Application of Mechanomyography for Examining Muscle Function

Application of Mechanomyography for Examining Muscle Function 2010

Editors Dr. Travis W. Beck Biophysics Laboratory Department of Health and Exercise Science University of Oklahoma, 1401 Asp Avenue Norman, Oklahoma, 73019-6081, USA

Application of Mechanomyography for Examining Muscle Function 2010 Published by Transworld Research Network 2010; Rights Reserved Transworld Research Network T.C. 37/661(2), Fort P.O., Trivandrum-695 023, Kerala, India Editors Dr. Travis W. Beck Managing Editor S.G. Pandalai Publication Manager A. Gayathri Transworld Research Network and the Editors assumes no responsibility for the opinions and statements advanced by contributors

ISBN: 978-81-7895-449-3

Preface Most researchers in the movement sciences are at least somewhat familiar with the technique of surface mechanomyography (MMG). Although the literature base for MMG did not start to develop until the 1980s, the field grew fairly rapidly along with the advent of the digital computer and sensors that were capable of detecting the signal. Eventually, the large electronic stethoscopes and condenser microphones that were used in most of the early MMG studies were replaced by lightweight accelerometers and laser displacement sensors that currently set the standard for MMG signal detection. These advances in technology allowed MMG to spread into many different fields, including physiology, biophysics, sports medicine, exercise physiology, and biomedical and rehabilitation engineering. Thus, the knowledge base for MMG has improved substantially over the past 25 years.

Two of the most important areas of MMG research are: (1) examining the mechanisms that generate the MMG signal, and (2) identifying practical applications for MMG. It could be argued that almost all MMG studies have been driven by one of these two areas. My purpose in writing this book is to summarize the literature with separate chapters for specific areas in MMG. These areas include: (1) the mechanisms underlying the MMG signal, (2) the MMG amplitude and frequency versus isometric force relationships, (3) MMG responses to muscle fatigue, (4) clinical applications of MMG, (5) MMG responses during dynamic muscle actions, (6) technical aspects of MMG, (7) processing the MMG signal, and (8) unique applications of MMG. It is important to point out that this book would not have been possible without the support of many people. The most notable, however, have been my parents. I will always be grateful for the

things that they have taught me and the opportunities that they have given me.

Travis Beck USA

Contents

Chapter 1 The mechanisms underlying the surface mechanomyogram Travis W. Beck Chapter 2 The mechanomyographic amplitude and frequency versus isometric force relationships Travis W. Beck Chapter 3 Surface mechanomyographic responses to muscle fatigue Travis W. Beck Chapter 4 Clinical applications of surface mechanomyography Travis W. Beck

1

17

37

53

Chapter 5 Mechanomyographic responses during dynamic muscle actions Travis W. Beck

69

Chapter 6 Technical aspects of surface mechanomyography Travis W. Beck

95

Chapter 7 Processing the surface mechanomyographic signal Travis W. Beck

109

Chapter 8 Unique applications of mechanomyography Travis W. Beck

117

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010, 1-16 ISBN :978-81-7895-449-3 Editor Travis.W. Beck.

1

The mechanisms underlying the surface mechanomyogram Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-608, USA

Abstract Muscle sounds were first discovered in the 1660s, when Francesco Grimaldi observed a rumbling sound as he placed his thumbs over the opening to his ears and clenched his fists. In 1810, William Hyde Wollaston reported that the sounds from a contracting muscle were very much like those from a carriage that was being pulled on the cobblestone streets of London at a speed of 8 miles per hour. By calculating the number of cobblestones that the carriage wheels contacted every second, Wollaston was able to conclude that muscle sounds were dominated by frequencies around 23 Hz. Until Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA E-mail: tbeck@ou.edu

2

Travis W. Beck

1948, however, research on muscle sounds was limited by the inability to accurately detect them and store the data for further analysis. Thus, the development of piezoelectric transducers and condenser microphones was important because it sparked a new interest in identifying the mechanisms that generate muscle sound and the information that it provides. To avoid confusion in the literature, the term “surface mechanomyogram” (MMG) was proposed to describe muscle sounds because it highlighted the fact that they are generated by the mechanical activities of motor units. In the late 1980s, carefully-controlled electrical stimulation studies with isolated muscle preparations showed that the MMG signal is generated by lateral muscle fiber oscillations, the amplitude of which were dependent on the length of the muscle and the intensity of the stimulation. Furthermore, the frequency content of the MMG signal was closely related to the muscle’s resonant frequency, and it was hypothesized that this property may make MMG useful for examining muscle stiffness. Additional research in the mid-1990s indicated that during most voluntary muscle actions, the activities from individual motor units are summated nonlinearly to form the MMG signal. Recent studies have also shown that the spike-triggered averaging technique allows the activities from individual motor units to be extracted from the MMG signal recorded during a voluntary muscle action. Furthermore, the amplitude and frequency contents of the signals produced by these motor units were influenced by their morphology. This is important because it provides a promising future for work in the area of examining muscle fiber type composition with MMG.

Introduction A discussion of the mechanisms underlying muscle sound [hereafter referred to as the surface mechanomyogram (MMG)] would not be complete without first providing a historical background of the field. The first acknowledgement of MMG was provided in 1663 by Francesco Grimaldi, who reported a rumbling noise when he placed his thumbs over the opening to his ears and clenched his fists. It was hypothesized that the rumbling was due to the “…continuous hurried motion of the spirits…” that caused tremors in the fingers, arms, and in the whole body (Orizio 1993). At the time, it was thought that muscle contraction was due to fluid movement that emanated from the brain and caused bulging of the muscles. Thus, even though Grimaldi’s explanation for MMG was incomplete, it is clear that he understood that muscle sound was related to muscle activity, and its properties were related to the properties of the contraction.

The mechanisms underlying the surface mechanomyogram

3

The first true experiment with MMG was conducted in 1810 by William Hyde Wollaston. Wollaston was a physicist, chemist, and physician, and was particularly interested in the frequencies of muscle sounds. Interestingly, he found that muscle sounds were similar in frequency to the sounds produced when a carriage was pulled on the cobblestone streets of London at a speed of 8 miles per hour. In 1810, the cobblestones were 6 inches in length. Thus, at a speed of 8 miles per hour (11.7 feet per second), the carriage wheels were hitting approximately 23 cobblestones per second. This allowed Wollaston to conclude that the dominant frequency of MMG was 23 Hz, which was a very accurate observation, considering the relatively primitive equipment that he had available. After 1810, studies on MMG were relatively scarce. In 1885, however, Herroun and Yeo reported that the muscle sounds produced during voluntary contractions were similar in frequency to those from electricallystimulated contractions, but only for stimulation rates less than 30 Hz. Although it was not immediately known at the time, this was an important finding because it was the first study to determine that a muscle does not produce any sound (i.e., it is silent) when its twitches become completely fused. This finding would eventually be supported by the results from future studies that used isolated muscle preparations and very sensitive MMG sensors. An important step was also taken by Gordon and Holbourn (1948), who were the first to create a record of the MMG signal with a microphone (Figure 1).

Although relatively primitive by today’s standards, this instrument allowed the authors to simultaneously record the surface electromyographic (EMG) and MMG signals from the orbicularis oculi muscle. As a result, they were able to conclude that the surface MMG signal is the mechanical counterpart of the motor unit electrical activity (Gordon and Holbourn 1948) as measured by EMG. Despite the advancements made by Herroun and Yeo (1885) and Gordon and Holbourn (1948), very little research was conducted in the area of muscle sounds until 1980, when Oster and Jaffe sparked a new interest in MMG that was at least partially due to the development of electronic sensors capable of detecting the signal. One of the most important aspects of the study by Oster and Jaffe (1980) was that the authors concluded that the MMG signal was actually generated by contracting muscle. Specifically, it had been argued that muscle sound could be caused by rubbing of the sensor on the skin surface during contraction, or even blood flow through vessels near the muscle.

However, these hypotheses were refuted by Oster and Jaffe (1980), who found that an MMG signal was still detectable when the muscle was contracted under water, and the sensor was 1 cm away from the skin surface. In addition, muscle sounds were generated by the biceps brachii muscle when an inflatable cuff

4

Travis W. Beck

Figure 1. The microphone used by Gordon and Holbourn (1948) to detect mechanomyographic (MMG) signals. Wire or surface electrodes (indicated by (B) in the figure) are optional. (A) shows the skin surface, (C) the microphone chamber with its end closed by a diaphragm (D). Movements of the skin surface cause pressure changes inside the microphone, that, in turn, result in movement of the diaphragm and deform a Rochelle salt crystal (E). The Rochelle salt crystal has piezoelectric properties, which allows it to create potentials that are transmitted through a screened cable to the amplifier (F). *Reprinted with permission from Gordon and Holbourn (1948).

was used to restrict blood flow through the arteries of the arm, thereby disproving the blood flow hypothesis. Perhaps the most important aspect of the Oster and Jaffe (1980) study, however, was that the authors used autocorrelation to determine that MMG signals were periodic with a dominant frequency of approximately 25 Hz. In addition, the amplitudes of the MMG signals increased with the intensity of the contractions, but the MMG signals were uncorrelated both in frequency and in phase with the corresponding surface EMG signal. The overall importance of the Oster and Jaffe (1980) investigation was reflected in the number of MMG studies conducted shortly

The mechanisms underlying the surface mechanomyogram

5

thereafter. Now that researchers knew the MMG signal was generated by muscle activity, they could focus on the mechanisms underlying it.

At this point, it is important to acknowledge the fact that various terms have been used to describe MMG, including soundmyography, phonomyography, acoustic-myography, accelerometermyography, and vibromyography.

Although the use of these terms was heavily influenced by the type of sensor used to detect the signal, it created confusion in the literature with regard to what was actually being measured. Thus, in 1995, the term “surface mechanomyogram” was suggested at a CIBA Foundation Symposium to distinguish the MMG signal from other mechanical signals that are unrelated to muscle activity (Orizio et al. 2003).

After the Oster and Jaffe (1980) study, three very important investigations into the mechanisms that generate the MMG signal were conducted in the laboratory of Dr. Daniel T. Barry (Barry 1987; Barry and Cole 1988, 1990). In the first study (Barry 1987), isolated frog gastrocnemius muscle was placed in a water bath and MMG signals were recorded during electrically-stimulated isometric twitches with two hydrophones that were placed on opposite sides of the muscle and perpendicular to its long axis. When the muscle was stimulated to contract with a single twitch, the resulting MMG signals from the two hydrophones were very similar in shape, but reversed in phase by 180°. In addition, each isometric twitch demonstrated an MMG signal with oscillations that had maximum amplitude at the beginning of the twitch, followed by oscillations that decayed in amplitude over time. The number of oscillations that were produced with each twitch was also influenced by muscle length, with longer muscle lengths resulting in more oscillations. Muscle length also affected the peak-to-peak amplitude of the MMG signal, since the maximum amplitude occurred at a muscle length that was approximately 90% of the optimal length for force production. Thus, it was suggested that muscle sound is likely generated by lateral muscle fiber oscillations, since the MMG signals from the two hydrophones were reversed in phase, thereby indicating side-toside movement of the muscle, as opposed to bulging in all directions. In addition, changes in muscle length can affect MMG amplitude, independent of the muscle activation level. The conclusions of Barry (1987) were supported by the results from a second study (Barry and Cole 1988) that used high speed cinematography to track the movement of isolated frog gastrocnemius muscle during electrically stimulated isometric twitches. Specifically, the high speed images showed lateral displacement of the muscle during each twitch that was consistent with the oscillating pressure waves detected by the hydrophones in the previous study. In addition, the frequency content of the MMG signal was closely related to the muscle’s resonant frequency, which suggested that MMG could potentially be used as a noninvasive measure of muscle stiffness (Barry

6

Travis W. Beck

and Cole 1988). This hypothesis lead to a third study (Barry and Cole 1990) that used time-frequency signal processing techniques to examine changes in MMG frequency during an electrically-stimulated isometric tetanic contraction. Specifically, isolated frog gastrocnemius muscle was electrically stimulated at a frequency of 150 Hz to induce tetanus, while an MMG signal was detected with a hydrophone. The results showed that the resonant frequency of the muscle was in the same range as the peak instantaneous MMG frequency. Thus, it was concluded (Barry and Cole 1990) that the frequency content of the MMG signal was heavily influenced by the resonant frequency of the muscle, which, in turn, was affected by muscle stiffness.

At roughly the same time that Dr. Barry’s laboratory was conducting research to examine the mechanisms underlying the MMG signal, similar research was being done in the Division of Applied Sciences at Harvard University. Specifically, Frangioni et al. (1987) were interested not only in the mechanisms underlying muscle vibration, but also in the directions that the vibrations were transmitted. The authors (Frangioni et al. 1987) electrically stimulated isolated frog gastrocnemius muscle in a saline bath and detected MMG signals with a hydrophone that was placed perpendicular to the long axis of the muscle. Like Barry (1987), Frangioni et al. (1987) found that during an isometric twitch, the MMG signal was generated by lateral movement of the whole muscle, and its maximum amplitude occurred when the muscle length was slightly less than that required for maximum force production. The authors (Frangioni et al. 1987) also hypothesized, however, that when stimulated to contract, the muscle vibrates like an axe handle, or any rigid object having a long elliptical shape. This was an important hypothesis because it was the first time that a model had been proposed to describe muscle sound. The finding that MMG amplitude was greatest when the muscle length was slightly less than that required for maximum force production was supported by Dobrunz et al. (1990) for isolated frog gastrocnemius muscle. In addition, it was hypothesized that the frequency content of the MMG signal may be closely related to muscle stiffness. Around the same time as the studies by Frangioni et al. (1987) and Barry (1987), Brozovich and Pollack (1983) performed an interesting investigation to determine if muscle sounds occurred in a discrete versus continuous manner. Specifically, isolated frog sartorius muscle was electrically stimulated to shorten under lightly loaded conditions while an MMG signal was detected with a piezoelectric sound transducer. The results showed that for each contraction, the MMG signal occurred in discrete bursts, rather than a continuous tone. Thus, the explanation provided by the authors was that since a muscle is a constant volume system, stepwise changes in fiber length should also cause stepwise changes in fiber radius. These stepwise changes in fiber radius, in turn, were thought to

The mechanisms underlying the surface mechanomyogram

7

generate the MMG signal (Brozovich and Pollack 1983). The results from these studies (Barry 1987; Barry and Cole 1988, 1990; Frangioni et al. 1987; Dobrunz et al. 1990; Brozovich and Pollack 1983) provided some very useful information regarding the mechanisms underlying the MMG signal during electrically stimulated isometric contractions.

It is important to point out that the mechanisms underlying the MMG signal during an electrically-stimulated contraction are very different from those during a voluntary muscle action (dynamic or isometric). Specifically, during an electrically stimulated contraction, all fibers are stimulated to contract simultaneously, and the response of the muscle is dependent on the stimulation frequency and the muscle’s ability to contract and relax at a rate that matches the stimulation rate. If the stimulation rate is too fast for the contraction and relaxation times of the muscle (e.g., due to muscle fatigue), then the twitches of the muscle eventually become fused, resulting in tetanus.

During a voluntary muscle action, however, the motor unit activities are usually not synchronized, and the twitches from each motor unit are summed to create a complex MMG signal (i.e., complex in the sense that it is generated by many motor units that are firing at different times, as opposed to one or a few motor units that are firing synchronously). An obvious question then, is how are the motor unit mechanical activities during a voluntary muscle action summed to form the MMG signal? Is this summation linear or nonlinear?

A very important study in this regard was conducted by Orizio et al.

(1996). To determine how the motor unit mechanical activities are summed to form the MMG signal, the authors electrically stimulated two separate motor units in the extensor digitorum communis muscle at different rates. During the first stimulation protocol, the first motor unit was stimulated at 3 Hz, and the second motor unit was stimulated separately at a frequency of 8 Hz. This same procedure was then performed a second time, but at stimulation rates of 9 and 20 Hz for the first and second motor units, respectively. The MMG signal from the 3 Hz stimulation was then linearly summed with that from the 8 Hz stimulation to create a new MMG signal. The same procedure was also followed for the MMG signals from the 9 Hz and 20 Hz stimulation rates. The second stimulation protocol involved simultaneous stimulation of the first and second motor units at 3 and 8 Hz. The resulting MMG signal was then compared with that from linearly summing the MMG signals from the 3 and 8 Hz stimulation rates. The MMG signal from the simultaneous 9 and 20 Hz stimulation was also compared with that from the linear sum of the separate 9 and 20 Hz stimulations. Figure 2 shows that the MMG signal from the simultaneous 3 and 8 Hz stimulation was nearly identical to that from linearly summing the signals from the separate 3 and 8 Hz stimulations, which suggested that when two separate motor units are active at these stimulation rates,

8

Travis W. Beck

Figure 2. One second time windows in which the mechanomyographic (MMG) signal was detected from the extensor digitorum communis during separate electrical stimulations of single motor units at 3 Hz (second graph from top) and 8 Hz (third graph from top). MMGg shows the MMG signal generated from mathematical (i.e., linear) summation of the 3 and 8 Hz stimulation signals. MMGs demonstrates the MMG signal from simultaneous stimulation of the separate motor units at 3 and 8 Hz. Notice that MMGg and MMGs are nearly identical, thereby indicating linear summation of the mechanical activities from the two motor units. The top graph shows the electrical stimulation protocol. *Reprinted with permission from Orizio et al. (1996).

The mechanisms underlying the surface mechanomyogram

9

their contributions are summed linearly to form the MMG signal.

As shown in Figure 3, however, the MMG signal from the simultaneous 9 and 20 Hz stimulation was very different from that of the linear sum of the separate 9 and 20 Hz signals. Thus, these findings suggested that when these higher stimulation rates were used, the contributions from the first and second motor units were not summed linearly. The practical importance of these results is that most voluntary muscle actions involve firing rates greater than 8 Hz. Therefore, during almost all voluntary activities, the contributions of each motor unit are summed nonlinearly to form the MMG signal.

Another important question is whether or not individual motor unit activities can be extracted from the MMG signal. Three important studies in this regard were conducted fairly recently. Specifically, Bichler and Celichowski (2001, p. 388) reported that during an electrically stimulated fatigue test of isolated motor units in rat medial gastrocnemius muscle, the motor units that were classified as “fast fatigable” (i.e., based on their MMG responses) demonstrated the greatest signs of fatigue, while the “fast resistant” and “slow-twitch” motor units were the most resistant to fatigue. Similar results were reported by Bichler (2000), who found that during electrically stimulated contractions (both single twitches and a fused tetanus) of isolated rat medial gastrocnemius muscle, the fast fatigable motor units demonstrated higher MMG amplitude values than both the fatigue resistant and slow-twitch motor units. In addition, Bichler and Celichowski (2001) found that during electrically stimulated contractions of isolated motor units in rat medial gastrocnemius muscle at frequencies that resulted in unfused tetanus, MMG amplitude was positively correlated with the velocity of the tension increase at the beginning of the contraction. Thus, it is possible that under these conditions, MMG amplitude could be used to identify different types of motor units (Bichler and Celichowski 2001).

Two interesting studies in this area have also been performed by Kaczmarek et al. (2005, in press). Specifically, the authors used both experimental and model MMG signals to examine the influence of motor unit location on the MMG signals from pennate muscles. In both investigations, the authors found that the characteristics of the MMG signal were influenced by the location of the MMG sensor over the muscle as well as the location of the active motor units within the muscle. These findings (Kaczmarek et al.

2005, in press) were important from a practical standpoint because they were the first to describe the influence of muscle architecture on the MMG signal.

In addition, muscle architecture can affect the MMG signal, independent of the twitch properties of the active motor units.

10

Travis W. Beck

Figure 3. One second time windows in which the mechanomyographic (MMG) signal was detected from the extensor digitorum communis during separate electrical stimulations of single motor units at 9 Hz (second graph from top) and 20 Hz (third graph from top). MMGg shows the MMG signal generated from mathematical (i.e., linear) summation of the 9 and 20 Hz stimulation signals. MMGs demonstrates the MMG signal from simultaneous stimulation of the separate motor units at 9 and 20 Hz. Notice that unlike the results shown in Figure 2, MMGg and MMGs are not identical, thereby indicating nonlinear summation of the mechanical activities from the two motor units. The top graph shows the electrical stimulation protocol. *Reprinted with permission from Orizio et al. (1996).

The mechanisms underlying the surface mechanomyogram

11

Another important question is whether or not the contributions from single motor units can be extracted from the MMG signal recorded during a voluntary muscle action. The results from the electrical stimulation studies indicated that fast-twitch motor units demonstrated greater MMG amplitude values than slow-twitch motor units, but the responses during voluntary muscle actions were still unclear. The first study to examine this issue was Yoshitake and Moritani (1999), who used the spike-triggered averaging technique to isolate single motor unit activities in MMG signals. Specifically, the authors (Yoshitake and Moritani 1999) detected intramuscular EMG signals from the biceps brachii muscle to identify individual motor unit action potentials. The time occurrences of these action potentials were then used to trigger the MMG signals, and the resulting peak-to-peak amplitudes of the MMG signals from the motor unit spikes were examined. The authors (Yoshitake and Moritani 1999) found that the peak-to-peak amplitude of the triggered MMG signal was positively correlated (r = 0.631-0.657) with that of the indwelling EMG signal.

In a second study, Yoshitake et al. (2002) electrically stimulated isolated motor units from the medial gastrocnemius muscle and reported that during single twitches, the duration of the MMG signal was positively correlated with both half relaxation time (r = 0.76) and twitch duration (r = 0.89). Furthermore, during repeated electrical stimulation of the motor units at frequencies of 5, 10, 15, and 20 Hz, the changes in force fluctuations with increases in stimulation rate were positively correlated (r = 0.76) with the changes in MMG amplitude.

Thus, it was suggested that the characteristics of the MMG signal are affected by the contractile properties of the activated motor units (Yoshitake et al.

2002). The surface EMG signal has also been used to trigger the motor unit mechanical activities in MMG signals (Cescon et al. 2004). Specifically, surface EMG and MMG signals were detected simultaneously from the abductor digiti minimi or first dorsal interosseous during 60-second sustained isometric muscle actions at 2% or 5% MVC, or with a selected motor unit firing at a rate of 8-10 Hz. The results showed that the peak-to-peak amplitudes of the EMG and MMG spikes were not significantly correlated. In addition, the peak-to-peak amplitude of the MMG signal from a single motor unit was different for the 2% versus the 5% MVC contractions. Thus, it was suggested (Cescon et al. 2004) that at these low force levels, the mechanical activities of the active motor units are summated nonlinearly to form the MMG signal.

This study (Cescon et al. 2004) was then followed up with a second investigation (Cescon et al. 2006) that used the surface EMG signal to trigger events in MMG signals from higher force levels (20%, 50%, and 80% MVC).

The authors (Cescon et al. 2006) found that the EMG motor unit spike amplitude increased with force, while the mean power frequency (MPF) decreased. Thus, it was concluded that the amplitude and frequency contents

12

Travis W. Beck

of the single motor unit EMG signals were not correlated with the corresponding contents from single motor unit MMG signals (Cescon et al.

2006).

The spike-triggered averaging technique has also been used to examine the influence of sensor location on the MMG signal. For example, Cescon et al.

(2007) used intramuscular EMG signals to trigger the MMG signals recorded from 15 different locations (i.e., a grid of accelerometers with 5 rows and 3 columns) over the tibialis anterior during isometric dorsiflexion muscle actions. The authors (Cescon et al. 2007) reported that for each motor unit action potential, the lateral MMG sensors showed signals that were primarily negative in phase, while the medial sensors demonstrated signals that were primarily positive in phase. Thus, it was suggested that a grid of MMG sensors may be capable of providing more complete information regarding the motor unit mechanical activities than when a single MMG sensor is used (Cescon et al. 2007). A similar study (Cescon et al. in press) was also performed with a slightly different arrangement of accelerometers. Specifically, 6 MMG sensors were placed along the long axis of the tibialis anterior muscle with a 30 mm separation. The purpose of this arrangement was to measure the longitudinal component of the MMG signal. Seven additional MMG sensors were also located around the circumference of the lower leg to measure the transverse vibrations, and intramuscular EMG electrodes were used to detect the firings of individual motor units. The EMG signal from the intramuscular electrodes was used as a trigger to identify the individual motor unit contributions to MMG signals. The authors (Cescon et al. in press) reported that the peak-topeak amplitude of the spike-triggered MMG signal was the same for all sensors located along the long axis of the tibialis anterior muscle, but it was different for the sensors placed around the circumference of the lower leg. In addition, the MMG sensors that were 180° apart around the circumference of the lower leg demonstrated MMG signals that were opposite in phase. Thus, it was suggested that the dependence of the MMG signal on the transverse, but not the longitudinal location of the sensor supported the hypothesis that lateral displacement of muscle fibers, rather than changes in muscle fiber diameter, is the primary mechanism generating the MMG signal (Cescon et al. in press).

Miyamoto and Oda (2003) performed an important study that investigated the effect of changes in knee and/or ankle joint angle on the MMG and EMG responses for the soleus and both the medial and lateral heads of the gastrocnemius muscle during an isometric MVC of the plantar flexors.

Specifically, the subjects were required to perform maximal isometric muscle actions of the plantar flexors at knee joint angles ranging from 60˚ to 180˚ (full extension) in steps of 30˚, and ankle joint angles ranging from 80˚ (10˚ of dorsiflexion) to 120˚ (30˚ of plantar flexion) in steps of 10˚. The results

The mechanisms underlying the surface mechanomyogram

13

showed that MMG amplitude for the medial and lateral heads of the gastrocnemius increased with increases in knee joint angle, but there were no changes in MMG amplitude for the soleus. In addition, as the ankle angle increased from 80˚ to 120˚, MMG amplitude decreased for all three muscles.

Finally, the changes in ankle and knee joint angles had no effect on the EMG amplitude values for any of the muscles. Thus, it was concluded that the MMG signals detected from individual muscles during an isometric MVC may be useful for providing information regarding the length-tension relationships (Miyamoto and Oda 2003). In addition, these findings supported those from previous in vitro studies that showed changes in MMG amplitude with changes in muscle length. This study was followed up by a second investigation (Miyamoto and Oda 2005) that examined the influence of changes in elbow joint angle on MMG amplitude for the biceps brachii during unfused and fused electrically-stimulated contractions. Specifically, the subjects were required to perform an electrical stimulation protocol in which the biceps brachii was stimulated at the motor point with single twitches, as well as at frequencies that caused unfused (10 Hz) and fused (30 Hz) contractions. This stimulation protocol was performed at elbow joint angles of 75˚, 90˚, 105˚, 120˚, 135˚, and 150˚, where 180˚ represents full extension of the forearm. The results showed that during the 10 Hz stimulation, decreases in force fluctuation were accompanied by reductions in MMG amplitude as the elbow joint angle increased. In addition, for the 30 Hz stimulation, the force fluctuations and MMG amplitude were similar at all elbow joint angles.

Finally, the contraction time and half relaxation time at each elbow joint angle were correlated with the force fluctuation and MMG amplitude during the 10 Hz stimulation, but not during the 30 Hz stimulation. Thus, it was concluded that the 30 Hz stimulation protocol caused fusion of motor unit twitches at all elbow joint angles, but the 10 Hz stimulation allowed for investigation of the motor units in the unfused state. Therefore, MMG could be a useful technique for studying the development of fusion and changes in the contractile properties of the muscle during unfused contractions (Miyamoto and Oda 2005). Farina et al. (2008) also performed an important study that examined the influence of motor unit location on the amplitude of the MMG signal. Specifically, 12 separate MMG signals were detected with a 3 × 4 grid of accelerometers that was placed over the tibialis anterior muscle, and three separate indwelling EMG signals were detected simultaneously with wire electrodes from different locations in the muscle. The subjects were instructed to voluntarily activate three different motor units with feedback from the indwelling EMG signals. The MMG signals from the 12 accelerometers were then averaged, and the indwelling EMG signals were used to spike trigger the MMG signal. The results showed that the peak-to-peak amplitude of the spike-

Travis W. Beck

14

triggered MMG signal was influenced both by the location of the motor unit within the muscle and the location of the accelerometer. Thus, it was concluded that like surface EMG, the volume conductor has an important influence on the MMG signal, and interpretations of MMG amplitude should take into account this influence. In addition, any relationship between twitch torque and single motor unit MMG responses from only one recording location has validity that is limited to motor units that are in a similar location (Farina et al. 2008).

In summary, the results from the studies described in this chapter showed that the MMG signal is generated primarily by lateral oscillations of muscle fibers. In addition, the MMG signal is highly complex, and reflects the mechanical activities of individual motor units. This information could be useful for describing differences in fiber type composition and identifying the mechanisms that underlie the force generation process. More research is needed, however, to determine the factors that can affect the individual motor unit contributions to the MMG signal.

References 1.

2.

3.

4.

5.

6.

7.

8.

Barry DT. Acoustic signals from frog skeletal muscle. Biophysical Journal 1988; 53:899-905.

Barry DT, Cole NM. Muscle sounds are emitted at the resonant frequencies of skeletal muscle. IEEE Transactions. On Biomedical Engineering 1990; 37:525531.

Bichler E. Mechanomyograms recorded during evoked contractions of single motor units in the rat medial gastrocnemius muscle. European Journal of Applied Physiology 2000; 83:310-319.

Bichler E, Celichowski J. Changes in the properties of mechanomyographic signals and in the tension during the fatigue test of rat medial gastrocnemius muscle motor units. Journal of Electromyography and Kinesiology 2001; 11:387-394.

Bichler E, Celichowski J. Mechanomyographic signals generated during unfused tetani of single motor units in the rat medial gastrocnemius muscle. European Journal of Applied Physiology 2001; 85:513-520.

Brozovich FV, Pollack GH. Muscle contraction generates discrete sound bursts.

Biophysical Journal 1983; 41:35-40.

Cescon C, Gazzoni M, Gobbo M, Orizio C, Farina D. Non-invasive assessment of single motor unit mechanomyographic response and twitch force by spike triggered averaging. Medical & Biological Engineering & Computing 2004; 42:496-501.

Cescon C, Madeleine P, Farina D. Longitudinal and transverse propagation of surface mechanomyographic waves generated by single motor unit activity.

Medical & Biological Engineering & Computing (In Press).

The mechanisms underlying the surface mechanomyogram

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

15

Cescon C, Madeleine P, Graven-Nielsen T, Merletti R, Farina D. Two-dimensional spatial distribution of surface mechanomyographical response to single motor unit activity. Journal of Neuroscience Methods 2007; 159:19-25.

Cescon C, Squazzi E, Merletti R, Farina D. Non-invasive characterization of single motor unit electromyographic and mechanomyographic activities in the biceps brachii muscle. Journal of Electromyography and Kinesiology 2006; 16:17-24.

Dobrunz L.E, Pelletin D.G, McMahon TA. Muscle stiffness measured under conditions simulating natural sound production. Biophysical Journal 1990; 58:557565.

Farina D, Li X, Madeleine P. Motor unit acceleration maps and interference mechanomyographic distribution. Journal of Biomechanics 2008; 41:2843-2849.

Frangioni JV, Kwan-Gett TS, Dobrunz LE, McMahon TA. The mechanism of low-frequency sound production in muscle. Biophysical Journal 1987; 51: 775-783.

Gordon G, Holbourn AHS. The sounds from single motor units in a contracting muscle. Journal of Physiology 1948; 107:456-464.

Herroun EF, Yeo GF. Note on the sound accompanying the single contraction of skeletal muscle. Journal of Physiology 1885; 6:287-292.

Kaczmarek P, Celichowski J, Drzymala-Celichowska H, Kasiński A. The image of motor units architecture in the mechanomyographic signal during the single motor units contraction: in vivo and simulation study. Journal of Electromyography and Kinesiology (In Press).

Kaczmarek P, Celichowski J, Kasiński A.Experimentally verified model of mechanomyograms recorded during single motor unit contractions. Journal of Electromyography and Kinesiology 2005;15:617-630.

Miyamoto N, Oda S. Mechanomyographic and electromyographic responses of the triceps surae during maximal voluntary contractions. Journal of Electromyography and Kinesiology 2003; 13:451-459.

Miyamoto N, Oda S. Effect of joint angle on mechanomyographic amplitude during unfused and fused tetani in the human biceps brachii muscle. European Journal of Applied Physiology 2005; 95:221-228.

Orizio C. Muscle sound: bases for the introduction of a mechanomyographic signal in muscle studies. Critical Reviews In Biomedical Engineering 1993; 21:201-243.

Orizio C, Gobbo M, Diemont B, Esposito F, Veicsteinas A. The surface mechanomyogram as a tool to describe the influence of fatigue on biceps brachii motor unit activation strategy. Historical basis and novel evidence. European Journal of Applied Physiology 2003; 90:326-336.

Orizio C, Liberati D, Locatelli C, DeGrandis D, Veicsteinas A.Surface mechanomyogram reflects muscle fibres twitches summation. Journal of Biomechanics 1996; 29:475-481.

Oster G, Jaffe JS. Low frequency sounds from sustained contraction of human skeletal muscle. Biophysical Journal 1980;30:119-128.

Wollaston WH. On the duration of muscle action. Philosophical Transactions of the Royal Society of London 1810; 1-5.

16

Travis W. Beck

25. Yoshitake Y, Moritani T. The muscle sound properties of different muscle fiber types during voluntary and electrically induced contractions. Journal of Electromyography and Kinesiology 1999;9: 209-217.

26. Yoshitake Y, Shinohara M, Ue H, Moritani T. Characteristics of surface mechanomyogram are dependent on development of fusion of motor unit twitches in humans. Journal of Applied Physiology 2002; 93:1744-1752.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010,17-35 ISBN : 978-81-7895-449-3 Editor: Travis W. Beck

2

The mechanomyographic amplitude and frequency versus isometric force relationships Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081, USA

Abstract The most common experimental design used in mechanomyographic (MMG) studies is to examine the patterns of responses for MMG amplitude and/or center frequency [mean power frequency (MPF) or median frequency] versus isometric force.The results from the studies that have used this approach have shown that these relationships can provide information regarding the motor control strategies (i.e., the relative contributions of motor unit recruitment and firing rate modulation) that are used by various muscles to increase isometric force Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA E-mail: tbeck@ou.edu

18

Travis W. Beck

production. It is important to point out, however, that other factors, such as the type of muscle action that is being performed (i.e., isometric ramp versus step) and the joint angle (which affects the muscle’s length) can affect these patterns. Several investigations have also examined these relationships in patients that suffer from neuromuscular disorders, as well as in preadolescents and the elderly. Recent studies, however, have shown that the patterns of responses for MMG amplitude and center frequency versus isometric force differ on a subject-by-subject and muscle-by-muscle basis. This obviously reduces the validity of drawing any general conclusions from these patterns regarding motor control strategies. Future research is needed to identify the mechanisms that cause differences in the MMG amplitude and center frequency versus isometric force relationships.

Introduction By far the most common experimental design used in mechanomyographic (MMG) studies is to record MMG signals at multiple isometric force levels, and then examine the patterns of responses for MMG amplitude and/or center frequency [mean power frequency (MPF) or median frequency] versus force.

The first investigation to use this approach was performed by Lammert et al.

(1976), who recorded MMG signals from the biceps brachii and rectus femoris during submaximal to maximal isometric muscle actions of the forearm flexors and leg extensors from 10% to 100% of the isometric maximum voluntary contraction (MVC). The authors (Lammert et al. 1976) reported that for the biceps brachii muscle, MMG amplitude increased linearly with force from 10% to 60% MVC, followed by a plateau from 60% to 80% MVC, and then another increase from 80% to 100% MVC. The patterns of responses for MMG amplitude from the rectus femoris muscle, however, were more complex. In particular, two of the subjects that had previously demonstrated high percentages (69.3% and 83.1%) of slow-twitch fibers showed almost no change in MMG amplitude from 10% to 50% MVC, followed by a linear increase from 50% to 100% MVC. Two other subjects with low percentages (28.6% and 35.8%) of slow-twitch fibers, however, demonstrated no change in MMG amplitude from 10% to 30% MVC, followed by a rapid increase from 30% to 50% MVC, and then a plateau from 50% to 100% MVC. In addition, five subjects with an unknown fiber type composition showed an almost linear increase in MMG amplitude from 10% to 100% MVC. Thus, it was suggested that the shape of the MMG amplitude versus isometric force relationship was affected not only by the muscle that was being investigated, but also by differences between subjects for fiber type composition. It was also hypothesized that increases in MMG amplitude with force likely reflected

The mechanomyographic amplitude and frequency versus isometric force relationships

19

recruitment of new motor units, while a plateau or decrease in MMG amplitude with increases in force is probably due to high motor unit firing rates and the subsequent fusion of twitches. This hypothesis became a popular explanation when describing the shapes of the MMG amplitude versus isometric force relationships in future studies. Specifically, as discussed in Chapter 1, firing of motor units at high rates increases fusion of twitches and reduces the lateral oscillations of the contracting muscle fibers. This results in decreases in MMG amplitude to the point where no muscle sound is present during fully fused tetanus.

An important issue when investigating the MMG amplitude and center frequency versus isometric force relationships is the experimental protocol that is used. The most common protocol is to have the subjects perform a separate muscle action (i.e., a step contraction) for each force level. The primary advantage of the step contraction methodology is that a relatively stable (i.e., stationary) MMG signal is generated for each force level. The main disadvantage is the time commitment required to perform all contractions, particularly when many force levels are being examined. The second most common protocol involves a continuous, linear (i.e., ramp) increase in isometric force throughout a given force range. The primary advantage of the ramp protocol is that the force range of interest can be examined with a single contraction that does not require a large time commitment from the subject.The biggest disadvantages are that the resulting MMG signal is not stationary, and most subjects need a few practice trials before they are capable of generating a sufficiently linear increase in isometric force. In addition, with some muscles, it is very difficult to achieve maximal force levels (i.e., near 100% MVC) during an isometric ramp, which limits the protocol to a submaximal force range.

It is important to point out that the force range being examined is a major factor that can affect the MMG amplitude and center frequency responses. For example, Oster and Jaffe (1980) reported that MMG amplitude for the biceps brachii increased linearly with force when the forearm was flexed at 90째 and progressively heavier weights were placed in the hand. The weights used, however, only ranged from approximately 0.25-9.0 kg, which resulted in submaximal efforts for all subjects. A similar approach was used by Barry et al. (1985), who recorded MMG signals from the biceps brachii during isometric muscle actions of the forearm flexors in which the forearm was at a 90째 angle and weights of 0, 5, 10, 12.5, 15, and 20 pounds were placed in the hand. The authors (Barry et al. 1985) reported that MMG amplitude increased linearly with force between 5 and 15 pounds, with nonlinearities occurring at forces below 5 pounds and above 15 pounds. Like Oster and Jaffe (1980), however, the 0-20 pound force range was submaximal for all subjects. Thus, a

20

Travis W. Beck

very important study was conducted by Orizio et al. (1989), who examined the MMG amplitude versus force relationship for the biceps brachii during isometric step muscle actions of the forearm flexors throughout the entire force range (0-100% MVC). The authors (Orizio et al. 1989) reported that MMG amplitude increased curvilinearly with force from 0% to 80% MVC, followed by a rapid decrease from 80% to 100% MVC. It was hypothesized (Orizio et al. 1989) that the increase in MMG amplitude from 0% to 80% MVC was probably due to motor unit recruitment, while the rapid decrease from 80% to 100% MVC likely reflected either high motor unit firing rates and subsequent fusion of motor unit twitches or high levels of muscle stiffness and/or intramuscular fluid pressure.

The study by Orizio et al. (1989) for MMG amplitude was followed up by a second investigation that used the same experimental protocol, but examined the MMG MPF versus force relationship for the biceps brachii (Orizio et al.

1990). The authors (Orizio et al. 1990) reported that MMG MPF remained relatively stable from 10% to 20% MVC, then increased in a linear fashion from 20% to 80% MVC, followed by a very rapid increase from 80% to 100% MVC. In addition, at force levels below 70% MVC, the MMG power spectrum was unimodal, with a single peak that shifted toward higher frequencies with increases in force. From 70% to 100% MVC, however, the MMG power spectrum became more bipolar, with a second peak that developed at higher frequencies. It was hypothesized that the rapid increases in MMG MPF and bimodal shape of the MMG power spectrum at force levels above 70% MVC was likely due to fusion of motor unit twitches, since the biceps brachii relies at least partially on firing rate changes to increase force production at high force levels (Orizio et al. 1990) (Figure 1).

This combination of studies that examined the MMG amplitude (Orizio et al. 1989) and MPF (Orizio et al. 1990) versus force relationships provided important evidence that the amplitude and frequency contents of the MMG signal are rich with information regarding motor control strategies.

The studies that followed the work of Orizio et al. (1989, 1990) began investigating the effects of biomechanical factors and various other influences on the MMG amplitude and center frequency versus force relationships. For example, Maton et al. (1990) examined the influence of both forearm position (pronated versus supinated) and MMG sensor location (proximal versus distal) on the patterns of responses for MMG amplitude and MPF versus isometric force for the biceps brachii. The results showed that for both forearm positions and sensor locations, MMG amplitude increased curvilinearly with force from 10% to 100% MVC. The slope of the MMG amplitude versus force relationship was greater, however, when the hand was supinated than when it was pronated. In addition, sensor location had very little effect on the shape

The mechanomyographic amplitude and frequency versus isometric force relationships

21

Figure 1. Mechanomyographic (MMG) mean power frequency (MPF) (indicated as MEAN FREQ. in this figure) for the biceps brachii muscle as a function of relative isometric forearm flexion force (i.e., %MVC). Values are mean Âą SEM. The open circles represent the values from the maximum entropy spectrum estimation method, and the solid circles reflect the values from the fast Fourier transform. Notice that the increases in MMG MPF with force were particularly rapid from 80% to 100% MVC. It was hypothesized that this pattern was due to high motor unit firing rates. *Reprinted with permission from Orizio et al. (1990).

of the MMG amplitude versus force relationship. The MMG MPF versus force relationship was, however, affected by both sensor location and forearm position. When the sensor was in the proximal location and the hand was supinated, MMG MPF increased with force from 10% to 40% MVC and then plateaued from 40% to 100% MVC. In the pronated position, however, MMG MPF increased with force from 10% to approximately 80% MVC and then plateaued from 80% to 100% MVC. When the sensor was in the distal location and the hand was supinated or pronated, MMG MPF remained relatively stable from 10% MVC to approximately 60% MVC, followed by an increase from 60% to 100% MVC. Thus, it was concluded that both sensor location and forearm position are important factors that can affect the shapes of the MMG amplitude and MPF versus isometric force relationships for the biceps brachii muscle (Maton et al. 1990). Zwarts and Keidel (1991) also provided important information regarding inter-subject differences in the

22

Travis W. Beck

MMG amplitude versus isometric force relationship. Specifically, the authors reported that for the biceps brachii muscle, some subjects demonstrated highly linear increases in MMG amplitude with force, while others showed an increase in MMG amplitude up to approximately 75% MVC, followed by a decrease from 75% to 100% MVC. This study made an important contribution because it was the first to suggest that the patterns of responses for MMG amplitude versus isometric force may be unique for each subject (Zwarts and Keidel 1991).

Up to this point, most studies had focused on the MMG amplitude and/or center frequency responses from large limb muscles such as the biceps brachii or rectus femoris, which rely primarily on motor unit recruitment for increasing isometric force. If, however, the MMG amplitude versus force relationship is influenced by motor control strategies, then it should have a unique shape for small hand and/or forearm muscles, which depend more heavily on firing rate modulation for increasing isometric force. Stokes and Cooper (1992) addressed this issue by examining the MMG amplitude versus isometric force relationship for the adductor pollicis muscle. The results showed that MMG amplitude increased curvilinearly with force from 10% to 100% MVC. Although this relationship differed from those in previous studies that reported linear patterns, it was similar to the results for the biceps brachii reported by Maton et al. (1990). Thus, it was still unclear if the patterns of responses for MMG amplitude and/or center frequency could be used to describe motor control strategies. An important consideration, however, is the potential for factors such as fiber type composition and/or muscle architecture to affect the MMG amplitude versus isometric force relationship. For example, Stiles and Pham (1991) examined the patterns of responses for MMG amplitude versus isometric force for the anterior temporalis and masseter muscles, both of which function in closing the jaw. The results indicated that MMG amplitude increased linearly with jaw closing force, or increased to a maximum value and then remained constant or decreased. In addition, for some subjects, the maximum MMG amplitude value occurred at a relatively low force level (5-10% MVC). An important consideration when discussing the results from this study is that the force range examined was only 0-30% MVC. The authors did, however, report that the peak frequency of the MMG power spectrum increased 2-4 Hz with increases in jaw closing force from 060% MVC. It was hypothesized (Stiles and Pham 1991) that motor unit recruitment may have been the primary mechanism underlying the increases in MMG amplitude, while increases in firing rate may have caused the decreases at higher force levels. In addition, Stokes et al. (1988) reported that during a 10-second isometric ramp muscle action of the back extensors, MMG amplitude for the erector spinae muscles increased curvilinearly from 0% to

The mechanomyographic amplitude and frequency versus isometric force relationships

23

100% MVC, while the electromyographic (EMG) amplitude versus force relationship was highly linear. In addition, the MMG amplitude versus isometric force relationship was slightly less reliable than the EMG amplitude versus isometric force relationship when repeated testing was performed on the same day. Thus, it was suggested (Stokes et al. 1988) that EMG may be a more reliable method than MMG for estimating muscle force production, but the information provided by the MMG signal is unique from that given by EMG. A second study (Stokes and Dalton 1991) reported that MMG amplitude for the rectus femoris increased linearly with isometric leg extension force from 10% to 100% MVC (Figure 2).

This pattern was obviously different from that reported for the biceps brachii (Orizio et al. 1989) and erector spinae (Stokes et al. 1988), so it was

Figure 2. Linear relationships for mechanomyographic (MMG) amplitude (indicated as IAMG in this figure) and electromyographic (EMG) amplitude (indicated as IEMG in this figure) for the rectus femoris versus relative isometric leg extension force (%MVC). Values shown are mean Âą SEM. The closed symbols reflect the MMG data, and the open symbols represent the EMG data. Notice that both relationships are highly linear. *Reprinted with permission from Stokes and Dalton (1991).

24

Travis W. Beck

suggested (Stokes and Dalton 1991) that differences in muscle architecture and joint angle (which determines muscle length) could affect the patterns of responses for MMG amplitude versus isometric force. In addition, it was hypothesized that since the rectus femoris showed highly linear increases in MMG amplitude with isometric force, MMG could potentially be useful for estimating force production in situations where force cannot be measured directly, such as for examining subjective weakness in the quadriceps femoris muscles after knee joint surgery (Stokes and Dalton 1991). Similar results were also reported by these authors (Stokes and Dalton 1991) in a second study of the MMG amplitude versus isometric force relationship for the rectus femoris. Zhang et al. (1992) then examined the MMG amplitude and MPF versus isometric force relationships for the rectus femoris muscle at knee joint angles of 30°, 60°, and 90° (where 0° represents full leg extension). The authors found that MMG amplitude increased linearly with force at all three knee joint angles. In addition, the 60° and 90° knee joint angles showed linear increases in MMG MPF with force, but there was no change in MMG MPF with increases in force at the 30° knee joint angle. Thus, it was suggested that MMG amplitude may be useful for estimating muscle force production, and the frequency content of the MMG signal is affected by changes in joint angle (Zhang et al. 1992). This study (Zhang et al. 1992) was followed up by a second investigation (Zhang et al. 1996) that found that during submaximal isometric muscle actions of the leg extensors from 20% to 80% MVC, MMG amplitude for the rectus femoris increased curvilinearly with leg extension force. Thus, even though MMG amplitude for the rectus femoris increased linearly with isometric force in the first study (Zhang et al. 1992), a curvilinear pattern was reported in the second investigation (Zhang et al. 1996). These findings (Zhang et al. 1992, 1996) indicated that even when MMG signals are detected from the same muscle and in the same laboratory, the force-related patterns for MMG amplitude may differ for each subject.

Akataki et al. (1996) examined the potential clinical applications of the MMG amplitude versus isometric force relationship by comparing the patterns of responses for the biceps brachii from healthy subjects with those from patients that suffered from spastic cerebral palsy. Although the authors only examined submaximal force levels (i.e., 10-50% MVC), the mean MMG amplitude values for the patients with spastic cerebral palsy were less than those for the healthy subjects at all force levels (Figure 3).

In addition, the isometric forearm flexion MVC for the patients with spastic cerebral palsy was approximately one-half of the corresponding value for the healthy subjects. Thus, it was suggested that MMG may be a useful tool for studying the degradation in muscle function that occurs with spastic cerebral palsy, and, possibly, for other diseases that affect the neuromuscular

The mechanomyographic amplitude and frequency versus isometric force relationships

25

Figure 3. Mechanomyographic (MMG) amplitude (indicated as RMSAMG in this figure) for the biceps brachii as a function of relative isometric forearm flexion force (%MVC).

The open symbols represent the values for normal, healthy subjects, and the closed symbols reflect the values for patients that suffer from spastic cerebral palsy. Notice that the values for the cerebral palsy patients are lower than those for the normal subjects at all force levels. *Reprinted with permission from Akataki et al. (1996).

system (Akataki et al. 1996). Akataki et al. (1999) also conducted an interesting investigation of the factors that can cause the plateau or decrease in MMG amplitude at high isometric force levels. Specifically, the authors (Akataki et al. 1999) used a spectral decomposition procedure to remove the longitudinal muscle fiber vibrations from the MMG signal. Their hypothesis was that MMG signals from pennate muscles are affected by longitudinal and lateral muscle fiber vibrations, and the longitudinal vibrations could be estimated by placing an MMG sensor over the patella. The vibration signal from the patella would then be used in the spectral decomposition procedure to remove the longitudinal vibrations from the MMG signal. Interestingly, the

26

Travis W. Beck

authors (Akataki et al. 1999) found that when the longitudinal vibrations were not removed from the MMG signal, 4 of the 12 subjects did not show a decrease in MMG amplitude at high force levels, while the remaining 8 subjects demonstrated a decrease in MMG amplitude for force levels above 70% MVC. When the spectral decomposition procedure was used to remove the longitudinal vibrations, however, all 12 subjects showed a decrease in MMG amplitude above 70% MVC. Thus, it was suggested that when subjects do not demonstrate a plateau or decrease in MMG amplitude at high force levels, the continued increases in MMG amplitude may be due to the influence of longitudinal vibrations on the MMG signal (Akataki et al. 1999).

In addition, Orizio et al. (1994) examined the effects of hypoxia on the MMG amplitude and MPF versus isometric force relationships for the biceps brachii muscle. Specifically, the authors (Orizio et al. 1994) recorded MMG signals from the biceps brachii muscle at 20%, 40%, 60%, 80%, and 100% MVC at 150 meters above sea level, as well as after 2, 15, and 40 days of exposure to an altitude of 5,050 meters above sea level (i.e., to induce hypoxia). The results showed that exposure to the high altitude condition had no effect on the patterns of responses for MMG amplitude and MPF versus force. Specifically, MMG amplitude increased with force from 20% to 60% MVC, followed by a decrease from 60% to 100% MVC. In addition, MMG MPF increased with force from 20% to 100% MVC. Thus, it was suggested (Orizio et al. 1994) that both acute and chronic exposure to high altitude did not affect the motor control strategy (as reflected in the patterns of responses for MMG amplitude and MPF) used by the biceps brachii to increase isometric force production. In addition, Esposito et al. (1996) used the MMG amplitude and MPF versus isometric force relationships to examine the influence of aging on the motor control strategy used by the biceps brachii to increase isometric force production. Specifically, the authors (Esposito et al. 1996) investigated the MMG amplitude and MPF versus isometric force relationships for the biceps brachii in young (age range = 20-34 years) and elderly (age range = 65-78 years) subjects. The results indicated that the young and elderly subjects showed similar patterns of responses for both MMG amplitude and MPF versus force. The mean MMG amplitude and MPF values for the elderly subjects were always lower, however, than those of the young subjects. Thus, it was hypothesized (Esposito et al. 1996) that the lower mean absolute MMG amplitude and MPF values for the elderly subjects may have been due to a loss of fast-twitch muscle fibers with aging. It was also suggested, however, that a thicker skinfold layer in the elderly subjects may have filtered the MMG signals to a greater extent than those for the young subjects (Esposito et al.

1996). Ebersole et al. (1999) also conducted a very interesting study that examined the influence of changes in knee joint angle on the patterns of

The mechanomyographic amplitude and frequency versus isometric force relationships

27

responses for MMG amplitude versus isometric torque for the vastus lateralis, rectus femoris, and vastus medialis muscles. The results showed that at 25째 of leg flexion (where 0째 represents full leg extension), MMG amplitude increased with torque from 25% to 100% MVC for both the vastus medialis and rectus femoris. The vastus lateralis, however, showed no change in MMG amplitude from 25% to 100% MVC. In addition, at 50째 of leg flexion, MMG amplitude increased with isometric torque from 25% to 100% MVC for all three muscles.

At 75째 of leg flexion, however, MMG amplitude increased with isometric torque from 25% to 75% MVC, followed by a plateau from 75% to 100% MVC. Thus, it was suggested that the differences among the three knee joint angles for the patterns of responses for MMG amplitude versus isometric torque were probably due to differences in muscle stiffness, intramuscular fluid pressure, and/or the motor control strategies used to increase isometric torque (Ebersole et al. 1999). This was an important study from a practical standpoint because previous investigations had not always used a standardized joint angle. Thus, the different patterns of responses shown in previous studies could have been due at least partially to testing at different joint angles.

Nonaka et al. (2000) used the MMG amplitude versus isometric force relationship for the biceps brachii to examine potential differences in muscle function between pre-adolescent boys (age range = 9-11 years) and young men (age range = 21-23 years). The results showed that MMG amplitude increased with isometric force from 10% to 80% MVC for both the pre-adolescent boys and the young men. At 60% and 80% MVC, however, the mean MMG amplitude values for the young men were significantly greater than those for the pre-adolescent boys. Thus, it was suggested (Nonaka et al. 2000) that greater absolute force production by the young men than the pre-adolescent boys was the most likely cause for the greater mean absolute MMG amplitude values at 60% and 80% MVC. Up to this point, no studies had attempted a detailed description of the factors underlying the MMG amplitude and MPF versus isometric force relationships. Thus, a very important investigation was performed by Akataki et al. (2001) to examine the characteristics of the MMG amplitude versus force relationship for the biceps brachii during an isometric ramp muscle action of the forearm flexors. Specifically, the MMG amplitude and MPF versus force relationships were decomposed into five separate regions that were thought to reflect the dominant motor control strategy (i.e.,recruitment versus firing rate modulation) being used. For example, from approximately 60% to 80% MVC, MMG amplitude decreased, while MMG MPF increased. Thus, it was suggested that above 60% MVC, force was increased primarily by increases in firing rates, rather than motor unit recruitment. From approximately 30% to 50% MVC, however, MMG amplitude increased rapidly, while MMG MPF increased slowly. These

28

Travis W. Beck

changes were hypothesized to reflect rapid motor unit recruitment, with small increases in firing rates. Therefore, the findings from this study (Akataki et al.

2001) were important because they were the first to suggest that the patterns of responses for MMG amplitude and MPF could provide detailed information regarding motor control strategies.

Akataki et al. (2003) also published a second paper that compared the biceps brachii and first dorsal interosseous for the MMG amplitude and MPF versus force relationships. The results showed that for the first dorsal interosseous, MMG amplitude remained relatively stable from 5% to approximately 20% MVC, increased from 20% to approximately 40% MVC, and then decreased from 40% to 70% MVC. The pattern for MMG MPF, however, showed an almost linear increase from 5% to 70% MVC. In addition, MMG amplitude for the biceps brachii increased from 5% to approximately 60% MVC, and then decreased from 60% to 70% MVC. The results for MMG MPF showed an increase from 5% to roughly 50% MVC, a decrease from 50% to approximately 60% MVC, and an increase from 60% to 70% MVC. Thus, it was hypothesized that the differences between the biceps brachii and first dorsal interosseous for the patterns of responses for MMG amplitude and MPF were probably due to differences in the motor control strategies used to increase isometric force (Akataki et al. 2003). Yoshitake and Moritani (1999), however, reported that during submaximal isometric step muscle actions of the plantar flexors from 20% to 80% MVC, both MMG amplitude and MMG MPF increased linearly for the medial gastrocnemius. The soleus, however, showed an increase in MMG amplitude from 20% to 60% MVC, and then a decrease from 60% to 80% MVC, while MMG MPF increased linearly from 20% to 80% MVC. These findings were important because the medial gastrocnemius and soleus are both innervated by the tibial nerve. Thus, the differences between the two muscles for the patterns of responses for MMG amplitude versus isometric force were probably not due to differences in motor control strategies. Instead, a more likely explanation is that the differences were due to discrepancies in fiber type composition, with the soleus being dominated by slow-twitch fibers and the medial gastrocnemius containing mainly fast-twitch fibers (Yoshitake and Moritani 1999).

Orizio et al. (2003) also conducted a very interesting study that examined the MMG amplitude and MPF versus isometric force relationships for the biceps brachii muscle when it was in both a fatigued and non-fatigued state.

The experimental protocol required the subjects to first perform a 6.75-second ramp from 15-85% MVC. The forearm flexors were then fatigued by having the subjects perform intermittent isometric muscle actions of the forearm flexors at 50% MVC until the required force could no longer be maintained.

Immediately after this fatiguing protocol, the subjects performed another 6.75-

The mechanomyographic amplitude and frequency versus isometric force relationships

29

second ramp from 15-85% of their new MVC (which corresponded to 50% of their original MVC). The results showed that when the biceps brachii muscle was not fatigued, MMG amplitude increased from 15% to 65% MVC, and then decreased from 65% to 85% MVC, while MMG MPF increased from 15% to 85% MVC. In the fatigued state, however, MMG amplitude decreased from 15% to 85% MVC, while MMG MPF increased. Thus, it was concluded that MMG amplitude was the parameter most affected by fatigue, and its unique pattern of response with increases in force in the fatigued state was probably due to fatigue of fast-twitch muscle fibers. In addition, the MMG amplitude and MPF versus isometric force relationships are affected not only by motor control strategies, but also by the fatigue status of the muscle (Orizio et al.

2003).

Ebersole et al. (2002) conducted an interesting study that examined the influence of an 8-week strength training program on the MMG amplitude versus isometric force relationship for the biceps brachii muscle. The results showed that the strength training program elicited significant increases in both flexed arm circumference and isometric forearm flexion strength. There were no differences, however, between the patterns of responses for MMG amplitude versus force for the biceps brachii from the pre-training versus posttraining tests. Thus, it was suggested (Ebersole et al. 2002) that the strength gains from the 8-week training program could have been due primarily to hypertrophic, rather than neural factors. In addition, increases in muscle stiffness with training may have counteracted the effects of hypertrophy on MMG amplitude, resulting in no changes in its patterns of response with force (Ebersole et al. 2002). Akataki et al. (2004) simultaneously examined the MMG and EMG amplitude and MPF versus force relationships for the biceps brachii to determine if MMG provided more information regarding motor control strategies than EMG. The results indicated that EMG amplitude increased curvilinearly with force from 5% to 80% MVC, whereas EMG MPF increased from 5% to 50% MVC and then plateaued from 50% to 80% MVC. In contrast, MMG amplitude increased curvilinearly with force from 5% to 60% MVC and then decreased from 60% to 80% MVC, while MMG MPF increased from 5% to 50% MVC, decreased from 50% to 60% MVC, and increased from 60% to 80% MVC. Thus, the authors (Akataki et al. 2004) hypothesized that the EMG MPF versus force relationship may be useful for determining when all motor units have been recruited, while the MMG amplitude versus force relationship showed when all the motor units had been recruited, as well as the beginning of fast-twitch motor unit recruitment.

However, these hypotheses still needed to be tested directly.

Our laboratory has also examined some of the inter-individual factors that can affect the MMG amplitude and MPF versus torque or force relationships.

30

Travis W. Beck

For example, Beck et al. (2008) investigated the patterns of responses for MMG amplitude versus isometric force for the vastus lateralis from resistancetrained (mean fast-twitch fiber type percentage = 59.1%), aerobically-trained (mean fast-twitch fiber type percentage = 27.4%), and sedentary subjects (mean fast-twitch fiber type percentage = 59.9%). At the beginning of the study, it was hypothesized that the subjects with a greater percentage of slowtwitch fibers may be more likely to demonstrate a plateau or decrease in MMG amplitude at high force levels than those with mostly fast-twitch fibers. The results indicated, however, that there was a large amount of variability between subjects for the patterns of responses for MMG amplitude versus force, and this variability was not related to the fiber type composition of the vastus lateralis (Beck et al. 2008). In addition, Beck et al. (2005) examined the influence of differences in gender on the MMG amplitude and MPF versus isometric torque relationships for the biceps brachii muscle. The authors (Beck et al. 2005) found that for young, college-aged men, MMG amplitude increased from 10% to 80% MVC and then plateaued from 80% to 100% MVC. The young college-aged women, however, showed a linear increase in MMG amplitude with force from 10% to 100% MVC. In addition, both the young men and young women showed linear increases in MMG MPF from 10% to 100% MVC. Thus, it was hypothesized that the differences between the men and women for the patterns of responses for MMG amplitude versus force were likely due to differences in absolute strength levels (since the mean forearm flexion strength value for the women was 43% that of the men), rather than differences in motor control strategies (Beck et al. 2005). Another study from our laboratory compared a piezoelectric crystal contact sensor with an accelerometer for the patterns of responses for MMG amplitude and MPF versus isometric torque for the biceps brachii (Beck et al. 2006). The results indicated that MMG amplitude increased linearly with isometric torque for both the piezoelectric crystal contact sensor and accelerometer. The patterns for MMG MPF, however, showed a linear decrease from 20% to 100% MVC for the accelerometer, but no change for the piezoelectric crystal contact sensor. Thus, it was concluded that the interpretation of the patterns of responses with regard to describing motor control strategies may be affected by the type of MMG sensor that is used to detect the signal (Beck et al. 2006).

Several recent studies have also compared isometric ramp versus step muscle actions for the MMG amplitude and MPF versus torque relationships.

For example, Ryan et al. (2008) found that during isometric ramp and step muscle actions of the leg extensors, there was a substantial degree of interindividual variability in the patterns of responses for MMG amplitude and MPF versus torque. These relationships were also different for each of the muscles that were investigated (the vastus lateralis and rectus femoris), as well

The mechanomyographic amplitude and frequency versus isometric force relationships

31

as for the step versus ramp muscle actions. Thus, it was concluded (Ryan et al.

2008) that the patterns for MMG amplitude and MPF versus torque should be interpreted with the understanding that they are mode (ramp versus step) and muscle-specific. In addition, the large degree of inter-individual variability suggested that these relationships may best be examined on a subject-bysubject basis (Ryan et al. 2008). Similar results were also reported in a previous study (Ryan et al. 2007) that compared the composite (i.e., mean) MMG amplitude and MPF versus isometric torque relationships for the vastus lateralis with the patterns of responses for individual subjects that were categorized into either a high strength or a low strength group. The results showed that for the composite patterns of responses, the MMG amplitude versus isometric torque relationship for the high strength group was best fit with a cubic model, while the corresponding relationship for the low strength group was best fit with a linear model. In addition, MMG MPF increased linearly with torque for both the high and low strength groups. Furthermore, only 66% and 33% of the subjects showed the same MMG amplitude versus force relationships as the composite patterns for the low strength and high strength groups, respectively. Only one of the twelve subjects showed the same linear increase in MMG MPF with isometric force that was seen for the composite patterns.Thus, it was concluded that differences in strength did not affect the patterns of responses for MMG amplitude or MPF versus isometric force. Instead, these relationships were different for each individual, which suggested that perhaps they should be investigated on a subject-by-subject basis (Ryan et al. 2007). An important question, however, is whether or not the patterns of responses for MMG amplitude and MPF demonstrate sufficient between-day reliability to be used for examining the effects of various interventions. Herda et al. (2008) recently examined this issue for the vastus lateralis during both isometric step and ramp muscle actions of the leg extensors. The results showed between-day reliability coefficients (intraclass correlation coefficients) ranging from 0.39-0.89 for MMG amplitude and 0.360.80 for MMG MPF. Thus, it was concluded that the reliability of the MMG signal was comparable to that of the EMG signal. It was also suggested, however, that future studies need to be done to determine if the patterns of responses (as opposed to the absolute values) for MMG amplitude and MPF demonstrate acceptable reliability (Herda et al. 2008).

Hwang (2007) investigated changes in the frequency contents of the MMG and EMG signals for the tibialis anterior during isometric load-varying contractions. Specifically, the subjects were required to perform two different tasks that included: (a) static isometric muscle actions of the dorsiflexors at four different force levels, and (b) load-varying isometric dorsiflexion where the subject was required to track a target sinusoidal force curve with three

32

Travis W. Beck

different amplitudes. The results showed that during both the static and loadvarying isometric muscle actions, the MPF of the MMG-EMG cross spectrum increased progressively with force, but the median frequency of the EMG signal remained constant. In addition, there was a positive, linear relationship between the MMG-EMG cross spectrum MPF and EMG amplitude, but the slope coefficient was significantly greater for the load-varying isometric muscle actions than for the static muscle actions. Thus, it was concluded that changes in motor unit firing rates for the tibialis anterior with increases in isometric force are dependent on the type of isometric muscle action that is being performed (i.e., static versus load-varying) (Hwang 2007). Ohta et al. (in press) investigated the potential relationship between changes in fascicle length and MMG amplitude during voluntary isometric muscle actions of the plantar flexors with a superimposed twitch. Specifically, the subjects were required to perform isometric muscle actions of the plantar flexors at 20%, 40%, 60%, 80%, and 100% MVC, and surface MMG signals were detected from the medial gastrocnemius muscle. During each muscle action, the plantar flexors were stimulated with a supramaximal stimulus that was sent through the posterior tibial nerve, and changes in fascicle length were measured with ultrasound.The results showed that MMG amplitude during the supramaximal stimulus and the change in fascicle length decreased curvilinearly from 20% to 80% MVC, but the superimposed twitch amplitude decreased linearly. In addition, there was a linear relationship between MMG amplitude and the change in fascicle length. Thus, it was concluded that MMG amplitude reflected changes in fascicle length better than the superimposed twitch amplitude. In addition, MMG amplitude may be useful for examining changes in muscle architecture (Ohta et al. in press).

Coburn et al. (2008) recently examined the MMG amplitude versus isometric torque relationships for the vastus lateralis, rectus femoris, and vastus medialis muscles. Specifically, the subjects performed submaximal to maximal isometric step muscle actions of the leg extensors, and the MMG amplitude responses were measured for the vastus lateralis, rectus femoris, and vastus medialis. The results showed that the composite (i.e., averaged across subjects) patterns for MMG amplitude were best fit with quadratic models for the vastus lateralis and vastus medialis, and a linear model for the rectus femoris. The individual responses, however, were very inconsistent, with some subjects showing linear patterns, others demonstrating a quadratic relationship, others showing cubic patterns, and for some subjects, there was no significant relationship between MMG amplitude and torque. Thus, it was concluded that when examining the patterns of responses for MMG amplitude versus isometric torque, the individual relationships should be considered, since they often differ from the composite patterns. In addition, caution should be used

The mechanomyographic amplitude and frequency versus isometric force relationships

33

when interpreting the patterns of responses for just one muscle, since these patterns often vary on a muscle-by-muscle basis (Coburn et al. 2008).

In summary, examination of the MMG amplitude and center frequency responses with increases in isometric force has played an important role in identifying the mechanisms that generate the signal and the information that it can provide. The results from the studies that have examined these responses clearly indicate that these relationships are rich with information. Future research in this area should examine the influence of factors such as changes in muscle stiffness, architecture, and intramuscular fluid pressure on these patterns. Although a great deal of research has been performed to identify the information that these patterns provide, there is still a lot that needs to be done to understand the factors that affect them.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

Akataki K, Mita K, Itoh Y. Relationship between mechanomyogram and force during voluntary contractions reinvestigated using spectral decomposition.

European Journal of Applied Physiology 1999; 80:173-179.

Akataki K, Mita K, Itoh K, Suzuki N, Watakabe M. Acoustic and electrical activities during voluntary isometric contraction of biceps brachii muscles in patients with spastic cerebral palsy. Muscle & Nerve 1996; 19:1252-1257.

Akataki K, Mita K, Watakabe M. Electromyographic and mechanomyographic estimation of motor unit activation strategy in voluntary force production.

Electromyography and Clinical Neurophysiology 2004; 44:489-496.

Akataki K, Mita K,Watakabe M, Itoh K.Mechanomyogram and force relationship during voluntary isometric ramp contractions of the biceps brachii muscle.

European Journal of Applied Physiology 2001; 84:19-25.

Akataki K, Mita K, Watakabe M, Itoh K. Mechanomyographic responses during voluntary ramp contractions of the human first dorsal interosseous muscle.

European Journal of Applied Physiology 2003; 89:520-525.

Barry DT, Geiringer SR, Ball RD. Acoustic myography: A noninvasive monitor of motor unit fatigue. Muscle & Nerve 1985; 8:189-194.

Beck TW, Housh TJ, Fry AC, Cramer JT, Weir JP, Schilling BK, Falvo MJ, Moore CA. The influence of myosin heavy chain isoform composition and training status on the patterns of responses for mechanomyographic amplitude versus isometric torque. Journal of Strength and Conditioning Research 2008; 22:818-825.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Gender comparisons of mechanomyographic amplitude and mean power frequency versus isometric torque relationships. Journal of Applied Biomechanics 2005; 21:96-109.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Comparison of a piezoelectric crystal contact sensor and an accelerometer for examining mechanomyographic amplitude and mean power frequency versus torque relationships during isokinetic and isometric muscle actions of the biceps brachii. Journal of Electromyography and Kinesiology 2006; 16:324-335.

34

Travis W. Beck

10. Coburn JW, Malek MH, Brown LE, Zinder SM. Mechanomyographic responses of the superficial quadriceps femoris muscles to incremental isometric muscle actions. Electromyography and Clinical Neurophysiology 2008; 48:97-102.

11. Ebersole KT, Housh TJ, Johnson GO, Evetovich TK, Smith DB, Perry SR. MMG and EMG responses of the superficial quadriceps femoris muscles. Journal of Electromyography and Kinesiology 1999; 9:219-227.

12. Ebersole KT, Housh TJ, Johnson GO, Perry SR, Bull AJ, Cramer JT. Mechanomyographic and electromyographic responses to unilateral isometric training. Journal of Strength and Conditioning Research 2002; 16:192-201.

13. Esposito F, Malgrati D, Veicsteinas A, Orizio C. Time and frequency domain analysis of electromyogram and sound myogram in the elderly. European Journal of Applied Physiology 1996; 73:503-510.

14. Herda TJ, Ryan ED, Beck TW, Costa PB, DeFreitas JM, Stout JR, Cramer JT. Reliability of mechanomyographic amplitude and mean power frequency during isometric step and ramp muscle actions. Journal of Neuroscience Methods 2008; 171:104-109.

15. Hwang IS. Physiological aspects of MMG and EMG spectra during load-varying isometric dorsiflexion. Electromyography and Clinical Neurophysiology 2007; 47:79-87.

16. Lammert O, Jorgensen F, Einer-Jensen N. Accelerometermyography (AMG) I: method for measuring mechanical vibrations from isometrically contracted muscles. In Biomechanics V-A. Komi, PV, Ed., University Park Press, Baltimore, 1976. p. 152-158.

17. Maton B, Petitjean M, Cnockaert JC. Phonomyogram and electromyogram relationships with isometric force reinvestigated in man. European Journal of Applied Physiology 1990; 60:194-201.

18. Nonaka H, Mita K, Akataki K, Watakabe M, Yabe K. Mechanomyographic investigation of muscle contractile properties in preadolescent boys.

Electromyography and Clinical Neurophysiology 2000; 40:287-293.

19. Ohta Y, Shima N, Yabe K. In vivo behaviour of human muscle architecture and mechanomyographic response using the interpolated twitch technique. Journal of Electromyography and Kinesiology (In Press).

20. Orizio C, Esposito F, Veicsteinas A. Effect of acclimatization to high altitude (5,050 m) on motor unit activation pattern and muscle performance. Journal of Applied Physiology 1994; 77:2840-2844.

21. Orizio C, Gobbo M, Diemont B, Esposito F, Veicsteinas A. The surface mechanomyogram as a tool to describe the influence of fatigue on biceps brachii motor unit activation strategy. Historical basis and novel evidence. European Journal of Applied Physiology 2003; 90:326-336.

22. Orizio C, Perini R, Diemont B, Figini MM, Veicsteinas A. Spectral analysis of muscular sound during isometric contraction of biceps brachii. Journal of Applied Physiology 1990; 68:508-512.

23. Orizio C, Perini R, Veicsteinas A. Muscular sound and force relationship during isometric contraction in men. European Journal of Applied Physiology 1989; 58:528-533.

The mechanomyographic amplitude and frequency versus isometric force relationships

35

24. Oster G, Jaffe JS. Low frequency sounds from sustained contraction of human skeletal muscle. Biophysical Journal 1980;30:119-128.

25. Ryan ED, Beck TW, Herda TJ, Hartman MJ, Stout JR, Housh TJ, Cramer JT. Mechanomyographic amplitude and mean power frequency responses during isometric ramp vs. step muscle actions. Journal of Neuroscience Methods 2008; 168:293-305.

26. Ryan ED, Cramer JT, Housh TJ, Beck TW, Herda TJ, Hartman MJ. Interindividual variability in the torque-related patterns of responses for mechanomyographic amplitude and mean power frequency. Journal of Neuroscience Methods 2007; 161:212-219.

27. Stiles R, Pham D. Acoustic- and surface electro-myography of human jaw elevator muscles. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society 1991; 13:0946-0947.

28. Stokes MJ, Cooper RG. Muscle sounds during voluntary and stimulated contractions of the human adductor pollicis muscle. Journal of Applied Physiology 1992; 72:1908-1913.

29. Stokes MJ, Dalton PA. Acoustic myographic activity increases linearly up to maximal voluntary isometric force in the human quadriceps muscle. Journal of the Neurological Sciences 1991; 101:163-167.

30. Stokes MJ, Dalton PA. Acoustic myography for investigating human skeletal muscle fatigue. Journal of Applied Physiology 1991; 71:1422-1426.

31. Stokes IAF, Moffroid MS, Rush S, Haugh LD, Comparison of acoustic and electrical signals from erectors spinae muscles. Muscle & Nerve 1988;11:331-336.

32. Yoshitake Y, Moritani T. The muscle sound properties of different muscle fiber types during voluntary and electrically induced contractions. Journal of Electromyography and Kinesiology 1999; 9:209-217.

33. Zhang Y-T, Frank CB, Rangayyan RM, Bell GD. A comparative study of simultaneous vibromyography and electromyography with active human quadriceps. IEEE Transactions On Biomedical Engineering 1992;39:1045-1052.

34. Zhang Y-T, Frank CB, Rangayyan RM, Bell GD. Relationships of the vibromyogram to the surface electromyogram of the human rectus femoris muscle during voluntary isometric contraction. Journal of Rehabilitation Research and Development 1996; 33:395-403.

35. Zwarts MJ, Keidel M. Relationship between electrical and vibratory output of muscle during voluntary contraction and fatigue. Muscle & Nerve 1991; 14: 756-761.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010, 37-51 ISBN : 978-81-7895-449-3 Editor: Travis W. Beck

3

Surface mechanomyographic responses to muscle fatigue Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081, USA

Abstract Many studies have examined the mechanomyographic (MMG) amplitude and/or center frequency [mean power frequency (MPF) or median frequency] responses during fatiguing isometric muscle actions.

An important common finding from these investigations is that the MMG responses are always dependent on the relative force level that is being examined, as well as the time duration of the muscle action that is being performed. There is also evidence to suggest that the muscle being examined and its corresponding fiber type composition can affect the MMG responses to fatigue. Simultaneous Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA E-mail: tbeck@ou.edu

38

Travis W. Beck

detection of MMG and surface electromyographic (EMG) signals has been used to examine the dissociation between the electrical and mechanical aspects of fatigue, as well as the motor control strategies used by various muscles to maintain isometric force production. There is still a great deal of work that needs to be done, however, to determine the mechanisms underlying the MMG amplitude and center frequency responses during various types of fatiguing isometric muscle actions. Future research should focus on the possible clinical applications of these patterns, and if they can be used in orthopaedic and/or rehabilitative settings.

Introduction A great deal of research has been performed to examine the behavior of the mechanomyographic (MMG) signal during fatiguing isometric muscle actions. The first study to directly investigate the MMG amplitude responses during muscle fatigue was performed by Barry et al. (1985). Specifically, the experimental protocol required the subjects to perform a sustained isometric muscle action of the forearm flexors at 75% of the maximum voluntary contraction (MVC). Once the 75% MVC force level could no longer be maintained, the subjects were required to continue the muscle action until their force production had decreased to 35% of the original MVC. The surface MMG signal was detected from the biceps brachii throughout the entire duration of the sustained muscle action. The results showed that MMG amplitude was highly correlated with force production (i.e., MMG amplitude remained stable at the beginning of the muscle action, and then decreased after the 75% MVC force level could not be maintained). In addition, even small increases in force production during the sustained muscle action produced large bursts of activity in the MMG signal. Thus, it was suggested that MMG amplitude may be a more sensitive indicator of changes in force when fatigue develops rapidly than electromyographic (EMG) amplitude, which does not always follow force production during fatiguing activities (Barry et al. 1985).

The investigation by Barry et al. (1985) was followed up with a very important study by Orizio et al. (1989), which examined the changes in MMG amplitude for the biceps brachii during sustained isometric muscle actions of the forearm flexors at 20%, 40%, 60%, and 80% MVC. Each muscle action was performed to exhaustion, and a surface EMG signal was detected from the biceps brachii simultaneously with the MMG signal. The results indicated that during the sustained muscle action at 20% MVC, MMG amplitude increased linearly across time, and it was hypothesized that this response was due to motor unit recruitment, increases in firing rates, and synchronization of firing times as the motor units that were recruited at the beginning of the muscle

Surface mechanomyographic responses to muscle fatigue

39

action became fatigued. At 40% MVC, however, MMG amplitude fluctuated around a steady value from the beginning of the muscle action up to exhaustion. It was suggested that at this force level, the information contained in the MMG signal regarding motor control strategies may be masked by changes in the muscle environment due to loss of perfusion. During the sustained muscle actions at 60% and 80% MVC, however, MMG amplitude decreased curvilinearly across time, and it was suggested that at these force levels, lengthening of the muscle fiber relaxation time could have resulted in fusion of motor unit twitches and a subsequent decrease in MMG amplitude (Figure 1).

It was also hypothesized, however, that high levels of muscle stiffness and intramuscular fluid pressure could have affected MMG amplitude at these force levels. In contrast, the results for EMG amplitude showed increases across time at all force levels. Thus, it was concluded that the MMG signal may provide more information than EMG regarding the motor control strategies used during fatiguing isometric muscle actions (Orizio et al. 1989).

This study was followed up by a second investigation (Orizio et al. 1992) that used the same experimental protocol, but examined the MMG and EMG mean power frequency (MPF) responses for the biceps brachii. The results showed that

Figure 1. Mechanomyographic (MMG) signals from the biceps brachii of one subject during sustained isometric muscle actions of the forearm flexors to exhaustion at 20%, 40%, 60%, and 80% MVC. The appropriate time and amplitude scales are shown for each signal. Notice that at 20% MVC, MMG amplitude increased over time, but it remained relatively stable at 40% MVC. In addition, MMG amplitude decreased over time during the sustained muscle actions at 60% and 80% MVC. *Reprinted with permission from Orizio et al. (1989).

40

Travis W. Beck

during the sustained muscle action at 80% MVC, MMG MPF increased during the first 30% of the muscle action and then decreased for the remainder. It was hypothesized that this response was due to initial increases in motor unit firing rates, followed by decreases due to increases in muscle fiber relaxation time that provided proprioceptive feedback to the central nervous system (i.e., muscle wisdom). During the sustained muscle action at 20% MVC, however, MMG MPF increased slightly during the first 30% of the muscle action and then remained relatively stable for the remainder. It was suggested that this response likely reflected relatively stable motor unit firing rates throughout the contraction. During the sustained muscle actions at 40% and 60% MVC, however, the MMG MPF responses were about halfway between those at 20% and 80% MVC. Specifically, at 40% MVC, MMG MPF increased slightly during the first 10% of the muscle action and then decreased slowly. At 60% MVC, however, MMG MPF remained relatively stable during the first 40% of the muscle action, and then decreased slowly. Thus, it was hypothesized that at these force levels, the MMG MPF responses reflected recruitment of new motor units with high firing rates, followed by no change in the average motor unit firing rate throughout the remainder of the muscle action. In addition, during all four sustained muscle actions, EMG MPF decreased throughout the duration of the contraction. Thus, the authors (Orizio et al. 1992) suggested that when combined with EMG, MMG may be useful for investigating the neural and peripheral mechanisms underlying muscle fatigue. These studies (Orizio et al. 1989, 1992) were followed up by Goldenberg et al. (1991), who examined the changes in MMG amplitude for the abductor digiti minimi during isometric muscle actions sustained to exhaustion at 15%, 25%, 50%, and 75% MVC. The results showed that during the sustained muscle action at 75% MVC, MMG amplitude remained relatively stable, but at 50% MVC, MMG amplitude decreased throughout the contraction. In addition, during the sustained muscle actions at 15% and 25% MVC, MMG amplitude increased over time. Thus, it was suggested that the increases in MMG amplitude during the sustained muscle actions at 15% and 25% MVC may have reflected recruitment of new motor units to maintain the required force level. In turn, the decrease in MMG amplitude during the sustained muscle action at 50% MVC may have been due to fatigue of fast-twitch motor units and increases in firing rates that resulted in fusion of motor unit twitches. In addition, the lack of a significant change in MMG amplitude during the sustained muscle action at 75% MVC could have been due to the relatively short duration of the muscle action and the inability of the slow-twitch fibers to maintain the required force as the larger fast-twitch fibers became fatigued. Thus, it was suggested (Goldenberg et al. 1992) that during a fatiguing isometric muscle action, MMG amplitude may be much more dependent on fiber type composition and motor

Surface mechanomyographic responses to muscle fatigue

41

control strategies than on absolute force production. Mealing et al. (1990) also conducted a very interesting study that examined changes in the shape of the MMG power spectrum for the rectus femoris during a sustained isometric muscle action of the leg extensors at 80% MVC. The authors (Mealing et al.

1990) reported that during the fatiguing muscle action, the MMG power spectrum tended to cycle between wide and narrow bandwidths, and it was hypothesized that this cycling may have been due to rotation of activity between different fiber types. Stokes and Dalton (1991) also used an interesting approach to examine the effects of muscle fatigue on the MMG signal. Specifically, the subjects in their study were required to perform sustained isometric muscle actions of the leg extensors at 10%, 25%, 50%, 60%, 75%, and 100% MVC while MMG and EMG signals were detected from the rectus femoris. When all muscle actions had been completed, the subjects performed a fatiguing protocol of repeated voluntary contractions (10-s on, 10s off) at 75% MVC until only 40% of the original MVC could be produced.

The subjects were then allowed to rest for 15 minutes, and the separate muscle actions at 10%, 25%, 50%, 60%, 75%, and 100% MVC were performed again.

The results showed that before and after the fatiguing protocol, the MMG and EMG amplitude versus force relationships were linear. After the fatiguing protocol, however, the linear slope of the EMG amplitude versus force relationship increased, while that for the MMG amplitude versus force relationship remained the same. Thus, it was hypothesized that MMG may be a more useful method for assessing force production in the fatigued state than EMG (Stokes and Dalton 1991).

Like Mealing et al. (1990), Herzog et al. (1994) examined the influence of fatigue on the frequency content of the MMG signal. Specifically, EMG and MMG signals were detected simultaneously from the rectus femoris and vastus lateralis during an isometric muscle action of the leg extensors sustained to exhaustion at 70% MVC. The results indicated that for both the vastus lateralis and rectus femoris, EMG and MMG median frequency decreased during the sustained muscle action. The authors reported, however, that at least part of the decrease in MMG median frequency was due to muscle tremor.

Nevertheless, it was suggested that muscle tremor was an important part of the MMG signal during fatigue, and it should not be removed because it originates from the muscle(s) of interest and is a result of the fatigue protocol (Herzog et al. 1994). Vaz et al. (1996) used a similar experimental protocol to examine the influence of fatigue on MMG amplitude and median frequency responses. The experimental protocol required the subjects to perform submaximal isometric muscle actions of the leg extensors at 70% MVC prior to, and 20 seconds, 50 seconds, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 15 minutes after a fatigue test. The fatigue test was an isometric muscle action of the leg extensors at 70% MVC

42

Travis W. Beck

that was sustained until 50% of the original MVC could no longer be achieved.

Surface EMG and MMG signals were detected simultaneously from the vastus lateralis and rectus femoris during the muscle actions performed before and after the fatigue test. The results showed that there were differences between the vastus lateralis and rectus femoris for the MMG amplitude responses to the fatigue protocol. Specifically, the rectus femoris showed a decrease in MMG amplitude during the fatigue protocol, but returned to baseline values fairly quickly during the recovery period. For the vastus lateralis, however, MMG amplitude increased significantly when the target force could no longer be maintained, and then it remained constant during the recovery period. The authors (Vaz et al. 1996) suggested that muscle tremor had an important influence on the MMG signal during fatiguing activities, and changes in the signal were independent of those in the EMG signal. Esposito et al. (1998) also examined the MMG and EMG time and frequency domain responses during fatiguing isometric muscle actions. Specifically, the subjects were required to perform a submaximal isometric muscle action of the forearm flexors at 80% MVC until the required force level could no longer be maintained. The subjects then rested for 10 minutes and performed repeated (6-s on, 4-s off) isometric muscle actions of the forearm flexors at 50% MVC. The subjects then rested for 9-min, followed by performance of an MVC. Immediately after the MVC, the subjects performed a second sustained isometric muscle action to exhaustion at 80% MVC. The results showed that during both sustained muscle actions at 80% MVC, EMG amplitude increased over time for the first 10 seconds and then remained relatively stable, while EMG MPF decreased continuously over time. The findings for MMG amplitude, however, showed that during the first fatiguing muscle action, MMG amplitude decreased during the first 7 seconds and then remained stable. For the second sustained muscle action, however, MMG amplitude remained relatively stable over time. In addition, the results for MMG MPF showed that during the first fatiguing muscle action, MMG MPF increased over time for the first 5 seconds and then decreased over time. During the second fatiguing muscle action, however, MMG MPF decreased throughout the entire muscle action. Thus, it was suggested that the effects of fatigue on the muscle may be more accurately reflected in the MMG signal than the EMG signal (Esposito et al. 1998).

Kouzaki et al. (1999) used a slightly different experimental design to examine the effects of muscle fatigue on the amplitude and frequency contents of the MMG and EMG signals. Specifically, the subjects were required to perform 50 consecutive maximal isometric muscle actions of the leg extensors (3-s on, 3-s off), and MMG and EMG signals were detected from the vastus lateralis, rectus femoris, and vastus medialis. The results showed that the average isometric leg extension MVC decreased 49.5% from the beginning to

Surface mechanomyographic responses to muscle fatigue

43

the end of the fatigue test. In addition, both EMG amplitude and EMG median frequency decreased curvilinearly at a similar rate for each muscle during the fatigue test. The results for MMG amplitude also showed curvilinear decreases over time for each muscle, although the rate at which it decreased was much greater for the rectus femoris than the vastus lateralis and vastus medialis. In addition, MMG median frequency decreased curvilinearly during the fatigue test for all three muscles. The initial value and the rate at which MMG median frequency decreased were much greater, however, for the rectus femoris than the vastus lateralis and vastus medialis. Thus, these findings were important from a practical standpoint because they indicated that both MMG amplitude and MMG median frequency were sensitive to differences in the fatigue characteristics of muscles that have a common innervation (i.e., the femoral nerve) but differ in architecture and, possibly, fiber type composition (Kouzaki et al. 1999).

Yoshitake et al. (2001) used a slightly different approach to examine the effects of fatigue on the MMG signal. Specifically, the authors recorded EMG and MMG signals simultaneously during a 60-second sustained isometric muscle action of the erector spinae as the subject supported their own body weight. Near-infrared spectroscopy was also used to assess muscle blood volume and tissue oxygenation levels. The results showed that EMG MPF decreased linearly throughout the fatigue test, while EMG amplitude increased during the first 36 seconds and then plateaued. In contrast, there was no change in MMG MPF during the test, and MMG amplitude increased during the first 20 seconds, and then decreased thereafter. Thus, it was hypothesized (Yoshitake et al. 2001) that restriction of blood flow due to high intramuscular mechanical pressure may be one of the most important factors underlying lumbar muscle fatigue and the subsequent lower back pain. In addition, MMG may be a useful method for examining the neural and mechanical aspects of muscle fatigue (Yoshitake et al. 2001).

Another very interesting study was performed by Søgaard et al. (2003) to determine if the MMG and EMG signals were sensitive to changes in muscle function induced by long term fatigue. In particular, the experimental protocol required the subjects to perform an isometric MVC of the forearm flexors, followed by separate muscle actions at 5% and 80% MVC. The subjects then performed intermittent isometric muscle actions at 30% MVC (6-s on, 4-s off) for a time period of 30 minutes to induce fatigue. Following the fatigue protocol, the subjects performed isometric MVCs to measure strength 10 minutes and 30 minutes after the fatiguing protocol. The subjects also performed submaximal isometric muscle actions at 80% of the new MVC and 5% of the pre-fatigue MVC. The results showed that even after 30 minutes of recovery, the isometric MVC was reduced by as much as 16%, and the MMG

44

Travis W. Beck

and EMG amplitude values were greater during the post-fatigue 5% MVC contraction than at the same force level before the fatiguing protocol. Thus, it was hypothesized (Søgaard et al. 2003) that MMG and EMG may be sensitive to subtle changes that occur in the muscle during long term fatigue. Another very important study was recently performed by Madeleine et al. (2006), who examined changes in the amplitude of the MMG signal, as well as the MPF, variance, and skewness of the MMG power spectrum for the biceps brachii during a 3-minute sustained isometric muscle action of the forearm flexors at 30% MVC. The results showed that MMG amplitude increased during the first 135 seconds of the muscle action and then plateaued. In contrast, MMG MPF and the skewness of the power spectrum decreased across time. The variance of the power spectrum, however, showed a complex behavior, where it decreased during the first 90 seconds of the muscle action, increased for the next 60 seconds, and then decreased for the remainder of the contraction. Thus, it was suggested that the fatigue-induced changes in the shape of the MMG power spectrum cannot be described exclusively by measures of center frequency (e.g., MPF or median frequency).Instead, the spectrum shows complex changes in bandwidth (i.e., variance) and skewness that are important when describing the effects of fatigue on the muscle being investigated. It is important to point out that these results were from an accelerometer, and a condenser microphone provided different patterns for all of the variables examined. Nevertheless, the findings from this study clearly indicated that the effects of fatigue on the MMG signal cannot be fully described by examining a single amplitude and/or center frequency parameter (Madeleine et al. 2006).

Previous studies have also used joint time-frequency signal processing techniques to examine the effects of muscle fatigue on the frequency content of the MMG signal. Itoh et al. (2004) used the short-time Fourier transform to investigate the changes in MMG MPF for the biceps brachii during sustained isometric muscle actions of the forearm flexors at 20% and 80% MVC. During both muscle actions, the subjects were required to continue contracting until their force production had dropped to 50% of the target force. The results showed that during the sustained muscle action at 20% MVC, MMG amplitude increased curvilinearly over time, while MMG MPF increased for the first 30% of the total contraction time, remained stable from 30% to 70% of the total contraction time, and decreased throughout the remainder of the muscle action.

During the sustained muscle action at 80% MVC, however, MMG amplitude decreased curvilinearly over time, and MMG MPF increased for the first 30% of the total contraction time and then decreased thereafter (Figure 2). Thus, it was suggested that both the amplitude and frequency contents of the MMG signal are useful when examining the neural and mechanical aspects of muscle fatigue (Itoh et al. 2004).

Surface mechanomyographic responses to muscle fatigue

45

Figure 2. Mechanomyographic (MMG) amplitude and mean power frequency (MPF) for the biceps brachii during submaximal isometric muscle actions of the forearm flexors that were sustained to exhaustion at 20% and 80% MVC. The relative force signals are shown in the bottom graphs. Notice the differences in the patterns of responses between the 20% and 80% MVC muscle actions for both MMG amplitude and MPF. *Reprinted with permission from Itoh et al. (2004).

46

Travis W. Beck

Weir et al. (2000) examined the influence of differences in muscle length on the MMG and EMG amplitude and MPF responses for the tibialis anterior during a 60-s sustained isometric muscle action of the dorsiflexors at 50% MVC. The results showed that at the long muscle length (when the foot was in 40째 of plantar flexion), the rate at which MMG and EMG amplitude increased over time was greater than at the short muscle length (when the foot was in 5째 of dorsiflexion). There were no differences, however, between the long and short muscle lengths for the changes in EMG and MMG MPF over time. Thus, it was hypothesized that the amplitudes of the MMG and EMG signals may be useful for examining the rate of motor unit recruitment during a fatiguing task at a submaximal force level. In addition, the rate of motor unit recruitment may be greater during a fatiguing muscle action at a long muscle length when compared to a short muscle length (Weir et al. 2000). Madeleine et al. (2002) also used an interesting experimental design to examine the effects of muscle fatigue on the MMG signal. Specifically, the subjects were required to perform sustained isometric muscle actions of the forearm flexors at 10% and 30% MVC in continuous or intermittent static format and with either visual or proprioceptive feedback.The results indicated that when the subjects were provided with proprioceptive feedback, the rate at which MMG and EMG amplitude increased over time was greater than when the subjects were given visual feedback. Thus, it was suggested that the combined use of MMG and EMG was helpful for identifying differences between the fatigue characteristics of visual versus proprioceptive feedback modes (Madeleine et al. 2002). In addition, Tarata (2003) used simultaneous recording of MMG and EMG signals to examine muscle fatigue for the biceps brachii and brachioradialis during an isometric muscle action of the forearm flexors sustained to exhaustion at 25% MVC. The results indicated that both MMG and EMG amplitude increased throughout the duration of the muscle action, while MMG and EMG MPF decreased (Figure 3). Thus, it was concluded that simultaneous examination of the MMG and EMG amplitude and MPF responses is useful for describing the neural and mechanical aspects of taskspecific muscle fatigue (Tarata 2003).

In addition, Blangsted et al. (2005) used simultaneous detection of MMG and EMG signals to examine low-frequency fatigue. Specifically, the experimental protocol required the subjects to perform a 10-minute sustained isometric muscle action of the wrist extensors at 10% MVC to induce fatigue.

Prior to this fatiguing activity, the subjects went through an electrical stimulation procedure (a 10-s train at 1 Hz, two 2.5-s trains at 20 Hz, and two2-s trains at 100 Hz), in addition to performing an isometric MVC, and separate 20-s muscle actions at 5% and 80% MVC. The same electrical stimulation protocol and series of voluntary muscle actions were performed 10,

Surface mechanomyographic responses to muscle fatigue

47

Figure 3. Contour plot of changes in electromyographic (EMG; top plot) and mechanomyographic (MMG; bottom plot) amplitude and frequency for the biceps brachii during a 450-second sustained muscle action of the forearm flexors at 25% of the isometric maximum voluntary contraction (MVC). Changes in the density of the contour plot reflect changes in MMG amplitude. *Reprinted with permission from Tarata (2003).

48

Travis W. Beck

30, 90, and 150 minutes after the fatiguing muscle action. The results showed that during both the 5% and 80% MVC muscle actions, the mean MMG amplitude values 30 minutes after fatigue were greater than those before fatigue. In addition, the MMG/EMG amplitude ratio (used as a measure of electromechanical efficiency) during the 80% MVC muscle action was elevated even 150 minutes after the fatigue protocol. The authors also reported decreased performance in many of the twitch parameters measured during electrical stimulation, including peak twitch force, time to peak force, maximum rate of force development, the rate of force decay, half-contraction time, half-relaxation time, and the force-time integral following the fatigue protocol. Each of these performance parameters was depressed 10, 30, 90, and 150 minutes after the fatigue protocol. Thus, it was concluded that MMG may be a useful technique for studying the mechanical aspects of low frequency fatigue. In addition, the combination of MMG and EMG could potentially be used to assess fatigue that develops over extended time periods, such as that associated with low back pain and fatigue that develops during a work day (Blangsted et al. 2005). A second study by Blangsted et al. (2005) used a similar experimental design to examine the potential relationships between changes in EMG and MMG amplitude and MPF during a sustained muscle action and the intramuscular pressure and tissue oxygenation levels.

Specifically, the subjects were required to perform a 10-minute sustained isometric muscle action of the forearm flexors at 10% MVC, and MMG and EMG signals were detected simultaneously from the biceps brachii. The results indicated that MMG amplitude increased over time during the sustained muscle action, but there was no change in EMG amplitude, and both MMG MPF and EMG MPF decreased. There was also a significant increase in the MMG amplitude/EMG amplitude ratio. Furthermore, all subjects were required to perform an isometric muscle action at 5% MVC before, as well as 10 and 30 minutes after the fatiguing muscle action. The results from these muscle actions showed that during the 5% MVC muscle action 10 and 30 minutes after fatigue, MMG amplitude was higher, and MMG MPF lower than before fatigue. In addition, EMG amplitude was elevated 30 minutes after fatigue, and EMG MPF was depressed 10 minutes after fatigue. Interestingly, however, the changes in tissue oxygenation and intramuscular pressure could not explain the corresponding patterns for MMG and EMG amplitude and MPF. In particular, the tissue oxygenation level returned to the resting level very quickly after fatigue, and intramuscular pressure did not change during the fatiguing muscle action. Thus, it was concluded that the combined use of MMG and EMG may be sensitive to fatigue-induced changes in skeletal muscle function that are not reflected in the tissue oxygenation and intramuscular pressure levels (Blangsted et al. 2005).

Surface mechanomyographic responses to muscle fatigue

49

Orizio et al. (1999) also used a unique approach to examine the effects of electrically-stimulated fatigue on several measures of twitch mechanics and MMG parameters. Specifically, the experimental protocol required the subjects to go through an electrical stimulation procedure that involved six single twitches separated by a 1-second time interval, followed by a 5-second period of repetitive stimulation where the frequency was increased by 1 Hz between one stimulation and the next in the range from 1-50 Hz. A fatigue protocol was then performed where the muscle was stimulated at 35 Hz for 40 seconds. The contractile properties of the muscle were then tested in the same manner as before the fatiguing stimulation. The results indicated that the fatiguing protocol reduced several of the twitch parameters, including the force peak, the peak rate of force production, the peak of the acceleration of force production, as well as increased contraction time and half-relaxation time. In addition, the peak-to-peak amplitude of the MMG signal decreased immediately after the fatiguing protocol, but returned to pre-fatigue levels within two minutes of recovery. Furthermore, MMG amplitude was highly correlated with the peak of the acceleration of force production during recovery. Thus, it was suggested that MMG may be a useful tool for examining changes in muscle mechanics due to fatigue, particularly when the force output of the muscle cannot be measured directly (Orizio et al. 1999). Al-Zahrani et al. (in press) examined the reliability of MMG amplitude, MPF, and median frequency during a fatiguing isometric muscle action. Specifically, the subjects were required to perform three separate 40-second sustained isometric muscle actions of the leg extensors at 75% MVC. The same experimental protocol was also followed on two more testing occasions that were separated by at least 48 hours. The results of the study generally indicated that the linear slope coefficients for the changes in MMG amplitude, MMG MPF, and MMG median frequency across time were not reliable, but the overall amplitude, MPF and median frequency values were. Thus, it was suggested that caution should be used when interpreting the linear slope coefficients for MMG amplitude, MMG MPF, and MMG median frequency during a fatiguing isometric muscle action (AlZahrani et al. in press).

Overall, the results from the studies that have examined the MMG amplitude and/or center frequency responses during fatiguing isometric muscle actions have shown that the behavior of the MMG signal is largely dependent on the duration and relative intensity of the sustained muscle actions. It is also dependent, however, on the muscle being examined and the experimental protocol that is used. It is important to point out that this variability likely reflects the sensitivity of the MMG signal to the demands of the fatiguing task.

Thus, the MMG amplitude and center frequency responses during fatigue reflect the interaction between the motor control strategies that are being used

50

Travis W. Beck

to maintain performance during the activity and the mechanical properties of the muscle.

References 1.

Al-Zahrani E, Gunascharan C, Callaghan M, Gaydecki P, Benitez D, Oldham J. Within-day and between-days reliability of quadriceps isometric muscle fatigue using mechanomyography on healthy subjects. Journal of Electromyography and Kinesiology (In Press).

2. Barry DT, Geiringer SR, Ball RD. Acoustic myography: A noninvasive monitor of motor unit fatigue. Muscle & Nerve 1985;8:189-194.

3. Blangsted AK, Sjøgaard G, Madeleine P, Olsen HB, Søgaard K. Voluntary lowforce contraction elicits prolonged low-frequency fatigue and changes in surface electromyography and mechanomyography. Journal of Electromyography and Kinesiology 2005;15:138-148.

4. Blangsted SK, Vedsted P, Sjøgaard G, Søgaard K. Intramuscular pressure and tissue oxygenation during low-force static contraction do not underlie muscle fatigue. Acta Physiologica Scandinavica 2005;183:379-388.

5. Esposito F, Orizio C, Veicsteinas A. Electromyogram and mechanomyogram changes in fresh and fatigued muscle during sustained contraction in men.

European Journal of Applied Physiology 1998;78:494-501.

6. Goldenberg MS, Yack HJ, Cerny FJ, Burton HW. Acoustic myography as an indicator of force during sustained contractions of a small hand muscle. Journal of Applied Physiology 1991;70:87-91.

7. Herzog W, Zhang Y-T, Vaz MA, Guimaraes ACS, Janssen C. Assessment of muscular fatigue using vibromyography. Muscle & Nerve 1994;17:1156-1161.

8. Itoh Y, Akataki K, Mita K, Watakabe M, Itoh K. Time-frequency analysis of mechanomyogram during sustained contractions with muscle fatigue. Systems and Computers In Japan 2004;35:26-36.

9. Kouzaki M, Shinohara M, Fukunaga T. Non-uniform mechanical activity of quadriceps muscle during fatigue by repeated maximal voluntary contraction in humans. European Journal of Applied Physiology 1999; 80:9-15.

10. Madeleine P, Ge H-Y, Jaskólska A, Farina D, Jaskólski A, Arendt-Nielsen L. Spectral moments of mechanomyographic signals recorded with accelerometer and microphone during sustained fatiguing contractions. Medical & Biological Engineering and Computing 2006;44:290-297.

11. Madeleine P, Jørgensen LV, Søgaard K, Arendt-Nielsen L, Sjøgaard G. Development of muscle fatigue as assessed by electromyography and mechanomyography during continuous and intermittent low-force contractions: effects of the feedback mode. European Journal of Applied Physiology 2002;87:28-37.

12. Mealing D, McBride JW, Khan AZ. Frequency domain analysis of muscle sound during fatigue of human quadriceps. Proceeding of the Annual International Conference of the IEEE Engineering In Medicine and Biology Society, 1990; 12:2215-2216.

Surface mechanomyographic responses to muscle fatigue

51

13. Orizio C, Diemont B, Esposito F, Alfonsi E, Parrinello G, Moglia A, Veicsteinas A. Surface mechanomyogram reflects the changes in the mechanical properties of muscle at fatigue. European Journal of Applied Physiology 1999;80:276-284.

14. Orizio C, Perini R, Diemont B, Veicsteinas A. Muscle sound and electromyogram spectrum analysis during exhausting contractions in man. European Journal of Applied Physiology 1992;65:1-7.

15. Orizio C, Perini R, Veicsteinas A. Changes of muscular sound during sustained isometric contraction up to exhaustion.

Journal of Applied Physiology 1989;66:1593-1598.

16. Søgaard K, Blangsted AK, Jørgensen LV, Madeleine P, Sjøgaard G. Evidence of long term muscle fatigue following prolonged intermittent contractions based on mechano- and electromyograms. Journal of Electromyography and Kinesiology 2003;13:441-450.

17. Stokes MJ, Dalton PA. Acoustic myography for investigating human skeletal muscle fatigue. Journal of Applied Physiology 1991;71:1422-1426.

18. Tarata MT. Mechanomyography versus Electromyography, in monitoring the muscular fatigue. BioMedical Engineering OnLine 2003;2:1-10.

19. Vaz MA, Zhang Y-T, Herzog W, Guimaraes ACS, MacIntosh BR. The behavior of rectus femoris and vastus lateralis during fatigue and recovery: an electromyographic and vibromyographic study. Electromyography and Clinical Neurophysiology 1996;36:221-230.

20. Weir JP, Ayers KM, Lacefield JF, Walsh KL. Mechanomyographic and electromyographic responses during fatigue in humans: influence of muscle length.

European Journal of Applied Physiology 2000;81:352-359.

21. Yoshitake Y, Ue H, Miyazaki M, Moritani T. Assessment of lower-back muscle fatigue using electromyography, mechanomyography, and near-infrared spectroscopy. European Journal of Applied Physiology 2001;84:174-179.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010, 53-67 ISBN : 978-81-7895-449-3 Editor: Travis W. Beck

4

Clinical applications of surface mechanomyography Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma, 73019-608, USA

Abstract Surface mechanomyography (MMG) has been used in two primary clinical applications: (1) as a control signal for an externally-powered prosthesis, and (2) for examining and/or diagnosing neuromuscular disorders. The results have generally shown that the MMG signal could be used for prosthesis control and has several advantages over the surface electromyographic (EMG) signal, including reduced sensitivity to exact sensor placement and robustness to changes in skin impedance. More research needs to be done, how ever, Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: tbeck@ou.edu

54

Travis W. Beck

to develop techniques that can be used to reduce movement artifacts and improve the classification accuracy with MMG signals. In addition, the results from several studies have shown that the MMG amplitude and/or frequency responses are different for normal, healthy subjects versus those that suffer from a neuromuscular disease, such as myotonic dystrophy and spastic cerebral palsy. Thus, MMG may be useful for diagnosing these disorders, as well as understanding how they affect muscle function. A particularly interesting application of MMG is for examining the effectiveness of anaesthesia. The recent studies that have investigated this issue have shown that the MMG signal is a useful alternative to the surface EMG and force signals. However, future studies are needed in this area to identify the feasibility of using MMG for this application in practical situations.

Introduction One of the first studies to examine the possibility of using mechanomyography (MMG) in clinical applications was Barry et al. (1986), who investigated the use of MMG as a control signal for an externally powered prosthesis. The authors found that MMG provided several advantages over using surface electromyography (EMG) for prosthetic control, including no need for direct skin contact, the fact that the MMG signal was unaffected by skin impedance, and its amplitude was high enough to produce a 50 mV output from a standard microphone.In addition, the MMG signal required less amplification and electrical shielding, and was qualitatively less sensitive to precise placement over the muscle than was EMG. Furthermore, the disadvantages of MMG, such as movement artifact and rubbing of the sensor on the skin were relatively easy to overcome. Thus, the authors concluded that MMG could be a reliable and inexpensive alternative to EMG for the control of an externally powered prosthesis. This hypothesis was supported by the finding that two subjects that had already been accustomed to EMG-controlled prostheses successfully learned to use an MMG-controlled prosthesis after only three minutes of practice (Barry et al. 1986). L’Estrange et al. (1993) investigated the possibility of using MMG to examine masseter muscle function in humans. Specifically, surface EMG and MMG signals were detected simultaneously from both the right and left masseter muscles during 4-second isometric jaw clenching. The results indicated that the within-day reliability for MMG amplitude was good, with a test-retest correlation of r = 0.70 for the right masseter muscle and r = 0.71 for the left muscle. It was suggested, however, that strict control of the contact pressure of the sensor over the muscle was important for ensuring reliability. In addition, frequency analysis of the MMG signals showed that they all had a mean power frequency

Clinical applications of surface mechanomyography

55

Figure 1. Frequency spectrum of the mechanomyographic (MMG) signal detected from the left masseter muscle of one subject. Notice that most of the power in the MMG signal is below 50 Hz. *Reprinted with permission from L’Estrange et al. (1993).

(MPF) of 9-10 Hz. Visual inspection of the frequency spectrum, however, indicated distinct peaks that were hypothesized to reflect exact harmonics of a fundamental frequency (Figure 1). Furthermore, it was suggested that MMG could potentially be useful for examining cranio-mandibular disorders, since tenderness of the masseter muscles is a common complaint of those that suffer from jaw pain (L’Estrange et al. 1993).

Wright and Stokes (1992) performed a similar study to examine the possibility of using EMG and MMG to investigate the etiology of low back pain. Specifically, the subjects were required to perform a 60-second sustained isometric muscle action of the back extensors in which they supported their body weight, and MMG and EMG signals were detected simultaneously from both the right and left erector spinae muscle groups. The results showed that EMG amplitude increased with time during the fatigue test for both the right and left erector spinae. There were no changes, however, for MMG amplitude during the fatigue test for either the right or left sides. In addition, the MMG amplitude / EMG amplitude ratio decreased during the fatigue test with similar slopes for both the right and left sides, thus indicating reduced efficiency of electrical activity. Finally, the within-day reliability of MMG amplitude was very good, with coefficients of variation ranging from 5.5-8.9%. Thus, it was concluded that for the subjects that participated in the study, the changes in erector spinae EMG and MMG amplitude were symmetrical on both the right

56

Travis W. Beck

and left sides, thus indicating normal paraspinal muscle function. In addition, since MMG amplitude did not change during the fatigue test, but EMG amplitude increased, MMG may be a better indicator of force production than EMG. Furthermore, when used in combination, the sensitivity of MMG and EMG may be useful in rehabilitative settings where a patient initially demonstrates asymmetry of the right and left paraspinal muscles. As the patient progresses through their rehabilitation program, however, their stages of recovery could be monitored with MMG and EMG to ensure that the program is developing symmetry between the paraspinal muscles of the right and left sides (Wright and Stokes 1992).

Lee et al. (1992) also examined the reliability of MMG amplitude for the erector spinae muscles during quiet standing, prone lying with the back halfextended, fully extended with no resistance, and fully extended against manual resistance. The results showed that the MMG signals recorded when the subject was in the prone position and the back fully extended were more reliable (as reflected by lower coefficients of variation) than those recorded in the other positions. In addition, it is important to point out that some of the subjects that participated in the study were patients that suffered from chronic, post-surgical low back pain. Thus, it was suggested (Lee et al. 1992) that MMG could be useful for monitoring the stages of recovery in patients that suffer from low back pain, since force production by the erector spinae is difficult to measure directly.

Dascalu et al. (1999) performed an interesting study to determine if MMG could be used to monitor the effectiveness of anaesthesia. Specifically, 25 patients that were undergoing either abdominal or orthopaedic surgery were anaesthetized, and three techniques were used to determine the effectiveness of the anaesthesia. The first method involved detecting EMG signals from the adductor pollicis muscle during supramaximal electrical stimulation of the ulnar nerve. Surface MMG signals were also detected simultaneously, as was the force signal from the muscle. The results indicated that the amplitude of the MMG signal was highly correlated with the force signal (r=0.86) and the EMG signal (r = 0.85). Thus, it was suggested that MMG may be an effective alternative to EMG and force signals for monitoring the effectiveness of neuromuscular block. It was also recommended, however, that future investigations still need to be performed before MMG can be widely accepted as a tool for monitoring anaesthesia (Dascalu et al. 1999). Lee et al. (1996) also examined the possibility of using MMG for determining the effectiveness of anaesthesia. Specifically, seven male pigs were subjected to Mg2+- induced neuromuscular block, and the EMG and MMG responses of the tibialis anterior muscle were investigated during electrical stimulation. The results showed that MMG was generally more sensitive than EMG for testing the effectiveness of

Clinical applications of surface mechanomyography

57

neuromuscular blockade. It was also suggested, however, that more investigations still needed to be performed before MMG could be used to gauge the effectiveness of anaesthesia in humans (Lee et al. 1996).

Orizio et al. (1997) were among the first to examine the possibility of using MMG to diagnose neuromuscular diseases. Specifically, the authors recorded MMG and EMG signals from the biceps brachii and flexor digitorum profundus during isometric muscle actions at 20%, 40%, and 60% of the maximum voluntary contraction (MVC) in ten subjects that suffered from myotonic dystrophy, as well as ten age-matched controls. The results indicated that the isometric strength values for the myotonic dystrophy patients were about 30% less than those of the controls for the forearm flexors, and approximately 74% less for the finger flexors. In addition, at any given force level, the mean MMG and EMG amplitude values for the myotonic dystrophy patients were significantly lower than those for the control subjects. Finally, the ratio of MMG amplitude / EMG amplitude (which was used as a measure of electromechanical efficiency) was similar for the myotonic dystrophy patients and control subjects for the forearm flexors, but much lower for the myotonic dystrophy patients for the finger flexors. Thus, it was suggested that the MMG amplitude / EMG amplitude ratio may be useful for estimating the efficiency of the mechanical contributions of the active motor units. In addition, the combined use of MMG and EMG may be able to differentiate between muscles that are more affected by muscular dystrophy (i.e., the finger flexors) and those that are less affected (Orizio et al. 1997).

The potential for MMG to be used in examining muscle atrophy is also an interesting clinical application. PisĹ?t et al. (2008) investigated the changes that occur in MMG amplitude and muscle contractile parameters during thirty-five days of bed rest in ten healthy men. Specifically, MMG amplitude for the biceps brachii, vastus medialis, biceps femoris, and gastrocnemius medialis were assessed during electrically stimulated isometric twitches both before and after the thirty-five days of bed rest. Changes in muscle thickness were also assessed by ultrasound imaging. The results showed that bed rest resulted in a significant increase (18%) in contraction time for the gastrocnemius muscles, and decreased muscle belly thickness (15%). In addition, the maximal displacement of the muscle (which is analogous to MMG amplitude) increased after bed rest for the vastus medialis (24%) biceps femoris (26%) and gastrocnemius medialis (30%). Thus, it was suggested (PisĹ?t et al. 2008) that MMG may be sensitive to changes in the viscoelastic properties of a muscle that is undergoing the process of atrophy. This, in turn, could have application to evaluating the decreases in muscle function that accompany space flight.

McKay et al. (2004) examined the MMG and EMG amplitude and MPF responses for the rectus femoris muscle immediately after aerobic exercise.

58

Travis W. Beck

Specifically, all subjects were required to cycle on a stationary cycle ergometer at a power output that corresponded to 70% of the power output at VO2 max for a duration of 30 minutes. Immediately before and after the exercise protocol, MMG and EMG signals were detected from the rectus femoris in the resting state. The results showed that MMG amplitude was elevated immediately after exercise (i.e., in relation to the pre-exercise value), but decayed over time with a time constant that was equivalent to that for VO2. In addition, there were no significant differences between the pre-exercise and immediate post-exercise values for EMG amplitude, EMG MPF, and MMG MPF. Thus, it was concluded that the elevated MMG activity after exercise reflected increased mechanical work that also caused greater oxygen consumption. In addition, it was suggested that MMG may be a more sensitive indicator of resting muscle activity than EMG, since EMG amplitude was not elevated after exercise. It was also recommended, however, that additional studies need to be performed to determine if post-exercise muscle sounds are related to exercise intensity, and whether these sounds can be affected by obesity or muscle disease (McKay et al. 2004). McKay et al. (2006) have also conducted some interesting studies that examine the MMG responses from resting muscle immediately after resistance exercise. Specifically, the subjects were required to perform maximal isokinetic muscle actions of the right leg extensors at a velocity of 60掳路s-1. These muscle actions were performed as separate sets of 1, 5, 10, 20, and 30 repetitions, and the resting MMG activity of the rectus femoris muscles of both the right and left legs were assessed with separate accelerometers. In addition, the resting MMG activity was measured at both full leg extension (i.e., a short muscle length) and full leg flexion (i.e., a long muscle length). The results indicated that MMG amplitude for the right leg was elevated after exercise, but only after the set that required 30 repetitions to be performed. In addition, there was a significant positive linear relationship between the total work performed and resting MMG amplitude after exercise for both the right (r = 0.61) and left (r = 0.67) legs. Furthermore, when the rectus femoris was at a long muscle length (i.e., when the leg was fully flexed), there was no relationship between MMG amplitude and total work, but the resting MMG amplitude before exercise was greater at a short muscle length than at a long length. Thus, it was concluded that resting MMG activity is likely neural in origin, since it was present in a contralateral muscle that had done no mechanical work. In addition, the greater resting MMG amplitude at the short muscle length than the long length was hypothesized to be due to a more compliant musculotendinous unit that allowed greater muscle fiber oscillations and larger MMG amplitude values.

Thus, these findings (McKay et al. 2006) were consistent with those from the previous study of aerobic exercise (McKay et al. 2004) and indicated that

Clinical applications of surface mechanomyography

59

resting MMG activity is elevated immediately after isokinetic exercise.

McKay et al. (2007) also measured resting MMG after conventional resistance exercise. Specifically, the subjects were required to perform one set of 10 repetitions with 50% of their one-repetition maximum (1-RM) as a warm-up, followed by five sets of 8 repetitions with 75% of their 1-RM on both the bilateral leg extension and leg press exercises.This workout took approximately 30 minutes to complete, after which resting MMG and EMG signals were recorded from the rectus femoris for 5.75 hours. The results showed that MMG amplitude increased from 3.0 卤 0.99 mm路s-2 pre-exercise to 10.1 卤 4.5 mm路s-2 post-exercise, which corresponded to an increase that ranged from 1.8 to 7.7 times for all subjects. In addition, the mean MMG amplitude value after exercise decayed over time with a time constant that was statistically equivalent to that for VO2. Finally, EMG amplitude was significantly greater after exercise than before exercise, but only four data points exceeded the lower limit of resolution for the EMG amplifier. Thus, it was concluded that resting MMG amplitude increased about threefold after strenuous resistance exercise and then decayed exponentially over time in a manner that was similar to that for VO2. The authors (McKay et al. 2007) also proposed an interesting hypothesis to explain the elevated MMG amplitude after exercise, with no change in EMG amplitude. Specifically, it was suggested that most commercially available EMG systems are simply not sensitive enough to detect resting muscle activity, but since muscle tissue is an excellent conductor of sound, MMG may be a more useful method than EMG for quantifying resting muscle activity. It was also recommended, however, that more research needs to be done to identify the mechanisms underlying resting muscle sound (McKay et al. 2007).

Hu et al. (2007) simultaneously recorded surface MMG and EMG signals from the biceps brachii during submaximal isometric muscle actions of the forearm flexors at 20%, 40%, 60%, and 80% MVC in normal healthy subjects and patients that had suffered a stroke. For the stroke patients, both the affected and unaffected limbs were tested. The results showed that the mean MMG and EMG amplitude and MPF values for the limb that was affected by the stroke were less than those for the unaffected limb, as well as those from the limbs of the healthy controls. Thus, it was suggested that a loss of fasttwitch motor units in the muscles of the affected limb and/or reduced neural drive to those muscles could have caused the MMG and EMG amplitude and MPF responses. In addition, MMG could potentially be used as a complement to EMG for examining the loss of muscle function that occurs following a stroke, as well as the return to normal function during rehabilitation (Hu et al.

2007).

60

Travis W. Beck

McAndrew et al. (2005) examined the potential for MMG to be used for several applications in musculoskeletal rehabilitation. Specifically, the authors used a laser-based MMG system that measured muscle displacement during voluntary or electrically-stimulated contractions. The use of the laser-based system allowed for calculation of contraction time, normalized contraction time, sustained time, and relaxation time, which are not usually assessed when MMG is detected with an accelerometer, condenser microphone, or piezoelectric device. One of the first tests that the authors conducted examined the effect of changes in pulse duration from the electrical stimulator on the twitch properties of the muscle and MMG responses. The results showed that a pulse duration of 100 Âľs was optimal for producing a maximal muscle contraction, as well as for optimizing the lateral displacement of the muscle and the contraction and relaxation speeds. The second part of the study involved a fatiguing stimulation protocol of isolated rat muscle (40-second duration at 2 Hz) that was designed to simulate the fatigue associated with chronic low back pain. Both before and after the fatiguing protocol, the twitch parameters and MMG responses of the muscle were tested. The results indicated that fatigue caused a decrease in peak displacement of the muscle and impairment of contraction time and half-relaxation time. Thus, the authors (McAndrew et al. 2005) concluded that MMG could be a useful method for examining chronic low back pain.

Nolan and de Paor (2004) investigated the possibility of using MMG to help disabled individuals communicate. For example, individuals that have suffered a stroke often find it difficult to communicate due to impairment of the muscles used for speech. Thus, the authors (Nolan and de Paor 2004) developed an accelerometer-based system that detected the MMG signals from the biceps brachii and sternocleidomastoid muscles. These signals were then used to control a software program that scanned through letters on a keyboard.

When the software program reached the letter that the patient wanted to select, they contracted the biceps brachii or sternocleidomastoid muscles, and the resulting MMG signals from the muscles were used to control the software and select the letter. Although the authors reported that facial movements sometimes caused unintentional triggering in the software, they also found that their system was fairly reliable and useful for people that needed an alternative method for communicating. They also suggested, however, that additional work needed to be done before the system could be used on a widespread basis (Nolan and de Paor 2004). Barry et al. (1990) also used MMG to examine various neuromuscular diseases. Specifically, the authors recorded surface MMG signals from the biceps brachii in both normal children (age range = 716 years) and kids that suffered from some form of neuromuscular disease (e.g., muscular dystrophy, myotonic dystrophy, dermatomyositis, etc.). The

Clinical applications of surface mechanomyography

61

results showed that the electromechanical efficiency (i.e., EMG amplitude / MMG amplitude ratio) of the diseased patients was less than that of the healthy subjects, and it was hypothesized that this was due to atrophied fibers in diseased muscle that generated electrical activity, but provided little mechanical contribution to force production. Thus, it was concluded that MMG allows mechanical data to be obtained from the muscle when force measurements are unavailable, such as when examining the paraspinal muscles, which are often one of the first muscle groups affected by myopathies (Barry et al. 1990). Akataki et al. (1996) examined the EMG and MMG responses from subjects that suffered from spastic cerebral palsy. Specifically, surface EMG and MMG signals were detected simultaneously from the biceps brachii during submaximal isometric muscle actions of the forearm flexors at force levels ranging from 10-50% MVC. The results indicated that for all force levels, the mean MMG amplitude values for the normal subjects were significantly greater than those for the cerebral palsy patients. In addition, both the MMG amplitude / muscle cross sectional area and MMG amplitude / EMG amplitude ratios were greater for the normal subjects than for the cerebral palsy patients at all force levels. The mean isometric forearm flexion strength value for the cerebral palsy patients was also roughly 50% that of the normal subjects. Thus, it was concluded that in addition to the EMG amplitude / force ratio, the MMG amplitude / EMG amplitude ratio may be a useful indicator of electromechanical efficiency. In addition, it is likely that the altered MMG amplitude / EMG amplitude ratio in the cerebral palsy patients was due to the deterioration of muscle contractile properties that accompanies muscle atrophy (Akataki et al. 1996).

Madeleine et al. (2007) investigated the possibility of using MMG to determine the etiology of chronic low back pain. Specifically, 12 separate MMG sensors were placed over the right and left erector spinae muscles (6 sensors on each side of the vertebral column). The subjects were then required to perform submaximal isometric muscle actions of the back extensors with different external loads, ranging from 0-15 kg in 2.5 kg increments. The subjects also performed a 6-minute sustained isometric muscle action of the back extensors with an external load of 7.5 kg. The results showed that MMG amplitude generally increased with the external load that was being supported, while MMG MPF decreased. Similar results were also reported for the MMG amplitude and MPF responses during the fatiguing muscle action, since MMG amplitude increased, and MMG MPF decreased over time. Perhaps the most important finding from the study, however, was that when the MMG amplitude and MPF values were expressed in absolute terms (i.e., m路s-2 and Hz, respectively), the patterns of responses were different for each sensor location.

The patterns for normalized MMG amplitude and MPF were the same,

62

Travis W. Beck

however, for all sensor locations. Thus, it was concluded that when measuring the mechanical activities of the paraspinal muscles with MMG, it is important to detect signals from multiple locations to account for differences that can be caused by sensor placement (Madeleine et al. 2007).

Madeleine and Arendt-Nielsen (2005) examined the effects of experimentally-induced muscle pain on MMG signals from the biceps brachii during submaximal isometric muscle actions of the forearm flexors.

Specifically, the experimentally-induced pain was administered by injecting a hypertonic saline solution into the right biceps brachii muscle, while injection of an isotonic saline solution was used as a control. Following injection of either the hypertonic saline or isotonic saline, the subjects were required to perform submaximal isometric muscle actions at 0%, 10%, 30%, 50%, and 70% MVC, as well as a 25-second isometric ramp muscle action from 0-50% MVC. The results showed that MMG amplitude increased significantly after the hypertonic saline injection during both the constant force and isometric ramp muscle actions. There were no changes, however, for the EMG variables.

Thus, it was concluded that the hypertonic saline injection not only caused increases in muscle pain, but it may also have changed the muscle’s contractile properties such that compensatory mechanisms (e.g., decreased firing rate and increased twitch force) were used to meet the force demands of the task. In addition, it was suggested that under well-controlled conditions, MMG may be more sensitive than EMG for detecting changes in the mechanical properties of the muscle due to experimentally-induced pain (Madeleine and Arendt-Nielsen 2005).

Jaskólska et al. (2006) examined the EMG and MMG responses for the triceps brachii, biceps brachii, and brachioradialis during maximal isometric muscle actions of the forearm flexors and extensors at different elbow joint angles (i.e., the optimal joint angle for force production, as well as 30° greater and 30° less than the optimal joint angle) in both young (mean ± SD age = 20.1 ± 1.1 years) and old (mean ± SD age = 64.9 ± 5.1 years) women. The results showed that the ratio between MMG amplitude and force was lower in the old women compared to the young women, but only for the triceps brachii. In addition, the MMG amplitude versus force ratio increased for both the biceps brachii and triceps brachii as the muscle was shortened for both the young and old women. Thus, it was concluded that the EMG and MMG responses to changes in muscle length were generally similar in the old and young women, but the optimal angle with respect to force production was greater in the older women.

Ng et al. (2006) conducted an interesting study to examine differences in force-related post-activation potentiation and MMG-related post-activation potentiation between normal, healthy subjects and those that suffered from

Clinical applications of surface mechanomyography

63

some type of myopathy. Force-related post-activation potentiation refers to the increase in electrically-stimulated force production that occurs after an isometric MVC, whereas MMG-related post-activation potentiation reflects the increase in MMG amplitude immediately after an isometric MVC. All subjects were required to undergo five single supramaximal electrically-stimulated twitches of the biceps brachii both before and immediately after a 10-second isometric MVC. The results showed that the MMG-related post-activation potentiation was positively correlated with the force-related post-activation potentiation for the normal subjects. In addition, the MMG-related postactivation potentiation was significantly lower in the subjects that suffered from a myopathy than it was in the normal or diseased control subjects (Figure 2). Furthermore, there was no relationship between the MMG-related postactivation potentiation and Type 2 muscle fiber atrophy.

Thus, it was concluded that although MMG may have some potential for differentiating between normal and non-dystrophic myopathies, more work still needed to be done before it could be used for diagnostic purposes (Ng et al. 2006). Orizio et al. (1997) investigated the MMG responses for the tibialis anterior in patients that suffered from myotonic dystrophy during electricallystimulated twitches, as well as repetitive stimulation at 5, 10, 15, or 20 Hz. The results showed that during the single twitches, MMG amplitude was 67% lower, MMG MPF was 44% lower, the twitch duration was 37% longer, and the electromechanical delay was 64% longer in the myotonic dystrophy patients than in the controls. In addition, the peak-to-peak amplitude of the MMG signal was less at each stimulation frequency for the myotonic dystrophy patients than for the controls. Thus, it was concluded that in addition to differences in sarcolemmal excitability, myotonic dystrophy is characterized by altered electromechanical coupling and a failure in the contractile machinery (Orizio et al. 1997).

Alonso et al. (2007) examined the MMG and EMG activities of the genioglossus, sternocleidomastoid, and diaphragm muscles during a high respiratory effort contraction in patients with obstructive sleep apnea syndrome as well as normal, healthy subjects. The purpose of the investigation was to determine if the combined use of MMG and EMG could be used to examine the characteristics of obstructive sleep apnea syndrome. The results showed that at progressively higher levels of respiratory effort, there was an increase in the nonlinear coupling of the respiratory muscles for both the patients and control subjects. That is, the genioglossus, sternocleidomastoid, and diaphragm muscles were more coordinated with their contractions. Thus, it was concluded that the combined use of MMG and EMG may be useful for examining the activities of respiratory muscles during labored breathing, as well as for investigating obstructive sleep apnea syndrome (Alonso et al. 2007). Madeleine

64

Travis W. Beck

Figure 2. Mechanomyographic (MMG) signals detected from the biceps brachii during a maximal electrically-stimulated isometric twitch immediately before (Pre-MVC) and after (Post-MVC) a 10-second isometric maximum voluntary contraction (MVC). (A) shows the data from a healthy subject, (B) demonstrates the signals from a subject with polymyositis, and (C) shows the data from a subject that suffers from Kennedy-AlterSung disease. Notice that in the healthy subject, the amplitude of the MMG signal for the Post-MVC twitch was potentiated, but this was not as apparent for the neuromuscular disease patients. *Reprinted with permission from Ng et al. (2006)

and Farina (2008) also performed an interesting study that examined the time and frequency domain characteristics of surface MMG signals detected from the upper trapezius muscle. Specifically, 12 accelerometers were placed over the dominant upper trapezius muscle, and the subjects were required to perform 10-second isometric shoulder elevation muscle actions at 10%, 20%, 40%, 60%, 80%, and 100% MVC. In addition, the subjects performed a sustained isometric muscle action at 20% MVC until the target force level could no longer be maintained. The results showed that the absolute MMG amplitude and MPF values were dependent on sensor location, but when the values were normalized relative to a standard value, there was no effect of sensor location. In addition, MMG amplitude at the end of the sustained muscle action at 20% MVC was expressed relative to the corresponding value at 100% MVC to calculate the activation ratio. The subjects that showed a greater increase in MMG amplitude over time, higher values for the activation ratio, and lower entropy values were associated with a longer time to task failure. The finding of lower entropy values is reflective of more homogeneous motor unit recruitment. Thus, it was concluded that there was a relationship between the time to task failure, the activation ratio, and MMG amplitude, which suggested that the spatial changes in MMG amplitude during a fatiguing muscle action reflected functional mechanisms that allowed for maintenance of

Clinical applications of surface mechanomyography

65

force production during fatiguing isometric muscle actions (Madeleine and Farina 2008). Huang et al. (2006) have also used MMG to investigate the mechanical and neural aspects of spastic hypertonia. Specifically, the subjects included patients that suffered from either a spinal cord lesion or a stroke, as well as normal subjects. Both the H-reflex and the maximum amplitude of the M-wave, in addition to the MMG amplitude and median frequency responses during the M-wave stimulation were assessed. The results showed that the patients that suffered from spastic hypertonia exhibited greater MMG amplitude values during the M-wave stimulation than the healthy control subjects. In addition, the amplitude of the MMG signal during the M-wave stimulation was correlated with the level of functional impairment displayed by the subjects. Thus, it was concluded that MMG may be a useful tool for assessing the impairment in the mechanical properties of the muscle that occurs with spastic hypertonia (Huang et al. 2006).

Overall, the results from the studies discussed in this chapter indicated that there are many potential clinical applications for MMG, particularly in the areas of neuromuscular disease assessment and control of externally-powered prostheses. There is still a great deal that needs to be done, however, before MMG can be used on a widespread basis for diagnostic purposes. In addition, research that examines MMG activity of resting muscle after exercise and to investigate the effectiveness of anaesthesia are particularly promising clinical applications. Future research should also continue to examine the MMG responses of the lower back muscles to determine if MMG can be used to investigate the etiology of chronic low back pain and asymmetry of the paraspinal muscles.

References 1.

2.

3.

4.

5.

Akataki K, Mita K, Itoh K, Suzuki N, Watakabe M. Acoustic and electrical activities during voluntary isometric contraction of biceps brachii muscles in patients with spastic cerebral palsy. Muscle & Nerve 1996;19:1252-1257.

Alonso JF, Ma単anas MA, Hoyer D, Topor ZL, Bruce EN. Evaluation of respiratory muscles activity by means of cross mutual information function at different levels of ventilatory effort. IEEE Transactions On Biomedical Engineering 2007; 54:1573-1582.

Barry DT, Gordon KE, Hinton GG. Acoustic and surface EMG diagnosis of pediatric muscle disease. Muscle & Nerve 1990;13:286-290.

Barry DT, Leonard Jr. JA, Geiter AJ, Ball RD. Acoustic myography as a control signal for an externally powered prosthesis. Archives of Physical Medicine and Rehabilitation 1986; 67:267-269.

Dascalu A, Geller E, Moalem Y, Manoah M, Enav S, Rucick Z. Acoustic monitoring of intraoperative neuromuscular block. British Journal of Anaesthesia 1999; 83:405-409.

66

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Travis W. Beck

Hu XL, Tong KY, Li L. The mechanomyography of persons after stroke during isometric voluntary contractions. Journal of Electromyography and Kinesiology 2007; 17:473-483.

Huang C-Y, Wang C-H, Hwang I-S. Characterization of the mechanical and neural components of spastic hypertonia with modified H reflex. Journal of Electromyography and Kinesiology 2006; 16:384-391.

Jaskólska A, Kisiel-Sajewicz K, Brzenczek-Owczarzak W, Yue GH, Jaskólski A. EMG and MMG of agonist and antagonist muscles as a function of age and joint angle. Journal of Electromyography and Kinesiology 2006; 16:89-102.

Lee DJ, Stokes MJ, Taylor RJ, Cooper RG. Electro and acoustic myography for noninvasive assessment of lumbar paraspinal muscle function. European Journal of Applied Physiology 1992; 64:199-203.

Lee C, Zhang X, Kwan W-F. Electromyographic and mechanomyographic characteristics of neuromuscular block by magnesium sulphate in the pig. British Journal of Anaesthesia 1996; 76:278-283.

L’Estrange PR, Rowell J, Stokes MJ. Acoustic myography in the assessment of human masseter muscle. Journal of Oral Rehabilitation 1993; 20:353-362.

Madeleine P,Arendt-Nielsen L.Experimental muscle pain increases mechanomyographic signal activity during sub-maximal isometric contractions.

Journal of Electromyography and Kinesiology 2005; 15:27-36.

Madeleine P, Farina D. Time to task failure in shoulder elevation is associated to increase in amplitude and to spatial heterogeneity of upper trapezius mechanomyographic signals. European Journal of Applied Physiology 2008; 102:325-333.

Madeleine P, Turker K, Arendt-Nielsen L, Farina, D.Heterogeneous mechanomyographic absolute activation of paraspinal muscles assessed by a two-dimensional array during short and sustained contractions. Journal of Biomechanics 2007; 40:2663-2671.

McAndrew D,Rosser N,Gorelick M,Philips K,Brown JMM. Mechanomyography for noninvasive clinical diagnosis in musculoskeletal rehabilitation.In: Proceedings of CybErg 2005. The Fourth International Cyberspace Conference On Ergonomics.Johannesburg: International Ergonomics Association Press Editors, Thatcher A, James J, Todd A. McKay WPS, Chilibeck PD, Chad KE, Daku BLF. Resting mechanomyography after aerobic exercise. Canadian Journal of Applied Physiology 2004; 29:743-757.

McKay WPS, Chilibeck PD, Daku BLF. Resting mechanomyography before and after resistance exercise. European Journal of Applied Physiology 2007;102:107-117.

McKay WPS, Jacobson P, Chilibeck PD, Daku BLF. Effects of graded levels of exercise on ipsilateral and contralateral post-exercise resting rectus femoris mechanomyography. European Journal of Applied Physiology 2006; 98:566-574.

Ng AR, Arimura K, Akataki K, Mita K, Higuaki I, Osame M. Mechanomyographic determination of post-activation potentiation in myopathies.

Clinical Neurophysiology 2006;117: 232-239.

Nolan Y, de Paor A. The mechanomyogram as a channel of communication and control for the disabled. Proceedings of the 26th Annual International Conference of the IEEE EMBS. San Francisco, CA. 2004. September 1-5.

Clinical applications of surface mechanomyography

67

21. Orizio C, Esposito F, Paganotti I, Marino L, Rossi B, Veicsteinas A. Electricallyelicited surface mechanomyogram in myotonic dystrophy. Italian Journal of Neurological Sciences 1997; 18:185-190.

22. Orizio C, Esposito F, Sansone V, Parrinello G, Meola G, Veicsteinas A. Muscle surface mechanical and electrical activities in myotonic dystrophy.

Electromyography and Clinical Neurophysiology 1997; 37:231-239.

23. Pišot R, Narici MV, Šimunič B, DeBoer M, Seynnes O, Jurdana M, Biolo G, Mekjavić IB. Whoe muscle contractile parameters and thickness loss during 35day bed rest. European Journal of Applied Physiology 2008;104:409-414.

24. Wright F, Stokes MJ. Symmetry of electro- and acoustic myography activity of the lumbar paraspinal muscles in normal adults. Scandinavian Journal of Rehabilitation Medicine 1992;24:127-131.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010, 69-93 Editor: Travis W. Beck ISBN : 978-81-7895-449-3

5

Mechanomyographic responses during dynamic muscle actions Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081

Abstract Surface mechanomyographic (MMG) signals have been detected during various types of dynamic muscle actions, including dynamic constant external resistance (DCER) muscle actions, concentric and eccentric isokinetic muscle actions, as well as incremental and constant power output cycle ergometry. Although the MMG amplitude and center frequency [mean power frequency (MPF) or median frequency] responses may differ slightly for different muscles and types of muscle actions, a particularly interesting relationship is that between MMG amplitude and muscle power output. Specifically, many Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: tbeck@ou.edu

70

Travis W. Beck

studies have shown that MMG amplitude is closely related to power output during cycle ergometry and maximal concentric and eccentric isokinetic muscle actions. These findings are important from a practical standpoint, because many training programs are designed to improve muscle power output, rather than just maximal torque. Thus, MMG could potentially be used to examine training-induced changes in the power output of individual muscles, which is useful since most joints are crossed by more than one muscle.

Introduction Although most previous studies have examined the mechanomyographic (MMG) amplitude and/or frequency responses during isometric muscle actions, there are several investigations that have used MMG to study muscle function during dynamic muscle actions. Important factors that must be considered when recording MMG signals during dynamic muscle actions include changes in torque production, muscle length, and the thickness of the tissue between the muscle and MMG sensor. Each of these factors can affect both the amplitude and frequency contents of the MMG signal detected during a dynamic muscle action, thereby affecting the validity of these parameters.

There are, however, several pieces of evidence that indicate that the MMG signal is generated by muscle vibrations during a dynamic muscle action.

First, MMG amplitude increases with torque production during both concentric and eccentric muscle actions, as well as with increases in power output during cycle ergometry. Furthermore, for a given torque level, MMG amplitude is lower during an eccentric muscle action than a concentric muscle action, thereby reflecting a reduced level of muscle activity. Perhaps the most convincing evidence, however, that muscle activity generates the MMG signal during a dynamic muscle action is that MMG amplitude decreases to levels that are similar to those at rest when the pedals on a cycle ergometer are driven by the investigator, rather than the subject. As soon as the subject begins driving the pedals, MMG amplitude increases. Thus, the majority of MMG research during dynamic muscle actions has been focused on determining whether or not the MMG signal can be used to evaluate differences in muscle between various types of dynamic muscle actions.

Evetovich et al. (1997) were the first to examine the MMG amplitude responses with increases in velocity during isokinetic muscle actions.

Specifically, MMG amplitude for the vastus lateralis increased with increases in velocity from 60-360掳路s-1 during maximal concentric isokinetic leg extensions, even though isokinetic peak torque decreased. Thus, it was hypothesized that decreases in muscle stiffness at high velocities may have

Mechanomyographic responses during dynamic muscle actions

71

allowed for greater muscle fiber oscillations and an MMG signal with greater amplitude. Theoretically, the velocity-related decrease in muscle stiffness could have been due to the fact that at high velocities, slow-twitch muscle fibers become unloaded, because they are unable to keep up with the speed of the movement, thereby resulting in less muscle stiffness and torque production.

Another interesting finding from the Evetovich et al. (1997) study was that the MMG amplitude values were highly reliable, with intraclass correlation coefficients (ICCs) ranging from R = 0.90-0.99, and no significant mean differences between the test and retest values at any velocity (Evetovich et al.

1997). Smith et al. (1998) also reported a velocity-related increase in MMG amplitude during maximal concentric isokinetic muscle actions, but the muscle examined was the biceps brachii, and the velocities ranged from 30-150掳路s-1.

Similar results were also reported for the biceps brachii during maximal eccentric isokinetic muscle actions of the forearm flexors (Smith et al. 1998).

Evetovich et al. (1998) then performed a second study that compared men and women for the velocity-related changes in MMG amplitude for the vastus lateralis during maximal concentric and eccentric isokinetic muscle actions.

The results showed that there were greater velocity-related increases in MMG amplitude for the men than the women, and the mean MMG amplitude values were greater for the men at all velocities. Interestingly, the velocity-related decreases in concentric isokinetic peak torque were greater for the women (33.3% decline) than the men (28.5% decline). Thus, it was suggested that the gender differences in the MMG amplitude responses may have been due to discrepancies in fiber type composition, muscle mass, and/or the thickness of the subcutaneous adipose tissue over the vastus lateralis. Interestingly, the velocity-related increases in MMG amplitude during the eccentric isokinetic muscle actions were the same for the men and women, but, like the concentric isokinetic muscle actions, the mean MMG amplitude values were higher at all velocities for the men than the women. Thus, it was concluded that MMG amplitude is related to velocity during maximal concentric and eccentric isokinetic muscle actions, and may be useful for differentiating between muscle function in men and women (Evetovich et al. 1998).

Ebersole et al. (2001) examined the MMG amplitude responses for the vastus lateralis with increases in velocity during maximal concentric isokinetic and passive leg extension muscle actions to determine if changes in MMG amplitude were a function of movement velocity, independent of muscle activation. During both the active and passive leg extension muscle actions, MMG amplitude for the vastus lateralis increased with velocity, but during the passive muscle actions, the vastus lateralis was inactive because the EMG amplitude values from the muscle were very small and did not change with

72

Travis W. Beck

increases in velocity. Thus, it was suggested that the velocity-related increases in MMG amplitude may have been due to greater turbulence of the intracellular and extracellular fluid mediums and/or cross-talk from the hamstring muscles (Ebersole et al. 2001). Ebersole et al. (2001) also performed the first study to examine the MMG center frequency responses with increases in velocity during maximal concentric isokinetic muscle actions. The results showed that with an increase in velocity from 60 to 300°·s-1, there was a significant decrease in leg extension peak torque and an increase in MMG amplitude for the vastus lateralis. There was, however, no change in the mean MMG MPF values for the vastus lateralis with increases in velocity. Thus, it was hypothesized that the tendonous iliotibial band that covers the vastus lateralis could have interfered with the muscle fiber oscillations that generate the MMG signal (Ebersole et al. 2000).

Cramer et al. (2000a) examined the relationship between MMG amplitude and power output during maximal concentric isokinetic leg extensions. The experimental protocol required the subjects to perform maximal concentric isokinetic muscle actions of the leg extensors at velocities of 60, 120, 180, 240, 300, 360, 420, and 480°·s-1. During each muscle action, MMG signals were detected from the vastus lateralis, rectus femoris, and vastus medialis. The results showed that with increases in velocity, isokinetic peak torque decreased, but mean power output increased up to approximately 180°·s-1, plateaued from 180 to 300°·s-1, and then decreased from 300 to 480°·s-1.

Interestingly, the patterns of responses for MMG amplitude were nearly identical to those for mean power output (Figure 1).

Thus, it was concluded that MMG amplitude may be more related to muscle power output than peak torque during maximal isokinetic muscle actions, and, therefore, could be useful for monitoring training-induced changes in power output (Cramer et al. 2000a). Cramer et al. (2000b) also examined muscle-specific differences in the MMG amplitude responses with increases in velocity during maximal concentric isokinetic muscle actions. All subjects were required to perform maximal concentric isokinetic leg extensions at velocities ranging from 60 to 300°·s-1, and MMG signals were detected from the vastus lateralis, rectus femoris, and vastus medialis. The results showed that MMG amplitude for each muscle increased from 60 to 180°·s-1. At velocities greater than 180°·s-1, however, MMG amplitude increased to 240°·s1 , and then plateaued from 240 to 300°·s-1 for the vastus lateralis, plateaued from 180 to 300°·s-1 for the rectus femoris, and increased from 180 to 300°·s-1 for the vastus medialis. Thus, it was concluded that the muscle-specific differences in the MMG amplitude responses could have been been due to differences in fiber type composition and/or muscle architecture (Cramer et al.

2000b)

Mechanomyographic responses during dynamic muscle actions

73

Figure 1. The top graph shows changes in mean power output with increases in velocity during maximal concentric isokinetic muscle actions of the leg extensors. The bottom graph demonstrates the corresponding changes in mechanomyographic (MMG) amplitude for the superficial quadriceps femoris muscles (data are collapsed across muscles). Notice that the patterns for mean power output and MMG amplitude were very similar. *Reprinted with permission from Cramer et al. (2000a).

A very important issue when recording MMG signals during dynamic muscle actions is cross-talk (i.e., contamination of the MMG signals from a muscle outside the muscle of interest). Cramer et al. (2003) quantified crosstalk among the superficial quadriceps femoris muscles during maximal

74

Travis W. Beck

concentric and eccentric isokinetic muscle actions of the leg extensors at a velocity of 60°·s-1. The authors used the cross-correlation technique to quantify the cross-talk, and found that the common variance shared between the MMG signals from any two muscles ranged from 14-27%. Thus, it was concluded that the chances for cross-talk were fairly small during maximal isokinetic muscle actions, even for muscles that are close to each other and have a common innervation (Cramer et al. 2003). In addition, the results from a more recent study (Cramer et al. 2004) suggested that the MMG mean power frequency (MPF) responses to increases in velocity may be muscle-specific. In particular, during maximal concentric isokinetic leg extensions, there were velocity-related increases in MMG MPF for the rectus femoris and vastus lateralis muscles, but not for the vastus medialis (Cramer et al. 2004). Thus, it was difficult to identify the exact mechanisms underlying the MMG MPF responses during maximal isokinetic muscle actions. In addition, the patterns of responses for MMG amplitude and MPF with increases in velocity during concentric isokinetic muscle actions were not always the same as those for eccentric isokinetic muscle actions. For example, Smith et al. (1997) reported that during maximal eccentric isokinetic muscle actions of the leg extensors at velocities of 60, 90, 120, and 180°·s-1, there was no change in maximal eccentric peak torque with increases in velocity, but MMG amplitude for the vastus lateralis increased (Figure 2). Furthermore, the MMG amplitude values were highly reliable, with intra-class correlation coefficients ranging from R = 0.97-0.98 and no significant differences between the mean MMG amplitude values for test versus retest at any velocity. Thus, three potential hypotheses were proposed to explain the MMG amplitude responses. First, an increased rate of cross bridge activity during high speed eccentric muscle actions could have resulted in larger vibrations of the myosin heads, thereby increasing MMG amplitude. It is also possible, however, that derecruitment of lowthreshold, slow-twitch motor units and selective activation of high-threshold, fast-twitch motor units during high velocity eccentric muscle actions could have resulted in greater MMG amplitude values. And finally, increases in MMG amplitude with velocity during eccentric muscle actions could have been due to faster movement of the limb and a subsequent greater overall disturbance of the intracellular and extracellular fluid mediums (Smith et al.

1997).

In addition, Evetovich et al. (1999) found that during maximal eccentric isokinetic muscle actions of the leg extensors at velocities of 60, 120, and 180°·s-1, MMG amplitude for the vastus lateralis increased with velocity, but there was no change in MMG MPF from 60 to 180°·s-1. Thus, it was hypothesized that the increase in MMG amplitude may have been due to factors

Mechanomyographic responses during dynamic muscle actions

75

Figure 2. The top graph shows changes in leg extension peak torque with increases in velocity during maximal eccentric isokinetic muscle actions of the leg extensors. The bottom graph demonstrates the corresponding changes in mechanomyographic (MMG) amplitude for the vastus lateralis. Notice that eccentric peak torque changed very little with increases in velocity, but MMG amplitude increased. *Reprinted with permission from Smith et al. (1997).

other than a velocity-related shift in the contributions of slow- and fast-twitch muscle fibers to torque production (Evetovich et al. 1999).

Furthermore, Cramer et al. (2002) reported that during maximal eccentric isokinetic muscle actions of the leg extensors, MMG MPF for the vastus lateralis, rectus femoris, and vastus medialis decreased with an increase in velocity from 60-120°·s-1, and then remained relatively stable from 120-180°·s1 . In addition, Cramer et al. (2002) recently found that mean power output and MMG amplitude for the vastus lateralis increased with velocity during maximal eccentric isokinetic muscle actions of the leg extensors at velocities of 30, 90, and 150°·s-1. These velocity-related increases in MMG amplitude were statistically equivalent for the men and women that participated in the study. Thus, it was suggested that like maximal concentric isokinetic muscle actions, MMG amplitude may be closely related to muscle power output during maximal eccentric isokinetic muscle actions (Cramer et al. 2002).

76

Travis W. Beck

Although much of the dynamic MMG literature has used isokinetic muscle actions, several studies have also examined the MMG amplitude and center frequency responses during dynamic constant external resistance (DCER) muscle actions. For example, Dalton and Stokes (1991) reported that during concentric and eccentric DCER muscle actions of the forearm flexors in which the subjects lifted and lowered weights ranging from 0-8.5 kg, MMG amplitude for the biceps brachii increased linearly from 0-8.5 kg during both the concentric and eccentric muscle actions (Figure 3).

In addition, during a separate study that used the same experimental protocol, MMG MPF for the biceps brachii increased from 0 to approximately 5.5 kg, and then decreased from 5.5-8.5 kg during the concentric muscle actions (Dalton and Stokes 1993). During the eccentric muscle actions, however, MMG MPF remained relatively stable from 0-8.5 kg. Thus, it was concluded that during concentric muscle actions of the forearm flexors, torque production may be increased by recruiting more motor units in the biceps brachii, as well as increasing their firing rates. During eccentric muscle actions,

Figure 3. Mechanomyographic (MMG) amplitude (indicated as IAMG in this figure) for the biceps brachii with increases in forearm flexion torque during concentric (solid symbols) and eccentric (open symbols) dynamic constant external resistance muscle actions. Notice that for every given torque level, MMG amplitude was lower during the eccentric than the concentric muscle actions. Also notice that the patterns of responses were very linear during both types of muscle actions. *Reprinted with permission from Dalton and Stokes (1991).

Mechanomyographic responses during dynamic muscle actions

77

however, torque may be increased primarily by motor unit recruitment (Dalton and Stokes 1991, 1993). Petitjean et al. (1992) also examined the MMG amplitude versus torque relationships for the biceps brachii and brachioradialis during submaximal concentric DCER muscle actions of the forearm flexors.

The results showed that MMG amplitude for both muscles increased linearly with concentric torque (Petitjean et al. 1992).

Several studies from our laboratory have investigated the MMG amplitude and/or MPF versus concentric or eccentric torque relationships, but during isokinetic, rather than DCER muscle actions. For example, Beck et al. (2006) reported that during submaximal to maximal eccentric isokinetic muscle actions of the forearm flexors at a velocity of 30°·s-1, MMG amplitude for the biceps brachii increased from 10% to approximately 60% of the isokinetic peak torque value, and then plateaued from 60-100% peak torque. In addition, MMG MPF for the biceps brachii increased linearly from 10-100% peak torque. Thus, it was concluded that eccentric torque production is increased by motor unit recruitment and firing rate modulation from 10-60% peak torque, followed by increases in firing rates from 60-100% peak torque (Beck et al.

2006). Coburn et al. (2004) also examined the MMG amplitude and MPF responses during submaximal to maximal eccentric isokinetic muscle actions at a velocity of 30°·s-1, but the muscle examined was the vastus medialis. The results showed that both MMG amplitude and MMG MPF increased linearly with eccentric torque from 10-100% peak torque.

Thus, it was concluded that for the vastus medialis muscle, motor unit recruitment and firing rate modulation may occur throughout the entire range of eccentric torque production (Coburn et al. 2004). Madeleine et al. (2001) reported that during both concentric and eccentric DCER muscle actions of the first dorsal interosseous, there were no changes in MMG amplitude or MPF with increases in torque. Since these responses were quite different from those of previous studies for the biceps brachii and vastus medialis, it is possible that they reflected muscle-specific differences in the motor control strategies that modulate eccentric torque production (Madeleine et al. 2001). Furthermore, in two separate studies, Beck et al. (2004a,b) found that MMG amplitude for the biceps brachii muscle increased linearly with torque during submaximal to maximal concentric isokinetic muscle actions of the forearm flexors at a velocity of 30°·s-1, but there was no change in MMG MPF. Thus, it was suggested that for the biceps brachii, concentric torque production may have been modulated primarily by motor unit recruitment, with little change in motor unit firing rates (Beck et al. 2004a,b). In addition, Coburn et al. (2004) examined the patterns of responses for MMG amplitude and MPF for the vastus medialis during submaximal to maximal concentric isokinetic muscle

78

Travis W. Beck

actions of the leg extensors at a velocity of 30°·s-1.The results indicated that both MMG amplitude and MPF increased linearly with torque, and it was suggested that these torque-related increases may have reflected concurrent motor unit recruitment and firing rate modulation (Coburn et al. 2004).

Several studies have also investigated the MMG amplitude and/or MPF responses during fatiguing concentric or eccentric isokinetic muscle actions.

Specifically, Perry-Rana (2002) examined the patterns for MMG amplitude for the vastus lateralis, rectus femoris, and vastus medialis muscles during 50 consecutive maximal concentric isokinetic leg extensions at velocities of 60, 180, and 300°·s-1. The results showed that at 60 and 300°·s-1, there were quadratic decreases in MMG amplitude for the vastus lateralis and vastus medialis muscles, but linear decreases for the rectus femoris (Figure 4). In addition, at 180°·s-1, MMG amplitude decreased in a quadratic fashion for the vastus medialis muscle, but there were linear decreases for the rectus femoris and vastus lateralis. Furthermore, the decreases in MMG amplitude at each velocity were greater for the rectus femoris than the vastus lateralis and vastus medialis. Finally, the decrease in concentric isokinetic leg extension torque at 60°·s-1 was best fit with a quadratic model, but at 180 and 300°·s-1, the PT patterns were best fit with cubic models.

Thus, it was hypothesized that the decreases in MMG amplitude with fatigue at the three velocities may have been due to reduced compliance and/or muscle wisdom, where the central nervous system decreases motor unit firing rates to compensate for the fatigue-induced increases in muscle fiber relaxation times (Perry-Rana et al. 2002). Beck et al. (2004) followed up the study by Perry-Rana et al. (2002) by examining the MMG amplitude and MPF responses for the biceps brachii during 50 consecutive maximal concentric isokinetic muscle actions of the forearm flexors at a velocity of 180°·s-1. The results showed that there were linear decreases in both MMG amplitude and MPF across the 50 repetitions, and a cubic decrease in isokinetic peak torque of approximately 70%. Thus, it was hypothesized that the decreases in MMG amplitude and MPF could have been due to fatigue-induced reductions in motor unit firing rates and/or decreased compliance in the biceps brachii muscle. It was also suggested, however, that de-recruitment of fast-twitch motor units with fatigue could, theoretically, have caused the reductions in MMG amplitude and MPF (Beck et al. 2004). Perry-Rana et al. (2003) also examined the patterns of responses for MMG amplitude from the rectus femoris, vastus lateralis, and vastus medialis during 25 consecutive maximal eccentric isokinetic muscle actions of the leg extensors at a velocity of 120°·s-1. The results showed that for the vastus lateralis and vastus medialis, MMG amplitude decreased linearly across the 25 repetitions. The rectus femoris muscle, however, showed a cubic pattern for MMG amplitude that was characterized

Mechanomyographic responses during dynamic muscle actions

79

Figure 4. Changes in mechanomyographic (MMG) amplitude for the vastus lateralis (solid line), rectus femoris (dotted line), and vastus medialis (dashed line) during 50 consecutive maximal concentric isokinetic muscle actions of the leg extensors at velocities of 60 (A), 180 (B), and 300掳路s-1 (C). Notice that MMG amplitude decreased at all velocities due to fatigue, but the patterns were different for each muscle. *Reprinted with permission from Perry-Rana et al. (2002).

by a slight increase during repetitions 1-5, a decrease during repetitions 5-20, and another increase during repetitions 20-25. In addition, the pattern for eccentric isokinetic peak torque was best fit with a cubic model, where peak torque increased during the first 10 repetitions, and then remained relatively stable from repetitions 11-25. Thus, it was suggested that the decreases in MMG amplitude for each muscle may have been due to the effects of muscle wisdom and/or reduced muscle compliance. Furthermore, the muscle-specific differences in the MMG amplitude patterns could have been due to differences among the three muscles for fiber type composition and/or architecture (PerryRana et al. 2003).

Several studies have also examined the MMG amplitude and/or center frequency responses during incremental or constant power output cycle ergometry. For example, Stout et al. (1997) reported that during an incremental cycle ergometer test, both MMG amplitude for the vastus lateralis and VO2 increased linearly with power output. Furthermore, 20 of the 24 subjects that participated in the study showed linear slope coefficients for the normalized MMG amplitude and VO2 versus power output relationships that were statistically equivalent. Thus, it was concluded that MMG may be useful for quantifying muscular activity and monitoring changes in exercise intensity during incremental cycle ergometry. In addition, the similar slope coefficients for the normalized MMG amplitude and VO2 versus power output relationships for most subjects suggested that there may be a close relationship

80

Travis W. Beck

between the metabolic and mechanical aspects of muscle activity during cycle ergometry (Stout et al. 1997). Interestingly, Shinohara et al. (1997) also found that MMG amplitude for the vastus lateralis increased linearly with power output during incremental cycle ergometry. In fact, MMG amplitude was more linearly related to power output than was EMG amplitude, which tended to increase curvilinearly at high power outputs (Shinohara et al. 1997). Perry et al. (2001b) reported similar results for the vastus lateralis during incremental cycle ergometry. In particular, the majority of the subjects that were tested (7 out of 9) showed linear increases in MMG amplitude with power output, but EMG amplitude increased curvilinearly. The authors also found that the linear slope coefficients for the increases in normalized heart rate and ratings of perceived exertion were statistically equivalent to that for the normalized MMG amplitude versus power output relationship. Thus, it was concluded that during incremental cycle ergometry, there may be close relationships among the mechanical (MMG), heart rate, and perception of effort aspects of muscle activity (Perry et al. 2001b). In addition, the results from a second study (Perry et al. 2001c) showed that there was no change in MMG MPF for the vastus lateralis, but a linear increase in MMG amplitude with increases in power output during an incremental cycle ergometer test that was performed to exhaustion. Thus, it was concluded that during incremental cycle ergometry, motor unit recruitment may be the primary mechanism for increasing power output, rather than changes in motor unit firing rates (Perry et al. 2001c).

Previous studies have also examined the MMG amplitude responses during continuous cycle ergometer workbouts performed at constant, submaximal power outputs. Specifically, Housh et al. (2000) investigated the patterns of responses for MMG amplitude for the vastus lateralis and vastus medialis during continuous, constant power output workbouts at 50, 65, 80, and 95% of the peak power achieved during an incremental cycle ergometer test performed to exhaustion. Interestingly, the MMG amplitude responses were dependent on the power output at which the workbout was performed, as well as the muscle that was examined. For example, MMG amplitude for the vastus lateralis and vastus medialis muscles decreased during the workbouts at 50% and 65% of the peak power, but remained stable at 80% of the peak power. At 95% of the peak power, however, MMG amplitude increased for the vastus medialis muscle, but remained relatively stable for the vastus lateralis.

Thus, it was suggested that the decreases in MMG amplitude for both muscles at 50% and 65% of the peak power may have been due to fatigue-induced decreases in motor unit firing rates. At 80% of the peak power, however, the lack of a significant change in MMG amplitude for either the vastus lateralis or vastus medialis may have been due to a balance between the influences of recruitment, which can increase MMG amplitude, and decreases in motor unit

Mechanomyographic responses during dynamic muscle actions

81

firing rates, which can decrease MMG amplitude. Furthermore, the increases in MMG amplitude for the vastus medialis muscle at 95% of the peak power indicated that recruitment may have had a greater influence on MMG amplitude than potential fatigue-induced decreases in motor unit firing rates. In contrast, the vastus lateralis showed no change in MMG amplitude at 95% of the peak power, and it was suggested that the tendonous iliotibial band that covers the vastus lateralis could have affected the muscle fiber oscillations that were being transmitted to the skin surface, thereby influencing MMG amplitude (Housh et al. 2000). Perry et al. (2001a) also examined the MMG amplitude responses from the vastus lateralis muscle during continuous, constant power output cycle ergometry. The results showed that MMG amplitude decreased during the continuous workbouts at 28%, 35%, and 42% of the peak power output. Thus, these results were consistent with those of Housh et al. (2000) for the vastus lateralis and vastus medialis muscles at 50% and 65% of the peak power output, and it was suggested that the decreases in MMG amplitude may have been due to the effects of muscle wisdom and/or reductions in muscle compliance (Perry et al. 2001a). Bull et al. (2000) investigated the MMG amplitude responses for the vastus lateralis muscle during continuous cycle ergometry at a submaximal workload known as critical power. Theoretically, critical power is the maximal power output that can be accomplished without fatigue, and, therefore, should be characterized by steady state VO2 and no change in muscle activation. Thus, MMG amplitude should also remain stable at critical power. Bull et al. (2000), however, reported that there was a quadratic decrease in MMG amplitude for the vastus lateralis during a 60-minute cycle ergometer workbout at critical power. It is important to note that critical power is often a fatiguing workload, and it is possible that the fatigue-related factors, such as muscle wisdom and/or reduced muscle compliance could have contributed to the decreases in MMG amplitude reported by Bull et al. (2000) for the vastus lateralis muscle.

Several studies have also investigated the possibility of using MMG to examine the neuromuscular adaptations that occur during a resistance training program. For example, Cerquiglini et al. (1973) investigated the effects of two months of DCER strength training on MMG frequency in two sedentary subjects and two Olympic weightlifters. The subjects were tested weekly for maximum isometric strength in two “typical lifting positions” while MMG signals were detected from the vastus lateralis and the medial head of the gastrocnemius. The results showed that the resistance training program caused a “…relative increase of higher frequencies (above 70 Hz)” in the MMG signals from the vastus lateralis and gastrocnemius muscles. Thus, it was suggested that the frequency content of the MMG signal could potentially be

82

Travis W. Beck

used by trainers and/or athletes to monitor the changes in muscle function that occur during a resistance training program (Cerquiglini et al. 1973). Evetovich et al. (1998, 2000) also examined the effects of resistance training on MMG amplitude and MPF. Specifically, the subjects were randomly assigned into either a training or a control group. The subjects in the training group performed maximal concentric isokinetic leg extensions of the non-dominant limb at a velocity of 90°·s-1 three times per week for twelve weeks. The subjects in both the training and control groups were tested for maximal concentric isokinetic leg extension peak torque at a velocity of 90°·s-1 prior to the training program and every four weeks thereafter, while MMG signals were detected from the vastus lateralis. The results showed that there was a significant increase in leg extension peak torque following the training for the training group, but not for the control group. There were, however, no changes in MMG amplitude or MPF for the vastus lateralis muscle from week 0 to week 12 for either group. Thus, it was hypothesized that the lack of significant changes in MMG amplitude and MPF for the vastus lateralis muscle following training could have been due to compression of hypertrophied fibers by the iliotibial band and/or training-induced adaptations in muscles other than the vastus lateralis (Evetovich et al. 1998, 2000). Esposito et al. (2005) examined the effects of a dynamic resistance training program on the MMG amplitude and MPF versus isometric torque relationships for the vastus lateralis muscle in elderly men. Specifically, the training program required the subjects to perform maximal concentric isokinetic muscle actions of the dominant leg extensors at velocities of 120 and 240°·s-1. This training was performed twice per week for twelve weeks. The results showed that the training program had no effect on the MMG amplitude values for the vastus lateralis at any of the relative torque levels (20, 40, 60, 80, and 100% MVC). At 80% and 100% MVC, however, the training program caused significant increases in MMG MPF. In addition, the MMG power density spectrum for the vastus lateralis during an isometric MVC before training was unimodal, with a well-defined peak at about 11 Hz. After training, however, the MMG power density spectrum became bimodal, with a large peak at approximately 15 Hz and a smaller peak at about 30 Hz. Thus, it was concluded that the training-induced increases in MMG MPF for the vastus lateralis at 80% and 100% MVC, combined with the changes in the shape of the MMG power density spectrum, reflected a “retrieval” of fasttwitch motor units, which can be lost in some muscles during the aging process (Esposito et al. 2005). Coburn et al. (2006) recently examined the effects of three days of velocity-specific isokinetic training on the MMG amplitude and MPF values from the rectus femoris, vastus lateralis, and vastus medialis muscles. Specifically, the subjects were randomly assigned to one of three

Mechanomyographic responses during dynamic muscle actions

83

groups: a) a control group, b) a slow velocity training group, or c) a fast velocity training group. The subjects in the two training groups performed three separate training sessions that consisted of maximal concentric isokinetic muscle actions of the nondominant leg extensors at a velocity of 30 (slow velocity training group) or 270°·s-1 (fast velocity training group). In addition, the subjects in the training groups were tested for maximal concentric isokinetic leg extension peak torque at velocities of 30, 150, and 270°·s-1 prior to, and following the training program. The subjects in the control group were also tested for leg extension peak torque at the same velocities, but did not perform any training. The results showed that there were training-induced increases in leg extension peak torque for the fast velocity training group at 270°·s-1 and for the slow velocity training group at 30, 150, and 270°·s-1. In addition, when compared to the control group, there were training-induced increases in MMG amplitude (averaged across the vastus lateralis, rectus femoris, and vastus medialis) for the fast velocity training group at 270°·s-1 and for the slow velocity training group at 150°·s-1. The isokinetic leg extension training had no effect, however, on the mean MMG MPF values for any of the muscles at any velocity. Thus, it was concluded that the training-induced increases in MMG amplitude may have been due to reduced muscle compliance. Specifically, it is possible that reduced coactivation in the biceps femoris, semitendinosus, and semimembranosus following the isokinetic training could have increased the net leg extension peak torque and reduced compliance in the quadriceps femoris muscles, thereby increasing MMG amplitude (Coburn et al. 2006).

Several recent studies have also examined the acute effects of stretching on MMG amplitude and/or MPF. For example, Evetovich et al. (2003) investigated the acute effects of static stretching of the forearm flexors on isokinetic peak torque, MMG amplitude, and EMG amplitude during maximal concentric isokinetic muscle actions of the forearm flexors at velocities of 30 and 270°·s-1. The results showed that the stretching caused a significant decrease in forearm flexion peak torque and increases in MMG amplitude for the biceps brachii muscle at both velocities. The stretching had no effect, however, on the EMG amplitude values for the biceps brachii muscle at either velocity. Thus, it was concluded that the stretching-induced decreases in forearm flexion peak torque and increases in MMG amplitude for the biceps brachii may have been due to reduced muscle stiffness (Evetovich et al. 2003).

Cramer et al. (2005) examined the acute effects of static stretching of the dominant leg extensors on peak torque, mean power output, MMG amplitude, and EMG amplitude during maximal concentric isokinetic leg extensions at velocities of 60 and 240°·s-1. The results showed that the stretching caused

84

Travis W. Beck

decreases in peak torque for the stretched limb at 60 and 240°·s-1, as well as for the unstretched limb at 60°·s-1. There were also stretching-induced decreases in EMG amplitude for the rectus femoris and vastus lateralis muscles of both the stretched and unstretched limbs at 60 and 240°·s-1. The stretching had no effect, however, on mean power output and MMG amplitude for the rectus femoris or vastus lateralis muscles for either limb at 60 and 240°·s-1. Thus, it was hypothesized that the decreases in leg extension peak torque and EMG amplitude following the static stretching may have been due, at least partially, to reduced muscle activation in the rectus femoris and vastus lateralis muscles (Cramer et al. 2005). Marek et al. (2005) examined the acute effects of static and proprioceptive neuromuscular facilitation (PNF) stretching on peak torque, mean power output, EMG amplitude, and MMG amplitude during maximal concentric isokinetic leg extensions at velocities of 60 and 300°·s-1. The results showed that both static and PNF stretching caused decreases in PT, mean power output, and EMG amplitude for the rectus femoris and vastus lateralis muscles at 60 and 300°·s-1. There was also an increase in MMG amplitude following the static stretching, but only for the rectus femoris muscle at 60°·s-1.

Thus, it was concluded that the stretching-induced decreases in peak torque, mean power output, and EMG amplitude for the rectus femoris and vastus lateralis may have been due to a combination of reduced muscle activation and decreases in muscle stiffness (Marek et al. 2005).

Several studies have also used dynamic muscle actions to test the effects of different interventions on the MMG signal. For example, Bajaj et al. (2002) examined the MMG amplitude responses for the first dorsal interosseous during a series of concentric, isometric, and eccentric muscle actions. For each series of muscle actions, the subjects performed a single abduction movement (concentric muscle action), followed immediately by a 2-sec isometric muscle action of the first dorsal interosseous with the index finger in the fully abducted position, and then a single adduction movement of the index finger (eccentric muscle action). These muscle actions were performed at four different relative torque levels (0, 25, 50, 75, and 100% of the isometric MVC) prior to, immediately following, and 24 and 48 hours after a series of maximal eccentric muscle actions of the first dorsal interosseous. The purpose of the eccentric muscle actions was to elicit muscle damage and cause delayed-onset muscle soreness. The results showed that the mean MMG amplitude values (averaged across the type of muscle action and percentage of the MVC) immediately after the eccentric muscle actions were greater than the values recorded before exercise and following 24 and 48 hours of rest. Thus, it was hypothesized that the increase in MMG amplitude immediately after the eccentric exercise could have been due to greater physiological tremor and/or increased edema in the muscle fibers of the first dorsal interosseous. The

Mechanomyographic responses during dynamic muscle actions

85

authors also reported increased subjective pain ratings immediately after the eccentric exercise. Thus, it was suggested that the eccentric exercise may have allowed for greater muscle fiber vibrations that resulted in activation of “mechanosensitive deep tissue nociceptors”. Activation of these receptors could then have contributed to the increased pain sensations felt immediately after the eccentric exercise (Bajaj et al. 2002). Petitjean et al. (1992) used MMG to examine “phonomechanical delay” (i.e., the time interval between the onsets of the MMG and acceleration signals) in the biceps brachii and brachioradialis during submaximal concentric DCER muscle actions of the forearm flexors performed against a 3-kg weight. The concentric muscle actions required the subjects to accelerate the weight at slow (20-120 radians·s2 ), intermediate (120-240 radians·s-2), and fast (240-360 radians·s-2) angular velocities. The results showed that the phonomechanical delay for the biceps brachii and brachioradialis muscles increased with acceleration, and it was hypothesized that the onsets of the MMG signals may have reflected the development of tension in the contractile components of the muscles. Thus, it was hypothesized that phonomechanical delay could potentially reflect the contributions of fast versus slow-twitch muscle fibers to dynamic torque production (Petitjean et al. 1992).

In addition, Vedsted et al. (2006) recently investigated the MMG amplitude responses for the biceps brachii during concentric, isometric, and eccentric DCER muscle actions of the forearm flexors at 10% and 20% of the isometric MVC. Muscle tissue oxygenation levels were also measured from the biceps brachii to provide information regarding the energy requirements of each type of muscle action.The results showed that the mean MMG amplitude values for the biceps brachii were greater during the concentric and eccentric muscle actions than during the isometric muscle actions. There were no differences, however, among the concentric, isometric, and eccentric muscle actions for muscle tissue oxygenation levels. Thus, it was suggested that during the concentric and eccentric muscle actions, low motor unit firing rates in the biceps brachii could have resulted in less fusion of motor unit twitches and a “…more distinct mechanical twitching…” that allowed larger muscle fiber oscillations and greater MMG amplitude values. For the isometric muscle actions, however, high motor unit firing rates could have resulted in greater fusion of motor unit twitches and reduced MMG amplitude values. Therefore, it was concluded that selective recruitment of superficially located fast-twitch motor units, particularly during the eccentric muscle actions, could have allowed the active motor units to “…influence more directly the MMG generation process”, thereby resulting in greater MMG amplitude values (Vedsted et al. 2006). Evetovich et al. (2002, 2004) recently examined the

86

Travis W. Beck

effects of dehydration and hyperhydration on concentric isokinetic forearm flexion peak torque, MMG amplitude, and MMG MPF for the biceps brachii muscle. Isokinetic peak torque was assessed at a velocity of 90°·s-1, and dehydration was achieved through restriction of water intake. The subjects were considered dehydrated if there was a change in body weight of more than 2.0%, and the urine specific gravity was greater than 1.020. In addition, hyperhydration was achieved through injection of glycerol, which causes fluid retention in all water compartments of the body. The results showed that neither dehydration nor hyperhydration had any effect on isokinetic peak torque, MMG amplitude, or MMG MPF. Thus, it was concluded that the MMG signal may be influenced more by motor control strategies and the intrinsic contractile properties of muscle than by the fluids that surround muscle fibers (Evetovich et al. 2002, 2004).

Ebersole and Malek (2008) recently used MMG to examine changes in electromechanical efficiency during fatiguing dynamic muscle actions. The subjects were required to perform 75 consecutive maximal concentric isokinetic muscle actions of the leg extensors at a velocity of 180°·s-1. During each muscle action, surface EMG and MMG signals were detected from the vastus lateralis and vastus medialis, and the ratio of MMG amplitude/EMG amplitude was calculated for both muscles as a measure of electromechanical efficiency. The results showed that the ratio of MMG amplitude/EMG amplitude decreased in a quadratic fashion for the vastus medialis, while the corresponding pattern for the vastus lateralis was best fit with a cubic model.

Thus, the authors performed a log transformation of the MMG amplitude/ EMG amplitude ratio data, which forced the patterns of responses during the fatiguing task to be linear. Even after the log transformation, however, the linear slope coefficients for the MMG amplitude/EMG amplitude ratio were the same for the vastus lateralis and vastus medialis. Thus, it was suggested that the decreases in the MMG amplitude / EMG amplitude ratios for the vastus lateralis and vastus medialis were due to fatigue of fast-twitch muscle fibers and their subsequent inability to produce oscillations and contribute to torque production with fatigue. It was also hypothesized, however, that the MMG and EMG responses could have reflected changes in motor control strategies due to fatigue. Regardless of the exact mechanism, these results suggested that the MMG amplitude / EMG amplitude ratio may be capable of discriminating between healthy and injured muscle in individuals that suffer from atrophy or some type of dysfunction of the vastus lateralis or vastus medialis that could potentially cause patellofemoral pain (Ebersole and Malek 2008).

Kawczyński et al. (2007) recently investigated the effects of eccentric exercise on MMG and EMG amplitude for the trapezius muscle. Specifically,

Mechanomyographic responses during dynamic muscle actions

87

the subjects were required to perform shoulder elevation muscle actions (both dynamic and isometric) prior to, and immediately following a bout of eccentric muscle actions that was designed to induce delayed-onset muscle soreness.

The concentric and isometric muscle actions were repeated 24 and 48 hours after the eccentric exercise. The results showed that MMG amplitude was higher immediately after the eccentric exercise when compared with the corresponding values before exercise, as well as 24 and 48 hours after exercise.

There were no changes, however, in the mean MMG MPF, EMG amplitude, or EMG MPF values after the eccentric exercise. Thus, it was suggested that the eccentric exercise likely changed the viscoelastic properties of the trapezius muscle, thereby allowing greater muscle fiber oscillations and increases in MMG amplitude (Kawczyński et al. 2007). Hendrix et al. (2008) performed an interesting study to examine the effects of pedaling cadence and power output on the MMG amplitude and MPF responses for the vastus lateralis during submaximal cycle ergometry. The subjects were required to perform an incremental cycle ergometer test to exhaustion, as well as continuous, 8-minute rides at power outputs that corresponded to 35%, 50%, and 65% of the peak power output achieved during the incremental test to exhaustion. In addition, during each 8-minute ride at a constant power output, the subjects pedaled at either 50 or 70 revolutions per minute during the first 4 minutes, and then switched to the alternate cadence during the second 4 minutes. The results showed that MMG amplitude was closely related to power output, but not pedaling cadence, and MMG MPF was not affected by power output or pedaling cadence. Thus, it was concluded that during cycle ergometry, MMG amplitude is not affected by pedaling cadence, but is greatly influenced by the changes in muscle activity that must accompany increases in power output (Hendrix et al. 2007).

Cramer et al. (2007c) recently examined the effects of three days of resistance training with or without creatine supplementation on isokinetic leg extension peak torque, mean power output, and EMG and MMG amplitude and median frequency. The subjects were required to perform maximal concentric isokinetic muscle actions of the leg extensors at velocities of 30, 150, and 270°·s-1 prior to, and following three days of resistance training (three sets of ten maximal muscle actions at a velocity of 150°·s-1). The resistance training was performed with either creatine supplementation or a placebo. The results showed that the creatine supplementation had no effect on isokinetic peak torque, mean power output, or any of the EMG and MMG parameters.

There were, however, significant increases in isokinetic peak torque, as well as EMG and MMG median frequency for both the creatine and placebo groups, but these changes were not consistent at all velocities. Thus, it was concluded

88

Travis W. Beck

that creatine supplementation combined with a short-term resistance training program had no effect on EMG and MMG amplitude and median frequency (Cramer et al. 2007c). A recent study by Cramer et al. (2007a) also examined the acute effects of static stretching on the isokinetic angle-torque relationship, as well as surface EMG and MMG amplitude values. The subjects were required to perform maximal concentric isokinetic muscle actions of the leg extensors at velocities of 1.04 and 5.23 radians per second prior to, and immediately following static stretching of the leg extensors. During all muscle actions, surface EMG and MMG signals were detected from the rectus femoris muscle. The results showed that stretching caused a decrease in isokinetic peak torque, acceleration time, and EMG amplitude, but there were no changes in work, the joint angle at peak torque, or MMG amplitude. Thus, it was concluded that the stretching-induced changes in peak torque and acceleration time were not consistently reflected in the corresponding EMG and MMG signals (Cramer et al. 2007a). This study was followed up by a second investigation (Cramer et al. 2007b) that examined the acute effects of static stretching of the leg extensors on maximal eccentric isokinetic peak torque (60 and 180掳路s-1 velocities), the joint angle at peak torque, mean power output, and MMG and EMG amplitude and MPF for the vastus lateralis and rectus femoris. The results showed that the stretching had no effect on peak torque, the joint angle at peak torque, mean power output, or EMG and MMG amplitude and MPF. Thus, it was concluded that static stretching did not affect maximal eccentric PT and mean power output, nor did it have any effect on muscle activation levels during eccentric muscle actions (Cramer et al.

2007b). Rana (2006) recently investigated the MMG and EMG amplitude responses for the vastus lateralis, rectus femoris, and vastus medialis during the Wingate Test (i.e., 30 seconds of maximal cycling). The results showed that although there were decreases in power output over the 30-second test, there was no change in EMG amplitude. The MMG amplitude values for each muscle, however, decreased during the test, with the rectus femoris showing the greatest decrease. Thus, it was concluded that MMG may provide more useful information regarding fatigue-related changes in the mechanical properties of muscle than EMG, and could be reflective of changes in power output for each muscle. Therefore, the results from this study provided yet another indication that during dynamic muscle actions, MMG amplitude is closely related to muscle power output (Rana 2006).

Kimura et al. (2008) examined the MMG amplitude responses for the vastus lateralis and rectus femoris during incremental cycle ergometry. The subjects were required to perform an incremental cycle test in which the power output started at 20 Watts and was increased by 30 Watts every minute until the subject could no longer continue. The results showed that MMG amplitude

Mechanomyographic responses during dynamic muscle actions

89

for both the vastus lateralis and rectus femoris increased nonlinearly with workload. In addition, the onset of the ventilatory threshold corresponded with abrupt changes in the MMG signals for both muscles, as MMG amplitude began to plateau for the vastus lateralis, but increased rapidly for the rectus femoris. Thus, it was concluded that MMG may reflect the fatigue-related changes in the muscle’s mechanical properties that occur with fatigue at the ventilatory threshold. In addition, since electret condenser microphones were used to detect the MMG signals from the rectus femoris and vastus lateralis, these sensors may be more appropriate than accelerometers when recording MMG signals during dynamic muscle actions (Kimura et al. 2008).

The results from the studies that were reviewed in this chapter indicated that MMG provides valuable information regarding muscle activity during dynamic muscle actions. A particularly interesting relationship is that between MMG amplitude and muscle power output. Given that many training programs are designed to increase power output, rather than just strength, it is possible that MMG may be useful for monitoring the training status of individual muscles or muscle groups. Certainly, more work needs to be done before this application can be used on a widespread basis, but it is suggested that future work in the area of dynamic MMG should focus on identifying the mechanisms underlying the relationship between MMG amplitude and muscle power output.

References 1.

2.

3.

4.

5.

Bajaj P, Madeleine P, Sjøgaard G, Arendt-Nielsen L. Assessment of postexercise muscle soreness by electromyography and mechanomyography. Journal of Pain 2002; 3:126-136.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Mechanomyographic amplitude and mean power frequency versus torque relationships during isokinetic and isometric muscle actions of the biceps brachii.

Journal of Electromyography and Kinesiology 2004; 14:555-564.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Mechanomyographic and electromyographic time and frequency domain responses during submaximal to maximal isokinetic muscle actions of the biceps brachii.

European Journal of Applied Physiology 2004; 92:352-359.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Mechanomyographic and electromyographic amplitude and frequency responses during fatiguing isokinetic muscle actions of the biceps brachii. Electromyography and Clinical Neurophysiology 2004; 44:431-441.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Mechanomyographic and electromyographic responses during submaximal to maximal eccentric isokinetic muscle actions of the biceps brachii. Journal of Strength and Conditioning Research 2006; 20:184-191.

90

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

Travis W. Beck

Bull AJ, Housh TJ, Johnson GO, Perry SR. Electromyographic and mechanomyographic responses at critical power. Canadian Journal of Applied Physiology 2000; 25:262270.

Cerquiglini S, Figura F, Marchetti M, Salleo A. Evaluation of athletic fitness in weight-lifters through biomechanical, bioelectrical, and bioacoustical data. In Biomechanics III Edited by: Cerquiglini S, Venerando A,Wartenweiler J. Baltimore: University Park Press; 1973; 189-195.

Coburn JW, Housh TJ, Cramer JT, Weir JP, Miller JM, Beck TW, Malek MH, Johnson GO. Mechanomyographic time and frequency domain responses of the vastus medialis muscle during submaximal to maximal isometric and isokinetic muscle actions. Electromyography and Clinical Neurophysiology 2004; 44:247-255.

Coburn JW, Housh TJ, Malek MH, Weir JP, Cramer JT, Beck TW, Johnson GO. Neuromuscular responses to three days of velocity-specific isokinetic training.

Journal of Strength and Conditioning Research 2006; 20:892-898.

Coburn JW, Housh TJ, Weir JP, Malek MH, Cramer JT, Beck TW, Johnson GO. Mechanomyographic responses of the vastus medialis to isometric and eccentric muscle actions. Medicine & Science In Sports & Exercise 2004; 36:1916-1922.

Cramer JT, Beck TW, Housh TJ, Massey LL, Marek SM, Dangelmeier S, Purkayastha S, Culbertson JY, Fitz KA, Egan AD. Acute effects of static stretching on characteristics of the isokinetic angle-torque relationship, surface electromyography, and mechanomyography. Journal of Sports Sciences 2007a; 25:687-698.

Cramer JT, Housh TJ, Evetovich TK, Johnson GO, Ebersole KT, Perry SR, Bull AJ.The relationships among peak torque, mean power output, mechanomyography, and electromyography in men and women during maximal, eccentric isokinetic muscle actions. European Journal of Applied Physiology 2002; 86:226-232.

Cramer JT, Housh TJ, Johnson GO, Ebersole KT, Perry SR, Bull AJ. Mechanomyographic amplitude and mean power output during maximal, concentric, isokientic muscle actions. Muscle & Nerve 2000a; 23:1826-1831.

Cramer JT, Housh TJ, Johnson GO, Ebersole KT, Perry SR, Bull AJ. Mechanomyographic and electromyographic responses of the superficial muscles of the quadriceps femoris during maximal, concentric isokinetic muscle actions.

Isokinetics and Exercise Science 2000b; 8:109-117.

Cramer JT, Housh TJ, Johnson GO, Weir JP, Beck TW, Coburn JW. An acute bout of static stretching does not affect maximal eccentric isokinetic peak torque, the joint angle at peak torque, mean power, electromyography, or mechanomyography. Journal of Orthopaedic and Sports Physical Therapy 2007b; 37:130-139.

Cramer JT, Housh TJ, Weir JP, Ebersole KT, Perry-Rana SR, Bull AJ, Johnson GO.Cross-correlation analyses of mechanomyographic signals from the superficial quadriceps femoris muscles during concentric and eccentric isokinetic muscle actions. Electromyography and Clinical Neurophysiology 2003;43:293-300.

Cramer JT, Housh TJ, Weir JP, Johnson GO, Berning JM, Perry SR, Bull AJ. Mechanomyographic and electromyographic amplitude and frequency responses from the superficial quadriceps femoris muscles during maximal, eccentric isokinetic muscle actions. Electromyography and Clinical Neurophysiology 2002; 42:337-346.

Mechanomyographic responses during dynamic muscle actions

91

18. Cramer JT, Housh TJ, Weir JP, Johnson GO, Berning JM, Perry SR, Bull AJ. Gender, muscle, and velocity comparisons of mechanomyographic and electromyographic responses during isokientic muscle actions. Scandinavian Journal of Medicine In Science In Sports 2004; 14:116-127.

19. Cramer JT, Housh TJ, Weir JP, Johnson GO, Coburn JW, Beck TW. The acute effects of static stretching on peak torque, mean power output, electromyography, and mechanomyography. European Journal of Applied Physiology 2005;93:530539.

20. Cramer JT, Stout JR, Culbertson JT, Egan AD. Effects of creatine supplementation and three days of resistance training on muscle strength, power output, and neuromuscular function. Journal of Strength and Conditioning Research 2007c; 21:668-677.

21. Dalton PA, Stokes MJ. Acoustic myography reflects force changes during dynamic concentric and eccentric contractions of the human biceps brachii muscle.

European Journal of Applied Physiology 1991; 63:412-416.

22. Dalton PA, Stokes MJ. Frequency of acoustic myography during isometric contraction of fresh and fatigued muscle and during dynamic contractions. Muscle & Nerve 1993; 16:255-261.

23. Ebersole KT, Housh TJ, Johnson GO, Evetovich TK, Smith DB. The mechanomyographic and electromyographic responses to passive leg extension movements. Isokinetics and Exercise Science 2001; 9:11-18.

24. Ebersole KT, Housh TJ, Weir JP, Johnson GO, Evetovich TK, Smith DB. The effects of leg angular velocity on mean power frequency and amplitude of the mechanomyographic signal. Electromyography and Clinical Neurophysiology 2000; 40:49-55.

25. Ebersole KT, Malek DM. Fatigue and the electromechanical efficiency of the vastus medialis and vastus lateralis muscles. Journal of Athletic Training 2008; 43:152-156.

26. Esposito F, Cè E, Gobbo M, Veicsteinas A, Orizio C. Surface EMG and mechanomyogram disclose isokinetic training effects on quadriceps muscle of elderly people. European Journal of Applied Physiology 2005; 94:549-557.

27. Evetovich TK, Boyd JC, Drake SM, Eschbach LC, Magal M, Soukup JT, Webster MJ, Whitehead MT, Weir JP. Effect of moderate dehydration on torque, electromyography, and mechanomyography. Muscle & Nerve 2002; 26:225-231.

28. Evetovich TK, Housh TJ, Johnson GO, Housh DJ, Ebersole KT, Smith DB. The effects of concentric isokinetic strength training of the quadriceps femoris on mechanomyography and muscle strength. Isokinetics and Exercise Science 1998; 7:123-128.

29. Evetovich TK, Housh TJ, Johnson GO, Smith DB, Ebersole KT, Perry SR. Gender comparisons of the mechanomyographic responses to maximal concentric and eccentric isokinetic muscle actions. Medicine & Science In Sports & Exercise 1998; 30:1697-1702.

30. Evetovich TK, Housh TJ, Stout JR, Johnson GO, Smith DB, Ebersole KT. Mechanomyographic responses to concentric isokinetic muscle contractions.

European Journal of Applied Physiology 1997; 75:166-169.

92

Travis W. Beck

31. Evetovich TK, Housh TJ, Weir JP, Housh DJ, Johnson GO, Ebersole KT, Smith DB. The effect of leg extension training on the mean power frequency of the mechanomyographic signal. Muscle & Nerve 2000; 23:973-975.

32. Evetovich TK, Housh TJ, Weir JP, Johnson GO, Smith DB, Ebersole KT. Mean power frequency and amplitude of the mechanomyographic signal during maximal eccentric isokinetic muscle actions.

Electromyography and Clinical Neurophysiology 1999; 39:123-127.

33. Evetovich TK, Nauman NJ, Conley DS, Todd JB. Effect of static stretching of the biceps brachii on torque, electromyography, and mechanomyography during concentric isokinetic muscle actions. Journal of Strength and Conditioning Research 2003; 17:484-488.

34. Evetovich TK, Whitehead MT, Webster MJ, Soukup JT, Magal M, Eschbach LC, Drake SM, Boyd SC, Weir JP, Hennerichs KR. The effect of glycerol on torque, electromyography, and mechanomyography. Journal of Strength and Conditioning Research 2004; 18:741-746.

35. Hendrix CR, Bull AJ, Housh TJ, Rana SR, Cramer JT, Beck TW, Weir JP, Malek MH, Mielke M. The effect of pedaling cadence and power output on mechanomyographic amplitude and mean power frequency during submaximal cycle ergometry. Electromyography and Clinical Neurophysiology 2008; 48:195-201.

36. Housh TJ, Perry SR, Bull AJ, Johnson GO, Ebersole KT, Housh DJ, de Vries HA. Mechanomyographic and electromyographic responses during submaximal cycle ergometry. European Journal of Applied Physiology 2000; 83:381-387.

37. Kawczyński A, Nie H, Jaskólska A, Jaskólski A, Arendt-Nielsen L, Madeleine P. Mechanomyography and electromyography during and after fatiguing shoulder eccentric contractions in males and females. Scandinavian Journal of Medicine and Science In Sports 2007; 17:172-179.

38. Kimura T, Fujibayashi M, Tanaka S, Moritani T. Mechanomyographic responses in quadriceps muscles during fatigue by continuous cycle exercise. European Journal of Applied Physiology 2008; 104:651-656.

39. Madeleine P, Bajaj P, Søgaard K, Arendt-Nielsen L. Mechanomyography and electromyography force relationships during concentric, isometric and eccentric contractions. Journal of Electromyography and Kinesiology 2001; 11:113-121.

40. Marek SM, Cramer JT Fincher AL, Massey LL, Dangelmeier SM, Purkayastha S, Fitz KA, Culbertson JY. Acute effects of static and proprioceptive neuromuscular facilitation stretching on muscle strength and power output. Journal of Athletic Training 2005; 40:94-103.

41. Perry-Rana SR, Housh TJ, Johnson GO, Bull AJ, Berning JM, Cramer JT. MMG and EMG responses during fatiguing isokinetic muscle contrations at different velocities. Muscle & Nerve 2002; 26:367-373.

42. Perry-Rana SR, Housh TJ, Johnson GO, Bull AJ, Cramer JT MMG and EMG responses during 25 maximal, eccentric, isokinetic muscle actions. Medicine & Science In Sports & Exercise 2003; 35:2048-2054.

43. Perry SR, Housh TJ, Johnson GO, Ebersole KT, Bull AJ. Mechanomyographic responses to continuous, constant power output cycle ergometry.

Electromyography and Clinical Neurophysiology 2001a; 41:137-144.

Mechanomyographic responses during dynamic muscle actions

93

44. Perry SR, Housh TJ, Johnson GO, Ebersole KT, Bull AJ, Evetovich TK, Smith DB. Mechanomyography, electromyography, heart rate, and ratings of perceived exertion during incremental cycle ergometry. Journal of Sports Medicine and Physical Fitness 2001b; 41:183-188.

45. Perry SR, Housh TJ, Weir JP, Johnson GO, Bull AJ, Ebersole KT. Mean power frequency and amplitude of the mechanomyographic and electromyographic signals during incremental cycle ergometry. Journal of Electromyography and Kinesiology 2001c; 11:299-305.

46. Petitjean M, Maton B, Cnockaert J-C.Evaluation of human dynamic contraction by phonomyography. Journal of Applied Physiology 1992; 73:2567-2573.

47. Rana SR. Effect of the Wingate Test on mechanomyography and electromyography. Journal of Strength and Conditioning Research 2006; 20:292-297.

48. Shinohara M, Kouzaki M, Yoshihisa T, Fukunaga T. Mechanomyography of the human quadriceps muscle during incremental cycle ergometry. European Journal of Applied Physiology 1997; 76(4):314-9.

49. Smith DB, Housh TJ, Johnson GO, Evetovich TK, Ebersole KT, Perry SR. Mechanomyographic and electromyographic responses to eccentric and concentric isokinetic muscle actions of the biceps brachii. Muscle & Nerve 1998; 21:1438-1444.

50. Smith DB, Housh TJ, Stout JR, Johnson GO, Evetovich TK, Ebersole KT. Mechanomyographic responses to maximal eccentric isokinetic muscle actions.

Journal of Applied Physiology 1997; 82:1003-1007.

51. Stout JR, Housh TJ, Johnson GO, Evetovich TK, Smith DB. Mechanomyography and oxygen consumption during incremental cycle ergometry. European Journal of Applied Physiology 1997; 76:314-319.

52. Vedsted P, Blangsted AK, Søgaard K, Orizio C, Sjøgaard G. Muscle tissue oxygenation, pressure, electrical, and mechanical responses during dynamic and static voluntary contractions. European Journal of Applied Physiology 2006; 96:165-177.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for examining muscle function, 2010, 95-107 Editor Travis.W. Beck. ISBN :978-81-7895-449-3

6

Technical aspects of surface mechanomyography Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081 USA

Abstract The technical aspects of detecting surface mechanomyographic (MMG) signals are important, particularly when a laboratory first decides to conduct MMG research, or when detecting MMG signals from muscles in new applications, such as for prosthesis control. Of particular importance is the amount of contact pressure applied over the sensor, since high contact pressures can attenuate the muscle fiber vibrations that generate the MMG signal. Recent studies have also shown that the MMG amplitude and frequency responses may be different, depending on whether an accelerometer, Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: tbeck@ou.edu

96

Travis W. Beck

piezoelectric contact sensor, or condenser microphone is used to detect the MMG signal. Several investigations have also used a laser displacement sensor to detect the MMG signal, since there is no contact pressure with this sensor, and it directly measures skin displacement in physiological units (i.e., millimeters or micrometers), rather than transducer-dependent units (i.e., volts or millivolts).

Introduction The technical aspects of detecting surface mechanomyographic (MMG) signals are important, particularly when a laboratory first decides to conduct MMG research, or when detecting MMG signals for a new application (e.g., controlling an externally-powered prosthesis). The first study to directly examine the technical aspects of MMG was Bolton et al. (1989). Specifically, the authors investigated the MMG responses from the thenar muscle group during supramaximal electrically-stimulated isometric twitches. The MMG signals were detected with either an electret condenser microphone or a piezoelectric crystal contact sensor. The results showed that the MMG responses for the electret condenser microphone were highly variable, as the sound wave amplitude varied 25% when different microphones of the same model were used. This finding highlighted the importance of accurately calibrating each MMG sensor prior to measuring any signals from contracting muscle. In addition, the authors found that contact pressure was very important in determining the response of the piezoelectric crystal contact sensor, as was the location of the sensor over the thenar muscles. Specifically, the investigators moved the electret condenser microphone over the center of the thenar muscle group, as well as in 1.5 cm increments that were parallel or at right angles to the long axis of the thumb. The results showed that MMG amplitude was greatest when the sensor was over the middle, or belly of the muscle, and then decreased rapidly as it was placed closer to the border of the muscle. In addition, the pressure of the microphone on the skin had a large effect on the amplitude of the MMG signal, as increases in pressure tended to result in greater MMG amplitude values. The authors also suggested that the entire MMG frequency range (i.e., 0-100 Hz) should be used, rather than highpass filtering the signal at an arbitrary cutoff of, for example, 5 or 10 Hz. In fact, the investigators found that even when frequencies below 1 Hz were eliminated, the resulting MMG signal was distinctly different from that when this frequency was passed. Thus, it was concluded that the contact pressure of the sensor and its location over the muscle, as well as the bandwidth of the MMG signal, are all important factors when collecting MMG data (Bolton et al. 1989). Accornero et al. (1989) were among the first to develop a single probe

The mechanisms underlying the surface mechanomyogram

97

Figure 1. Example of the sensor designed by Accornero et al. (1989) to simultaneously detect mechanomyographic (MMG) and electromyographic (EMG) signals. The sensor is a piezoceramic transducer with two active electrodes (E1 and E2) at its border and a ground electrode in the center. *Reprinted with permission from Accornero et al.

(1989).

that could be used to simultaneously detect surface MMG and electromyographic (EMG) signals. Specifically, the probe consisted of a piezoceramic disc that was glued to a flexible printed circuit board with three copper strips that contacted the skin surface. The piezoceramic disc served as the MMG sensor, and the flexible printed circuit board with the three copper strips was used to detect the surface EMG signal. In addition, the surface EMG electrodes were preamplified, which reduced the need for skin abrasion prior to recording (Figure 1).

The authors found that this device worked well for detecting MMG and EMG signals, and suggested that the probe could be useful when conducting studies on muscle fatigue, where MMG and EMG signals often show different responses (Accornero et al. 1989). Smith and Stokes (1993) also performed a very important study to examine the influence of contact pressure on the MMG amplitude responses at different submaximal isometric force levels.

Specifically, the subjects were required to perform separate isometric muscle actions of the leg extensors at 10%, 25%, 50%, 75%, and 100% of the maximum voluntary contraction (MVC), and an MMG signal was detected

98

Travis W. Beck

from the rectus femoris during each muscle action. In addition, three different contact pressures (180, 790, and 1200 Pascals) were applied to the MMG sensor to determine if contact pressure affected the patterns of responses for MMG amplitude versus isometric force. The results showed that the MMG amplitude versus isometric force relationships were very similar for the 180 and 790 Pascal contact pressures. When the contact pressure was increased to 1200 Pascals, however, the mean MMG amplitude values increased significantly at all force levels, and the patterns of responses became more curvilinear. Thus, it was concluded that contact pressure is a very important factor when examining the patterns of responses for MMG amplitude versus isometric force.In addition, there may be a threshold contact pressure that should not be exceeded when recording MMG signals, but this pressure is likely to be different for different subjects due to discrepancies in subcutaneous skinfold thickness (Figure 2).

Finally, the authors recommended that absolute MMG amplitude should not be used to predict force production by the muscle, since contact pressure and

Figure 2. Changes in mechanomyographic (MMG) amplitude (indicated as Integrated Acoustic Myography in this figure) for the rectus femoris muscle with increases in isometric leg extension force for two separate subjects (a and b, respectively). The solid symbols represent the data when the contact pressure of the sensor was 180 Pascals, and the open symbols show the data for a much higher contact pressure (790 Pascals in (a) and 1200 Pascals in (b)). Notice that for the low contact pressure, the increases in MMG amplitude with force were highly linear, but MMG amplitude decreased from 80-100% MVC for the higher contact pressures. *Reprinted with permission from Smith and Stokes (1993).

The mechanisms underlying the surface mechanomyogram

99

differences in skinfold thickness can have a large influence on these values (Smith and Stokes 1993). Watakabe et al. (1998) also addressed an important technical aspect of detecting MMG signals with a piezoelectric contact sensor and an accelerometer. Specifically, the primary purpose of the investigation was to determine exactly what physiological parameters (i.e., acceleration, velocity, or displacement) were being measured by a piezoelectric crystal contact sensor and an accelerometer.Thus, both the piezoelectric crystal contact sensor (Hewlett Packard, model 21050A) and accelerometer (Nihon Koden MT-3T) were tested on a sinusoidal vibration system, as well as during voluntary isometric muscle actions of the rectus femoris. The results showed that when the double integral of the accelerometer signal was calculated (i.e., to convert acceleration in m路s-2 to displacement in 碌m), the resulting MMG signal was very similar in shape to that obtained by the piezoelectric crystal contact sensor. Thus, it was concluded that the piezoelectric crystal contact sensor measures skin displacement during muscle contraction. In addition, the amplitude of the MMG signal detected during a voluntary muscle action increased progressively with contact pressures ranging from 0.5-1.5 Newtons, to the point where the MMG amplitude values for a contact pressure of 1.5 Newtons were nearly twice as large as those for a contact pressure of 0.5 Newtons. Thus, it was concluded that both the accelerometer and piezoelectric crystal contact sensor are adequate devices for detecting MMG signals.

However, the advantage of the accelerometer is that it provides a measurement in physical units (i.e., m路s-2), rather than transducer-dependent units (Watakabe et al. 1998). This study was followed up by a second investigation (Watakabe et al. 2001) that examined the characteristics of a condenser microphone (Matsushita Communication Industrial model WM-034B) and an accelerometer (Kistler Instrument model 8352A2). Specifically, the authors investigated the frequency response of the condenser microphone during mechanical sinusoidal vibration, as well as a voluntary muscle action of the forearm flexors. The results showed that during the sinusoidal vibration, the frequency response of the condenser microphone was highly dependent on the length of the air chamber used in the microphone, such that lengths of 15, 20, and 25 mm resulted in cutoff frequencies of 10, 8, and 4 Hz, respectively. In addition, during the voluntary muscle actions, the MMG signal from the condenser microphone was very similar in shape to the double integral of the corresponding signal from the accelerometer. Thus, it was concluded that with the condenser microphone, muscle vibrations cause skin displacement, resulting in pressure changes in the air chamber of the microphone. In addition, when the limb was moved to purposefully create movement artifact, the resulting amplitude of the MMG signal increased 7.7 to 12.3 times the

100

Travis W. Beck

baseline value. The movement artifact was much more severe, however, for the accelerometer than the condenser microphone. Thus, it was concluded that the condenser microphone is probably a much better sensor for detecting the MMG signal during dynamic muscle actions than is the accelerometer. In addition, the condenser microphone should have a diameter of at least 10 mm and a length of 15 mm in order to accurately cover the frequency range of MMG signals (Watakabe et al. 2001). The last study in the series that examined the responses of different MMG sensors investigated the MMG responses of an accelerometer (Kistler Instrument 8352A2) and a laser displacement sensor (Keyence Corporation LK-080). These devices were compared during sinusoidal mechanical vibration, as well as voluntary isometric muscle actions of the rectus femoris at 20%, 40%, and 60% of the isometric leg extension MVC. The results showed that the double integral of the accelerometer signal was very similar to the corresponding signal from the laser displacement sensor, and, therefore, the accelerometer accurately measured acceleration of the skin surface. In addition, the MMG signal from the accelerometer was gradually distorted when weights of 2, 4, 10, and 50 grams were applied over it. Thus, it was suggested that when detecting MMG signals, the weight of the accelerometer should not be greater than 5 grams (Watakabe et al. 2003).

Wee and Ashley (1990) investigated the possibility of cross-talk with surface MMG signals. One MMG sensor was placed over the biceps brachii muscle and a second sensor over the triceps brachii muscle. The subjects were then required to perform a sustained isometric muscle action of the forearm flexors in which they supported a load of 2.3 kg, and MMG signals were detected simultaneously from both the biceps brachii and triceps brachii. The results indicated that there was significant transmission of muscle vibrations from the biceps brachii to the triceps brachii during the forearm flexion muscle action. Thus, it was suggested that the transmission of muscle vibrations to distant tissues should be considered when recording MMG signals, particularly when the muscles of interest are in close proximity to each other (Wee and Ashley 1990). Inoue et al. (2000) examined the possibility of using an accelerometer to develop a portable MMG system that could be used to record MMG signals during daily life activities. Specifically, the system incorporated multiple sensors, including sensors for the electrocardiogram, limb acceleration, and angular speed, as well as the MMG signal. The results showed that the portable MMG system accurately measured MMG data during a variety of daily life activities, including walking, quiet sitting, getting on and off of a train, as well as isometric and dynamic training. Thus, it was concluded that the portable MMG system could be used to examine muscle function during various activities of daily living away from the laboratory

The mechanisms underlying the surface mechanomyogram

101

setting (Inoue et al. 2000). Madeleine et al. (2006) examined the influence of sensor location on the patterns of responses for MMG amplitude and center frequency [both mean power frequency (MPF) and median frequency] versus isometric force for the tibialis anterior muscle. A 5 × 3 grid of accelerometers was placed over the tibialis anterior, and the subjects were required to perform submaximal to maximal isometric muscle actions of the dorsiflexors. The results showed that different sensor locations caused different MMG amplitude and center frequency values, as well as different patterns of responses across isometric force. Thus, it was concluded that it would be difficult to describe motor control strategies from MMG amplitude and center frequency patterns when only one sensor is used to detect the signal. In addition, it may be necessary to record MMG signals from multiple locations, and use a composite, or some other type of weighted response when describing motor control strategies (Madeleine et al. 2006).

Jaskólska et al. (2007) compared a condenser microphone with an accelerometer for examining the MMG amplitude and MPF responses during submaximal concentric, isometric, and eccentric muscle actions. The experimental protocol required the subjects to perform submaximal concentric, isometric, and eccentric muscle actions of the forearm flexors at 10%, 30%, 50%, and 70% of the isometric MVC. During each muscle action, two separate surface MMG signals were detected from the biceps brachii with an accelerometer and a condenser microphone. The results showed that during the concentric, eccentric, and isometric muscle actions, the overall shape of the MMG amplitude versus torque relationships were similar for the condenser microphone and accelerometer. The MMG MPF versus torque relationships were different, however, for the two sensors, particularly during the eccentric muscle actions. Thus, it was concluded that the accelerometer and condenser microphone may provide different MMG amplitude and/or center frequency responses, and, therefore, the type of sensor that is used could affect the interpretation of the patterns of responses (Jaskólska et al. 2007). These findings were similar to those of Beck et al. (2006), who compared a piezoelectric crystal contact sensor with an accelerometer for the patterns of responses for MMG amplitude and MPF versus torque for the biceps brachii during both concentric and isometric muscle actions of the forearm flexors (Figure 3).

The subjects were required to perform submaximal to maximal concentric isokinetic and isometric muscle actions of the forearm flexors, and two separate surface MMG signals were detected simultaneously from the biceps brachii with an accelerometer and a piezoelectric crystal contact sensor. The results showed that during both the dynamic and isometric muscle actions, there

102

Travis W. Beck

Figure 3. Mechanomyographic (MMG) signals detected from the biceps brachii during a 6-second isometric muscle action of the forearm flexors of one subject at 80% MVC. The top graph shows the signal from a piezoelectric crystal contact sensor, the middle graph is the signal from an accelerometer, and the bottom signal is forearm flexion torque. Notice that the MMG signals from the contact sensor and accelerometer had very different shapes. *Reprinted with permission from Beck et al. (2006).

were linear increases in MMG amplitude with torque for the accelerometer and piezoelectric crystal contact sensor. However, the linear slope for the normalized MMG amplitude versus isokinetic torque relationship for the accelerometer was less than that for the contact sensor. In addition, there was no significant relationship for normalized MMG MPF versus isokinetic and isometric torque for the contact sensor, but the accelerometer demonstrated a quadratic or linear relationship for the isokinetic and isometric muscle actions, respectively. There were also several significant mean differences between the contact sensor and accelerometer for normalized MMG amplitude and MPF values. Thus, it was concluded that in some cases involving dynamic and isometric muscle actions, the contact sensor and accelerometer resulted in different torque-related responses for MMG amplitude and MPF that may affect the interpretation of the motor control strategies involved (Beck et al.

2006). Ouamer et al. (1999) examined the influence of sensor orientation (i.e., parallel versus perpendicular to the long axis of the muscle) on MMG responses from the biceps brachii during a voluntary isometric muscle action.

The mechanisms underlying the surface mechanomyogram

103

Specifically, two separate surface MMG sensors were placed either parallel or perpendicular to the long axis of the biceps brachii muscle as the subjects were required to perform an isometric MVC of the forearm flexors, as well as submaximal muscle actions at 30% and 50% MVC. The results indicated that when the MMG sensors were placed in line with the muscle, the resulting signals were in phase with each other. When the sensors were placed perpendicular to the long axis of the muscle, however, the resulting signals were reversed in phase. Thus, it was concluded that sensor location relative to the long axis of the muscle is an important consideration when detecting MMG signals. It was also hypothesized, however, that the resonant bending vibrations may not be the only movement that takes place in the muscle.

Specifically, frequencies between 10 and 16 Hz may reflect non-resonant movements that could be due to local (i.e., at the muscle fiber or motor unit level) factors, rather than factors at the whole muscle level (Ouamer et al.

1999). Courteville et al. (1998) performed an interesting study that described the use of a high sensitivity microphone for detecting MMG signals.

Specifically, the authors used a condenser microphone to detect MMG signals from the flexor digitorum superficialis during submaximal isometric muscle actions of the wrist flexors at 15%, 25%, 30%, 35%, and 45% MVC. The results showed that both MMG amplitude and MMG MPF increased with isometric force. In addition, the authors reported that there were several advantages of their MMG sensor when compared to other devices, such as accelerometers and piezoelectric crystal contact sensors. These advantages included a high sensitivity to low frequency vibrations, a value that is expressed in displacement units, and insensitivity to global movements of the muscle and/or limb (Courteville et al. 1998).

Mito et al. (2007) investigated the effects of changes in skin temperature on MMG and EMG amplitude during voluntary isometric muscle actions of the forearm flexors. The subjects were required to perform isometric forearm flexion muscle actions at 20%, 40%, 60%, and 80% MVC, and surface MMG and EMG signals were detected simultaneously from the biceps brachii. This series of muscle actions was performed at a control temperature (34Ëš Celsius), as well as at higher (40Ëš Celsius) and lower (28Ëš Celsius) temperatures that were achieved by heating and cooling the muscle, respectively. The results showed that for all temperatures, both MMG and EMG amplitude for the biceps brachii increased with isometric force. In addition, there were increases in MMG amplitude for the high temperature condition from 20-60% MVC, but then MMG amplitude plateaued from 60-80% MVC. These responses were different, however, under the control and low temperature conditions, and the changes in temperature had no effect on the patterns of responses for EMG

104

Travis W. Beck

amplitude. Thus, it was concluded that the changes in muscle temperature likely altered the mechanical properties of the muscle fibers in the biceps brachii, thereby affecting the MMG amplitude responses (Mito et al. 2007).

Kim et al. (2008) examined the influence of force tremor on the MMG signals detected with a condenser microphone and an accelerometer. Specifically, the subjects performed isometric muscle actions of the forearm flexors and extensors at 20%, 40%, 60%, 80%, and 100% MVC, and surface MMG signals were detected from the biceps brachii and triceps brachii with a condenser microphone and an accelerometer. The results showed that during agonist muscle activity (either biceps brachii or triceps brachii), the amplitude of the MMG signal from the condenser microphone was greater for the agonist than for the antagonist. In contrast, however, MMG amplitude for the accelerometer was often similar in the agonist and antagonist muscles. Thus, it was concluded that the MMG signal detected by an accelerometer is more susceptible to force tremor than the corresponding signal from a condenser microphone, since the force tremor was transmitted to the antagonist muscle and detected by the accelerometer. It was also recommended, however, that future studies need to be done to determine the mechanisms that cause an accelerometer to be more susceptible to force tremor than a condenser microphone (Kim et al. 2008).

Kim et al. (2008) compared the MMG signals detected with a condenser microphone versus an accelerometer during fatiguing isometric muscle actions of the forearm flexors. Specifically, the MMG signals were detected from both the biceps brachii (i.e., an agonist muscle) and the triceps brachii (i.e., an antagonist muscle) during a sustained isometric muscle action of the forearm flexors at 30% MVC. The results showed that the condenser microphone provided different MMG amplitude and frequency responses when compared to the accelerometer. In addition, the MMG signal detected by the condenser microphone was less affected by physiological tremor than the corresponding signal from the accelerometer. Thus, it was concluded that in situations where physiological tremor could affect the MMG signal, it may be more appropriate to use a condenser microphone to detect the signal than an accelerometer (Kim et al. 2008).

Silva and Chau (2005) examined the use of a mathematical model for removing movement artifacts from MMG signals for the purpose of controlling an externally powered prosthesis. Specifically, a microphone and accelerometer were coupled into a single sensor that could be used not only for prosthesis control, but also for removing the sounds associated with swallowing and respiration, as well as heart sounds. The results showed that the mathematical model accurately separated the sounds associated with muscle contraction from the movement artifacts, and, therefore, helped to improve the accuracy of controlling an externally powered prosthesis. In

The mechanisms underlying the surface mechanomyogram

105

addition, it was suggested that future studies should use the model to try to remove other noise sources from MMG signals (Silva and Chau 2005).

Gregori et al. (2003) investigated the feasibility of using a differential probe to reject movement artifacts from MMG signals. Specifically,two piezoelectric membranes were spaced 25 mm apart and integrated into a differential circuit board that allowed for detection of the differential MMG signal. The results showed that the MMG signal from the differential probe was very similar in amplitude and frequency to that of the signal from a non-differential, or single MMG sensor. In addition, the differential probe allowed for a much better rejection of movement artifacts, and, therefore, improved the signal-to-noise ratio. Thus, it was concluded that the differential MMG probe could be useful for recording MMG signals in noisy environments (Gregori et al. 2003).

Rafolt and Gallasch (2002) examined the use of a scanner galvanometer to detect the muscle vibrations associated with MMG signals. The precision of the galvanometer was first tested on a piezo-electric disc actuator, which showed that when the contact pressure forces ranged from 0.1-2.0 Newtons, the galvanometer was capable of detecting skin deflections of 1 Âľm. In the second experiment, the galvanometer was placed over the gastrocnemius muscle, and the posterior tibial nerve was stimulated at progressively higher voltages to elicit different twitch responses. The results showed that for contact pressure forces between 0.1 and 20 Newtons, the amplitude of the skin displacement detected by the galvanometer increased with stimulation voltage.

Finally, the responses of the galvanometer were compared with those of an accelerometer during a submaximal isometric muscle action of the plantar flexors. The results showed that when the galvanometer displacement signal was double differentiated, the resulting acceleration signal was very similar to that recorded simultaneously from an accelerometer. Thus, it was concluded that the scanner galvanometer is an accurate instrument for detecting the skin fluctuations associated with the MMG signal. It was also recommended, however, that future investigations should test the responses of the galvanometer under different testing conditions and with various muscles (Rafolt and Gallasch 2002). Akataki et al. (1999) examined the within-day and between-day reliability of MMG and EMG amplitude for the biceps brachii during submaximal to maximal isometric muscle actions of the forearm flexors. The results showed that the coefficient of variation values for withinday and between-day reliability were very similar for the MMG and EMG signals. In addition, the within-day intraclass correlation coefficient for MMG amplitude was R = 0.95, with a between-day coefficient of R = 0.80. Thus, it was concluded that the repeatability of MMG amplitude for the biceps brachii was similar to that for EMG amplitude (Akataki et al. 1999).

106

Travis W. Beck

Overall, the results from the studies discussed in this chapter indicated that accelerometers, piezoelectric crystal contact sensors, condenser microphones, and laser displacement sensors can all be used to detect MMG signals.

However, contact pressure, sensor location, and an understanding of the measurement units provided by the MMG sensor are important when interpreting both MMG amplitude and frequency responses. These factors become particularly important when using MMG in clinical applications, such as for control of externally powered prostheses. In addition, sensors that provide an MMG signal in volts or millivolts must be calibrated if the signal is to be converted to physiological units, such as micrometers or m路s-2.

References 1.

2.

3.

4.

5.

6.

7.

8.

Accornero N, Berardelli A, Manfredi M. A composite probe for acoustic and electromyographic recording of muscular activity. Electroencephalography and Clinical Neurophysiology 1989; 72:548-549.

Akataki K, Mita K, Itoh Y.Repeatability study of mechanomyography in submaximal isometric contractions using coefficient of variation and intraclass correlation coefficient. Electromyography and Clinical Neurophysiology 1999; 39:161-166.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Comparison of a piezoelectric contact sensor and an accelerometer for examining mechanomyographic amplitude and mean power frequency versus torque relationships during isokinetic and isometric muscle actions of the biceps brachii.

Journal of Electromyography and Kinesiology 2006; 16:324-335.

Bolton CP, Parkes, A, Thompson TR, Clark MR, Sterne CJ. Recording sound from human skeletal muscle: technical and physiological aspects. Muscle & Nerve 1989; 12:126-134.

Courteville A, Gharbi T, Cornu J-Y.MMG measurement: A high-sensitivity microphone-based sensor for clinical use. IEEE Transactions On Biomedical Engineering 1998; 45:145-150.

Gregori B,Gali茅 E,Accornero N.Surface electromyography and mechanomyography recording: a new differential composite probe. Medical & Biological Engineering & Computing 2003; 41:665-669.

Inoue T, Naumura K, Hosaka H, Itao K. Non-contact muscle sound sensing by laser displacement meter and development of a wearable muscle sound recorder for daily life. Proceedings of the 22nd Annual EMBS International Conference, July 23-28, 2000, Chicago, Illinois.

Jask贸kolska A, Madeleine P, Jask贸lski A, Kisiel-Sajewicz K, Arendt-Nielsen L.A comparison between mechanomyographic condenser microphone and accelerometer measurements during submaximal isometric, concentric and eccentric contractions. Journal of Electromyography and Kinesiology 2007; 17:336-347.

The mechanisms underlying the surface mechanomyogram

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

107

Kim T-K, Shimomura Y, Iwanaga K, Katsuura T. Influence of force tremor on mechanomyographic signals recorded with an accelerometer and a condenser microphone during measurement of agonist and antagonist muscles in voluntary submaximal isometric contractions. Journal of Physiological Anthropology 2008; 27:33-42.

Kim T-K, Shimomura Y,Iwanaga K, Katsuura T.Comparison of an accelerometer and a condenser microphone for mechanomyographic signals during measurement of agonist and antagonist muscles in sustained isometric muscle contractions: The influence of the force tremor. Journal of Physiological Anthropology 2008; 27:121-131.

Madeleine P,Cescon C,Farina D.Spatial and force dependency of mechanomyographic signal features. Journal of Neuroscience Methods 2006; 1 58:89-99.

Mito K, Kitahara S, Tamura T, Kaneko K, Sakamoto K, Shimizu Y.Effect of skin temperature on RMS amplitude of electromyogram and mechanomyogram during voluntary isometric contraction. Electromyography and Clinical Neurophysiology 2007; 47:153-160.

Ouamer M, Boiteaux M, Petitjean M, Travens L, Salès A. Acoustic myography during voluntary isometric contraction reveals non-propagative lateral vibration.

Journal of Biomechanics 1999; 32:1279-1285.

Rafolt D, Gallasch E.Surface myomechanical responses recorded on a scanner galvanometer. Medical & Biological Engineering & Computing 2002; 40:594-599.

Silva J, Chau T.A mathematical model for source separation of MMG signals recorded with a coupled microphone-accelerometer sensor pair. IEEE Transactions On Biomedical Engineering 2005;52:1493-1501.

Smith T.G, Stokes M.J.Technical aspects of acoustic myography (AMG) of human skeletal muscle: contact pressure and force/AMG relationships. Journal of Neuroscience Methods 1993; 47:85-92.

Watakabe M, Itoh Y,Mita K,Akataki K.Technical aspects of mechanomyography recording with piezoelectric contact sensor. Medical & Biological Engineering & Computing 1998; 36:557-561.

Watakabe M, Mita K, Akataki K, Itoh Y. Mechanical behaviour of condenser microphone in mechanomyography. Medical & Biological Engineering & Computing 2001; 39:195-201.

Watakabe M, Mita K, Akataki K, Ito K.Reliability of the mechanomyogram detected with an accelerometer during voluntary contractions. Medical and Biological Engineering & Computing 2003; 41:198-202.

Wee AS, Ashley RA. Transmission of acoustic or vibratory signals from a contracting muscle to relatively distant tissues. Electromyography and Clinical Neurophysiology 1990; 30:303-306.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for examining muscle function, 2010, 109-116 Editor Travis.W. Beck. ISBN :978-81-7895-449-3

7

Processing the surface mechanomyographic signal Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081

Abstract Signal processing is an important issue that should be considered prior to recording mechanomyographic (MMG) signals. The first studies in this area found that the fast Fourier transform and maximum entropy spectrum estimation techniques provided MMG frequency spectra that were very similar in shape. The shape of the spectrum from the maximum entropy spectrum estimation technique, however, was highly dependent on the model order that was chosen. Recent investigations have also used joint time-frequency signal processing techniques, such as wavelet- based methods and the short-time Fourier transform to analyze Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: tbeck@ou.edu

110

Travis W. Beck

MMG signals. The results from these studies have shown that when the wavelet-based techniques were used, the patterns of responses for MMG center frequency were very similar to those from Fourier-based methods. An advantage of the wavelet-based methodology, however, is that it provides information in both the time and frequency domains, and, therefore, is ideal for processing nonstationary MMG signals, such as those recorded during dynamic muscle actions. Additional research still needs to be done to identify what type of information can be provided by joint time-frequency processing of MMG signals.

Introduction There has been a great deal of work performed in the area of mechanomyographic (MMG) signal processing. One of the first studies was done by Diemont et al. (1988). The authors detected surface MMG signals from the biceps brachii during submaximal isometric muscle actions of the forearm flexors at 20% and 80% MVC. The signals were then processed with the fast Fourier transform (FFT) and maximum entropy spectrum estimation (MESE) algorithms, both of which provide power spectra. An important consideration, however, is that the MESE technique does not assume signal stationarity like the FFT method.Thus, the MESE technique could potentially be useful when processing nonstationary MMG signals. The results showed that the power spectra from the FFT and MESE algorithms were very similar in shape, as long as the correct model order was used with the MESE technique. Specifically, when the model order was too low, the power spectrum was excessively smooth, but when the model order was too high, the power spectrum was often noisy with a great deal of spurious information.

Thus, it was suggested that the model order should be chosen carefully when using the MESE algorithm. In addition, the power spectra from both algorithms were usually positively skewed, and, therefore, the resulting MMG mean power frequency (MPF) values were generally greater than the MMG median frequency values. Therefore, it was recommended that the MPF or median frequency values could be used to characterize the shape of the MMG power spectrum (Diemont et al. 1988). This study was followed up by a second investigation (Figini and Diemont 1989) that used both the FFT and MESE methods to calculate the cross spectrum of MMG and EMG signals.

Specifically, the cross spectrum provides information regarding the common frequency components in MMG and EMG signals, and the authors suggested that this information must be related to the activation pattern of the muscle. In addition, it was concluded that appropriate selection of the signal processing

Processing the surface mechanomyographic signal

111

technique was just as important as the choice of a proper experimental design (Figini and Diemont 1989).

Goddard et al. (2005) performed an interesting study that used the discrete wavelet transform to examine the MMG signals from the soleus during quiet resting in either a supine (i.e., with the subject lying on their back) or sitting (i.e., with the feet motionless on the ground) position. The results showed that the resting muscle activity from the soleus was greater in the seated than the supine position. In addition, the amplitude of the MMG signal increased throughout the 20 minute period in the sitting position. The results from the discrete wavelet transform also showed that the increase in MMG amplitude over the 20 minute period occurred primarily in the 16-62 Hz frequency range.

Thus, it was concluded that resting MMG activity could be due to skeletal muscle contraction for the purpose of returning blood back to the heart (i.e., the “skeletal muscle pump” phenomenon). In addition, it was hypothesized that MMG could potentially be used as a diagnostic device for identifying those with inadequate skeletal muscle pump activity, thereby putting them at risk for complications from hypotension and reduced blood flow during orthostasis (Goddard et al. 2005). Torres et al. (2005) recently used the wavelet transform to process MMG signals from the diaphragm of two mongrel dogs.

Specifically, the MMG sensor was placed in the 8th intercostal space, and an MMG signal was recorded from the diaphragm during normal breathing. As acknowledged by the authors, this task is challenging because movement of the thoracic wall creates large motion artifacts in the MMG signal. Thus, the purpose of the wavelet analysis was to separate the MMG signals of the diaphragm from the motion artifacts caused by movement of the thoracic wall.

The results showed that for the first dog tested, the cutoff frequency needed to eliminate motion artifact was between 6.6 and 13.52 Hz, but for the second dog, the best cutoff frequency was between 1.64 and 3.3 Hz. Thus, it was concluded that the wavelet technique was a useful method for separating the low- and high-frequency components of the MMG signal, although different cutoff frequencies may be needed for different signals. In addition, waveletbased signal processing methods may be useful for analyzing the MMG signals detected from the diaphragm during normal breathing (Torres et al. 2005).

Our laboratory has performed several investigations to compare the patterns of responses for MMG center frequency obtained with wavelet- and Fourier-based methods. Specifically, Beck et al. (2005) reported that during 50 consecutive maximal concentric isokinetic muscle actions of the forearm flexors at a velocity of 180°·s-1, there were quadratic decreases in MMG MPF, MMG median frequency (both obtained with the discrete Fourier transform),

112

Travis W. Beck

Figure 1. Patterns of responses for mechanomyographic (MMG) mean power frequency (mpf), median frequency (mdf), and wavelet center frequency (cf) for the biceps brachii muscle during 50 consecutive maximal concentric isokinetic muscle actions of the forearm flexors at a velocity of 180掳路s-1. Notice that the patterns of responses for MMG frequency were very similar among the three methods. *Reprinted with permission from Beck et al. (2005).

and MMG wavelet center frequency (obtained with the discrete wavelet transform) (Figure 1).

These findings were important from a practical standpoint because it has been suggested that Fourier-based methods should not be used to process nonstationary signals (e.g., MMG signals recorded during dynamic muscle actions). Wavelet-based methods, however, do not assume signal stationarity, which may make them more appropriate for processing MMG signals recorded during dynamic muscle actions. Despite the concerns over signal stationarity, the results from our study showed that the patterns of responses for MMG center frequency were very similar for the Fourier- and wavelet-based methods. Thus, it was concluded that the discrete Fourier transform is an acceptable method for determining the patterns of responses for MMG center frequency during fatiguing dynamic muscle actions (Beck et al. 2005). Similar results were also reported for the biceps brachii during submaximal to maximal concentric (Beck et al. 2005) and eccentric (Beck et al. 2006) isokinetic muscle

Processing the surface mechanomyographic signal

113

actions of the forearm flexors. Ryan et al. (2008) performed an interesting study that compared the patterns of responses for MMG center frequency from the short-time Fourier and continuous wavelet transforms. Specifically, the subjects were required to perform a 6-second isometric ramp muscle action of the leg extensors from 5-100% MVC, and surface MMG signals were detected from the vastus lateralis and rectus femoris. The results showed that the shorttime Fourier and continuous wavelet transforms provided similar patterns of responses for MMG center frequency (Figure 2).

Figure 2. Mechanomyographic (MMG) amplitude (top graph), mean power frequency (MPF) from the short-time Fourier transform (STFT), and MPF from the continuous wavelet transform (CWT) for the vastus lateralis during a 6-second isometric ramp muscle action of the leg extensors. Notice that the patterns of responses for MMG MPF from the STFT were very similar to those from the CWT. *Reprinted with permission from Ryan et al. (2008).

114

Travis W. Beck

Thus, these findings were similar to the results from our previous investigations, and suggested that Fourier- and wavelet-based methods provided similar information regarding the patterns of responses for MMG center frequency (Ryan et al. 2008).

The interest in wavelet-based methods for processing MMG signals was important because it provided the impetus for our laboratory to develop a filter bank of wavelets designed specifically for MMG signals (Beck et al. 2008).

Specifically, these wavelets cover the entire frequency range for MMG and are designed to provide the best possible combination of time and frequency resolution. The unique aspect of the wavelets, however, is that they are nonlinearly scaled, which allows them to provide equal weight to high and low frequency components in MMG signals (Figure 3).

Figure 3. Filter bank of wavelets used for the nonlinear scaling wavelet analysis designed by Beck and von Tscharner (2008) for mechanomyographic (MMG) signals.

(A) shows the frequency domain representation of the wavelets and their sum (dashed line). (B) shows two of the wavelets in the time domain. Notice that in (A), the sum of the wavelets is almost perfectly flat across the entire frequency range for MMG (i.e., 5100 Hz). Also, notice that in (B), the two wavelets have a different number of oscillations, which is a property of their nonlinear scaling. *Reprinted with permission from Beck et al. (2008).

Processing the surface mechanomyographic signal

115

In addition, the results from the MMG wavelet analysis are in the form of a time-frequency distribution known as an intensity pattern. This pattern is similar in principle to the spectrogram from the short-time Fourier transform and scalogram from linearly scaled wavelets. An important distinction, however, is that the wavelet analysis is designed specifically for MMG signals, whereas the short-time Fourier and linearly scaled wavelet transform are intended for many different types of signals (Beck et al. 2008).

Overall, the results from these studies have shown that both Fourier- and wavelet-based techniques are adequate for examining changes in MMG center frequency. An advantage of the wavelet-based methodologies, however, is that they provide much more information, and they are more appropriate for nonstationary MMG signals than the Fourier-based techniques. More research still needs to be done to identify applications of the results from wavelet analysis of MMG signals.

References 1.

2.

3.

4.

5.

6.

7.

Beck TW, Housh TJ, Johnson GO, Cramer JT, Weir JP, Coburn JW, Malek MH. Comparison of the fast Fourier transform and continuous wavelet transform for examining mechanomyographic frequency versus eccentric torque relationships.

Journal of Neuroscience Methods 2006; 150:59-66.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Comparison of Fourier and wavelet transform procedures for examining the mechanomyographic and electromyographic frequency domain responses during fatiguing isokinetic muscle actions of the biceps brachii. Journal of Electromyography and Kinesiology 2005; 15:190-199.

Beck TW, Housh TJ, Johnson GO, Weir JP, Cramer JT, Coburn JW, Malek MH. Comparison of Fourier and wavelet transform procedures for examining mechanomyographic and electromyographic frequency versus isokinetic torque relationships. Electromyography and Clinical Neurophysiology 2005; 45:93-103.

Beck TW, von Tscharner V, Housh TJ, Cramer JT, Weir JP, Malek MH, Mielke M. Time/frequency events of surface mechanomyographic signals resolved by nonlinearly scaled wavelets. Biomedical Signal Processing and Control 2008; 3:255-266.

Diemont B, Figini MM, Orizio C, Perini R, Veicsteinas A. Spectral analysis of muscular sound at low and high contraction level. International Journal of Biomedical Computing 1988;23:161-175.

Figini MM, Diemont B. Mathematics of the muscle sound. Proceedings of the IEEE Engineering In Medicine & Biology Society 11th Annual International Conference. 1989.

Goddard AA, Madhavan G, Cole JP, Fowler ML, McLeod KJ.Time-frequency analysis of endogenous calf skeletal muscle vibromyograms. Proceedings of the 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.

Shanghai, China, September 1-4, 2005.

116

8.

9.

Travis W. Beck

Ryan ED, Cramer JT, Egan AD, Hartman MJ, Herda TJ. Time and frequency domain responses of the mechanomyogram and electromyogram during isometric ramp contractions: A comparison of the short-time Fourier and continuous wavelet transforms. Journal of Electromyography and Kinesiology 2008; 18:54-67.

Torres A, Fiz JA, Galdig B, Gea J, Morera J, JanĂŠ R. A wavelet multiscale based method to separate the high and low frequency components of mechanomyographic signals. Proceedings of the 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference. Shanghai, China, September 1-4, 2005.

Transworld Research Network 37/661 (2), Fort P.O., Trivandrum-695 023, Kerala, India

Applications of Mechanomyography for Examining Muscle Function, 2010, 117-136 Editor Travis.W. Beck. ISBN :978-81-7895-449-3

8

Unique applications of mechanomyography Travis W. Beck Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, Norman, Oklahoma 73019-6081

Abstract Recent investigations have examined some unique applications of surface mechanomyography (MMG).

For example, several studies have shown different MMG amplitude and frequency responses for muscles that differed in fiber type composition. Other investigations have used MMG for biofeedback to enhance muscle relaxation. Typically, surface electromyographic (EMG) signals are used for biofeedback, but there are several disadvantages to this method, including changes in skin impedance and a high sensitivity to electrode placement. Previous studies have also used MMG to examine the effects Correspondence/Reprint request: Dr. Travis W. Beck, Biophysics Laboratory, Department of Health and Exercise Science, University of Oklahoma, 1401 Asp Avenue, Norman, Oklahoma, 73019-6081, USA. E-mail: tbeck@ou.edu

118

Travis W. Beck

of static and dynamic stretching on the neural and mechanical aspects of muscle function, as well as the effects of hypothermia. Generally speaking, these studies have shown that MMG is a very sensitive indicator of changes in muscle function, and, in many cases, provides information that is unique from EMG and force.

Introduction There are several recent studies that have examined some unique applications of mechanomyography (MMG). For example, Shima et al. (2006) investigated the effects of postactivation potentiation on the MMG signal. The experimental protocol involved a supramaximal electrical stimulation of the tibial nerve to measure peak twitch torque, the peak acceleration of twitch torque development, as well as peak-to-peak electromyographic (EMG) and MMG amplitude values for the gastrocnemius muscle. This initial twitch was followed by a 10-second isometric maximum voluntary contraction (MVC) of the plantar flexors and supramaximal twitches that were elicited 2, 15, 30, 60, and 180 seconds after the isometric MVC. The results showed that the peak twitch torque, peak acceleration of twitch torque development, and peak-topeak MMG amplitude values were greater 2, 15, and 30 seconds after the MVC when compared to the corresponding values recorded before the MVC. In addition, the peak twitch torque remained elevated 60 and 180 seconds after the MVC, and the peak acceleration of twitch torque development remained elevated 60 seconds after the MVC. Furthermore, the peak-to-peak MMG amplitude was positively correlated with both the peak twitch torque (r = 0.79) and the peak acceleration of twitch torque development (r = 0.78).

The 10-second isometric MVC had no effect, however, on the peak-topeak EMG amplitude values. Thus, it was suggested that the peak-to-peak amplitude of the MMG signal from an electrically-stimulated isometric twitch may provide useful information regarding mechanical changes in the muscle that are associated with increased force production after an isometric MVC. It has been hypothesized that this phenomenon is related to facilitation of excitation-contraction coupling mechanisms through the Ca2+ - dependent process of phosphorylation of the regulatory light chains of myosin. Thus, MMG may be a useful technique for investigating the mechanisms that underlie post-activation potentiation (Shima et al. 2006). Yoshitake et al.

(2005) investigated the relationship between changes in MMG amplitude and alterations in fascicle length (measured with ultrasound) during electrical stimulation of the gastrocnemius muscle. Specifically, the posterior tibial nerve was stimulated at frequencies that increased linearly from 1 to 20 Hz over a 4 second time period. During the stimulation, the surface MMG signal

Unique applications of Mechanomyography

119

was measured from the gastrocnemius muscle, and ultrasound was used to detect changes in fascicle length. The results showed that with increases in stimulation rate, both MMG amplitude and the changes in fascicle length decreased. These decreases followed similar patterns, and the changes in MMG amplitude were positively correlated (r = 0.94) with the changes in fascicle length. Thus, it was concluded that the origin of the MMG signal is the pressure wave resulting from architectural changes in contracting muscle, and these pressure changes cause mechanical movement of the skin surface (Yoshitake et al. 2005).

Dahmane et al. (2006) investigated the possibility of using MMG as an indicator of muscle fiber type composition. Specifically, MMG was used to measure the contraction time of the biceps femoris muscle during a supramaximal electrically-stimulated twitch. The subjects that participated in the study included 15 sedentary men and 15 high level sprinters. In addition, muscle biopsies were taken from 15 sedentary men that had died due to various causes (e.g., suicide, traffic accident, etc.). The results showed that the contraction time was less for the sprinters than the sedentary men, and was negatively correlated (r = -0.72) with running speed during a flying 20 meter running trial. In addition, the muscle biopsies showed that the biceps femoris consisted of approximately 49% Type I, or slow-twitch fibers. Thus, it was suggested that the biceps femoris may have a large capacity for changing its fiber type composition with training (i.e., it could be a very “plastic� muscle), and that these changes could potentially be tracked with MMG (Dahmane et al.

2006). Beck et al. (2007) also examined the potential for MMG to be used in estimating muscle fiber type composition. The subjects included 5 resistancetrained and 5 aerobically-trained men, and all subjects were required to perform a 30-second sustained isometric muscle action of the leg extensors at 50% MVC as the surface MMG signal was detected from the vastus lateralis.

In addition, immediately after the sustained muscle action, a biopsy was taken from the vastus lateralis and analyzed for myosin heavy chain isoform composition. The results from the myosin heavy chain analyses showed that the vastus lateralis of the resistance-trained subjects consisted primarily of fasttwitch muscle fibers, while that of the aerobically-trained subjects contained mainly slow-twitch fibers. In addition, the mean MMG amplitude and MPF values for the resistance-trained subjects were greater than those for the aerobically-trained subjects at all time points during the sustained muscle action. Thus, it was suggested that the MMG amplitude and MPF responses for the resistance-trained subjects could have been due to a greater percentage of fast-twitch muscle fibers in the vastus lateralis when compared to that of the aerobically-trained subjects (Beck et al. 2007). This study was followed up by

120

Travis W. Beck

a second investigation (Beck et al. 2008) that compared the patterns of responses for MMG amplitude versus isometric torque for the vastus lateralis among resistance-trained, aerobically-trained, and sedentary subjects. Like the previous investigation, the resistance-trained subjects had mostly fast-twitch muscle fibers in their vastus lateralis, while the aerobically-trained subjects had primarily slow-twitch fibers. The vastus lateralis of the sedentary subjects had roughly equal proportions of fast- and slow-twitch fibers. Despite these differences in fiber type composition, however, there were no consistent patterns of responses for MMG amplitude versus isometric torque. Some subjects showed increases in MMG amplitude throughout the entire force production range, while others demonstrated a plateau at high force levels.

Thus, it was concluded that differences in fiber type composition and training status did not explain the unique torque-related patterns of responses for MMG amplitude (Beck et al. 2008).

Although there were no consistent fiber type-related differences in the patterns of responses for MMG amplitude, that does not necessarily mean that MMG cannot be used to estimate fiber type composition. Beck et al. (in press) used multiple regression in an attempt to estimate the percentage of fast-twitch muscle fibers in the vastus lateralis based on the MMG median frequency and maximal isometric leg extension strength values. The results showed that neither isometric leg extension strength nor MMG median frequency alone were significantly correlated with the fast-twitch fiber type content. The combination of these two variables, however, explained a significant proportion (59.8%) of the variance in fast-twitch fiber type content, with a multiple correlation of R = 0.773 and a standard error of the estimate of 15.4%.

Thus, it was concluded that a simple, time-efficient, and noninvasive test that simultaneously measures isometric strength and MMG median frequency could be useful for estimating the fast-twitch fiber type content in well trained men (Beck et al. in press). Mealing et al. (1996) also conducted an interesting study that examined the MMG frequency responses from muscles that had a different fiber type composition. Surface MMG signals were detected from the soleus and biceps brachii during an isometric muscle action at 50% MVC. The results showed that the MMG signal for the soleus muscle was dominated by power between 5 and 10 Hz, whereas the MMG power spectrum for the biceps brachii had a much larger bandwidth, with power not only between 5 and 10 Hz, but also between 10 and 25 Hz (Figure 1).

Thus, it was concluded that the discrepancies between the two muscles for MMG power spectra were likely due to differences in fiber type content. It was also suggested that a possible application of MMG is in sport, where individual differences in fiber type composition could be important for determining performance (Mealing et al. 1996). Marchetti et al. (1992) compared

Unique applications of Mechanomyography

121

Figure 1. Average (i.e., across subjects) mechanomyographic (MMG) power spectra for the soleus (top graph) and biceps brachii (bottom graph) muscles. The signals were detected during submaximal isometric muscle actions of the plantar flexors or forearm flexors at 50% of the isometric maximum voluntary contraction (MVC). Notice that the spectrum for the soleus was dominated by power below 10 Hz, while the spectrum for the biceps brachii showed most of its power above 10 Hz. *Reprinted with permission from Mealing et al. (1996).

the MMG frequency responses of the vastus lateralis and soleus during supramaximal electrically-stimulated isometric twitches. The results showed that the MMG median frequency values from the vastus lateralis were significantly greater than those for the soleus in all subjects that participated in the study. In addition, the time required to reach peak MMG amplitude was significantly less for the vastus lateralis than the soleus. Thus, it was concluded that the different MMG responses for the two muscles were most likely due to differences in fiber type composition. In addition, the authors recommended that future studies should compare the MMG amplitude and frequency responses from electrical stimulation in both endurance and power athletes (Marchetti et al. 1992).

122

Travis W. Beck

Orizio et al. (1999) also examined the MMG responses for the tibialis anterior during electrical stimulation. The experimental protocol included six single supramaximal twitches, followed by a 5-second ramp in which the stimulation frequency was increased from 1 to 50 Hz. After the 1 to 50 Hz ramp, a fatiguing protocol was induced that consisted of continuous stimulation at 35 Hz for 40 seconds. The 6 single twitches and 1 to 50 Hz ramp were then repeated immediately after the fatiguing stimulation, as well as 0.5, 1, 2, 3, 4, and 6 minutes after the fatigue protocol. The results showed that the fatigue protocol caused reductions in peak twitch force, the peak rate of force development, and the peak of the acceleration of force development, while both the contraction time and half-relaxation time increased. In addition, the peak-to-peak MMG amplitude decreased with fatigue. With the exception of half-relaxation time, all of the force and MMG parameters were restored to normal values within two minutes of recovery. Furthermore, the peak-to-peak MMG amplitude was highly correlated with the peak of the acceleration of force development in fresh muscle and during recovery.

It was concluded that MMG can be used to track changes in the mechanics of individual muscles with fatigue. This is an important advantage when compared to the force signal because there are very few joints that are crossed by just one muscle. Thus, examination of the force signal during fatigue provides information regarding fatigue-induced changes in the properties of several muscles, whereas the MMG signal is muscle-specific, and, therefore, can be used to examine fatigue of individual muscles (Orizio et al. 1999).

Jask贸lska et al. (2004) conducted an interesting study that examined the effect of skinfold thickness on MMG frequency. Specifically, surface MMG signals were detected simultaneously from the biceps brachii, triceps brachii, and brachioradialis during maximal isometric muscle actions of the forearm flexors and extensors, and skinfold thicknesses were measured in the same locations that the MMG sensors were placed for each muscle. The results showed that skinfold thickness had a large effect on the median frequency of the MMG signal and a smaller influence on the peak frequency. Specifically, greater skinfold thicknesses were associated with lower MMG median frequency values. Thus, it was concluded that studies that report absolute MMG frequency values should consider using skinfold thicknesses as a covariate (Jask贸lska et al. 2004). Mamaghani et al. (2002) investigated the influence of changes in joint angle on the MMG amplitude and MPF responses of the upper trapezius, anterior deltoid, biceps brachii, and brachioradialis during sustained submaximal isometric muscle actions at 20%, 40%, and 60% MVC. The results showed that MMG amplitude followed force production, except at 20% MVC. In addition, MMG MPF remained relatively constant for all muscles during the sustained muscle actions. Thus, it was suggested that MMG

Unique applications of Mechanomyography

123

amplitude may be useful for tracking changes in force production during fatiguing isometric muscle actions (Mamaghani et al. 2002).

Nonaka et al. (2006) examined gender differences in the MMG amplitude and MPF versus isometric torque relationships for the biceps brachii. The subjects were required to perform an isometric ramp muscle action of the forearm flexors from 5-80% MVC at a rate of 10% MVC/second while surface MMG signals were detected from the biceps brachii. In addition, the crosssectional area of the biceps brachii was estimated with ultrasound imaging.

The results showed that the mean isometric forearm flexion strength and crosssectional area values for the men were 56.3% and 56.1% greater than those for the women, respectively. Furthermore, the mean MMG amplitude values for the men were greater than those for the women at all force levels. Even when the MMG amplitude values were expressed relative to the isometric MVC, the men showed greater values than the women at all force levels, and the mean difference became larger at higher force levels. Finally, the mean MMG MPF values were significantly greater for the men when compared to those for the women. Although muscle biopsies were not taken from the biceps brachii, the authors hypothesized that the differences between the men and women for the patterns of responses for MMG amplitude versus isometric torque were probably due to discrepancies in muscle fiber type composition. Specifically, it was suggested that since women generally have a greater percentage of slowtwitch fibers than men, the different torque-related patterns of responses for men and women were likely due to fusion of twitches from slow-twitch motor units at a lower force level for women versus men (Nonaka et al. 2006).

Yamamoto and Takano (1994) investigated the possibility of using MMG to assess muscle tonus during functional electrical stimulation. The authors cited the fact that in many cases, it is difficult to directly measure muscle strength during electrical stimulation. Thus, they recorded MMG signals from the tibialis anterior during electrical stimulation, and found that the amplitude of the MMG signal was linearly related to dorsiflexion force production.

Therefore, it was concluded that the MMG signal may be a useful method for estimating force production from individual muscles during electrical stimulation (Yamamoto and Takano 1994). Harba and Chee (1997) conducted an interesting investigation that examined the propagation velocities of MMG and EMG signals. The subjects were required to perform a sustained “light� isometric muscle action of the forearm flexors, and two separate surface MMG and EMG signals were detected simultaneously from the biceps brachii muscle. The MMG and EMG sensors were placed in line with the long axis of the muscle, and were separated by the same distance. Cross-correlation was used to determine the time delay between the signals (both MMG and EMG)

124

Travis W. Beck

detected at the two different sensor locations, which allowed for calculation of propagation velocity. The results showed that the propagation velocities of the MMG and EMG signals were nearly the same. These findings have important practical implications in terms of identifying the mechanisms that generate the MMG signal. Specifically, they indicated that the mechanical twitches that underlie the MMG signal may propagate along the muscle fibers in a wave-like fashion, rather than occurring simultaneously at all points along the muscle fiber (Harba and Chee 1997).

Keidel and Keidel (1989) also performed an important study that simultaneously investigated the MMG and indwelling EMG responses from the biceps brachii, masseter, wrist extensor, and tibialis anterior muscles during relaxation, as well as during graded isometric muscle actions. The results showed that the MMG signal provided important information regarding motor control strategies, and could potentially reflect the activity of central reflex loops. It was also suggested, however, that more studies needed to be performed to identify the mechanisms that cause the various MMG responses in different experimental protocols (Keidel and Keidel 1989). Petitjean and Maton (1995) performed a similar study that applied the spike-triggered averaging technique to MMG signals. The subjects were required to perform submaximal isometric muscle actions of the forearm extensors as surface MMG and indwelling EMG signals were detected simultaneously from the anconeus muscle. The intramuscular EMG signal was used for the spiketriggered averaging, and provided audio feedback to the subjects to help them achieve a steady muscle activation level. The results showed that by using the spike-triggered averaging technique, it was possible to isolate the MMG signal from individual motor units. Furthermore, the MMG signal from each motor unit had a characteristic shape, and, therefore, could be used to identify the contributions of individual motor units to force production. These findings were important because they indicated that the mechanical activities of individual motor units made predictable contributions to the MMG signal and could be extracted. It was recommended, however, that future studies were needed to determine whether or not the spike-triggered averaging method could be used at high, as well as at low force levels (Petitjean and Maton 1995). The spike-triggered averaging technique was also used by Jørgensen and Lammert (1976) to investigate the MMG responses from individual motor units in the rectus femoris during submaximal isometric muscle actions of the leg extensors at 20%, 30%, and 40% MVC. The results showed that the size of the action potential detected with the indwelling EMG electrodes was correlated with the amplitude of the mechanical contribution of the motor unit to the MMG signal. In addition, the mechanical contributions of individual motor units to the MMG signal were very reliable. Thus, it was concluded that

Unique applications of Mechanomyography

125

the MMG signal is generated by the mechanical activities of individual motor units, and these activities are different, based on the motor unit’s size and location within the muscle (Jørgensen and Lammert 1976). The study by Yoshitake et al. (2002) was also very important for determining the mechanisms that generate the MMG signal. Specifically, the authors electrically stimulated the medial gastrocnemius at rates of 5, 10, 15, and 20 Hz, and intramuscular EMG and surface MMG signals were detected from the muscle. The indwelling EMG electrodes allowed eight separate motor units to be identified. The results showed that when each motor unit was stimulated, there was a significant positive correlation between the duration of the MMG signal and that of the force twitch. In addition, both MMG amplitude and the magnitude of the force fluctuations decreased with increasing stimulation rates.

Finally, MMG amplitude was negatively correlated with several twitch parameters, including twitch duration and half-relaxation time. Thus, it was concluded that the characteristics of the surface MMG signal were dependent on those of the active motor units during the contraction (Yoshitake et al.

2002).

Celichowski et al. (1998) examined the relationship between force fluctuations and MMG amplitude for the forefinger flexor muscles. All subjects were required to perform triangular or trapeziform contractions at different force levels, and MMG signals were detected from the anteromedial part of the forearm. The results showed that MMG amplitude was dependent on the force that was being produced. Specifically, both the rate and amplitude of changes in force were correlated with those for MMG amplitude. Thus, it was suggested that MMG could be used to examine the force production of individual muscles (Celichowski et al. 1998). Rodriquez et al. (1993) examined the MMG and EMG amplitude and median frequency responses from the rectus femoris during sustained isometric muscle actions of the leg extensors at 20%, 40%, and 80% MVC. Each muscle action was performed until exhaustion, and the results showed that for each force level, EMG amplitude increased and EMG median frequency decreased. The results for MMG amplitude, however, demonstrated increases over time only during the 20% and 40% MVC muscle actions. Thus, it was concluded that MMG and EMG amplitude behave similarly, but only at low force levels. Furthermore, in many situations, MMG provides information that is unique from that of EMG (Rodriquez et al. 1993).

Mitchell et al. (2008) examined the effects of diathermy on muscle temperature, EMG amplitude, and MMG amplitude. This was an important study from a practical standpoint because exercise often results in increases in muscle temperature. The subjects were randomly assigned to one of three

126

Travis W. Beck

groups: (a) a diathermy group that was subjected to a 20-minute treatment, (b) a sham-diathermy group that thought they were receiving the treatment, but actually were not, and (c) a control group that did not receive the treatment.

Prior to, and immediately following the diathermy treatment, all subjects were required to perform a 6-second isometric ramp muscle action of the leg extensors from 10-90% MVC, and surface MMG signals were detected from the vastus lateralis. The results showed that the diathermy treatment provided a significant increase (approximately 2째 Celsius) in muscle temperature, MMG amplitude, and MMG MPF at all torque levels, but had no effect on leg extension force, EMG amplitude, or EMG MPF (Figure 2).

Thus, it was concluded that increases in muscle temperature may decrease musculotendonous stiffness, thereby affecting both MMG amplitude and MMG MPF. The changes in muscle temperature did not, however, affect leg extension strength or the motor control strategies that influence isometric force production (Mitchell et al. 2008). Kimura et al. (2003) examined the effects of decreases in muscle temperature on the MMG amplitude responses for the soleus and medial gastrocnemius muscles. Surface MMG signals were detected simultaneously from both muscles during supramaximal electrically-stimulated

Figure 2. Changes in mechanomyographic (MMG) amplitude for the vastus lateralis with increases in isometric leg extension torque from 10% to 90% of the isometric maximum voluntary contraction (MVC). The closed symbols reflect the values before diathermy treatment, and the open symbols show the data after diathermy treatment.

Notice that the diathermy treatment caused an increase in MMG amplitude at nearly all force levels. *Reprinted with permission from Mitchell et al. (2008).

Unique applications of Mechanomyography

127

twitches both before, and immediately after manipulating muscle temperature with ice packs. The control temperature was 34° Celsius, and the administration of ice packs allowed data to be collected at temperatures of 25°, 20°, and 15° Celsius. The results showed that the decreases in muscle temperature caused reductions in peak twitch force and the maximal rate of force development, increased contraction time, half-relaxation time, and the maximal rate of force relaxation. The decrease in muscle temperature resulted in increased force production during the sustained 10 Hz stimulation, but the magnitude of the force fluctuations decreased, as did MMG amplitude for both the medial gastrocnemius and soleus muscles (Figure 3).

Thus, it was concluded that the significant decrease in muscle contractile properties caused by the reduced muscle temperature was reflected in the MMG responses to electrical stimulation. Therefore, MMG could be useful for examining muscle contractile properties in a variety of physiological conditions (Kimura et al. 2003).

Evetovich et al. (2007) investigated the use of MMG as a biofeedback method for improving relaxation and delaying fatigue. The subjects were required to perform as many repetitions as possible with 85% of their one-repetition

Figure 3. Mechanomyographic (MMG) signals for the soleus muscle during electrically-stimulated contractions at 10 Hz when the muscle was at room temperature (control) and cooled to 25° Celsius. Notice the dramatic decrease in MMG amplitude when the muscle was cooled. *Reprinted with permission from Kimura et al. (2003).

128

Travis W. Beck

maximum (1-RM) during the bilateral forearm flexion exercise. During all muscle actions, surface MMG and EMG signals were detected from the biceps brachii, but only the MMG signal was used for biofeedback. In addition, the subjects were required to use the MMG signal during separate visits in an attempt to achieve complete muscle relaxation. The results showed that the use of MMG biofeedback did not increase the number of repetitions that the subject could perform with 85% of their 1-RM, but it did help to achieve more complete relaxation of the biceps brachii muscle. Thus, it was concluded that MMG can enhance the development of muscle relaxation, but it is not useful for delaying fatigue (Evetovich et al. 2007).

Silva et al. (2003) have conducted some very interesting work in the area of prosthesis control. Specifically, the authors investigated the use of siliconembedded accelerometers to detect MMG signals for the purpose of controlling an externally-powered prosthesis. An important limitation of many EMG-powered prostheses is that it is often difficult to keep EMG sensors placed reliably within a silicon-based socket. Thus, the signal-to-noise ratio of embedded accelerometers was compared with that of non-embedded accelerometers, and the authors found that the embedded accelerometers provided a much better signal quality. In addition, it was recommended that the softest or hardest silicon types should be used to embed the accelerometers, since they provided the least variability. It was also suggested, however, that additional work needs to be done before silicon-embedded accelerometers can be used on a widespread basis for prosthesis control (Silva et al. 2003). Silva et al. (2004) then performed a second investigation that examined the use of a pattern classification procedure in which MMG amplitude was used to control an externally powered prosthesis. Three silicon-embedded accelerometers were placed in separate locations on the end of the stump, and the MMG amplitude values from the accelerometers were used to classify the movements of hand extension and flexion. The results showed that for the two subjects that participated in the study, the classification accuracy values were 70.90% and 70.01%. Thus, it was concluded that MMG could be a useful method for controlling an externally-powered prosthesis, and future studies are needed to determine its feasibility for grading different contraction levels (Silva et al.

2004). Silva et al. (2005) continued their work in the area of prosthesis control by examining the possibility of using the conventional 2-site electromyography system, but with MMG sensors in place of the EMG sensors. The results showed accuracy rates of 88% and 71% for the two subjects that participated in the study. The authors also described several advantages of MMG over EMG for prosthesis control, including less sensitivity to sensor placement, robustness to changing skin impedance levels, and reduced sensor costs (Silva et al. 2005). Alves and Chau (2006) continued their work in prosthesis control

Unique applications of Mechanomyography

129

by using a vision-based segmentation procedure to separate the MMG signals generated during different types of hand grasping sequences. The authors noted that a major challenge with continuous collection of MMG signals for prosthesis control is the subsequent separation of the MMG data stream into segments that reflect individual contractions. Thus, MMG data acquisition was used in conjunction with transverse plane video identification of the gripping activities. The results showed that this system could recognize two different grips with an average accuracy of 97.8% and seven different grips with an average accuracy of 73.0%. Thus, it was concluded that this procedure could be useful for the development of an MMG-based multi-function prosthetic hand (Alves and Chau 2006).

Alves and Chau (2008) also examined the stationarity of MMG signals during different types of gripping activities. Specifically, MMG signals were detected from the wrist extensors during various types of gripping activities, and the signals were tested for weak stationarity using the reverse arrangements test. The results indicated that 20% of the MMG signals recorded during the gripping activities were nonstationary. Thus, it was suggested that time-frequency techniques may be necessary in order to accurately control externally powered prostheses with MMG. In addition, for 47% of the nonstationary signals, the source of the nonstationarity was undetermined, but in 38% of the cases, the nonstationarity was due to a changing variance. The remaining 15% of nonstationary signals were due to time-varying mean, variance and mean, frequency, or frequency and variance values. Therefore, it was concluded that the distribution of the stationary test statistic could reflect the time course of muscle activity used to generate functional grasps (Alves and Chau 2008). Alves and Chau (2008) then conducted a second study that examined vision-based segmentation of MMG signals during gripping activities. Specifically, MMG signals were detected from the wrist extensors and flexors, and a video-based system was used along with MMG to segment the signals. The results showed that the automatic signal segmentation method was capable of tolerating extraneous hand movements, as well as differentiating among different types of grips and estimating their transition times. The accuracy of this technique was 97.8% for two grips and 73.0% for seven grips.

In addition, the detection procedure identified contraction and termination times that were within 173 milliseconds of the times measured with manual segmentation. Thus, it was concluded that this technique could be very useful when using large collections of signals for the purpose of training MMG-based prostheses (Alves and Chau 2008).

Marusiak et al. (in press) recently investigated differences between the EMG and MMG signals from normal, healthy subjects versus those that

130

Travis W. Beck

suffered from Parkinson’s disease. Specifically, the subjects were required to perform submaximal and maximal isometric muscle actions of the forearm flexors while MMG and EMG signals were detected simultaneously from the biceps brachii. The results showed that when compared to the normal subjects, the patients with Parkinson’s disease showed higher MMG amplitude, and lower MMG median frequency values. In addition, there were no consistent differences between the normal subjects and patients with Parkinson’s disease for EMG amplitude or EMG median frequency values. Thus, it was concluded that the MMG signal can be used to identify differences in muscle function between normal subjects and patients with Parkinson’s disease, and, therefore, it could be a useful tool for assessing those with a neuromuscular disease (Marusiak et al. in press). Garcia et al. (2008) investigated the possibility of using MMG to determine arm dominance. The subjects were required to perform submaximal to maximal isometric muscle actions of both the right and left forearm flexors, and surface MMG signals were detected from the biceps brachii. The results showed that there were no significant differences between the dominant and nondominant arms for the MMG amplitude and MPF responses with increases in isometric force. Thus, it was concluded that time and frequency domain parameters from the MMG signal cannot be used to determine arm dominance (Garcia et al. 2008). Orizio et al. (2008) recently investigated the MMG responses for the tibialis anterior during electricallystimulated isometric muscle actions of the dorsiflexors. The electrical stimulation protocol involved both a short (12.5 seconds) and a long duration procedure (approximately 1 hour). Both procedures used various stimulation frequencies, and MMG signals were detected from the tibialis anterior with a laser displacement sensor. The results showed that detecting MMG signals during electrically stimulated contractions could be used to develop a noninvasive technique for identifying the muscle-tendon-joint transfer function. In addition, there were many similarities between the transfer functions obtained from the torque signal and the laser-detected MMG signal.

Thus, MMG could be used to study the properties of the muscle-tendon-joint unit when torque cannot be measured directly (Orizio et al. 2008). TocaHerrera et al. (2008) also examined the MMG and EMG responses following one training session that involved maximal electrical stimulation. Specifically, the untrained subjects were required to perform a maximal unilateral isometric muscle action of the leg extensors, and surface EMG signals were detected from the rectus femoris and biceps femoris both before and after the electrical stimulation training. The training consisted of 100 Hz cycles with a 300µs pulse duration, with cycles of 10-seconds on and 10-seconds off for a total duration of 10 minutes. In addition, the surface MMG signal was detected from the rectus femoris during each maximal muscle action. The results

Unique applications of Mechanomyography

131

showed that when compared to a control group that performed no training, the electrical stimulation resulted in significant increases in isometric leg extension strength, EMG amplitude for the rectus femoris, and a decrease in EMG amplitude for the biceps femoris. There were no changes, however, in MMG amplitude for the rectus femoris for the training or control groups.

Thus, it was concluded that electrical stimulation training could be useful in rehabilitative settings during the first stages of injury when the affected limb cannot be used to perform work. It was also suggested, however, that future work needs to be done to determine if electrical stimulation training is beneficial for those that already perform resistance training on a regular basis (Toca-Herrera et al. 2008).

Kouzaki and Fukunaga (2008) performed an important study that investigated the MMG frequency responses for the soleus muscle during quiet standing. Specifically, the subjects were required to stand barefoot in the upright position with the eyes open or closed, and MMG and EMG signals were detected simultaneously from the soleus.

In addition, a laser displacement sensor was used to detect changes in the location of the subject’s center of mass. The results showed that there was significant coherency between the MMG signal and the displacement of the center of mass. Thus, it was concluded that the kinematic and physiological measures of postural control during quiet standing can be examined with the frequency content of the MMG signal (Kouzaki and Fukunaga 2008). Esposito et al. (in press) investigated the MMG and EMG responses from the finger flexor muscles in elite rock-climbers and control subjects. All subjects were required to perform submaximal to maximal isometric muscle actions of the finger flexors, and MMG and EMG signals were detected during each muscle action. The results showed that the rock climbers demonstrated significantly greater MVC values for the finger flexors than the controls. In addition, EMG amplitude increased with force for both groups, but it was significantly greater for the climbers when compared to the controls at 60%, 80%, and 100% MVC. Furthermore, MMG amplitude increased with force up to 80% MVC for the climbers and 60% MVC for the controls, beyond which, it decreased. The mean EMG MPF values increased from 20-80% MVC in both the climbers and controls, followed by a plateau from 80-100% MVC. Finally, MMG MPF increased from 20-100% MVC for both the climbers and controls, but it was significantly greater for the climbers at 60%, 80%, and 100% MVC. Thus, it was concluded that the combined analysis of MMG and EMG signals provided information regarding potential differences in motor control strategies between the two groups. In addition, the strenuous training performed by the rock climbers may have caused conversion of some slow-twitch fibers to fast-twitch

132

Travis W. Beck

fibers (Esposito et al. in press). Herda et al. (2008) recently examined the acute effects of static or dynamic stretching on isometric leg flexion strength, as well as MMG amplitude and EMG amplitude for the biceps femoris muscle.

Specifically, unilateral isometric leg flexion strength was measured at knee joint angles of 41°, 61°, 81°, and 101° below full leg extension prior to, and immediately following static or dynamic stretching exercises. The results showed that static stretching caused decreases in strength at the 81° and 101° knee joint angles, but not at the 41° and 61° joint angles. In addition, the dynamic stretching had no effect on isometric leg flexion strength. The findings for EMG amplitude indicated that there was no change after the static stretching, and EMG amplitude actually increased after the dynamic stretching, but only for the 101° and 81° knee joint angles. In addition, the static stretching caused increases in MMG amplitude, but only at the 101° knee joint angle, and the dynamic stretching caused increases in MMG amplitude at all joint angles. Thus, it was suggested that the decreases in strength after static stretching may have been due to mechanical, rather than neural factors. In addition, dynamic stretching could be less detrimental to muscle strength than static stretching (Herda et al. 2008).

Jaskólski et al. (2007) recently examined the EMG and MMG responses of agonist and antagonist muscles after eccentric exercise. All subjects performed 25 submaximal eccentric muscle actions of the forearm extensors at 50% of their isometric MVC. Prior to, immediately following, and 24, 48, 72, and 120 hours after the eccentric exercise, the subjects were tested for isometric forearm flexion strength and EMG and MMG amplitude and median frequency of the biceps brachii and triceps brachii muscles. The results showed that the eccentric exercise caused a 34% decrease in isometric forearm flexion strength immediately after exercise, and strength levels did not return to their resting values within the 120 hour time period that was measured. In addition, EMG median frequency decreased for both the biceps brachii and triceps brachii after eccentric exercise, and the MMG amplitude values for both muscles were lower 24, 48, 72, and 120 hours after the eccentric exercise when compared to those immediately after the eccentric exercise. Thus, it was concluded that the similar electrical and mechanical changes in the agonist and antagonist muscles were reflective of a common drive that controlled the agonist and antagonist motor unit pools. In addition, the eccentric exercise-induced changes in EMG and MMG amplitude and frequency parameters may have been due to increased tremor and contractile impairments, such as a reduced rate of calcium release from the sarcoplasmic reticulum, changes in motor control of the agonist and antagonist muscles, and increased muscle stiffness (Jaskólski et al. 2007). Madeleine et al. (2006) investigated the possibility of using MMG for biofeedback during standardized computer work. Specifically,

Unique applications of Mechanomyography

133

MMG and EMG sensors were placed on the upper portion of the trapezius muscle, and the subjects were required to perform work on a computer with either MMG or EMG as a source of biofeedback. The results showed that when MMG was used for biofeedback, muscle activity was significantly lower than when EMG was used. In addition, the MMG-based biofeedback decreased the rating of perceived exertion during the computer work, but it also reduced the total amount of work that could be performed, because the subjects were forced to concentrate on reducing muscle activity, rather than completing their work. Thus, it was concluded that MMG-based biofeedback may be a reliable alternative to the EMG-based methodology, and, therefore, it could be helpful for decreasing the risk of work-related musculoskeletal disorders (Madeleine et al. 2006). Reza et al. (2005) performed an interesting study that investigated the MMG and EMG responses during transcranial magnetic stimulation.

Specifically, surface MMG and EMG sensors were placed over the biceps brachii muscle, and the subjects were required to perform submaximal to maximal isometric muscle actions of the forearm flexors at 5%, 10%, 20%, 30%, 40%, 60%, and 100% of the isometric MVC. The transcranial magnetic stimulation was then used to elicit motor evoked potentials at rest and during the submaximal to maximal muscle actions. The results showed that MMG amplitude increased with force up to 60% MVC, and then decreased. In addition, both the onset latency and length of the silent period decreased with an increase in the isometric force level. Thus, it was concluded that MMG could be very useful for examining the mechanical responses of the muscle during transcranial magnetic stimulation, particularly when surface EMG is not feasible, such as when recording signals in environments that are contaminated with electromagnetic noise (Reza et al. 2005). Drake et al. (2003) examined the effects of oral contraceptives on isometric leg extension strength and MMG and EMG amplitude for the rectus femoris. Specifically, submaximal (25%, 50%, and 75% MVC) and maximal isometric muscle actions of the leg extensors were performed either with or without the use of oral contraceptives.

The results showed that the oral contraceptives had no effect on leg extension strength, MMG amplitude, or EMG amplitude for the rectus femoris. Thus, it was concluded that the use of oral contraceptives did not influence strength or muscle function during isometric muscle actions. It was also recommended, however, that future studies needed to be performed to determine if the use of oral contraceptives could affect strength in other muscle groups (Drake et al.

2003).

In summary, the results from the studies reviewed in this chapter showed the variety of applications for MMG research. It is important for future investigations to carefully and systematically examine the validity of these

134

Travis W. Beck

applications to ensure the accuracy and reliability of the information provided by the MMG signal. To this end, studies are still needed to determine which factors have the greatest influence on the MMG signal and the limitations that they impose.

References 1.

Alves N, Chau T. Vision-based segmentation of continuous mechanomyographic grasping sequences for training multifunction prostheses. Proceedings of the 28th IEEE EMBS Annual International Conference. New York City, USA, August 30September 3, 2006.

2. Alves N, Chau T. Stationarity distributions of mechanomyogram signals from isometric contractions of extrinsic hand muscles during functional grasping.

Journal of Electromyography and Kinesiology 2008; 18:509-515.

3. Alves N, Chau T. Vision-based segmentation of continuous mechanomyographic grasping sequences. IEEE Transactions on Biomedical Engineering 2008; 55:765-773.

4. Beck TW, Housh TJ, Fry AC, Cramer JT, Weir JP, Schilling BK, Falvo MJ, Moore CA. The influence of muscle fiber type composition on the patterns of responses for electromyographic and mechanomyographic amplitude and mean power frequency during a fatiguing submaximal isometric muscle action.

Electromyography and Clinical Neurophysiology 2007; 47:221-232.

5. Beck TW, Housh TJ, Fry AC, Cramer JT, Weir JP, Schilling BK, Falvo MJ, Moore CA. The influence of myosin heavy chain isoform composition and training status on the patterns of responses for mechanomyographic amplitude versus isometric torque. Journal of Strength and Conditioning Research 2008; 22:818-825.

6. Beck TW, Housh TJ, Fry AC, Cramer JT, Weir JP, Schilling BK, Falvo MJ, Moore CA. An examination of the relationships among myosin heavy chain isoform content, isometric strength, and mechanomyographic median frequency. Journal of Strength and Conditioning Research (In Press).

7. Celichowski J, Grottel K, Bichler E. Relationship between mechanomyogram signals and changes in force of human forefinger flexor muscles during voluntary contraction. European Journal of Applied Physiology 1998; 78:283-288.

8. Dahmane R, DjordjeviÄ? S, Smerdu V. Adaptive potential of human biceps femoris muscle demonstrated by histochemical, immunohistochemical and mechanomyographical methods.Medical & Biological Engineering & Computing 2006; 44:999-1006.

9. Drake SM, Evetovich T, Eschbach C, Webster M. A pilot study on the effect of oral contraceptives on electromyography and mechanomyography during isometric muscle actions. Journal of Electromyography and Kinesiology 2003; 13:297-301.

10. Esposito F, Limonta E, Emiliano C, Gobbo M, Veicsteinas A, Orizio C. Electrical and mechanical response of finger flexor muscles during voluntary isometric contractions in elite rock-climbers. European Journal of Applied Physiology (In Press).

11. Evetovich TK, Conley DS, Todd JB, Rogers DC, Stone TL. Effect of mechanomyography as a biofeedback method to enhance muscle relaxation and performance. Journal of Strength and Conditioning Research 2007; 21:96-99.

Unique applications of Mechanomyography

135

12. Garcia MAC, Vargas CD, de Souza MN, Imbiriba LA, de Oliveira LF. Evaluation of arm dominance by using the mechanomyographic signal. Journal of Motor Behavior 2008; 40:83-89.

13. Harba MIA, Chee GE. Muscle mechanomyographic and electromyographic signals compared with reference to actions potential average propagation velocity.

Proceedings of the 19th International Conference of IEEE/EMBS, October 30 – November 2, 1997, Chicago, Illinois, USA. 14. Herda TJ, Cramer JT, Ryan ED, McHugh MP, Stout JR. Acute effects of static versus dynamic stretching on isometric peak torque, electromyography, and mechanomyography of the biceps femoris muscle. Journal of Strength and Conditioning Research 2008; 22:809-817.

15. Jaskólska A, Brzenczek W, Kisiel-Sajewicz K, Kawczyński A, Marusiak J, Jaskólski A.The effect of skinfold on frequency of human muscle mechanomyogram. Journal of Electromyography and Kinesiology 2004; 14:217225.

16. Jaskólski A, Andrzejewska R, Marusiak J, Kisiel-Sajewicz K, Jaskólska A. Similar response of agonist and antagonist muscles after eccentric exercise revealed by electromyography and mechanomyography. Journal of Electromyography and Kinesiology 2007; 17:568-577.

17. Jørgensen F, Lammert O. Accelerometermyography (AMG) II: contribution of the motor unit. In: Biomechanics V-A. Komi PV, Ed., University Park Press, Baltimore, 1976, 159-164.

18. Keidel M, Keidel W-D. The computer-vibromyography as a biometric program in studying muscle function. Biomedizinische Technik 1989; 34:107-116.

19. Kimura T, Hamada T, Ueno LM, Moritani T. Changes in contractile properties and neuromuscular propagation evaluated by simultaneous mechanomyogram and electromyogram during experimentally induced hypothermia. Journal of Electromyography and Kinesiology 2003; 13:433-440.

20. Kouzaki M, Fukunaga T. Frequency features of mechanomyographic signals of human soleus muscle during quiet standing. Journal of Neuroscience Methods 2008; 173:241-248.

21. Madeleine P, Vedsted P, Blangsted AK, Sjøgaard G, Søgaard K. Effects of electromyographic and mechanomyographic biofeedback on upper trapezius muscle activity during standardized computer work. Ergonomics 2006; 49:921-933.

22. Mamaghani NK, Shimomura Y, Iwanaga K, Katsuura T. Mechanomyogram and electromyogram responses of upper limb during sustained isometric fatigue with varying shoulder and elbow postures. Journal of Physiological Anthropology and Applied Human Science 2002; 21:29-43.

23. Marchetti M, Felici F, Bernardi M, Minasi P, Di Filippo L.Can evoked phonomyography be used to recognize fast and slow muscle in man? International Journal of Sports Medicine 1992; 13:65-68.

24. Marusiak J, Jaskólska A, Kisiel-Sajewicz K, Yue GH, Jaskólski A. EMG and MMG activities of agonist and antagonist muscles in Parkinson’s disease patients during absolute submaximal load holding. Journal of Electromyography and Kinesiology (In Press).

136

Travis W. Beck

25. Mealing D, Long, G, McCarthy PW. Vibromyographic recording from human muscles with known fibre composition differences. British Journal of Sports Medicine 1996; 30:27-31.

26. Mitchell SM, Trowbridge CA, Fincher AL, Cramer JT. Effect of diathermy on muscle temperature, electromyography, and mechanomyography. Muscle & Nerve 2008; 38:992-1004.

27. Nonaka H, Mita K, Akataki K, Watakabe M, Itoh Y. Sex differences in mechanomyographic responses to voluntary isometric contractions. Medicine & Science In Sports & Exercise 2006; 38:1311-1316.

28. Orizio C, Diemont B, Esposito F, Alfonsi E, Parrinello G, Moglia A, Veicsteinas A. Surface mechanomyogram reflects the changes in the mechanical properties of muscle at fatigue. European Journal of Applied Physiology 1999; 80:276-284.

29. Orizio C, Solomonow M, Diemont B, Gobbo M. Muscle-joint unit transfer function derived from torque and surface mechanomyogram in humans using different stimulation protocols. Journal of Neuroscience Methods 2008; 173:59-66.

30. Petitjean M, Maton B. Phonomyogram from single motor units during voluntary isometric contraction. European Journal of Applied Physiology 1995; 71: 215-222.

31. Reza MF, Ikoma K, Chuma T.Mano Y.Mechanomyographic response to transcranial magnetic stimulation from biceps brachii and during transcutaneous electrical nerve stimulation on extensor carpi radialis. Journal of Neuroscience Methods 2005; 149:164-171.

32. Rodriquez AA, Agre JC, Knudtson ER, Franke TM,Ng AV. Acoustic myography compared to electromyography during isometric fatigue and recovery. Muscle & Nerve 1993; 16:188-192.

33. Shima N, Rice CL, Ota Y, Yabe K. The effect of postactivation potentiation on the mechanomyogram. European Journal of Applied Physiology 2006; 96:17-23.

34. Silva J, Chau T, Naumann S, Heim W. Systematic characterization of siliconembedded accelerometers for mechanomyography. Medical & Biological Engineering & Computing 2003; 41:290-295.

35. Silva J, Heim W,Chau T. MMG-based classification of muscle activity for prosthesis control. Proceedings of the 26th Annual International Conference of the IEEE EMBS. 2004; September 1-5, San Francisco, CA, USA. 36. Silva J, Heim W, Chau T.A self-contained, mechanomyography-driven externally powered prosthesis. Archives of Physical Medicine and Rehabilitation 2005;86: 2066-2070.

37. Toca-Herrera JL, Gallach JE, G贸mis M, Gonz谩lez LM. Cross-education after one session of unilateral surface electrical stimulation of the rectus femoris. Journal of Strength and Conditioning Research 2008; 22:614-618.

38. Yoshitake Y, Kawakami Y, Kanehisa H, Fukunaga T. Surface mechanomyogram reflects length changes in fascicles of human skeletal muscles. International Journal of Sport and Health Science 2005; 3: 280-285.

39. Yoshitake Y, Shinohara M, Ue H, Moritani T.Characteristics of surface mechanomyogram are dependent on development of fusion of motor units in humans. Journal of Applied Physiology 2002; 93:1744-1752.