Does the frequency content of the surface mechanomyographic signal reflect motor unit firing rates? A brief review

2.1. Repetitive electrical stimulation studies 

There are several studies that have demonstrated that during repetitive electrical stimulation of both isolated and intact skeletal muscle, the mean power frequency (MPF) or peak frequency of the surface MMG signal is very similar to the stimulation rate, but only for unfused tetani [14], [15], [41], [82], [83], [84], [88]. This has suggested that unlike EMG, MMG reflects the mechanical response of the muscle to electrical activation, and this response is largely dependent on the contractile characteristics (i.e. contraction and relaxation times) of the muscle [88], [89]. For example, both surface and intramuscular EMG measure the electrical currents (action potentials) that travel along the muscle fiber membranes, and during most voluntary muscle actions, the action potentials from different motor units are not synchronized [7]. During repetitive electrical stimulation, however, the EMG signal is generated by the electrical currents from the stimulator and, therefore, its frequency content matches the stimulation rate [47]. This is not always the case for MMG. For example, Yoshitake and Moritani [88] reported that during repetitive electrical stimulation of the tibial nerve, MMG MPFs for the medial gastrocnemius and soleus muscles nearly matched the stimulation rate of 5Hz. When the stimulation rate was increased to 20Hz, MMG MPF corresponded closely with the stimulation rate for the medial gastrocnemius, but not for the soleus [88]. It was suggested [88] that a greater proportion of slow-twitch muscle fibers in the soleus than in the medial gastrocnemius allowed its twitches to be fused at a lower stimulation rate. Furthermore, Bichler and Celichowski [14] reported that when compared to fast-twitch motor units, slow-twitch motor units in the rat medial gastrocnemius muscle had longer contraction and half-relaxation times, and, at a given stimulation rate, their twitches were more fully fused and had lower MMG amplitude (peak-to-peak) values. Thus, these findings [14], [88] suggested that the MMG signal may contain information regarding muscle-specific differences in the mechanical response to a given rate of electrical stimulation.

Interestingly, the characteristics of the MMG signal appear to be more closely related to the fluctuations in force that occur during repetitive electrical stimulation than to the net force that is produced by the muscle. Specifically, Stokes and Cooper [82] reported that during electrical stimulation of the adductor pollicis muscle at frequencies between 10 and 20Hz, the amplitudes of the MMG and force oscillation signals decreased similarly with an increase in stimulation rate, but the net force produced by the muscle increased. In addition, Kimura et al. [47] examined the effects of experimentally-induced hypothermia on contractile function in the medial gastrocnemius and soleus muscles during electrical stimulation of the tibial nerve at 10Hz. The authors [47] reported that when the muscles were cooled with ice packs to a temperature of 24°C, there were significant increases in electrically-stimulated plantar flexion force, but decreases in the amplitudes of the MMG signals from both muscles, as well as in the amplitude of the force fluctuation signal. The peak frequencies of the MMG signals from the medial gastrocnemius and soleus, however, nearly matched the stimulation rate in both the control and hypothermic conditions [47]. It was suggested that increases in contraction and relaxation times during hypothermia resulted in greater fusion of motor unit twitches, thereby influencing the amplitudes of both the MMG and force fluctuation signals [47]. In addition, Vaz et al. [84, p. 118] found that during repetitive electrical stimulation of isolated cat soleus muscle, each electrical impulse resulted in a “distinct vibratory signal” from the muscle. The amplitude of the MMG signal decreased, however, when the muscle was held at longer lengths. It was suggested that changes in muscle length may affect the active and passive properties of muscle contraction, thereby influencing the characteristics of the surface MMG signal [84]. Furthermore, Vaz et al. [83] used a similar experimental design with isolated cat soleus muscle, and reported that increases in the stimulation rate resulted in increases in MMG median frequency and decreases in MMG amplitude. It was suggested [83] that the muscle vibrations that generate the MMG signal are produced by the force fluctuations of unfused contractions of motor units.

An important distinction must be made, however, between the MMG signals recorded during repetitive electrical stimulation and voluntary muscle actions. In particular, during supramaximal electrical stimulation, all muscle fibers are activated simultaneously, and the muscle oscillates laterally as a single unit [82]. This situation is very different from a voluntary muscle action, where the mechanical motor unit activities, which are typically not synchronized, are summated at the skin surface [67], [70], [71]. Nevertheless, the findings from electrical stimulation studies have provided two important pieces of evidence that support the use of MMG for examining changes in motor unit firing rates: (a) the surface MMG is a mechanical signal [82] that is generated by motor unit responses to electrical activation [71], and (b) the frequency of the MMG signal increases with the stimulation rate, but only when the twitches are not completely fused [47], [88].

2.2. Muscle fiber type studies 

Several studies [57], [61], [62], [75] have reported that the frequency of the MMG signal is higher for muscles composed of a large percentage of fast-twitch fibers than in those that consist primarily of slow-twitch fibers. For example, Mealing et al. [61] recorded surface MMG signals from the biceps brachii and soleus during voluntary isometric muscle actions at 50% MVC. The authors [61, p. 29] reported that for the biceps brachii muscle, the MMG power density spectrum was “distinctly bimodal,” with significant signal power between 10 and 25Hz. The MMG signal from the soleus muscle, however, was dominated by frequencies between 5 and 10Hz, with only a single peak in the power spectrum at approximately 6Hz [61]. In addition, the authors [61] reported that some individuals demonstrated MMG signals with wide frequency bands for both the soleus and biceps brachii muscles. It was suggested that the frequency content of the MMG signal may reflect the greater proportion of fast-twitch motor units in the biceps brachii when compared to the soleus [44], as well as differences in fiber type composition between individuals [61]. In addition, Mealing and McCarthy [62] reported similar results when the MMG power density spectra from the orbicularis oris and soleus muscles were compared. Specifically, the MMG signals from the orbicularis oris muscle contained a wide range of frequencies, with a MPF of approximately 22Hz [62]. The soleus muscle, however, demonstrated MMG signals that consisted of one “predominant frequency” of approximately 10.8Hz [62, p. 0948]. It was hypothesized that the higher frequency MMG signals for the orbicularis oris than for the soleus may have been due to a greater number of fast-twitch motor units with higher firing rates [62]. In addition, potential differences in fiber type composition may also be reflected in the MMG signals recorded during electrically-stimulated isometric twitches. For example, Marchetti et al. [57] found that when the vastus lateralis and soleus muscles were electrically stimulated with single supramaximal pulses, the vastus lateralis demonstrated higher MMG median frequency values than the soleus. Thus, it was suggested that MMG may also contain information regarding the overall contraction speed of the muscle [57].

Orizio and Veicsteinas [75] provided perhaps the most compelling evidence that MMG frequency is influenced by muscle fiber type composition. The authors [75] examined the patterns of response for MMG MPF from the vastus lateralis muscle during a sustained maximal isometric leg extension in sprinters, long distance runners, and sedentary subjects. For all subjects, MMG MPF followed a “bilinear trend” that was characterized by a rapid decrease during approximately the first 30s of the fatigue test, followed by a slower decline to exhaustion [75, p. 598]. The initial rapid decrease in MMG MPF occurred in approximately 65%, 40%, and 35% of the total contraction time for the sprinters, sedentary subjects, and long distance runners, respectively [75]. In addition, the sprinters demonstrated more power in the high frequency (i.e. 15–60Hz) region of the MMG power density spectrum than the sedentary subjects and long distance runners at all time points during the fatigue test [75]. It was suggested that the duration (expressed as a percentage of the total contraction time) of the initial rapid decrease in MMG MPF may be related to the percentage of fast-twitch muscle fibers in the vastus lateralis [75]. In addition, the high frequency (i.e. 15–60Hz) portion of the MMG power density spectrum could potentially contain information regarding the contribution of fast-twitch motor units to the surface MMG signal [75].

Thus, it is apparent that during voluntary muscle actions, the frequency content of the surface MMG signal is influenced, at least partially, by the fiber type composition of the muscle being investigated. This has important implications in terms of the potential relationship between MMG frequency and motor unit firing rates because previous investigations have reported that high threshold fast-twitch motor units have higher firing rates than low threshold slow-twitch motor units and require greater stimulation rates to achieve complete fusion of motor unit twitches [14], [17], [39]. Theoretically, both of these factors could contribute to the muscle-specific differences in MMG frequency that have been reported in previous studies [57], [61], [62], as well as to the unique MMG MPF patterns demonstrated during fatiguing exercise in individuals with different athletic backgrounds [75].

2.3. MMG MPF versus isometric and dynamic torque relationships 

One of the most promising applications of MMG is its potential use as a tool for investigating motor control strategies. This was recognized relatively quickly by many researchers, and by the mid-1990s, several studies had examined the MMG amplitude and/or MPF versus torque relationships during dynamic or isometric muscle actions [24], [60], [69], [72], [74]. Orizio et al. [72] were among the first investigators to report that MMG MPF increased with isometric torque production. In particular, the authors [72] found that MMG MPF for the biceps brachii muscle remained relatively stable with increases in forearm flexion torque from 10% to 20% MVC, and then increased steadily from 20% to approximately 80% MVC. From 60% to 100% MVC, however, the MMG signal demonstrated an increase in the amount of power present at higher frequencies (i.e. 15–45Hz; Fig. 1), and MMG MPF increased rapidly from approximately 80% to 100% MVC. It was suggested that the increases in MMG MPF from 20% to approximately 80% MVC likely reflected recruitment of motor units with progressively higher firing rates [72]. In addition, the changes in the shape of the MMG power density spectrum, and the rapid increases in MMG MPF at higher torque levels (i.e. above 60–80% MVC) may have been due to a progressively greater reliance on increases in motor unit firing rates for generating additional torque [72]. This motor control strategy is typical for large limb muscles (such as the biceps brachii), in which isometric torque is increased primarily through concurrent modulation of the number of active motor units and their firing rates [29]. At very high torque levels, however, increasing motor unit firing rates becomes more important for generating additional torque when almost all motor units have been recruited [50]. In addition, Akataki et al. [3] recently reported different torque-related patterns for MMG MPF from the biceps brachii and first dorsal interosseous during isometric ramp muscle actions from 5% to 70% MVC. Specifically, the authors [3] found that for the biceps brachii muscle, MMG MPF increased from 5% to approximately 53% MVC, decreased slightly from 53% to approximately 62% MVC, and then increased rapidly from 62% to 70% MVC. For the first dorsal interosseous muscle, however, MMG MPF increased linearly with torque from 5% to 70% MVC [3]. It was suggested that the differences between the MMG MPF patterns from the biceps brachii and first dorsal interosseous muscles may have been due to the unique motor control strategies that are used to increase isometric torque in the two muscles [3]. Specifically, the biceps brachii muscle relies heavily on motor unit recruitment, and, to a lesser extent, on increases in firing rates for modulating isometric torque [17], [50]. For the first dorsal interosseous muscle, however, all motor units are recruited at approximately 50% MVC, and, therefore, firing rate modulation is the primary method for increasing torque from 50% to 100% MVC [26].

View Large Image Download to PowerPoint

Fig. 1. Example of the increases in the high frequency portion (HF%; total power in the 15–45Hz bandwidth relative to the total power in the 6–45Hz bandwidth) of the mechanomyographic (MMG) power density spectrum with increases in isometric torque. The MMG signals were recorded from the biceps brachii in seven healthy male subjects (age range=21–25years) during voluntary incremental step isometric muscle actions of the forearm flexors at a joint angle of 115° between the arm and the forearm. Data reported are from Orizio et al. [72].

In addition, recent studies have reported no changes in MMG MPF for the biceps brachii [8] and vastus medialis [19] during submaximal to maximal concentric isokinetic muscle actions. It was suggested that recruitment, with little change in motor unit firing rates, may be the primary motor control strategy for increasing concentric torque production [8], [19]. This hypothesis is consistent with the results from previous studies, in which motor unit firing rates (detected with intramuscular EMG electrodes) for the biceps brachii muscle remained relatively stable with increases in concentric forearm flexion torque [48], [53]. In addition, recent studies have reported that during eccentric isokinetic muscle actions, MMG MPF increased linearly with torque for both the biceps brachii [10] and vastus medialis [20] muscles. These findings were consistent with those from previous investigations that have suggested that unlike concentric muscle actions, torque production during eccentric muscle actions may be modulated by changes in firing rates, in addition to motor unit recruitment [29], [48], [53].

Collectively, the results from these studies indicated that MMG MPF patterns may provide information regarding the potential differences that exist within a muscle, as well as between muscles, for the motor control strategies that are used to modulate torque production during isometric, concentric, and eccentric muscle actions [3], [8], [10], [19], [20], [72]. In particular, the MMG power density spectrum appears to be influenced primarily by the “global” motor unit firing rate [70, p. 329]. Thus, increases in MMG MPF may reflect recruitment of fast-twitch motor units, which have higher initial firing rates than slow-twitch motor units [17], [39], [70], and/or increases in motor unit firing rates.

2.4. MMG frequency domain responses during fatiguing activities 

Fatiguing exercise provides one of the best situations for examining the potential relationship between MMG frequency and motor unit firing rates. Specifically, it has been hypothesized [58] that during fatiguing activities, the muscle provides proprioceptive feedback to the central nervous system regarding increases in muscle fiber relaxation times. The central nervous system then decreases motor unit firing rates to allow for optimal fusion of motor unit twitches [58]. This concept was termed “muscular wisdom” as a motor control strategy that “minimizes fatigue during prolonged effort” [58, p. 169]. In addition, it has been suggested [77] that during sustained high intensity exercise, some motor units may stop firing (i.e. be de-recruited) prior to the end of the task. Thus, it is possible that muscular wisdom and/or motor unit de-recruitment could result in decreases in MMG frequency (Fig. 2), and several studies have examined the patterns for MMG MPF or median frequency during sustained [32], [42], [55], [56], [66], [73], [85], [86] or repeated [49] isometric muscle actions, as well as during repeated concentric isokinetic muscle actions [9]. For example, Orizio [66] reported that during sustained isometric muscle actions of the forearm flexors at 80% and 100% MVC, the MMG MPF pattern for the biceps brachii was dependent on the intensity of the contraction. Specifically, when the muscle action was performed at 80% MVC, MMG MPF increased throughout approximately the first 30% of the total contraction time (TCT), and then decreased from 30% to 100% TCT [66]. When the muscle action was performed at 100% MVC, however, MMG MPF decreased from 0% to 100% TCT [66]. It was suggested that at 80% MVC, the increase in MMG MPF from 0% to 30% TCT may have been due to increases in motor unit firing rates to sustain the torque output [66]. Furthermore, the decreases in MMG MPF from 30% to 100% TCT, as well as during the sustained isometric muscle action at 100% MVC, may have reflected fatigue-induced decreases in motor unit firing rates (i.e. muscular wisdom) [58], [66]. In addition, Kouzaki et al. [49] examined the MMG median frequency responses from the rectus femoris, vastus lateralis, and vastus medialis muscles during 50 consecutive isometric MVCs of the leg extensors. Each MVC was 3s in duration, followed by a 3-s relaxation period [49]. The authors [49] reported that MMG median frequencies for the rectus femoris, vastus lateralis, and vastus medialis muscles decreased during the first 25 repetitions of the fatigue test and then remained relatively stable during muscle actions 25–50. The initial decreases in MMG median frequency were much greater, however, for the rectus femoris than for the vastus lateralis and vastus medialis muscles [49]. It was suggested [49, p. 14] that the decreases in MMG median frequencies for each muscle may have reflected a fatigue-induced “dropout from recruitment” of fast-twitch muscle fibers, which are more abundant in the rectus femoris than in the vastus lateralis and vastus medialis [44].

View Large Image Download to PowerPoint

Fig. 2. Example of the linear decrease in mechanomyographic (MMG) mean power frequency (MPF) for the vastus lateralis muscle during a sustained isometric leg extension at 50% of the isometric maximum voluntary contraction (MVC). The data shown are the mean (±SEM) values for five healthy male subjects (mean±SD age=32.6±5.2years).

In addition, Esposito et al. [32] recently hypothesized that MMG MPF patterns may provide information regarding fatigue-induced alterations in the motor control strategies that are used during sustained muscle actions. Specifically, the authors [32] reported that during a sustained isometric muscle action of the forearm flexors at 80% MVC, MMG MPF for the biceps brachii increased during the first 6–8 seconds of the muscle action and then decreased to exhaustion. Immediately following this initial fatigue test, the subjects rested for 10min and then performed fatiguing exercise that consisted of repetitive cycles (6-s contraction followed by 4-s rest) of isometric muscle actions of the forearm flexors at 50% MVC [32]. This exercise was performed until the subjects could no longer produce the required 50% MVC torque [32]. After this exercise, the subjects rested for 9-min, and then performed an isometric MVC of the forearm flexors [32]. Following this strength assessment, the subjects performed another sustained isometric muscle action of the forearm flexors at 80% of the new MVC [32]. The authors [32] reported that unlike the first fatigue test (i.e. before the repetitive fatiguing exercise), MMG MPF decreased from the beginning of the second fatigue test to exhaustion. It was suggested that the increase in MMG MPF during the first fatigue test was likely due to increases in motor unit firing rates to maintain the required torque [32]. During the second fatigue test, however, elongation of motor unit twitches in the biceps brachii muscle may have resulted in proprioceptive feedback to the central nervous system that prevented an increase in motor unit firing rates (i.e. muscular wisdom) [32], [58].

Collectively, these findings suggested that MMG MPF and median frequency patterns may provide information regarding the motor control strategies that are used during fatiguing exercise. In particular, changes in the frequency content of the MMG signal could potentially reflect fatigue-induced decreases in motor unit firing rates [32], [58], [66], and/or de-recruitment of fast-twitch muscle fibers [49], [77], both of which could reduce the global motor unit firing rate. Furthermore, the results from these studies also provided support for two key points that are important in examining the potential relationship between MMG frequency and motor unit firing rates: (a) the surface MMG signal is generated by mechanical motor unit activities, but it provides different information from the torque signal during fatiguing exercise (because MMG MPF can increase or decrease during a sustained isometric muscle action, despite a constant torque level) [32], [66], and (b) decreases in MMG MPF or median frequency may reflect reductions in the global motor unit firing rate, rather than decreases in the firing rates of individual motor units (because motor unit de-recruitment could also reduce MMG MPF or median frequency) [49].

2.5. MMG frequency domain responses in young versus elderly subjects and diseased versus healthy muscle 

Previous studies [68], [79] have reported that the frequency content of the surface MMG signal may be influenced by neuromuscular disease. For example, Rhatigan et al. [79] found that during an isometric MVC of the forearm flexors, the MMG MPF values for the biceps brachii muscle were lower for subjects with a neuromuscular disorder (such as Myotonic Dystrophy, Spinal Muscular Atrophy, Myasthenia Gravis, Parkinson’s Disease, Friedreich’s Ataxia, Polymyositis, motorneuron diseases, or facioscapulohumeral and limb-girdle disorders) than for healthy individuals. In addition, Orizio et al. [68] reported that during supramaximal electrically-stimulated isometric twitches of the tibialis anterior muscle, subjects with Myotonic Dystrophy demonstrated surface MMG signals that had lower frequency contents than those from healthy individuals. Specifically, the MMG power density spectra from the subjects with Myotonic Dystrophy had peaks at approximately 22 and 42Hz, but the corresponding spectra for the healthy individuals had peaks at approximately 32 and 48Hz. In addition, the mean MMG MPF value for the subjects with Myotonic Dystrophy was approximately 21Hz less than the mean MMG MPF value for the healthy individuals [68]. It was suggested that an impaired ability to generate muscular tension for the subjects with Myotonic Dystrophy may have reduced the resonant frequency of the muscle, resulting in surface MMG signals with lower frequency contents [68].

There is also evidence to suggest that aging may influence MMG frequency. For example, Akataki et al. [2] reported that during an isometric ramp muscle action of the forearm flexors from 10% to 80% MVC, the MMG MPF versus isometric torque relationship for the biceps brachii was slightly different in young and elderly subjects. Specifically, although both the young and elderly subjects demonstrated a general increase in MMG MPF with isometric torque, the absolute MMG MPF values were typically higher in the young subjects than in the elderly subjects [2]. Furthermore, the MMG MPF versus isometric torque relationship for the young subjects consisted of five “regions,” each of which may have reflected different relative contributions of fast- and slow-twitch motor units to the surface MMG signal [2, p. 508]. The elderly subjects, however, demonstrated an MMG MPF versus isometric torque relationship that consisted of only three regions [2]. It was hypothesized [2, p. 510] that the MMG MPF responses in each of these three regions may reflect a “…functional deterioration in the FT-MUs [fast-twitch motor units], perhaps leading to a greater proportion of force production capability by the ST-MUs [slow-twitch motor units] in the elderly.” Interestingly, there is also evidence to suggest that the MMG MPF versus isometric torque relationship may provide information regarding retrieval of fast-twitch motor units following a resistance training program. Specifically, Esposito et al. [30] reported that before a 12 week isokinetic training program for the leg extensors, the MMG power density spectrum for the vastus lateralis muscle in 10 elderly men (age range=62–78years) during a maximum isometric leg extension was unimodal, with a well-defined peak at approximately 11Hz. After the isokinetic training program, however, the MMG power density spectrum was bimodal, with a large peak at about 15Hz and a second peak at approximately 30Hz. Esposito et al. [31] also examined the MMG MPF versus isometric torque relationships for the biceps brachii muscle in young and elderly subjects. The authors [31] reported that the absolute MMG MPF values for the young subjects were significantly higher than those for the elderly subjects at several torque levels. In addition, the MMG power density spectrum for the young subjects was bimodal at 80% MVC, with peaks at approximately 12Hz and 38Hz [31]. The elderly subjects, however, demonstrated a unimodal MMG power density spectrum at 80% MVC, with a single peak at approximately 12Hz [31]. It was suggested that when compared to the young subjects, the lower frequency MMG signals for the elderly subjects may have been due to a smaller number of fast-twitch motor units in the biceps brachii muscle [31].

Collectively, the results from these studies [2], [31], [68], [79] have provided indirect evidence that the frequency domain of the surface MMG signal contains information regarding motor unit firing rates. Specifically, the presence of neuromuscular disease could potentially influence MMG frequency by: (a) directly reducing motor unit firing rates, and/or (b) interfering with the processes of contraction and relaxation, resulting in fusion of motor unit twitches at lower firing rates. Theoretically, both of these factors could contribute to lower MMG MPF values. In addition, an age-related decrease in the number of fast-twitch motor units could also influence the MMG power density spectrum because fast-twitch motor units have higher initial firing rates [17], [39] and require greater stimulation rates to achieve complete fusion of motor unit twitches [14] than slow-twitch motor units.

3.1. MMG MPF does not always increase with isometric torque 

Previous investigations [1], [3], [70], [72] have suggested that rapid increases in MMG MPF with isometric torque from approximately 60% to 100% MVC may reflect increases in motor unit firing rates. Several other studies, however, have found no change [24], [54], [60] or even a decrease [8], [59] in MMG MPF at these same torque levels. For example, Maton et al. [60] reported that MMG MPF for the biceps brachii increased with isometric forearm flexion torque from 10% to approximately 30% MVC, and then plateaued from 30% to 100% MVC. In addition, Dalton and Stokes [24] found that during isometric forearm flexion muscle actions, MMG MPF for the biceps brachii increased with torque from 10% to approximately 75% MVC and then plateaued from 75% to 100% MVC. It was suggested [24] that the plateau in MMG MPF may have been due to inter-subject variability in the motor control strategy that was used to increase isometric torque in the biceps brachii muscle (i.e. for some subjects, firing rate modulation may be important for increasing torque, whereas for others, it may not). Matheson et al. [59] examined the MMG MPF versus isometric torque relationship for the rectus femoris muscle in high (leg extension torque range=254–330ft.–lbs.) and low (leg extension torque range=101–198ft.–lbs.) force subjects. The authors [59] found that MMG MPF for the rectus femoris muscle increased similarly for the two groups from 20% to 80% MVC. From 80% to 100% MVC, however, MMG MPF for the rectus femoris muscle decreased in the high force group and increased in the low force group [59]. In addition, Beck et al. [8] found that during submaximal to maximal isometric muscle actions of the forearm flexors, MMG MPF for the biceps brachii muscle remained relatively stable from 10% to approximately 50% MVC, increased from 50% to approximately 80% MVC, and then decreased from 80% to 100% MVC. It was suggested that the increase in MMG MPF from 50% to approximately 80% MVC may have reflected recruitment of fast-twitch motor units and/or increases in motor unit firing rates [8]. The decrease in MMG MPF from 80% to 100% MVC, however, may have been due to high levels of muscle stiffness and/or intramuscular fluid pressure [8]. Specifically, muscle stiffness is primarily a function of the number of attached cross bridges [33], and intramuscular fluid pressure increases with isometric torque [80]. Thus, it is possible that at high levels of isometric torque production, muscle stiffness and/or intramuscular fluid pressure could impair the lateral muscle fiber oscillations that generate the MMG signal, thereby influencing MMG frequency [1], [8], [19], [67], [74]. Furthermore, Madeleine et al. [54] reported that there was no significant change in MMG MPF during isometric muscle actions of the first dorsal interosseous from 0% to 100% MVC. The authors [54] hypothesized that in addition to motor control strategies, the frequency content of the MMG signal may also be influenced by the fiber type composition of the muscle being investigated.

Thus, the results from these studies [8], [24], [54], [59], [60] suggested that: (a) MMG frequency may not always reflect potential changes in the global motor unit firing rate, (b) factors such as muscle stiffness and/or intramuscular fluid pressure could interfere with a potential relationship between MMG frequency and the global motor unit firing rate, and/or (c) the global motor unit firing rate may not always increase with isometric torque. However, several studies have reported increases in motor unit firing rates with isometric torque, and there is evidence to support a relationship between the frequency content of the MMG signal and the global motor unit firing rate [17], [50], [61], [88]. Thus, it is more likely that in some cases, factors such as muscle stiffness and/or intramuscular fluid pressure could potentially impair the lateral muscle fiber oscillations that generate the surface MMG signal, thereby interfering with a relationship between MMG frequency and the global motor unit firing rate.

3.2. MMG MPF does not always increase with velocity during maximal concentric or eccentric isokinetic muscle actions 

It has been suggested [37], [65] that during maximal concentric isokinetic muscle actions, there is a velocity-related shift in the contributions of slow-twitch muscle fibers to torque production. Specifically, at low velocities, both slow- and fast-twitch muscle fibers contribute to the torque that is produced by the muscle [37]. With increases in velocity, however, slow-twitch muscle fibers become “unloaded” because they are unable to contract rapidly enough to keep up with the speed of the movement [65, p. 452]. Thus, it is possible that with increases in velocity, the contributions of slow-twitch motor units to the MMG signal decrease, potentially resulting in greater MMG MPF values. Cramer et al. [23] provided tentative support for this hypothesis by demonstrating that during maximal concentric isokinetic leg extensions, MMG MPF for the rectus femoris, vastus lateralis, and vastus medialis muscles was greater at 300°s−1 than at 60, 120, 180, and 240°s−1. Furthermore, Cramer et al. [22] reported that during maximal concentric isokinetic leg extensions at velocities ranging from 60–480°s−1, there were velocity-related increases in MMG MPF for both the rectus femoris and vastus lateralis, but not for the vastus medialis. It was suggested [22] that the muscle-specific differences in the MMG MPF responses may have been due to potential differences in fiber type composition among the rectus femoris, vastus lateralis, and vastus medialis muscles [44]. Ebersole et al. [28], however, reported that during maximal concentric isokinetic leg extensions, there was no change in MMG MPF for the vastus lateralis muscle with an increase in velocity from 60 to 300°s−1. Thus, it is unclear if movement velocity has a significant effect on MMG MPF during maximal concentric isokinetic muscle actions. It is possible, however, that any potential influence could be related to muscle fiber type composition [22].

In addition, it has been suggested that there may be de-recruitment of slow-twitch motor units and selective recruitment of fast-twitch motor units with increases in velocity during eccentric muscle actions [65]. Theoretically, this could also result in a velocity-related increase in MMG MPF. Evetovich et al. [34], however, reported that during maximal eccentric isokinetic muscle actions of the leg extensors at velocities of 60, 120, and 180°s−1, there were no changes in MMG MPF for the vastus lateralis with increases in velocity. Furthermore, Cramer et al. [21] found that MMG MPF for the rectus femoris, vastus lateralis, and vastus medialis actually decreased approximately 12% with an increase in velocity from 60 to 120°s−1 during maximal eccentric isokinetic muscle actions of the leg extensors. In addition, there was no change in MMG MPF with an increase in velocity from 120 to 180°s−1 [21]. It was suggested [21] that during maximal eccentric isokinetic muscle actions, the velocity-related patterns for MMG MPF may be similar to those for eccentric isokinetic peak torque (which did not change with increases in velocity from 60 to 180°s−1).

Thus, the results from the studies that have examined the MMG MPF responses during maximal concentric [23], [28] or eccentric [21], [34] isokinetic muscle actions indicated that: (a) there is not always a velocity-related increase in the global motor unit firing rate during maximal isokinetic muscle actions, and/or (b) there is a velocity-related increase in the global motor unit firing rate, but it is not always reflected in MMG MPF. The possibility for slow-twitch muscle fibers to become unloaded during concentric muscle actions [37], or even de-recruited during eccentric muscle actions [65] suggested, however, that in some cases, MMG MPF may not be influenced by changes in the global motor unit firing rate. As hypothesized by Cramer et al. [22], it is possible that any potential relationship between MMG MPF and the global motor unit firing rate may also be influenced by muscle fiber type composition.