The muscle sound properties of different muscle fiber types during voluntary and electrically induced contractions

Journal of Electromyography and Kinesiology 9 (1999) 209–217

The muscle sound properties of different muscle fiber types during voluntary and electrically induced contractions Yasuhide Yoshitake, Toshio Moritani

*

Laboratory of Applied Physiology, The Graduate School of Human and Environmental Studies, Kyoto University, Kyoto, Japan Received 2 August 1998; accepted 11 August 1998

Abstract Soundmyogram (SMG) and electromyogram signals were recorded simultaneously from the relatively fast medial gastrocnemius (MG) and slow soleus (SOL) during voluntary and electrically induced contractions. Using a spike-triggered averaging technique, the averaged elementary sound and corresponding MU spikes were also obtained from about 35 different MUs identified. The rmsSMG of MG increased as a function of force (P ⬍ 0.01). On the contrary, these values for SOL increased up to 60% MVC (P ⬍ 0.01), but decreased at 80% MVC. The relationship between the peak to peak amplitude of SMG and MU spike indicated significant positive correlations (r ⫽ 0.631 苲 0.657, P ⬍ 0.01). During electrical stimulation at 5 Hz, the SMG power spectral peak frequency (PF) was matched with stimulation frequency in both muscles. At higher stimulation frequencies, e.g., > 15 Hz, only in the MG was SMG-PF synchronized with stimulation frequency; the slow SOL did not show such synchronization. Our data suggest that the SMG frequency components might reflect active motor unit firing rates, and that the SMG amplitude depends upon mechanical properties of contraction, muscle fiber composition, and firing rate during voluntary and electrically induced contractions.  1999 Elsevier Science Ltd. All rights reserved. Keywords: Soundmyogram; Surface EMG; Motor units; Electrical stimulation

1. Introduction Muscle contractile properties, motor unit (MU) recruitment, and MU firing frequency (rate coding) have been investigated by means of electromyography (surface or intramuscular EMG). Investigators have suggested that the relative contributions of MU recruitment and firing rate were different in different muscles [19,25]. These authors indicated that the differences may depend upon the muscle fiber compositions, contractile properties, and mechanism of force output control. Muscle sound has recently gained considerable attention as a new non-invasive tool to investigate MU activity. The suggested main generating mechanisms of muscle sound are: (1) a slow bulk lateral movement of the muscle related to the different regional distribution of the contractile elements, (2) the excitation into ringing of the muscle at its own resonant frequency, and (3) the

* Corresponding author. Tel.: ⫹ 81-75-753-6888; fax: ⫹ 81-75753-6888; e-mail: [email protected]

pressure waves generated by the dimensional changes of the fibers of active MUs [32]. Therefore this phenomenon might reflect the intrinsic mechanical activity of muscle contractions. By analogy to the electromyogram (EMG), the muscle sound (soundmyogram; SMG) was analyzed in the time and frequency domain. Many investigators have demonstrated that the SMG amplitude vs. force relationship was different in different muscles. The relationship between SMG amplitude and % maximal voluntary contraction (MVC) was found to be linear in the quadriceps [39], quadratic in the lumbar erector spinae [38], curvilinear in the adductor pollicis [40], and parabolic from 10 to 80% MVC followed by a decrease at 90 and 100% MVC in the biceps brachii [30]. Orizio [32] and Stokes and Dalton [39] suggested that these non-uniform relationships were largely due to the differences of muscle fiber compositions and their contractile properties. In previous studies, the most sound relevant frequency parameters found during isometric contractions of human muscles were reported to be almost below 50 Hz. Recently the SMG frequency content vs. force relation-

1050-6411/99/$ - see front matter  1999 Elsevier Science Ltd. All rights reserved. PII: S 1 0 5 0 - 6 4 1 1 ( 9 8 ) 0 0 0 3 5 - 2

210

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

ship was clearly demonstrated by Orizio et al. [31] in the biceps brachii. They showed that increasing the intensity of contraction resulted in the enlargement of the SMG spectrum. Moreover, they suggested that SMG mean power frequency was similar to the underlying MU firing rate, as detected by intramuscular electrodes at different intensities of contraction reported by Moritani and Muro [27]. While the growing studies in muscle sound have been performed, the uncertain data were also reported in the literature about the relationship between SMG characteristics, MU size, and muscle fiber composition [21,30,31,38,39]. The purpose of the present investigation was therefore to clarify the relationship among MU activity, muscle fiber composition, and muscle sound characteristics. This was accomplished by recording the surface EMG and SMG simultaneously during voluntary and electrically induced contractions so that surface EMG and SMG frequency power spectral analysis could be directly compared in the muscles possessed of considerably different muscle fiber compositions: the slow soleus (SOL) and the relatively fast medial gastrocnemius (MG). The SOL is generally composed of 70–100% slow twitch fibers, whereas the MG possesses > 50% fast twitch fibers, respectively [4,37]. Moreover, we have decomposed single SMGs from individual motor units in order to compare SMG properties with MU size by means of a spike-triggered averaging technique which has been recognized as a highly suitable method to study motor units during voluntary contractions [17,22].

2. Methods 2.1. Subjects and testing protocol Six male subjects (29.7 ⫾ 8.36 yrs, 176 ⫾ 3.87 cm, 66.3 ⫾ 3.83 kg, mean ⫾ SD) with no history of neuromuscular disorders volunteered for this investigation. Each subject was fully informed about the possible risks and nature of the experiments and signed the informed consent. Each subject was then seated on an insulated, straight-back chair with wide belts crossing his chest and abdomen to tightly immobilize his body to isolate the planter flexion movement of the ankle. An additional strap was also used to secure his thigh to the chair. The exact position of the force measurement device was carefully adjusted so that the experimental leg was fully extended with the foot at 90°. 2.2. Force and EMG measurements

amplified through a DC amplifier (SA-100; TEAC). The force signal was displayed on an oscilloscope (V-209; Hitachi) located in front of the subject for visual feedback. The surface EMG signal was picked up by bipolar silver–silver chloride electrodes (9 mm pick-up diameter, 35 mm inter-electrode distance) filled with conducting jelly were applied over the belly of the head of each muscle. Electrode placement was preceded by abrasion of skin surface to reduce the source impedance to less than 3 k⍀. The myoelectric signal was band-pass filtered (5–1000 Hz) and differentially amplified (Nihonkoden MEG-6100; gain: 1000 times, input impedance: > 100 M⍀, CMRR: > 80 dB). In some cases (N ⫽ 2), the intramuscular spikes of gastrocnemius and soleus muscles were recorded to decompose single MU and corresponding SMG by means of a spike-triggered averaging technique (Experiment 3) from high impedance bipolar fine wire electrodes (100 ␮m diameter; one with 100 ␮m and the other with 500 ␮m uninsulated areas) with a hook for fixation. These electrodes were passed through a 24-gauge, 3 cm steel hypodermic needle and inserted approximately 1.5 cm from the skin surface just below the microphone sensor being attached. The MU spike signals were fed through a high impedance probe (300 M⍀ input impedance) with a 100 dB common mode rejection ratio and amplified (Nihon-koden MEG6100, band-pass filter 100 Hz to 10 kHz) by the methods of Moritani et al. [26]. 2.3. Soundmyogram The SMG was detected by a specially designed microphone sensor (10 mm diamater, mass 5 g, Daia Medical, Japan) with a flat frequency response between 5 and 2000 Hz. The microphone sensors were attached to the center of the belly of the SOL and MG muscles with equal distance from EMG electrodes that were positioned to line up the microphone sensor on the longitudinal axis of each muscle. 2.4. Signals recordings The EMG, SMG, and force signals were simultaneously and continuously stored on a desk-top computer (DOS/V) after analog-to-digital conversion (HTB 410; 13 bits resolution). The sampling frequency was 1 kHz (Experiment 1), 2 kHz (Experiment 2) or 10 kHz (Experiment 3) for all signals. Because the SMG maximal frequency range is about 60 Hz, the SMG signal resolution was reduced to 128 Hz sampling interval for subsequent frequency power spectral analyses. 2.5. Experimental designs and analysis

For force recording, a strain gauge transducer (TUBR200K; TEAC) was placed between the base metal plate and the force lever plate. The force signal was

Prior to the experiment, we performed the intramuscular microstimulations to determine the differences of

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

mechanical properties between slow SOL and relatively fast MG muscles. A bipolar needle electrode was inserted into the resting MG and SOL muscles of each subject approximately 2.0 cm from the skin surface. Supramaximal single rectangular pulse of 1 ms duration from the electro-stimulator (SEN-7203; Nihon-koden) were delivered through a stimulator isolation unit (SS 102J; Nihon-koden). The force signals were amplified and stored on a desk-top computer (DOS/V) after analog-to-digital conversion (see, Signals Recordings, above). The following parameters were assessed off-line for evoked muscle twitches: peak twitch force (Fmax), contraction time (CT), half-relaxation time (RT1/2), and maximal rate of force development ( ⫹ dF/dtmax), and relaxation ( ⫺ dF/ditmax) by the methods of Moritani et al. [28]. 2.6. Experiment 1 Isometric plantar flexion was performed. The maximal voluntary contraction (MVC) was defined as the greatest force recorded in three brief maximal efforts. Then the subjects were instructed to maintain isometric muscle action at 20, 40, 60, and 80% MVC in random order for about 4 s while actual recordings of the signals were recorded for 2 s after the force signal reached the target force level. For each force level, the experiments were performed twice with a 3 min rest between contractions to avoid fatigue. The digitized EMG and SMG data were processed with Hamming window function and 2048 and 256 points fast Fourier transform (FFT) were performed to obtain mean power frequency (MPF) and the root mean square values of EMG and SMG amplitude (rmsEMG, SMG), respectively. MPF was defined as the ratio between spectral moments of orders one and zero [27].

211

nals were recorded for 2 s in the middle portion of the trials. The digitized SMG data were processed with the Hamming window function and 256 points FFT to obtain peak power frequency (PF) and rms amplitude values. 2.8. Experiment 3 Isometric plantar flexion was again performed to obtain the SMG signals from the MG muscles composed of a well-balanced population of slow- and fast-twitch fibers at submaximal force levels for 40 s. The subjects (n ⫽ 2) who gave the informed consent were asked to maintain force output as steady as possible so as to recruit the same MUs (about 5 MUs in each trial) as identified by means of intramuscular EMG signals displayed on the oscilloscope. Intramuscular MU spike recordings were made at least four times for each subject from the bipolar fine-wire electrodes at different depths and locations in an effort to obtain different MUs. The SMG and intramuscular EMG were trigger-averaged by means of a window discrimination technique [25–27], which continuously analyzed the intramuscular EMG signal, and allowed a single chosen motor unit action potential peak to be used as a trigger. By this computeraided MU decomposition analysis method, each SMG and MU spike samples were then averaged with the previous one, respectively. The specific SMG from each isolated motor unit was obtained by averaging at least over 100 successive samples. The averaged SMG and MU spikes were calculated for its peak to peak amplitude. SMG signal is influenced by the distance between the microphone sensor and the contracting motor unit, thus the obtained results were presented for each individual subject and muscle, respectively.

2.7. Experiment 2

3. Results

Electrically-evoked potentials were measured by the methods of deVries et al. [8] in two subjects who gave the informed consent to participate in these electrical stimulation experiments. The EMG and SMG recording procedures were identical to those mentioned above. The posterior tibial nerve was stimulated percutaneously with a small surface electrode in the popliteal fossa. Single rectangular pulses of 1 ms duration from the electrostimulator (SEN-7203; Nihon-koden) were delivered through a stimulator isolation unit (SS 102J; Nihonkoden). To determine the optimum stimulation point, several weak stimulations were delivered. In this study, the stimulus intensity was set at one level for which only M wave appeared while stimulation frequencies were changed randomly at six different levels (5, 10, 15, 20, 25, and 30 Hz). These brief stimulations were delivered for 10 s. In order to ensure that transient signals at the beginning and end of the contraction were excluded, sig-

3.1. Test–retest reliability To prove the preciseness of microphone sensor and the computer subroutines for estimating the Fourier spectra and rms-SMG, the test–retest reproducibility was determined during voluntary contractions performed on two separate occasions. The test–retest reliability of our SMG analysis methods was such that the obtained MPF and rms-SMG values recorded during force maintenance at different levels, correlated r ⫽ 0.984 (P ⬍ 0.001) and r ⫽ 0.989 (P ⬍ 0.001) with no significant differences in the mean values (P > 0.05), respectively. Fig. 1 represents a typical set of computer output demonstrating electrically evoked single twitch force characteristics observed from MG and SOL muscles, respectively. When mechanical parameters obtained from MG muscle were compared with those of SOL muscle, statistically significant differences were observed in Fmax

212

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

Fig. 1. A typical set of computer outputs showing twitch contraction force curves obtained from the gastrocnemius (MG) and soleus (SOL) muscles induced by a supramaximal intramuscular stimulation.

(207 ⫾ 17.9 vs. 156 ⫾ 7.5 mN, P ⬍ 0.01), CT (85.8 ⫾ 1.3 vs. 110.7 ⫾ 3.5 ms, P ⬍ 0.01), RT1/2 (78.2 ⫾ 2.0 vs. 110.7 ⫾ 3.5 ms, P ⬍ 0.01), ⫹ dF/dtmax (3.8 ⫾ 0.37 vs. 2.7 ⫾ 0.14 N s−1, P ⬍ 0.01), ⫺ dF/dtmax (1.9 ⫾ 0.11 vs. 1.42 ⫾ 0.08 N s−1, P ⬍ 0.01), respectively. Fig. 2 represents the group data (mean ⫾ SEM) of the surface EMG root mean square amplitude (rms-EMG) of the SOL and MG muscles at 20, 40, 60, and 80% MVC. It was found that there were significant increases in rmsEMG (MG: 597 ⫾ 150 to 2380 ⫾ 154 ␮V, P ⬍ 0.01, SOL: 532 ⫾ 98.1 to 1770 ⫾ 224 ␮V, P ⬍ 0.01) when force output increased from 20 to 80% MVC. Fig. 3 represents a typical data showing the recorded raw SMG data of a subject at 20, 40, 60 and 80% MVC. It can be noted that amplitude of SMG signal of the MG increased up to 80% MVC, but at this 80% MVC level it decreased in SOL. Figs. 4 and 5 show the group data for the rms-SMG and MPF-SMG of the MG and SOL muscles at 20, 40, 60, and 80% MVC, respectively. The rms-SMG of the MG muscles increased as a function of force (MG: 24.4 ⫾ 5.26 to 61.4 ⫾ 13.1 mV, P ⬍ 0.05), but, that of the

SOL muscle did not show any significant increase (P > 0.05) with force of contraction. At 80% MVC the SOL rms-SMG was actually decreased. On the contrary, MPF-SMG increased as a function of relative force for both muscles (MG: 9.6 ⫾ 0.89 to 13.5 ⫾ 0.65 Hz, P ⬍ 0.01, SOL: 6.71 ⫾ 0.50 to 12 ⫾ 1.49 Hz, P ⬍ 0.01). Fig. 6 represents a typical set of 2 s SMG recordings of the MG and SOL muscles recorded simultaneously at 5 and 20 Hz stimulations and the corresponding MPF, peak frequency (PF-), and rms-SMG values for comparison. These results demonstrated that PF- and MPF-SMG corresponded well with the stimulation frequency at 5 Hz for both MG and SOL muscles, respectively. However these values at 20 Hz stimulation were quite different for the MG and SOL; SMG power spectrum peak frequency appeared nearly identical to the stimulus frequency for the MG, but this was way below the stimulation frequency in the case of SOL. Nearly identical data were observed for two other subjects volunteered for this part of the experiments. These data clearly indicated that there existed critical stimulation frequencies for these two different muscles above which the SMG frequency component could not match the stimulation frequency (MG 苲 25 Hz, SOL 苲 10 Hz). In both muscles, SMG amplitude decreased dramatically as stimulation frequency increased. Fig. 7(A) shows typical single MU spike potentials of the MG muscle recorded from a subject during a well controlled weak isometric planter flexion. The top tracing shows two extracted MU spike data. The average of the elementary sound and MU spikes of two subjects were obtained from 16 and 19 motor units identified, respectively. Fig. 7(B) shows typical averaged MU spikes and the corresponding SMG signals. The majority of the averaged SMG demonstrated a negative deflection and its duration was within approximately 150 ms. In Fig. 8, the relationship between P–P amplitude of MU spike and SMG of two different subjects are shown. These results suggested that P–P amplitude of SMG

Fig. 2. Changes in surface EMG root-mean-square amplitude (RMS) of the soleus (SOL) and gastrocnemius (MG) muscles during isometric planter flexion at different force levels.

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

213

Fig. 3. A typical set of computer outputs showing raw SMG of the soleus (SOL) and gastrocnemius (MG) muscles during different levels of contractions.

Fig. 4. Changes in SMG root-mean-square amplitude (RMS) of the soleus (SOL) and gastrocnemius (MG) muscles during isometric contractions at four different force levels.

Fig. 5. Changes in SMG mean power frequency (MPF) of the soleus (SOL) and gastrocnemius (MG) muscles during isometric contractions at four different force levels.

214

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

Fig. 6. A typical set of computer outputs showing raw SMG, normalized power spectra, calculated peak frequency (PF), mean power frequency (MPF), and root mean square amplitude (RMS) values of the gastrocnemius (MG) and soleus (SOL) muscles during electrical stimulations at 5 and 20 Hz, respectively.

increased nearly parallel with MU spike amplitude as indicated by a significant positive correlation (SUB: 1, r ⫽ 0.631, P ⬍ 0.01; SUB: 2, r ⫽ 0.657, P ⬍ 0.01). 4. Discussion

Fig. 7. (A) A typical sample of intramuscular MU spike recording during weak voluntary contraction. (B) Typical averaged MU spikes and the corresponding SMG signals obtained by MU spike-triggered averaging technique.

Mammalian fast and slow muscle can be identified by a significant difference in electro-mechanical properties (e.g. contraction time, half relaxation time, maximum velocity of shortening, etc.) [5,6,44]. Garnett et al. [12] investigated human gastrocnemius muscle motor unit organization by performing controlled intramuscular microstimulation and by measuring glycogen depletion with muscle biopsy technique. The results strongly suggest that the critical contractile parameters enabling motor units to be separated into classes are contraction time or fatigue resistance. It is well established that there is a correlation between twitch speed and relative proportions of type I and type II fibers [4,20,41]. Our data indicated that MG had significantly faster and larger twitches than SOL (Fig. 1) in every subject tested. These observed significant differences in twitch mechanical properties are anticipated in view of the much higher percentage of Type II (fast-twitch) fibers in the MG than in SOL. This study compared the SMG amplitude and frequency components in the muscles possessed of considerably different muscle fiber compositions between relatively “fast-twitch” MG and its “slow-twitch” synergist SOL during voluntary isometric muscle actions. In the past, experimental data on the relationship

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

215

Fig. 8. Relationship between MU spike and SMG peak to peak amplitudes obtained from 16 MUs (SUB: 1) and 19 MUs (SUB: 2) identified in the present study.

between SMG amplitude and force for various muscles have often led to reports of linear or curvilinear relationships [2,38–40], but Orizio et al. [30] have shown a parabolic increase of integrated SMG from 10 to 80% MVC followed by a decrease at 90 and 100% MVC in the biceps brachii. These differences could be due to the muscle fiber type distribution and composition within the muscle [32,39]. Our results on the SMG amplitude changes for the muscles possessing different muscle fiber compositions are in good agreement with previous results reported by Orizio [32]. In the present study we observed an increase of rms-SMG from 20 to 60% MVC followed by a decrease at 80% MVC in the slow SOL. However, in the fast MG muscle, rms-SMG and EMG significantly increased linearly with force output. Taking into account that when all the MU are active, an increase of the firing frequency causes a reduction of the sound amplitude due possibly to force fusion [30,32–34,40]. The reduction of the SMG amplitude for slow SOL muscle may be attributed to the fact that at high contraction levels the high firing frequency of MUs could have resulted in a fusion-like contractile state in which dimensional changes of activated muscle fibers may be greatly reduced. This would in turn lead to diminution of generated pressure wave detectable at the muscle surface as muscle sound. Differences in the SMG amplitude changes for the MG and SOL could also be attributed to the possible difference in the relative involvement of gastrocnemius and soleus even at the same level of force outputs, i.e, 20, 40, 60, and 80% MVC used in the present study. In other words, at the identical % MVC level, actual force contribution from these different muscle groups may vary as not only the muscle composition but also muscle structure (biarticular vs. mono-articular muscle) does differ to a considerable extent. Previous studies [33,40,42] have indicated that SMG signal during stimulated activity reflected the stimulation frequency and force oscillation of muscle fibers. A simi-

lar relationship between muscle sound amplitude and stimulation rate was also described in isolated frog muscle [9]. These authors attributed the reduction of the muscle sound observed at subtetanic and tetanic stimulation rates to the concomitant reduction in output force fluctuations and therefore in the “lateral movement”. Kossev et al. [18] have recently shown that force fusion could easily take place in slow-twitch fibers with increasing firing frequency. Our data support these previous findings on experimental animals in that the SOL SMG amplitude decreased at higher force levels. Our SMG frequency spectral data on SOL not being able to follow high stimulation frequency due to a greater proportion of slow twitch fibers substantiated this finding. The dominant frequency of the SMG signal during stimulated activity is dependent on the stimulation frequency [3,33,40]. Our results also indicated that the MPF-SMG almost completely matched the stimulation frequency until critical values (MG 苲 25 Hz, SOL 苲 10 Hz, see Fig. 6). Moreover, in voluntary contractions, Orizio et al. [31] have suggested that the SMG mean frequency was similar to MU firing rate as detected by intramuscular electrode during voluntary contractions [27]. Results obtained during electrical stimulation showed that the SMG signal frequency matched the stimulation frequency [33,40]. Our SMG data are in good agreement with these findings. However, some investigators have reported higher SMG frequencies during electrical stimulations [35,42,43]. The differences in these experimental results could be largely accounted for by the different properties of the transducers (air-coupled microphone, piezoelectric contact sensor, electronic condenser microphone, accelerometer). Especially, great differences in signal shape are evident when signals from accelerometers and other devices are compared [32]. The detailed discussion on the differences in the waveforms obtained from each device has been thoroughly presented by Bolton et al. [3] and more recently by Ori-

216

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

zio [32]. Our data have consistently demonstrated that SMG frequency components may represent the active MU firing rate and that SMG is the mechanical counterpart of the electrical activity of the muscle fibers during both voluntary and electrically induced contractions. In comparing EMG and SMG during isometric muscle actions, Gordon and Holbourn [14] were the first to investigate the correlation between motor unit action potentials and muscle sound. More recently, Barry et al. [2] have recorded motor unit sound waves and indicated a good correspondence between the motor unit action potentials and the long polyphasic sound waves. Petitjean and Maton [36] have recently demonstrated the decomposition of single sound by using the method similar to the present procedure. However, previous studies have not yet clearly shown the correlation between MU and SMG characteristics. Our data showed a wide range of amplitudes for the decomposed muscle sound with a ratio of about 5 between the largest and the smallest wave. From our data it seems most likely that muscle sound amplitude could be, at least in part, influenced by the size of the activated motor units. Since the “size principle” of Henneman et al. [15] was first proposed based upon results from cat motoneurones, strong evidence has been presented that in muscle contraction there is a specific sequence of recruitment in order of increasing motoneuron and MU size [7,10,11,16,22,23]. Goldberg and Derfler [13] have shown a positive correlation among recruitment order, spike amplitude, and twitch tension of single MUs in human masseter muscle. From our data showing a relatively high significant correlation between SMG P–P amplitude and corresponding single MU amplitude and previous reports [7,10,11,13,16,22,23], it can be suggested that single SMG amplitude increases with increasing MU size. The positive correlation between SMG amplitude and % MVC was also clear from our data in the MG and the foregoing data [21,30,38–40]. Thus, it is most likely that newly recruited MU sound generates larger than already recruited MU sound during a progressive increase of muscle contractions. On the contrary, in the slow SOL observed in our study, SMG amplitude decreased at high contraction levels. Thus it can be suggested that firing rate, type of activated muscle fiber, and muscle fiber compositions are critical for determining the SMG characteristics. Lastly, although we carefully recorded SMG in this experiment while motor units were firing at a fairly constant frequency, we could not eliminate the variability of their individual firing rate during voluntary contraction. Previous studies [1,24,29] have indicated the limitation of spike-trigger averaging technique that averaged signals, even when averaged at low firing rates of about 8 Hz, may be distorted by a partial fusion. Thus, this could explain in part the distorting of sound amplitude we

found in the present study. If confirmed for any given motor unit increasing its firing rate, this relationship would also imply that the elementary sound would merely vanish if its firing frequency increased and reached a critical value. In fact, Vaz et al. [42] have demonstrated that SMG disappears completely at high stimulation frequencies. In conclusion, it appears that the SMG frequency components reflect active motor unit firing rates, and that the SMG amplitude mainly represents mechanical properties of contraction, muscle fiber composition, and firing rate during voluntary and electrically induced contractions.

References [1] Andreassen S, Bar-on E. Estimation of motor unit twitches. IEEE Trans Biomed Eng 1983;30:742–8. [2] Barry DT, Geiringer SR, Ball RD. Acoustic myography: a noninvasive monitor of motor unit fatigue. Muscle and Nerve 1985;8:189–94. [3] Bolton C, Parkes T, Thompson R, Clark M, Sterne J. Recording sounds from human skeletal muscle: technical and physiological aspects. Muscle and Nerve 1989;12:126–34. [4] Buchthal F, Schmalbruch H. Contraction times and fibre types in intact human muscle. Acta Physiol Scand 1970;79:435–52. [5] Burke RE, Edgerton VR. Motor unit properties and selective involvement in movement. Exerc Sport Sci Rev 1975;3:31–81. [6] Close R. Dynamic properties of mammalian skeletal muscles. Physiol Rev 1972;52:129–97. [7] De Luca CJ, LeFever RS, McCue MP. Behaviour of human motor units in different muscles during linearly varying contractions. J Physiol 1982;329:113–28. [8] DeVries HA, Wiswell RA, Bulbulian R, Moritani T. Tranquilizer effect of exercise: acute effects of moderate aerobic exercise on spinal reflex activation level. Am J Phys Med 1981;60:57–66. [9] Frangioni JV, Kwan-Gett TS, Dubrunz LE, McMahon TA. The mechanism of low frequency sound production in muscle. Biophys 1987;51:775–83. [10] Freund HJ, Budingen H, Dietz V. Activity of single motor units from human forearm muscles during voluntary isometric contractions. J Neurophysiol 1975;38:939–46. [11] Freund HJ. Motor unit and muscle activity in voluntary motor control. Physiol Rev 1983;63:387–436. [12] Garnett RAF, O’Donovan MJ, Stephens JA, Taylor A. Motor unit organization of human medial gastrocnemius. J Physiol 1979;287:33–43. [13] Goldberg L, Derfler B. Relationship among recruitment order, spike amplitude, and twitch tension of single motor units in human masseter muscle. J Neurophysiol 1977;40:879–90. [14] Gordon G, Holbourn A. The sound from single motor units in a contracting muscle. J Physiol 1948;107:456–64. [15] Henneman E, Somjem G, Carpenter D. Functional significance of cell size in spinal motoneurons. J Neurophysiol 1965;28:560–80. [16] Henneman E, Olson CB. Relations between structure and function in the design of skeletal muscle. J Neurophysiol 1965;28:581–97. [17] Jorgensen F, Lammert O. Acceleromyography (AMG) 2: contribution of the motor unit. In: Komi PV, editor. Biomechanics VA. Illinois: Haman Kinetics Publishers, 1976:159–64. [18] Kossev A, Elek JM, Wohlfarth K, Schubert M, Dengler R, Wolf W. Assessment of human motor unit twitches—a comparison of

Y. Yoshitake, T. Moritani / Journal of Electromyography and Kinesiology 9 (1999) 209–217

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

spike-triggered averaging and intramuscular microstimulation. Electroenceph Clin Neurophysiol 1994;93:100–5. Kukulka C, Clamann H. Comparison of the recruitment and discharge properties of motor units in human brachial biceps and adductor pollics during isometric contractions. Brain Res 1981;219:45–55. LeBozec S, Maton B. Differences between motor unit firing rate, twitch characteristics and fibre type composition in an agonistic muscle group in man. Eur J Appl Physiol 1987;56:350–5. Maton B, Petitjean M, Cnockaert JC. Phonomyogram and electromyogram relationships with isometric force reinvestigated in man. Eur J Appl Physiol 1990;60:194–201. Milner-Brown HS, Stein RB, Yemm R. The contractile properties of human motor units during voluntary isometric contractions. J Physiol 1973;230:285–306. Milner-Brown HS, Stein RB, Yemm R. Changes in firing rate on human motor units during linearly changing voluntary contractions. J Physiol 1973;230:371–90. Monster AW, Chan H. Isometric force production by motor units of extensor digitorum communis muscle in man. J Neurophysiol 1977;40:1432–43. Moritani T, Muro M, Kijima A, Berry M. Intramuscular spike analysis during ramp force and muscle fatigue. Electromyogr Clin Neurophysiol 1986;26:147–60. Moritani T, Muro M, Nagata A. Intramuscular and surface electromyogram changes during muscle fatigue. J Appl Physiol 1986;60:1179–85. Moritani T, Muro M. Motor unit activity and surface electromyogram power spectrum during force of contraction. Eur J Appl Physiol 1987;56:260–5. Moritani T, Berry MJ, Bacharach DW, Nakamura E. Gas exchange parameters, muscle blood flow and electromechanical properties of the plantar flexors. Eur J Appl Physiol 1987;56:30–7. Nordstrom MA, Miles T, Veale J. Effect of motor unit firing pattern on twitches obtained by spike-triggered averaging. Muscle and Nerve 1989;12:556–67. Orizio C, Perini R, Veicsteeinas A. Muscular sound and force relationship during isometric contraction in man. Eur J Appl Physiol 1989;58:528–33. Orizio C, Perini R, Diemont B, Fingini MM, Veicsteinas A. Spectral analysis of muscular sound during isometric contraction of biceps brachii. J Appl Physiol 1990;68:508–12. Orizio C. Muscle sound: bases for the introduction of mechanomyographic signal in muscle studies. Crit Rev Biomed Eng 1993;21:201–43. Orizio C, Solomonow M, Baratta R, Veicsteinas A. Influence of motor units recruitment and firing rate on the soundmyogram and EMG characteristics in cat gastrocnemius. J Electromyogr Kinesiol 1993;2:232–41. Orizio C, Liberati D, Locatelli C, Grandis DD, Veicsteinas A. Surface mechanomyogram reflects muscle fibres twitches summation. J Biomech 1996;29:475–81. Orizio C, Esposito F, Paganotti I, Mario L, Rossi B, Veicsteinas A. Electrically elicited surface mechanomyogram in myotonic dystrophy. Ital J Neuro Sci 1997;18:185–90. Petitjean M, Maton B. Phonomyogram from single motor units during voluntary isometric contraction. Eur J Appl Physiol 1995;71:215–22.

217

[37] Saltin B, Gollnick PD. Skeletal muscle adaptability: significance for metabolism and performance. In: Peachey LD, Adrian RH, Geiger SR, editors. Handbook of physiology, sect. 10. Bethesda: American Physiological Society, 1983:555–631. [38] Stokes MJ, Moffroid MS, Rush S, Haugh LD. Comparison of acoustic and electrical signals from erectores spinae muscles. Muscle and Nerve 1988;2:331–6. [39] Stokes MJ, Dalton PA. Acoustic myographic activity increases linearly up to maximal voluntary isometric force in the human quadriceps muscle. J Neuro Sci 1991;101:163–7. [40] Stokes MJ, Cooper RG. Muscle sounds during voluntary and stimulated contractions of the human adductor pollicis muscle. J Appl Physiol 1992;72:1908–13. [41] Vandervoot AA, McComas AJ. A comparison of the contractile properties of the human gastrocnemius and soleus muscles. Eur J Appl Physiol 1983;51:435–40. [42] Vaz MA, Herzog W, Zhang Y-T, Leonard TR, Ngyuyen H. Mechanism of electrically elicited muscle vibrations in the insitu cat soleus muscle. Muscle and Nerve 1996;19:774–6. [43] Vaz MA, Herzog W, Zhang Y-T, Leonard TR, Ngyuyen H. The effect of muscle length on electrically elicited muscle vibrations in the in-situ cat soleus muscle. J Electromyogr Kinesiol 1997;7:113–21. [44] Vrbova G. Influence of activity on same characteristic properties of slow and fast mammalian muscles. Exerc Sport Sci Rev 1979;7:181–213. Yasuhide Yoshitake was born in 1971 in Ehime, Japan. He received his B.Sc. degree in chemical engineering from the Osaka City University in 1994. He received the M.Sc. degree in 1996 from the Graduate School of Human and Environmental Studies at Kyoto University under the guidance of Dr Moritani and is currently working towards his Ph.D. degree. His major research interests focus on muscle sound, electromyography and on-line computer signal processing. He is currently investigating the physiological mechanism(s) of muscle sound generation and “catch-like” properties of human skeletal muscle while employing intramuscular stimulation and near infrared spectroscopy (NIRS) for assessing electrophysiological and metabolic responses, respectively. Toshio Moritani was born in Japan in 1950. He received his Ph.D. degree in exercise physiology from the University of Southern California in 1980 under the direction of Dr Hebert A. deVries. In 1985, following faculty appointments at the University of Texas at Arlington and Texas A&M University, he returned to Japan and joined the Department of Integrated Human Studies at Kyoto University. In 1992, he was appointed as Associate Professor of Applied Physiology at the Graduate School of Human and Environmental Studies, Kyoto University, where he is currently Director of the Applied Physiology Laboratory. Dr Moritani has been elected as a Fellow of the American College of Sports Medicine. Dr Moritani is the Editor of the Journal of Electromyography and Kinesiology. He is also serving as a member of the Editorial Board for the European Journal of Applied Physiology and Editorial Consultants for the Journal of Biomechanics. He is also one of the Council Members of the International Society of Biomechanics and the Japanese Society of Electrophysiology and Kinesiology.