Abstract
The purpose of this study was to investigate the influence of changes in ankle joint angle on the mechanomyogram (MMG) amplitude of the human medial gastrocnemius (MG) muscle during voluntary isometric plantarflexion contractions. Ten healthy individuals were asked to perform voluntary isometric contractions at six different contraction intensities (from 10% to 100%) and at three different ankle joint angles (plantarflexion of 26°; plantarflexion of 10°; dorsiflexion of 3°). MMG signals were recorded from the surface over the MG muscle, using a 3-axis accelerometer. The relations between root mean square (RMS) MMG and isometric plantarflexion torque at different ankle joint angles were characterized to evaluate the effects of altered muscle mechanical properties on RMS MMG.
We found that the relation between RMS MMG and plantarflexion torque is changed at different ankle joint angles: RMS MMG increases monotonically with increasing the plantarflexion torque but decreases as the ankle joint became dorsiflexed. Moreover, RMS MMG shows a negative correlation with muscle length, with passive torque, and with maximum voluntary torque, which were all changed significantly at different ankle joint angles.
Our findings demonstrate the potential effects of changing muscle mechanical properties on muscle vibration amplitude. Future studies are required to explore the major sources of this muscle vibration from the perspective of muscle mechanics and muscle activation level, attributable to changes in the neural command.
Keywords: muscle surface vibration, mechanomyography, muscle mechanics, isometric voluntary contraction
Introduction
Muscle is known to generate low-frequency physiological vibrations during contraction, called the mechanomyogram (MMG). While the electromyogram (EMG) records electrical signals generated by excitations of active motor units (MUs) recruited during muscle contraction (Mills, 2005), MMG records the subsequent mechanical response that is typically detectable from the skin surface over a muscle of interest (Beck et al., 2005). These mechanical vibrations appear to be mainly attributable to neural excitations of a muscle, as supported by previous findings that the MMG amplitude is linked to MU recruitment (Akataki et al., 2001) and that the MMG frequency band is believed to represent the global firing rate of active but partially fused MUs (Orizio et al., 2003). Furthermore, it has been proposed that muscle vibrations are impacted by both the active and passive properties of muscles. Likening active muscle to a vibration beam, there has been evidence that the lateral vibrational properties of skeletal muscle can be influenced by muscle tension, by the elastic modulus, and by muscle length (Cole and Barry, 1994; Frangioni et al., 1987). Perhaps surprisingly, the potential effects of changes in muscle length on the MMG amplitude have not been explored extensively, and there are discrepancies in the MMG-joint angle relations (Miyamoto and Oda, 2003; 2005). Given that changes in muscle length are accompanied by changes in other factors such as muscle stiffness and maximum force-generating capacity (Gordon et al., 1966; Yamauchi and Koyama, 2019), the purpose of this study is to explore the potential effects of these various factors on the MMG amplitude.
The impact of changes in muscle length on MMG was examined in frog gastrocnemius muscles during electrically stimulation-induced isometric contractions, showing that the MMG amplitude is a function of muscle length. For instance, the MMG amplitude increases before reaching optimal muscle length (i.e., length at where the maximum isometric force is generated) but decreases afterward (Frangioni, 1987). Other animal studies have also supported the idea that changes in muscle mechanical properties induced by changes in muscle length can affect the MMG amplitude (Vaz et al., 1997).
A few studies explored the potential effects of changes in muscle length on MMG in human preparations, but there is still controversy. For example, the MMG amplitude decreases with increasing muscle length during evoked contractions of the biceps brachii muscle (Miyamoto and Oda, 2005), whereas it tends to increase during maximal voluntary contractions for triceps surae (Miyamoto and Oda, 2003). Other findings within human rectus femoris muscle during isomeric contractions also demonstrated that for the same contraction level, there tends to be a higher MMG amplitude at longer muscle length via the manipulation of knee joint angles (Ebersole et al., 1998). One possible reason for this inconsistency is that the MMG amplitude may be highly impacted by the way that a muscle is activated. Moreover, it is likely that the muscle mechanical vibrations may be impacted by other mechanical properties related to muscle length, which may offset the effects of muscle length change by itself. Till now, however, there is no systematic research to identify the effects of changes in muscle length and other associated mechanical properties on muscle surface vibrations in human preparations during voluntary contractions.
Motivated by this need, we set out to explore the influence of changes in muscle mechanical properties which result from changes in muscle length, on root mean square (RMS) MMG of human MG muscles during voluntary isometric contractions. Changes in muscle length were induced by changing ankle joint angle. Relations between RMS MMG and isometric plantarflexion torque were derived at different ankle joint angles and were then fit with a power function to characterize these relations. Our primary hypothesis is that the associations between RMS MMG and plantarflexion torque would be affected by ankle joint angle, potentially due to changes in muscle length and associated changes in muscle mechanical properties. A correlation analysis was further performed to evaluate the potential effects of these passive muscular factors on RMS MMG, to evaluate the potential mechanism of MMG generation from the perspective of muscle internal mechanics.
Methods
Participants
Ten healthy young individuals (age: 25.8 ± 2.15 years; height: 170.0 ± 10.3 cm; weight: 66.9 ± 12.5 kg; F/M: 4/6) participated in this study. They had no relevant medical history of neuromuscular disorders in their lower limb. Before testing, an explanation of the purpose and procedures of the experiment were provided, and informed consent was then obtained from all participants. All procedures complied with Helsinki declaration compliance and were approved by the Northwestern University’s Institutional Review Board.
Experimental set-up and signal detection
Subjects were asked to sit upright in a Biodex chair with the knee joint angle kept at full extension. The right foot was fixed on a footplate attached with a 6-axis force measuring device (Omega160, ATI Industrial Automation, Apex, NC, USA). Surface EMG electrodes (Bagnoli, Delsys Inc., Boston, MA, USA) were placed on the middle part of the lateral gastrocnemius (LG), soleus (SOL), and tibialis anterior, and on the distal part of the MG muscle belly. A ground electrode was placed on the lateral malleolus, and the attachment area used for all the electrodes was cleaned with alcohol pads before electrode placement. MMG signals were recorded from the distal MG muscle using a 3-axis accelerometer (ADXL335, Analog Devices, Wilmington, MA, USA; sensitivity: 300 mV/g), with z-direction perpendicular to the skin surface over MG muscle (Figure 1). MMG signals were also collected from the belly part of both LG and SOL muscles with other accelerometers (ADXL354, Analog Devices, Wilmington, MA, USA; sensitivity: 400 mV/g). EMG, MMG, and torque signals were recorded simultaneously at a sampling frequency of 2 kHz (NI USB-6259 BNC, National Instrument, Austin, TX, USA).
Figure 1.
Experimental set-up.
Experimental protocol
Experiments were performed at three different ankle joint angles: plantarflexion of 26° (−26°, P1), plantarflexion of 10° (−10°, P2), and dorsiflexion of 3° (3°, P3), which were measured relative to the neutral position (0° defined as perpendicular between the shank and foot). Each ankle joint angle was set by adjusting the foot-plate position for all the participants. At each ankle angle and in a random order, the subjects performed one trial passively and three maximum voluntary isometric contractions (MVICs).
Average value of the three MVIC trials (details in Data Analysis section) was used as the calibration level for estimating the submaximal contraction intensities. Each subject performed three isometric plantarflexions at five different levels: 10%, 20%, 30%, 50%, and 70%MVIC. At each ankle joint angle, the order of the contraction intensities was randomized for each subject. The different intensities were provided using torque visual feedback to subjects in real-time on a monitor. They were encouraged to reach the target level from resting state and then maintain the target level ± 3%MVIC for at least 6 s. A 30-s break was given between two consecutive trials to minimize muscle fatigue. To estimate the MG muscle length of each subject, the shank length was measured three times from the lateral femur epicondyle to the lateral malleolus, the average value of which was recorded for further analysis.
Data Analysis
The torque signals were lowpass filtered by 4th order Butterworth filter with a cutoff frequency of 6 Hz. To determine representative level of muscle contractions for each trial and to minimize the motion artifacts in MMG signals at the beginning and end of the muscle contractions, a 4-s segment in the middle of a sustained contraction was selected with minimum standard deviation criterion. The mean value of the torque segment was calculated to estimate the actual contraction level (hereafter called Torque) in each trial.
Raw EMG signals and MMG signals were band-pass filtered using 4th order Butterworth filters with a passband of 20–450 Hz and 4–40 Hz (Cè et al., 2015; Lei et al., 2011), respectively. For all the trials, RMS MMG values of the z-axis MMG signals (i.e., perpendicular to the skin) were calculated, using the same time segment, and these were then converted to acceleration (mm/s2). To evaluate the reliability of RMS MMG, the standard deviation of the record at a given torque was calculated for all contraction intensities. Peak RMS MMG was determined using the average value of RMS MMG from MVIC trials. Similarly, RMS EMG values of the MG, LG, and SOL muscles were also calculated, using the same time segment, and the load-sharing ratio of the MG muscle was then determined by calculating the average value of the ratio of RMS EMG of the MG muscle to the summation of RMS EMG of the three muscles across different torque levels.
Since changes in ankle joint angle changed muscle length, but also resulted in different passive and active muscle tension levels (Ateş et al., 2018), further analyses were performed to calculate passive torque, maximum voluntary torque, and MG muscle length at each ankle angle. For each subject, the passive torque was estimated by calculating an average value of the first 1-s of the lowpass filtered torque signal in all trials, followed by the subtraction of the passive torque value at P1. Thus, the passive torque data at P1 were excluded in the further analyses. The MG muscle length was estimated from the generic model ‘3DGaitModel2392’ in OpenSim v3.3 (Arnold et al., 2010) after subject-specific scaling, and was then normalized by shank length.
Statistical Analysis
First, a linear mixed-effects model was used to test hypotheses: (1) RMS MMG changes with Torque; and (2) the relation between RMS MMG and Torque is impacted by changing ankle joint angle. The relation between RMS MMG and Torque was well characterized using a power function (i.e., y = Ax3 + B where y is RMS MMG, A the scaling factor, x Torque, and B RMS MMG at resting state or the noise level of MMG signals), by qualitative visual observation. Fixed effects were thus the intercept, Torque3, and interaction between ankle joint angle and Torque3. Subjects were treated as a random effect. The mixed-effects model revealed that the B value was not significantly changed by changing ankle joint angles (p = 0.067), and thus only the A value was used for further analyses to infer the magnitude of RMS MMG at different ankle joint angles when Torque is constant. Moreover, the magnitude of the A value was quite small due to the exponent of 3, and the power function above was thus rewritten as y = 0.001·Ax3 + B to make the A value more readable.
To compare key variables (i.e., the peak RMS MMG, the normalized muscle length, the passive torque, the maximum torque) at different ankle joint angles, other linear mixed-effects models were designed, treating ankle joint angle as a fixed effect and subject as a random effect.
The Kolmogorov-Smirnov test was performed to assess the normality of the A value and the key variables. Since these variables did not satisfy the assumption (p < 0.05), the Spearman ranked correlation (r) analysis was applied to evaluate potential associations between the A value and the key variables. The p-values of the correlation coefficients were then adjusted via controlling the false discovery rate with the Benjamini–Hochberg procedure (Benjamini and Hochberg, 1995).
Results
On average, the standard deviation of RMS MMG at a given torque was smaller than 6 mm/s2 (~4% of the peak RMS MMG). Moreover, the linear mixed-effects model examining the relations between RMS MMG and Torque including the MVIC trials at each ankle joint angle showed a high determination coefficient (R2 = 0.880), indicating that our model explains the relations well.
Figure 2 shows a representative relation between RMS MMG and Torque at each ankle joint angle from one participant. For all ten participants, RMS MMG increased significantly with increasing ankle plantarflexion torque at all ankle joint angles (F(1,534) = 39.622, p < 0.0001). At different ankle joint angles, altered relations between RMS MMG and Torque were detected (F(1,534) = 7.127, p < 0.0001).
Figure 2.
Representative relations between RMS MMG and ankle plantarflexion torque at three different positions (P1: plantarflexion of 26°; P2: plantarflexion of 10°; P3: dorsiflexion of 3°).
Contrast analysis revealed that the A value at P1 was significantly greater than that of either P2 (F(1,534) = 8.197, p = 0.006, dz = 0.959 ) or P3 (F(1,534) = 12.879, p = 0.001, dz = 1.154) (Figure 3). However, no significant difference was found between P2 and P3 (F(1,534) = 0.545, p = 0.461, dz = 0.598). In summary, the A value was the greatest at P1 with a median of 1.225 (IQR = 1.612–0.281), followed by P2 with a median of 0.342 (IQR = 0.568–0.120) and P3 with a median of 0.132 (IQR = 0.392–0.066).
Figure 3.
Comparison of the A value at different ankle joint angles. The greatest A value was found at P1 with the shortest muscle length (p < 0.001).
The repeated measures ANOVA test further revealed that the peak RMS MMG at P3 was significantly smaller than that of P1 (F(1,27) = 9.498, p = 0.014, dz = 0.712) and P2 (F(1,27) = 5.817 p = 0.035, dz = 0.563), although no significant difference was found between P1 and P2 (F (1,27) = 12.597, p = 0.502, dz = 0.156; Figure 4).
Figure 4.
Comparison of the peak RMS MMG at three different ankle joint angles.
As the dominant factor covaried with ankle joint angle, there was a significant increase in the normalized MG muscle length (F(2,27) = 1588.062, p < 0.0001; Figure 5A) with the ankle joint became dorsiflexed. The similar trend was also found in the passive torque (F(2,27) = 25.853, p < 0.0001; Figure 5B) and the maximum torque (F(2,27) = 17.983, p < 0.0001; Figure 5C).
Figure 5.
Comparison of the muscle passive mechanical properties, Normalized MG muscle length (A), passive torque (B), and maximum torque (C), at three different ankle joint angles.
Negative correlations were found between the A value and normalized muscle length (r = −0.346, p = 0.061), passive torque (r = −0.661, p = 0.002), and maximum torque (r = −0.900, p < 0.0001). When evaluating the correlations between the A value and the associated mechanical properties at each ankle joint angle: there was a positive correlation between the A value and normalized muscle length (P1: r = 0.872, p = 0.001; P2: r = 0.659, p = 0.038; P3: r = 0.622, p = 0.055); for the correlation between the A value and maximum torque, there was a significant negative relationship (P1: r = −0.746, p = 0.018; P2: r = −0.806, p = 0.008; P3: r = −0.830, p = 0.006); and no significant correlation was found between the A value and passive torque (P2: r = −0.547, p = 0.102; P3: r = −0.462, p = 0.179).
Discussion
This study aimed to evaluate the effects of changes in ankle joint angle on the RMS MMG of the human MG muscle during voluntary isometric contractions. Our main findings were that: (1) Altered relations between RMS MMG and Torque were observed at different ankle joint angles, detected by significant changes in the parameter A; (2) The A value at P1 (plantarflexion of 26°) was significantly greater than that at P2 (plantarflexion of 10°) and P3 (dorsiflexion of 3°); (3) The peak RMS MMG at P3 (dorsiflexion of 3°) was significantly smaller than that at P2 (plantarflexion of 10°) and P1 (plantarflexion of 26°), indicating that the peak RMS MMG was highly affected by ankle joint angle, and thus ankle joint angle might be the main contributor to differences in the A values; (4) The A values were significantly correlated with the passive torque and the maximum torque, implying that the muscle vibration amplitudes are likely to be affected by changes in muscle mechanical properties induced by changes in muscle length. These findings indicate that the MMG generation may also be strongly impacted by muscle internal mechanics.
Physiological implications of the model parameter A
Varied relations between RMS MMG and Torque were observed at different ankle joint angles, suggesting that muscle vibration amplitude could be impacted by changes in muscle mechanical properties, which are induced in turn, by changes in ankle joint angle. Given the fact that at different ankle joint angles, the B value estimated from the power function (RMS MMG= A*Torque3 + B) is consistent, while the parameter A changes significantly, we can treat the parameter A as a surrogate of muscle vibration amplitude at each ankle joint angle. Collectively, it is plausible to evaluate the effects of muscle mechanical properties (i.e., muscle length, muscle tension, and maximum force-capacity) on muscle vibration amplitude by performing correlation analysis between the parameter A and these factors.
Potential effects of changes in muscle length on RMS MMG
The A value at P1 was significantly greater than that at both P2 and P3, suggesting that altered muscle vibration amplitude at different ankle joint angles may be linked to different muscle lengths. Interestingly, the A value was negatively correlated with the normalized muscle length when considering the pooled data but was positively correlated with the normalized muscle length at each ankle joint angle. It is reported that the amplitude of the acoustic sound (a different recording technique of muscle mechanical vibrations), recorded from the belly of frog gastrocnemius muscles, increases with increasing muscle length on the ascending limb of the tension-length curve for that muscle (Frangioni, 1987). These findings may potentially be extended to human MG muscles because human MG muscles also operate primarily on the ascending limb (Arnold and Delp, 2011; Ishikawa et al., 2007).
Furthermore, likening active muscle to a vibrating beam (Cole and Barry, 1994; Frangioni, 1987), the fact that the transverse deflection of a vibrating beam can be expressed as a function of the beam length (Cain and Hulse, 1990a) can also support our findings that the A value is positively correlated with the normalized muscle length at each ankle joint angle. Collectively, the decrease in the A value with ankle dorsiflexion appears to be attributable to other factors (e.g., muscle stiffness, passive muscle tension, and muscle maximum torque), which covaried with changes in muscle length.
Potential effects of changes in passive torque on RMS MMG
Changes in ankle joint angle lead to changes not only in muscle length but also in passive torque (or tension). Given that the passive tension is associated with MTU stiffness (Koo et al., 2013), the decrease in the A value from P1 to P3 may be largely attributable to an increase in the MTU stiffness. This can also be supported by an earlier study that RMS MMG decreases as MTU stiffness increases(Longo et al., 2014). Accordingly, it appears that a lower MTU stiffness may allow the muscle to oscillate to a greater extent, thus leading to a higher RMS MMG at shorter muscle lengths.
Our results also suggest that peak RMS MMG at P1 is larger than at P2 and P3, demonstrating that increasing muscle vibration amplitude is expected to occur with ankle joint angle plantarflexed, specifically during maximal voluntary contractions. However, an earlier finding in human triceps surae during maximal voluntary contractions suggested a decreased RMS MMG when muscle stiffness is reduced with ankle joint angle plantarflexed (Miyamoto and Oda, 2003). This inconsistency may be due to different postures in the experimental set-up (i.e., prone position vs. seating position) and thus different muscle activation strategies. Considering that at shorter muscle length, the reduction in half-relaxation time is more significant than that in contraction time (Gandevia and McKenzie, 1988), it is plausible that unfused tetanus at shorter muscle length likely induces greater force ripples and that these fluctuations could be manifested as an increase in the MMG amplitude (Yoshitake et al., 2008). Further studies are needed to consider both muscle activation strategies and muscle mechanical properties during voluntary contractions.
Potential effects of changes in maximum torque on RMS MMG
Overall, the A value was negatively correlated with the maximum torque. Given that maximum voluntary strength of plantar flexors is positively correlated with muscle anatomical cross-sectional area (Bamman et al., 2000), our findings indicate that RMS MMG may also be associated with the cross-sectional area or muscle size.
According to beam theory, the bending stiffness of a beam is a function of Young’s modulus, the area moment of inertia of the beam cross-section about the axis of interest, and the length of the beam (Cain and Hulse, 1990b). Likening active muscle to a vibrating beam, it is plausible that the greater the muscle cross-sectional area, the greater the bending stiffness (i.e., the resistance against transverse or lateral vibrations). Since the transverse deflection of a vibrating beam is inversely proportional to the bending stiffness (Cain and Hulse, 1990a), the decrease in the A value or the peak RMS MMG with increasing the maximum torque may be attributable to such increase. However, special care should be taken because the deflection is a function of not only the bending stiffness but also the bending moment. Manifested in the shrinking of the myofilament lattice, the radial force can be generated with longitudinal muscle force (Brenner B and LC., 1985). This radial force can be regarded as a potential source of variations in the bending moment during active muscle contraction, also contributing to lateral deflections of a contracting muscle. Further studies are required to describe these complicated internal muscle mechanics.
Potential Clinical Applications
Based on our finding that the RMS MMG-Torque relationship appears to be determined mainly by the peak RMS MMG at 100%MVIC, we performed further analyses to explore whether the RMS MMG-Torque relationship under submaximal contractions can be affected by ankle joint angle as well. As a result, the relation between RMS MMG and submaximal torque can be well characterized with a linear model (R2 = 0.882), demonstrating a strong positive correlation at each ankle joint angle; however, this relation is not significantly affected by ankle joint angle (p = 0.793). This linear RMS MMG-Torque relationship under submaximal contractions (up to 70%MVIC) suggests that it is acceptable to predict the absolute plantarflexion torque level from MMG RMS regardless of ankle joint angle. Considering that the force generation in many clinical populations (i.e., stroke survivors) is considered submaximal contraction levels, MMG could be a potential tool for the assessment of muscle force/joint torque capacity during manual muscle testing (MMT) in routine clinical settings, although further studies are required to test in clinical populations.
Limitations
There are other factors to be considered. First, MMG signals generated by muscle contraction can be attenuated by subcutaneous tissues, which act as a low-pass filter (Herda et al., 2010; Scheeren et al., 2016). Considering that the thickness of the subcutaneous layer over the MG muscle was significantly different between two knee angles (i.e., fully extended and 90° flexed) (Avancini et al., 2015), the difference in the A value between ankle joint angles might result from the changes in subcutaneous tissue thickness, as also mentioned in an earlier study (Cescon et al., 2004).
Given that the load-sharing strategy among these muscles can be changed by different ankle angles (Alrowayeh et al., 2011), the RMS MMG-Torque relations may be affected by the load-sharing between all muscles across the ankle joint. Further analyses revealed that the load-sharing ratio of the MG muscles was significantly different by ankle angles (p = 0.018). However, this significant difference may be attributable to changes in RMS EMG of the SOL muscles by ankle angle (p < 0.001) because changes in RMS EMG of the MG (p = 0.152) and LG (p = 0.858) muscles are relatively consistent. Collectively, it appears that changes in RMS MMG of the MG muscles were not affected by the load-sharing, but the net ankle joint torques might be influenced by the angle-dependent SOL muscle activities.
Considering that crosstalk could potentially impact MMG signals, the crosstalk effects from LG and SOL were evaluated by cross-correlation analyses between MG and each of these muscles (i.e., SOL-MG, LG-MG) (Beck et al., 2010). Given that the peak cross-correlation coefficients for the SOL-MG muscle pair were not changed significantly with ankle joint angle (p = 0.238), it appears that the SOL activation may not contribute materially to changes in the A value at different ankle joint angles. However, there was a significant effect of changes in ankle joint angle on the peak cross-correlation for the LG-MG muscle pair (p < 0.05), while only showing a significant difference between P1 and P3 (ηp2 = 0.072; p = 0.044). This indicates that the relative contribution of LG may alter the ankle joint angle-dependent A value, but these significant differences were only observed within 20% MVIC, meaning that during moderate to high contractions, the cross-correlation coefficients for the LG-MG muscle pair were not significantly impacted by changes in ankle joint angle. Considering that a measurable difference in RMS MMG with changes in ankle joint angle was found mainly at 100% MVIC, the potential contamination of the MMG signals from both LG and SOL might be small.
Changes in muscle architecture during contraction of the MG muscle, a pennate muscle, may also alter muscle vibratory properties (Cescon, 2004) because changes in pennation angle (i.e., the angle between fascicle orientation and the long axis of the muscle) depend on both contraction intensities and ankle joint angles (Kawakami et al., 1998; Kim et al., 2013; Maganaris et al., 1998). Given that the MMG is generated by lateral oscillations of active muscle fibers (Orizio, 1993), such different pennation angles may presumably change the principal axis of the muscle vibration and thus potentially change the skin surface vibration. In this study, the MMG sensor was located at the distal part of the MG muscle, and the effect of the pennation angle on RMS MMG may thus be minimized because changes in the pennation angle were relatively less at the distal location compared to the proximal location (Héroux et al., 2016). Future studies are required to understand the effect of changes in muscle architecture on muscle vibratory properties.
Lastly, this study did not consider potential sex-dependent RMS MMG-torque relationship (Yoshino and Shimomura, 2017). Based on our preliminary analyses with the dataset in this study, the A value was independent of sex at each ankle joint angle (P1: p = 0.326; P2: p = 0.142; P3: p = 0.147). It would be interesting to investigate the joint angle-dependent RMS MMG-torque relationship in both males and females.
Conclusions
The present study showed that RMS MMG of the MG muscle increased markedly with increasing ankle plantarflexion torque at three different ankle joint angles, ranging from plantarflexion to dorsiflexion. The A values estimated from our fitting model and peak RMS MMG, were different at different ankle joint angles, suggesting that the muscle vibration amplitude can be altered depending on ankle angles. Besides, correlations between the A value and muscle length, passive torque as well as maximum torque indicate that there are significant effects of muscle mechanical properties on muscle vibration amplitude. Future studies are required to study how motor unit behavior and intrinsic properties such as material properties and muscle architecture can alter muscle vibration properties, to understand the dominant source of muscle surface vibrations.
Acknowledgments
We thank all participants in this study. This study was supported by grants from the National Institute on Disability, Independent Living, and Rehabilitation Research (90SFGE0005), the Davee Foundation Stroke Research Seed Grant initiative and the Northwestern University Department of Neurology, Division of Stroke and Neurocritical Care, and the National Institutes of Health (R01HD089952).
Author Biography
Fandi Shi obtained her master’s degree in biomedical engineering from Northwestern University, Illinois, United States. She is currently a Ph.D. candidate in the Biomedical Engineering department at Northwestern University and works as a research student in Single Motor Unit Laboratory at the Shirley Ryan AbilityLab (formerly the Rehabilitation Institute of Chicago, or RIC). Her research focus is to understand muscle physiological changes following stroke via the technique of mechanomyogram (MMG). Currently, her research goal is to explore the potential relations between the MMG response with motor impairment-related measurements, such as force fluctuation. For the long-term goal, she plans to understand the underlying generation mechanism of MMG and apply her findings to clinical settings to provide a readily accessible approach to monitor muscle motor impairment and to track the rehabilitation progress of clinical populations.
William Zev Rymer, MD, PhD is Director of the Single Motor Unit Laboratory at the Shirley Ryan AbilityLab (formerly the Rehabilitation Institute of Chicago, or RIC). He served as the former Vice President for Research and the John G. Searle Chair of Rehabilitation Research at the RIC. Dr. Rymer has appointments as Professor of PM&R, Physiology, and Biomedical Engineering at Northwestern University. He received his medical training from the University of Melbourne, graduating with honors, and his PhD in Neuroscience from Monash University. His research concerns the neural control and biomechanics of movement in human and animal models, and the disturbances of voluntary movement and their origins in people with neurological disabilities, particularly those with stroke and spinal cord injury. He currently holds grants from the NIH, NIDILRR, and several foundations. He has published more than 300 papers, with more than 150 in the fields of biomechanics and control of movement. Dr. Rymer has conducted large and successful pre- and post-doctoral training programs in bioengineering and physiology for many years. He is currently Project Director of a NIDILRR-funded multi-center clinical trial to evaluate the effectiveness of intermittent hypoxia therapy in individuals with spinal cord injury.
Jongsang Son obtained a Ph.D. in neuromuscular biomechanics and rehabilitation engineering from Yonsei University, Wonju, South Korea. He is currently a Research Scientist at Shirley Ryan AbilityLab and a Research Assistant Professor in the Department of Physical Medicine and Rehabilitation at Northwestern University, Chicago, IL, United States. His research priorities are to understand the underlying neuromuscular mechanisms of motor impairments in various clinical populations and to investigate neuromuscular adaptations to potential rehabilitation interventions. Ultimately, he hopes to translate his discoveries into practical interventions that can prevent the development of motor impairments and help people with chronic disease improve motor functions.
Footnotes
Conflicts of interests
The authors have no conflicts of interest
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 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(1–2):19–25. [DOI] [PubMed] [Google Scholar]
- Alrowayeh HN, Sabbahi MA, Etnyre B. Similarities and differences of the soleus and gastrocnemius H-reflexes during varied body postures, foot positions, and muscle function: multifactor designs for repeated measures. BMC Neurology. 2011;11(1):1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Arnold EM, Delp SL. Fibre operating lengths of human lower limb muscles during walking. Philosophical Transactions of the Royal Society B: Biological Sciences. 2011;366(1570):1530–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Arnold EM, Ward SR, Lieber RL, Delp SL. A model of the lower limb for analysis of human movement. Annals of biomedical engineering. 2010;38(2):269–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ateş F, Andrade RJ, Freitas SR, Hug F, Lacourpaille L, Gross R, et al. Passive stiffness of monoarticular lower leg muscles is influenced by knee joint angle. European journal of applied physiology. 2018;118(3):585–93. [DOI] [PubMed] [Google Scholar]
- Avancini C, de Oliveira LF, Menegaldo LL, Vieira TM. Variations in the spatial distribution of the amplitude of surface electromyograms are unlikely explained by changes in the length of medial gastrocnemius fibres with knee joint angle. PLoS One. 2015;10(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bamman MM, Newcomer BR, Larson-Meyer DE, Weinsier RL, Hunter GR. Evaluation of the strength-size relationship in vivo using various muscle size indices. Medicine and science in sports and exercise. 2000;32(7):1307–13. [DOI] [PubMed] [Google Scholar]
- Beck TW, DeFreitas JM, Stock MS. An examination of cross-talk among surface mechanomyographic signals from the superficial quadriceps femoris muscles during isometric muscle actions. Human movement science. 2010;29:165–71. [DOI] [PubMed] [Google Scholar]
- Beck TW, Housh TJ, Cramer JT, Weir JP, Johnson GO, Coburn JW, et al. Mechanomyographic amplitude and frequency responses during dynamic muscle actions: a comprehensive review. Biomedical engineering online. 2005;4:1–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological). 1995;57(1):289–300. [Google Scholar]
- Brenner B, LC. Y Equatorial x-ray diffraction from single skinned rabbit psoas fibers at various degrees of activation. Changes in intensities and lattice spacing. Biophysical journal. 1985;48(5):829–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cain J, Hulse R. Beam Deflections and Rotations. In: Structural mechanics: Macmillan International Higher Education; 1990a. [Google Scholar]
- Cain J, Hulse R. Shearing Forces and Bending Moments. In: Structural mechanics: Macmillan International Higher Education; 1990b. [Google Scholar]
- Cè E, Rampichini S, Venturelli M, Limonta E, Veicsteinas A, Esposito F. Electromechanical delay components during relaxation after voluntary contraction: reliability and effects of fatigue. Muscle & nerve 2015;51:907–15. [DOI] [PubMed] [Google Scholar]
- Cescon C, Farina D, Gobbo M, Merletti R, Orizio C. Effect of accelerometer location on mechanomyogram variables during voluntary, constant-force contractions in three human muscles. Medical and Biological Engineering and Computing. 2004;42(1):121–7. [DOI] [PubMed] [Google Scholar]
- Cole NM, Barry DT. Muscle sound frequencies of the frog are modulated by skeletal muscle tension. Biophysical journal. 1994;66(4):1104–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ebersole KT, Housh TJ, Johnson GO, Evetovich TK, Smith DB, Perry SR. The effect of leg flexion angle on the mechanomyographic responses to isometric muscle actions. European journal of applied physiology occupational physiology. 1998;78:264–9. [DOI] [PubMed] [Google Scholar]
- Frangioni JV, Kwan-Gett TS, Dobrunz LE, McMahon TA. The mechanism of low-frequency sound production in muscle. Biophysical journal. 1987;51(5):775–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gandevia S, McKenzie D. Activation of human muscles at short muscle lengths during maximal static efforts. The Journal of physiology. 1988;407:599–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gordon AM, Huxley AF, Julian FJ. The variation in isometric tension with sarcomere length in vertebrate muscle fibres. The Journal of physiology. 1966;184:170–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Herda TJ, Housh TJ, Fry AC, Weir JP, Schilling BK, Ryan ED, et al. A noninvasive, log-transform method for fiber type discrimination using mechanomyography. Journal of Electromyography and Kinesiology. 2010;20(5):787–94. [DOI] [PubMed] [Google Scholar]
- Héroux ME, Stubbs PW, Herbert RD. Behavior of human gastrocnemius muscle fascicles during ramped submaximal isometric contractions. Physiological Reports. 2016;4(17):e12947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ishikawa M, Pakaslahti J, Komi PV. Medial gastrocnemius muscle behavior during human running and walking. Gait & posture. 2007;25:380–4. [DOI] [PubMed] [Google Scholar]
- Kawakami Y, Ichinose Y, Fukunaga T. Architectural and functional features of human triceps surae muscles during contraction. Journal of applied physiology. 1998;85(2):398–404. [DOI] [PubMed] [Google Scholar]
- Kim S, Son J, Yi C, Kim D, Kim Y. Pennation angles of ankle dorsiflexor and plantarflexors depending on muscle contraction intensity. International Journal of Precision Engineering and Manufacturing. 2013;14(5):855–8. [Google Scholar]
- Koo TK, Guo JY, Cohen JH, Parker KJ. Relationship between shear elastic modulus and passive muscle force: an ex-vivo study. Journal of biomechanics. 2013;46(12):2053–9. [DOI] [PubMed] [Google Scholar]
- Lei KF, Tsai WW, Lin WY, Lee MY. MMG-torque estimation under dynamic contractions. 2011 IEEE International Conference on Systems, Man, and Cybernetics: IEEE; 2011. p. 585–90. [Google Scholar]
- Longo S, Cè E, Rampichini S, Devoto M, Limonta E, Esposito F. Mechanomyogram amplitude correlates with human gastrocnemius medialis muscle and tendon stiffness both before and after acute passive stretching. Experimental physiology. 2014;99(10):1359–69. [DOI] [PubMed] [Google Scholar]
- Maganaris CN, Baltzopoulos V, Sargeant AJ. In vivo measurements of the triceps surae complex architecture in man: implications for muscle function. The Journal of physiology. 1998;512(2):603–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mills KR. The basics of electromyography. 2005;76:ii32–ii5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miyamoto N, Oda S. Mechanomyographic and electromyographic responses of the triceps surae during maximal voluntary contractions. Journal of Electromyography and Kinesiology. 2003;13:451–9. [DOI] [PubMed] [Google Scholar]
- 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;95:221–8. [DOI] [PubMed] [Google Scholar]
- Orizio C Muscle sound: Bases for the. Journal of Critical Reviews in Biomedical Engineering. 1993;21(3):201–43. [PubMed] [Google Scholar]
- 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(3–4):326–36. [DOI] [PubMed] [Google Scholar]
- Scheeren EM, Mineiro L, Neves EB, Krueger E, Nogueira Neto GN, Nohama P. Influence of subcutaneous fat on mechanomyographic signals at three levels of voluntary effort. Research on Biomedical Engineering. 2016;32(4):307–17. [Google Scholar]
- Vaz MA, Herzog W, Zhang Y-T, Leonard TR, Nguyen H. The effect of muscle length on electrically elicited muscle vibrations in the in-situ cat soleus muscle. Journal of Electromyography and Kinesiology. 1997;7:113–21. [DOI] [PubMed] [Google Scholar]
- Yamauchi J, Koyama K. Relation between the ankle joint angle and the maximum isometric force of the toe flexor muscles. Journal of biomechanics. 2019;85:1–5. [DOI] [PubMed] [Google Scholar]
- Yoshino C, Shimomura Y. Sex-Related Differences in Mechanomyography of Stepwise Loading during Elbow Flexion. The Japanese Journal of Ergonomics. 2017;53:1–7. [Google Scholar]
- Yoshitake Y, Masani K, Shinohara M. Laser-detected lateral muscle displacement is correlated with force fluctuations during voluntary contractions in humans. Journal of neuroscience methods. 2008;173:271–8. [DOI] [PubMed] [Google Scholar]