al, 1968; CHAFFIN, 1969 and LLOYD, 1971), and some investigators used the autocorrelation and its Fourier transformation technique for analysis of the EMG led by subcutaneous electrodes (SCOTT, 1967)or unipolar surface electrodes (KWATNY et al, 1970). In my previous study (SATO, 1976), it was demonstrated that the autocorrelation and its Fourier transformation technique yielding power spectrum of the bipolar surface EMG was useful for the fundamental studies on muscle functions, and that there existed different EMG power spectrum patterns for different human skeletal muscles. The subject of this study is an analysis of the bipolar surface EMG power spectra during isometric contractions with particular emphasis on the

2 106 H. SATO intea- and inter-individual variations, the relationship between the right- and lef t- side and the variations due to contraction level and muscle length. and the cumulative power (CP) by METHODS The methods were similar to those described in detail previously (SATO, 1976). The bipolar surface EMG was led from a pair of silver disc electrodes (10 mm in diameter) attached to the skin over the examined muscle, parallel to muscle fibers and spaced about 30 mm between centers. The EMG signal was amplified with a time constant of 0.1 or 0.3 sec by a differential amplifier (Nihon Kohden Kogyo Co., RB-5), whose frequency response was down less than 3 db at 3 khz, and recorded on a magnetic tape recorder (TEAC Co., R-200) at a speed of 15.2 cm/sec with simultaneous ink-written record on a polygraph (Nihon Kohden Kogyo Co., RM-85). The power spectral density function, PrH, =2, 3,..., N-2, of the EMG was calculated r from the data sampled at 2.56 msec intervals, utilizing a record length of 5 sec, for the frequency range from 6.2 to Hz in steps of 3.1 Hz. The method of calculation was to enter the EMG signal stored on analog magnetic tape into ATAC Medical Data Processing Computer (Nihon Kohden Kogyo Co.) which determined the autocorrelation function,* (i**), i=0, 1,..., N-1, and from this, calculate the power spectral density function smoothed by Hamming window using a digital computer. For comparison of the power spectra of the EMG, the relative power density (RPD) was calculated by The subjects were nine normal males ranging in age from 21 to 25 years. RESULTS 1. Intra-individual variation in the EMG power spectra of the same muscle The EMG power spectra of the left biceps brachii were investigated for four subjects. The subject contracted the biceps brachii against a load of 2 kg applied to the wrist downwards with the elbow at a right angle and the forearm horizontally in supination for about 10 sec. The method to obtain isometric contractions of the biceps brachii was similar in all experiments except the experiments concerning the effect of the muscle length on the EMG power spectra. The contraction was repeated about every twenty minutes for about twenty-four hours with the electrodes kept attached to the skin. The EMG power spectra were obtained for the contractions performed about every hour and those for good EMG recordings were used as data. Fig. la shows four typical curves of the relative power density (RPD) out of twenty-seven curves for one subject (S), and Fig. lb shows the cumulative power (CP) curves corresponding to Fig. 1A.

3 Some factors affecting the power spectra of surface electromyograms 107 Fig. 2. Intra-individual variation in the EMG power spectra of left biceps brachii for four subjects. The frequency range where the power was above half of the maximum (a) and the frequency ranges for the maximum power (b), Q1 (c), Qz (d) and Q3 (e) are shown. Fig. 1. Intra-individual variation in the relative power density (A) and the cumulative power (B) of the EMG of left biceps brachii. Subject S. In order to analyze intra-individual variation in the EMG power spectra, the frequency range where the power density was above half of the maximum and the frequency range for the maximum power Fig. 3. Inter-individual variation in the EMG power spectra of right biceps brachii. Same symbols (a, b, c, d and e) as for Fig. 2. density were obtained. In addition the frequency ranges for Q1, Q2 and Q3, where CP(Q1)=25%, CP(Q2)=50% and CP(Q3) =75%, were obtained. Fig. 2 shows the frequency ranges for these five values in four subjects (S, 0, Y and D). The differences between maximum and minimum frequency of Qi, Q2 and Q3 were 7-13 Hz, 8-15 Hz and Hz respectively for four subjects. 2. Inter-individual variation in the EMG power spectra of the same muscle

4 108 H. SATO The EMG power spectra of the right biceps brachii were obtained for seven subjects using a load of 6 kg. The data used in this study were also included in the previous study (SATO, 1976). Fig. 3 shows the inter-individual variation with the help of five values similar to Fig. 2. The surface EMG power spectra of the biceps brachii were relatively constant, which is also shown in Fig Relationship between the right- and left-side Six subjects contracted the right and left biceps brachii against a load of 2 kg separately. The EMG power spectra of the right and left biceps were similar in five subjects (Fig. 4) and somewhat duff e- rent in one subject, K (Fig. 5). However, the difference in the EMG power spectra between the right and left biceps brachii was not larger than the above-mentioned inter-individual variation. 4. The effect of the contraction level on the EMG power spectra The EMG power spectra of the biceps brachii were obtained for loads of 3, 6 and 9 kg. In order to eliminate the effects of fatigue, the subject took sufficient rest between contractions and the contractions were performed in random order. Three subjects were used and one of them was examined twice. There was no systematic variation in the EMG power spectra with the contraction level as shown in Fig. 6. In addition, the rectus f emoris, vastus medialis and vastus lateralis were investigated for one subject. The subject was Fig. 4. The relative power density of the EMG of right (R) and left (L) biceps brachii. Subject S. Fig. 5. The relative power density of the EMG of right (R) and left (L) biceps brachii. Subject K.

5 Some factors affecting the power spectra of surface electromyograms 149 Fig. 6. The relative power density of the EMG of the biceps brachii for loads 3, 6 and 9 kg. Subject D. seated and contracted these muscles against loads of 3, 6 and 9 kg suspended the ankle with the knee extended from and the thigh held horizontal. No systematic variation was also observed in the EMG power spectra with the contraction level for these muscles. 5. The effect of the muscle length on the EMG power spectra The biceps brachii was studied for three subjects in this analysis. In order to change the muscle length the angle between the upper arm and the forearm was changed. The subject held a load of 6 kg applied to the wrist (directly or through a pulley) perpendicularly to the forearm maintaining at 45, 90 and 135 degrees, the angle of the elbow joint with the upper arm vertical and the forearm supinated. The tension exerted by the biceps brachii probably varied with the angle of the elbow joint. But it seems to be no problem since the EMG power spectra did not Fig. 7. The relative power density of the EMG of the biceps brachii for the elbow angles of 45, 90 and 135 degrees. Subject D. change with the contraction level as mentioned above. As shown in Fig. 7, no systematic variation was observed in the EMG power spectra with the muscle length. DISCUSSION KOGI & HAKAMADA (1962a) and SATO (1964) reported that the frequency spectra of the bipolar surface EMG were obviously influenced by electrode conditions such as the diameter of surface electrodes, the interelectrode distance, their relative position to the muscle belly and their angle to the direction of muscle fibers and suggested the preference of using the same electrode condition in comparative studies of frequency spectra. Therefore attention was paid to maintaining constant electrode condition in this study. KAISER & PETERSEN (1963) reported the frequency that profile of the EMG led by a coaxial needle electrode was reproduci-

6 110 H. SATO ble in the same muscle during constant contraction and time in relation to the beginning of the contraction. The variations in the power spectra of the bipolar surface EMG of the biceps brachii during isometric contractions were investigated in this study. There were variations in the EMG power spectra not only between subjects or right- and left-side but even at the same location in a muscle on repeated runs. Part of the variations may be due to the random variations among the waves of the interference pattern, but there also appeared to be real differences between subjects or right- and left-side. This was shown in the EMG power spectra of the right and left biceps brachii in subject K (Fig. 5) who had been doing karate practice. However, the EMG power spectra of the biceps brachii showed relatively definite pattern with some ultra- and interindividual variations. Although some investigators reported the changes of the EMG frequency spectra with the tension level (FEX & KRAKAU, 1957; KAISER & PETERSEN, 1963 and KWA- TNY et al, 1970), the majority of the investigators reported no systematic change in the EMG frequency spectra except increased amplitude with the tension level from the studies by surface electrodes (NIGHTINGALE, 1959; HAYES, 1960; KOGI & HAKAMADA, 1962b and SATO, 1964) and subcutaneous electrodes (WALTON, 1952; KOPEC & HAUSMAN-PETRUSEWICZ, 1966; SCOTT, 1967 and GERSTEN et al, 1967) which agrees with the finding that the rate of zero crossing in a full interference pattern of the EMG led by a monopolar needle electrode did not vary with muscle tension (FUSFELD, 1971). The result in this study that no systematic variation was observed in the EMG power spectra with the tension level may be explained by assuming that the effects of increased frequency of motor unit discharges and motor unit recruitment with increased tension level (ADRIAN & BRONK, 1929 ; LINDSLEY,1935; BIG LAND & LIP POLD, 1954; MORRIS & GASTEIGER, 1955; PERSON & KUDINA, 1972; MILNER-BROWN et al, 1973a, 1973b and MILNER-BROWN & STEIN, 1975) are cancelled in the synchronization by the effects of an increase of motor units discharges with increased intensity of contraction (PERSON & KUDINA, 1968 and PER- SON & LIBKIND, 1970). The stretch reflex would play an important role in maintaining isometric contraction. Although muscle spindle discharges responding to changes in muscle length cause the desynchronization of motor units discharges (KUBOTA & OSHIMA, 1959), f usimotor activity adjusts the sensitivity of muscle spindle independently of absolute muscle length (ELDRED, 1967). This may account for no systematic variation in the EMG power spectra with the muscle length. The result in this study disagrees with the findings of SATO (1964, 1965) and KIKUCHI (1968) that the lower frequency components of the bipolar surface EMG increased as the muscle length increased. The explanation for this discrepancy may be in part related to the differences in methods of contraction and frequency analysis technique of the EMG data.

7 Some factors affecting the power spectra of surface electromyograms 111 It was concluded from this study that the power spectra of the bipolar surface EMG of a muscle show relatively definite pattern irrespective of subjects, right- or left-side, contraction level and muscle length at least for normal biceps brachii if constant electrode condition is maintained. Since it seems likely that this is true for other muscles, the difference of the EMG power spectra among various muscles (SATO, 1976) may reflect different characteristics of muscles. ACKNOWLEDGMENT I wish to thank Professor M. SATO for his encouragement and Mr. M. Ismi for his cooperation in the experiments on the intra-individual variation in the EMG power spectra. I also wish to thank the subjects who were students of Kyushu Institute of Design. This work was supported in part by a grant from Science Foundation, the Ministry of Education in Japan. REFERENCES ADRIAN, ED. & D.W. BRONK, 1929: The discharge of impulses in motor nerve fibres. Part II. The frequency of discharge in reflex and voluntary contractions. J. Physiol., 67: BIGLAND, B. & 0. C. J. LIPPOLD, 1954: Motor unit activity in the voluntary contraction of human muscle. J. Physiol., 125: CHAFFIN, D.B., 1969: Surface electromyography frequency analysis as a diagnostic tool. J. Occupat. Med., 11: ELDRED, E., 1967: Functional implications of dynamic and static components of the spindle response to stretch. Amer. J. Phys. Med., 46: FEx, J. & CE. T. KRAKAU, 1957: Some experiences with Walton's frequency analysis of the electromyogram. J. Neurol. Neurosurg. Psychiat., 20: FUSFELD, R. D., 1971: Analysis of electromyographic signals by measurement of wave duration. Electroenceph. din. Neurophysiol., 30: GERSTEN, J. W., D.M. SE, 1967: Harmonic analysis in carriers of Duchenne muscular dystrophy. Arch. Phys. Med., 48: HAYAMI, A., H. SATO & M. SATO, 1967: STILLWELL & N. A. Ro- Cooling effects on the muscular activity. J. Anthrop. Soc. Nippon, 75: HAYES, K. J., 1960: Wave analyses of tissue noise and muscle action potentials. J. Appi. Physiol., 15: KADEFORS, R., E. KAISER & I. PETERSEN,1968: Dynamic spectrum analysis of myo-potentials and with special reference Electromyography, 8: KAISER, E. & I. PETERSEN, 1963: to muscle fatigue. Frequency analysis of muscle action potentials during tetanic contraction. Electromyography, 3: KIKUCHI, Y., 1968: The effect of alcohol on the frequency components of the surface electromyograms. J. Anthrop. Soc. Nippon, 76: (In Japanese with English abstract), KOGI, K. & T. HAKAMADA, 1962a: Frequency analysis of the surface electromyogram in muscle fatigue. J. Sci. Labour, 38: (In Japanese with English abstract). KOGI, K. & T. HAKAMADA,1962b : Slowing of surface electromyogram and muscle strength in muscle fatigue. Rep. Inst. Sci. Labour, 60: KONDO, S., M. SATO, Y, KIKUCHI, M. TOMITA, M. OKADA & A. HAYAMI, 1968: Electromyographic studies on facial muscles of Japanese- American hybrids. J. Anthrop. Soc. Nippon, 76: KOPE6, J, & I. HAUSMAN-PETRUSEWICZ, 1966: Application of harmonic analysis to the electromyograms evaluation. Acta Physiol. Polonica, 17: KUSOTA, K. & T. OSHIMA, 1959: Effects of

8 112 H. SATo gamma blocking on muscular activity and their relation to myasthenic state. Neurologia medico-chirurgica, 1: KWATNY, E., D.H. THOMAS & HG. KWATNY, 1970: An application of signal processing techniques to the study of myoelectric signals. IEEE Trans. Bio-Med. Eng., BME-17 : LINDSLEY, D.B., 1935: Electrical activity of human motor units during voluntary contraction. Am. J. Physiol., 114: LLOYD, A. J., 1971: Surface electromyography during sustained isometric contractions. J. Appl. Physiol., 30: MILNER-BROWN, H.S., R.B. STEIN & R. YEMM. 1973a: The Orderly recruitment of human motor units during voluntary isometric contractions. J. Physiol., 230: MILNER-BROW N, H.S., R. B. STEIN & R. YEMM, 1973b: Changes in firing rate of human motor units during linearly changing voluntary contractions. J. Physiol., 230: MILNER-BROWN, H. S. & R. B. STEIN, 1975: The relation between the surface electromyogram and muscular force. J. Physiol., 246: NIGHTINGALE, A., 1959: "Background noise" in electromyography. Phys. Med. Biol., 3 : NORRIS, F. H. Jr., & E. L. GASTEIGER, 1955: Action potentials of single motor units in normal muscle. Electroenceph. din. Neurophysiol., 7: PERSON, R. S. & L. P. KUDINA, 1968: Crosscorrelation of electromyograms showing interference patterns. Electroenceph, din. Neurophysiol., 25: PERSON, R. S. & L. P. KUDINA,1972: Discharge frequency and discharge pattern of human motor units during voluntary contraction of muscle. Electroenceph. din. Neurophysiol., 32: PERSON, R.S. & M. S. LIBKIND, 1970: Simulation of electromyograms showing interference patterns. Electroenceph. din. Neurophysiol., 28: SATO, H., 1976: Power spectral analysis of surface electromyograms during isometric contractions. J. Anthrop. Soc. Nippon, 84: SATO, M., 1964: Frequency components of the electromyogram led with the bipolar surface electrodes. J. Anthrop. Soc. Nippon, 72: (In Japanese with English abstract), SATO, M., 1965: Some problems in the quantitative evaluation of muscle fatigue by frequency analysis of the electromyogram. J. Anthrop. Soc. Nippon, 73: SATO, M., A. HAYAMI & H. SATO, 1965: Differential fatiguability between the one- and two-joint muscles. J. Anthrop. Soc. Nippon, 73: SATO, M. & S. TSURUMA, 1967: A scope of the frequency analysis of the electromyogram. Ann. Rep. Phys. Education, 1: 7-28 (In Japanese with English abstract), SCOTT, R.N., 1967: Myo-electric energy spectra. Med. Biol. Eng., 5 : WALTON, J.N., 1952: The electromyogram in myopathy, analysis with the audio-frequency spectrometer. J. Neurol. Neurosurg. Psychiat., 15: (Received December 1, 1975)

9 Some factors affecting the power spectra of surface electromyograms 113