Mechanomyographic Signals Research Articles

An end-to-end hand action recognition framework based on cross-time mechanomyography signals

The susceptibility of mechanomyography (MMG) signals acquisition to sensor donning and doffing, and the apparent time-varying characteristics of biomedical signals collected over different periods, inevitably lead to a reduction in model recognition accuracy. To investigate the adverse effects on the recognition results of hand actions, a 12-day cross-time MMG data collection experiment with eight subjects was conducted by an armband, then a novel MMG-based hand action recognition framework with densely connected convolutional networks (DenseNet) was proposed. In this study, data from 10 days were selected as a training subset, and the remaining data from another 2 days were used as a test set to evaluate the model’s performance. As the number of days in the training set increases, the recognition accuracy increases and becomes more stable, peaking when the training set includes 10 days and achieving an average recognition rate of 99.57% (± 0.37%). In addition, part of the training subset is extracted and recombined into a new dataset and the better classification performances of models can be achieved from the test set. The method proposed effectively mitigates the adverse effects of sensor donning and doffing on recognition results.

Are performance and perceived fatigability dependent on the anchor scheme of fatiguing isometric tasks in men?

Ratings of perceived exertion (RPE) can be used to regulate exercise intensity. This study examined the effect of anchor scheme on performance fatigability and neuromuscular responses following fatiguing forearm flexion tasks. Twelve men (age 20.9±2.2 years; height 179.8±5.3 cm; body mass 80.2±9.9 kg) performed sustained, isometric forearm flexion tasks to failure anchored to RPE=6 (RPEFT) and the torque (TRQFT) that corresponded to RPE=6. Pre-test and post-test maximal voluntary isometric contractions (MVIC) were performed to quantify changes in the amplitude (AMP) and mean power frequency (MPF) of the electromyographic (EMG) and mechanomyographic (MMG) signals. Neuromuscular efficiency (NME) was calculated by dividing normalized torque by normalized EMG AMP. A dependent t-test was used to assess the mean difference for time to task failure (TTF). Repeated measures ANOVAs were used to compare mean differences for performance fatigability and normalized neuromuscular parameters. The RPEFT had a greater TTF than the TRQFT (P<0.001). MVIC and NME decreased from pre-test to post-test following the RPEFT and TRQFT (P<0.05) with no differences between anchor schemes. Following the TRQFT, normalized EMG MPF decreased from pre-test to post-test (P=0.004). Following the RPEFT, normalized MMG MPF increased from pre-test to post-test (P=0.021). There were no changes in normalized EMG AMP or MMG AMP (P>0.05). These findings indicated anchor scheme-specific neuromuscular responses and TTF, despite no difference in performance fatigability. Furthermore, performance fatigability was likely due to peripheral fatigue (based on normalized EMG MPF and NME) following the TRQFT, but peripheral and central fatigue (based on normalized MMG MPF and NME) following the RPEFT.

Estimating Rotational Acceleration in Shoulder and Elbow Joints Using a Transformer Algorithm and a Fusion of Biosignals.

In the present study, we used a transformer model and a fusion of biosignals to estimate rotational acceleration in elbow and shoulder joints. To achieve our study objectives, we proposed a mechanomyography (MMG) signal isolation technique based on a variational mode decomposition (VMD) algorithm. Our results show that the VMD algorithm delivered excellent performance in MMG signal extraction compared to the commonly used technique of empirical mode decomposition (EMD). In addition, we found that transformer models delivered estimates of joint acceleration that were more precise than those produced by mainstream time series forecasting models. The average R2 values of transformer are 0.967, 0.968, and 0.935, respectively. Finally, we found that using a fusion of signals resulted in more precise estimation performance compared to using MMG signals alone. The differences between the average R2 values are 0.041, 0.053, and 0.043, respectively. Taken together, the VMD isolation method, the transformer algorithm and the signal fusion technique described in this paper can be seen as supplying a robust framework for estimating rotational acceleration in upper-limb joints. Further study is warranted to examine the effectiveness of this framework in other musculoskeletal contexts.

Effects of a Sustained, Isometric Forearm Flexion Task to Failure on Torque and Neuromuscular Responses at 3 Elbow Joint Angles

Abstract Ortega, DG, Housh, TJ, Smith, RW, Arnett, JE, Neltner, TJ, Anders, JPV, Schmidt, RJ, and Johnson, GO. The effects of a sustained, isometric forearm flexion task to failure on torque and neuromuscular responses at 3 elbow joint angles. J Strength Cond Res 38(1): e25–e33, 2024—This study examined the effects of a sustained, isometric forearm flexion task anchored to torque to task failure on maximal voluntary isometric contraction (MVIC) and neuromuscular responses at 3 elbow joint angles. Eleven women (mean ± SD: age = 20.8 ± 2.7 years, height = 169.3 ± 7.4 cm, body mass = 67.7 ± 6.9 kg) performed two 3s forearm flexion MVICs at elbow joint angles (JAs) of 75°, 100°, and 125° before and after a sustained, isometric forearm flexion task to failure at a fatiguing joint angle of 100° anchored to a torque value that corresponded to a rating of perceived exertion of 8 (RPE = 8). The amplitude (AMP) and mean power frequency (MPF) of the electromyographic (EMG) and mechanomyographic (MMG) signals were recorded from the biceps brachii. Repeated-measures ANOVAs were used to compare mean differences for MVIC and neuromuscular parameters. Collapsed across JAs, MVIC (p &lt; 0.001) and EMG MPF (p = 0.006) pretest values were greater than posttest values. Collapsed across time, EMG MPF at JA75 was greater than JA100 (p &lt; 0.001) and JA125 (p &lt; 0.001), and JA100 was greater (p = 0.007) than JA125. For EMG AMP, there was a fatigue-induced decrease at JA75 (p = 0.003). For neuromuscular efficiency (NME = normalized torque/normalized EMG AMP), there were decreases from pretest to posttest at JA100 (p = 0.002) and JA125 (p = 0.008). There were no significant interactions or main effects for MMG AMP and MMG MPF. From these findings, it was hypothesized that the decline in MVICs at JA75, JA100, and JA125 was due to fatigue-induced metabolic perturbations that resulted in JA-specific neuromuscular responses. Thus, neuromuscular parameters may provide insight into the JA-specific mechanisms of fatigue.

Mechanomyography signals pattern recognition in hand movements using swarm intelligence algorithm optimized support vector machine based on acceleration sensors

On the basis of extracting mechanomyography (MMG) signal features, the classification of hand movements has certain application values in human-machine interaction systems and wearable devices. In this paper, pattern recognition of hand movements based on MMG signal is studied with swarm intelligence algorithms introduced to optimize support vector machine (SVM). Time domain (TD) features, wavelet packet node energy (WPNE) features, frequency domain (FD) features, convolution neural network (CNN) features were extracted from each channel to constitute different feature sets. Three novel swarm intelligence algorithms (i.e., bald eagle search (BES), sparrow search algorithm (SSA), grey wolf optimization (GWO)) optimized SVM is proposed to train the models and recognition of hand movements are tested for each MMG feature extraction method. Using GWO as the optimization algorithm, time consumption is less than using the other two swarm algorithms. Using GWO with TD+FD features can obtain the classification accuracy of 93.55 %, which is higher than other methods while using CNN to extract features can be independent of domain knowledge. The results confirm GWO-SVM with TD + FD features is superior to some other methods in the classification problem for tiny samples based on MMG.

Fatiguing Joint Angle Does Not Influence Torque and Neuromuscular Responses Following Sustained, Isometric Forearm Flexion Tasks Anchored to Perceptual Intensity in Men.

This study examined the effects of joint angle (JA) on maximal voluntary isometric contraction (MVIC) and neuromuscular responses following fatiguing tasks anchored to RPE. Nine men (mean ± SD: age = 20.7 ± 1.2 yrs) performed forearm flexion MVICs at elbow JAs of 75° and 125° before and after sustained, isometric forearm flexion tasks to failure at fatiguing joint angles (FJA) of 75° and 125° anchored to RPE = 8. The amplitude and frequency of the electromyographic and mechanomyographic signals were recorded. Neuromuscular efficiency was calculated by dividing normalized torque by normalized electromyographic amplitude. A dependent t-test was used to assess the mean difference for time to task failure (TTF) between FJA. Repeated measure ANOVAs were used to assess mean differences for pre-test to post-test MVIC and neuromuscular responses. There was no significant difference between FJA for TTF (p = 0.223). The MVIC (collapsed across FJA and MVIC JA) decreased from pre-test to post-test (51.1 ± 5.0 vs. 45.3 ± 5.6 Nm, p < 0.001). Normalized neuromuscular parameters remained unchanged (p > 0.05). The FJA resulted in similar torque and neuromuscular responses, and the decreases in MVIC were not tracked by changes in the neuromuscular parameters. Thus, the neuromuscular parameters were not sensitive to fatigue, and pre-test to post-test measures may be compared between different FJA.

Mechanomyography Signal Pattern Recognition of Knee and Ankle Movements Using Swarm Intelligence Algorithm-Based Feature Selection Methods.

Pattern recognition of lower-limb movements based on mechanomyography (MMG) signals has a certain application value in the study of wearable rehabilitation-training devices. In this paper, MMG feature selection methods based on a chameleon swarm algorithm (CSA) and a grasshopper optimization algorithm (GOA) are proposed for the pattern recognition of knee and ankle movements in the sitting and standing positions. Wireless multichannel MMG acquisition systems were designed and used to collect MMG movements from four sites on the subjects thighs. The relationship between the threshold values and classification accuracy was analyzed, and comparatively high recognition rates were obtained after redundant information was eliminated. When the threshold value rose, the recognition rates from the CSA fluctuated within a small range: up to 88.17% (sitting position) and 90.07% (standing position). However, the recognition rates from the GOA drop dramatically when increasing the threshold value. The comparison results demonstrated that using a GOA consumes less time and selects fewer features, while a CSA gives higher recognition rates of knee and ankle movements.

ViT-LLMR: Vision Transformer-based lower limb motion recognition from fusion signals of MMG and IMU

One of the key problems in lower limb-based human–computer interaction (HCI) technology is to use wearable devices to recognize the wearer's lower limb motions. The information commonly used to discriminate human motion mainly includes biological and kinematic signals. Considering that unimodal signals do not provide enough information to recognize lower limb movements, in this paper, we proposed a Vision Transformer (ViT)-based architecture for lower limb motion recognition from multichannel Mechanomyography (MMG) signals and kinematic data. Firstly, we applied the self-attention mechanism to enhance each input channel signal. Then the data was fed into ViT model. Vision Transformer-based Lower Limb Motion Recognition (ViT - LLMR) architecture proposed in this paper can avoid the model training problems such as autonomous feature extraction and feature selection for machine learning, and the model can recognize eight lower limb motions containing six subjects with an accuracy of 94.62%. In addition, we analyzed the generalization ability of the model when undersampling and only collecting fragment signals. In conclusion, the proposed ViT - LLMR architecture could provide a basis for practical applications in different HCI fields.

Application of the Ratings of Perceived Exertion-Clamp Model to Examine the Effects of Joint Angle on the Time Course of Torque and Neuromuscular Responses During a Sustained, Isometric Forearm Flexion to Task Failure.

Abstract Smith, RW, Housh, TJ, Anders, JPV, Neltner, TJ, Arnett, JE, Schmidt, RJ, and Johnson, GO. Application of the ratings of perceived exertion-clamp model to examine the effects of joint angle on the time course of torque and neuromuscular responses during a sustained, isometric forearm flexion to task failure. J Strength Cond Res 37(5): 1023–1033, 2023—The present study used the ratings of perceived exertion (RPE)-clamp model during a sustained, isometric forearm flexion task to failure to determine the effects of joint angle on torque and neuromuscular responses. Twelve college-aged women performed two 3-second maximal voluntary isometric contractions at elbow joint angles (EJ) of 75°, 100°, and 125° before sustained, isometric, forearm flexions anchored to RPE = 7 to task failure (defined as RPE &gt; 7, or the torque was reduced to zero) at EJ75, EJ100, and EJ125. The amplitude (AMP) and frequency (MPF) of the electromyographic (EMG) and mechanomyographic (MMG) signals from the biceps brachii were recorded. Repeated-measures ANOVAs and post hoc comparisons were used to examine differences across time and between joint angles for torque and neuromuscular parameters. A p-value ≤0.05 was considered significant. For each joint angle, there was a decrease (p &lt; 0.05) in torque across the sustained, isometric task. Collapsed across joint angle, there were decreases (p &lt; 0.001, = 0.378) in EMG AMP from 30 to 100% time to task failure. No changes were observed, however, for the other neuromuscular responses. The subjects voluntarily reduced torque to maintain RPE = 7 for each joint angle, and we hypothesize that RPE was maintained by afferent feedback from group III/IV motor neurons and corollary discharge (efferent copy of an internal signal that develops from central motor commands). The RPE-Clamp Model may be extended to athletes recovering from injury because a self-selected exercise intensity may be perceived as more enjoyable and promote adherence.

Coactivation does not contribute to fatigue-induced decreases in torque during reciprocal, isokinetic muscle actions

BACKGROUND: Studies of coactivation have typically utilized single movement isometric or isokinetic fatiguing muscle actions. OBJECTIVE: The purpose of the current study was to examine coactivation of the biceps brachii (BB) and triceps brachii (TB) in response to a maximal, reciprocal, isokinetic fatiguing task of the forearm flexors and extensors at slow (60∘/s) and moderate (180∘/s) isokinetic velocities in men. METHODS: Ten men (mean ± SD: age = 21.6 ± 1.3 years) completed 50 consecutive, maximal, reciprocal, isokinetic muscle actions of the right forearm flexors and extensors at 60 and 180∘/s. The amplitude (AMP) and mean power frequency (MPF) contents of the electromyographic (EMG) and mechanomyographic (MMG) signals from the BB and TB were recorded simultaneously throughout the fatiguing task. Repeated measures ANOVAs with Tukey post hocs were used to determine mean differences for the torque and neuromuscular parameters across repetitions. RESULTS: The torque analyses indicated greater fatigability at 180∘/s, compared to 60∘/s (p= 0.02). There were no significant changes in EMG AMP for either muscle during flexion or extension at 60∘/s (p&gt; 0.05). At 180∘/s, there were significant increases in agonist EMG AMP (p= 0.01 to 0.004), however, no changes in antagonist EMG AMP (p&gt; 0.05). For EMG MPF, there were significant decreases during flexion and extension (p&lt; 0.001 to p= 0.02) at both velocities, collapsed across muscle. There were no significant (p&gt; 0.05) changes across repetition for MMG AMP or MPF. CONCLUSIONS: This study indicated velocity-specific responses to fatigue, with a greater magnitude of fatigability at 180∘/s. Furthermore, despite increases in EMG AMP of the agonist muscles at 180∘/s only, it was not sufficient to alter the ratio of coactivation, likely due to common neural drive between muscles. Thus, the decreases in torque in the present study were not attributable to increases in coactivation.

Is the Interpolated-Twitch Technique-Derived Voluntary Activation Just Neural? Novel Perspectives from Mechanomyographic Data.

Voluntary activation (VA) determined by interpolation-twitch technique could be affected by the characteristics of the in-series elastic components. To overcome this possible bias, a novel approach based on the mechanomyographic (MMG) signal to detect voluntary activation (VA MMG ) has been proposed. We examined the changes in VA and VA MMG after passive stretching to check the influence of neural and mechanical factors in the force output. Twenty-six healthy men underwent VA assessment using the interpolated-twitch technique before and after unilateral passive stretching of the plantarflexors (five 45-s on + 15-s off). In addition to the force signal, the MMG signal was detected on gastrocnemius medialis, gastrocnemius lateralis, and soleus. From the force and MMG signal analysis, VA and VA MMG were calculated in the stretched and contralateral nonstretched limbs. Joint passive stiffness was also defined. In the stretched limb, passive stretching increased dorsiflexion range (mean ± SD = +18% ± 10%, P < 0.001, ES = 1.54) but reduced joint passive stiffness (-22% ± 8%, P < 0.001, ES = -1.75), maximum voluntary contraction (-15% ± 7%, P < 0.001, ES = -0.87), VA (-7% ± 3%, P < 0.001, ES = -2.32), and VA MMG (~-5% ± 2%, P < 0.001, ES = -1.26/-1.14). In the contralateral nonstretched limb, passive stretching increased dorsiflexion range (+10% ± 6%, P < 0.001, ES = 0.80) but reduced joint passive stiffness (-3% ± 2%, P = 0.041, ES = -0.27), maximum voluntary contraction (-4% ± 3%, P = 0.035, ES = -0.24), VA (-4% ± 2%, P < 0.001, ES = -1.77), and VA MMG (~- 2% ± 1%, P < 0.05, ES = -0.54/-0.46). The stretch-induced changes in VA correlated with VA MMG ( R ranging from 0.447 to 0.583 considering all muscles) and with joint passive stiffness (stretched limb: R = 0.503; contralateral nonstretched limb: R = 0.530). VA output is overall influenced by both neural and mechanical factors, not distinguishable using the interpolated-twitch technique. VA MMG is a complementary index to assess the changes in VA not influenced by mechanical factors and to examine synergistic muscles.

Analysis of anthropometrics and mechanomyography signals as forearm flexion, pronation and supination torque predictors

This study aimed to analyze anthropometrics and mechanomyography (MMG) signals as forearm flexion, pronation, and supination torque predictors. 25 young, healthy, male participants performed isometric forearm flexion, pronation, and supination tasks from 20 to 100% maximal voluntary isometric contraction (MVIC) while maintaining 90° at the elbow joint. Nine anthropometric measures were recorded, and MMG signals from the biceps brachii (BB), brachialis (BRA), and brachioradialis (BRD) muscles were digitally acquired using triaxial accelerometers. These were then correlated with torque values. Significant positive correlations were found for arm circumference (CA) and MMG root mean square (RMS) values with flexion torque. Flexion torque might be predicted using CA (r = 0.426–0.575), a pseudo for muscle size while MMGRMS (r = 0.441), an indication of muscle activation.

A Novel Noise Suppression and Artifact Removal Method of Mechanomyography Based on RLS, IGWO-VMD, and CEEMDAN

Mechanomyography (MMG) signals have extensive applications in muscle function assessment and human intention recognition. However, during signal acquisition, MMG signals are easily contaminated by noise and artifacts, which seriously affects the recognition of their characteristics. To address these issues, a novel noise suppression and artifact removal method based on recursive least square (RLS), improved Gray Wolf Optimizer-optimized variable mode decomposition (IGWO-VMD), and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is proposed. In this paper, the RLS algorithm is first applied to adaptively filter out the power line interference (PLI). Then, IGWO is designed to select the appropriate VMD parameters and use the VMD to decompose the noisy signal into band-limited intrinsic mode functions (BLIMFs). In addition, the BLIMFs are classified into the low-frequency part and high-frequency part according to the given correlation coefficient (CC) threshold value. The effective components of the low-frequency part are identified by the center frequency. Meanwhile, the high-frequency part is decomposed by CEEMDAN, and its effective components are obtained according to the proposed sample entropy threshold range. Finally, the effective components of the low and high-frequency parts are reconstructed to obtain the denoised signal to realize the extraction of useful signals. Simulation experiment results demonstrate that the proposed method outperforms the classical methods and the designed IGWO-VMD method in terms of denoising performance. The effectiveness of the proposed method is verified through the measured MMG signal experiments. The proposed method not only effectively suppresses noise and artifacts but also overcomes the limitations of VMD and CCEMDAN.

Estimation of Knee Extension Force Using Mechanomyography Signals Based on GRA and ICS-SVR

During lower-extremity rehabilitation training, muscle activity status needs to be monitored in real time to adjust the assisted force appropriately, but it is a challenging task to obtain muscle force noninvasively. Mechanomyography (MMG) signals offer unparalleled advantages over sEMG, reflecting the intention of human movement while being noninvasive. Therefore, in this paper, based on MMG, a combined scheme of gray relational analysis (GRA) and support vector regression optimized by an improved cuckoo search algorithm (ICS-SVR) is proposed to estimate the knee joint extension force. Firstly, the features reflecting muscle activity comprehensively, such as time-domain features, frequency-domain features, time–frequency-domain features, and nonlinear dynamics features, were extracted from MMG signals, and the relational degree was calculated using the GRA method to obtain the correlation features with high relatedness to the knee joint extension force sequence. Then, a combination of correlated features with high relational degree was input into the designed ICS-SVR model for muscle force estimation. The experimental results show that the evaluation indices of the knee joint extension force estimation obtained by the combined scheme of GRA and ICS-SVR were superior to other regression models and could estimate the muscle force with higher estimation accuracy. It is further demonstrated that the proposed scheme can meet the need of muscle force estimation required for rehabilitation devices, powered prostheses, etc.

Velocity-Specific Coactivation and Neuromuscular Responses to Fatiguing, Reciprocal, Isokinetic, Forearm Flexion, and Extension Muscle Actions.

Abstract Neltner, TJ, Anders, JPV, Keller, JL, Smith, RW, Housh, TJ, Schmidt, RJ, and Johnson, GO. Velocity-specific coactivation and neuromuscular responses to fatiguing, reciprocal, isokinetic, forearm flexion, and extension muscle actions. J Strength Cond Res 36(3): 649–660, 2022—The purpose of this study was to examine the effects of fatiguing, reciprocal, maximal, isokinetic, forearm flexion, and extension on coactivation. Ten recreationally trained women (mean ± SD: age = 21 ± 1.63 years) performed 50 maximal, reciprocal, isokinetic, forearm flexion, and extension muscle actions at 60 and 180°·s−1 on separate days. Electromyographic (EMG) and mechanomyographic (MMG) signals from the biceps brachii (BB) and triceps brachii (TB) were recorded. A 2 (time) × 2 (velocity) × 2 (movement) and 4 separate 2 (time) × 2 (velocity) × 2 (muscle) × 2 (movement) repeated-measures analysis of variances were used to examine normalized torque and normalized EMG and MMG amplitude (AMP) and mean power frequency, respectively, with an alpha of p &lt; 0.05. There were no significant interactions for torque but a main effect for time (p &lt; 0.001, d = 3.29) such that it decreased. There was a significant four-way interaction for EMG AMP (p = 0.004, = 0.61) which revealed parallel responses between the BB and TB during both movements, with no changes at 60°·s−1 and increases at 180°·s−1. The current findings suggested a common drive between agonist and antagonist muscles during fatiguing forearm flexion and extension muscle actions, with velocity-specific activation strategies. In addition, both EMG and MMG are viable techniques for assessing coactivation during fatiguing exercise.

Punctual mechanical oscillation in modulation of muscular tonus in children with spasticity

Spasticity is a motor condition present in 75 to 88% of children with Cerebral Palsy (CP). One form of treatment is called punctual mechanical oscillation (PO). The current study aimed to study different protocols for the application of PO and the magnitude of their effects. In total, 7children with medical diagnosis of CP and ICD (International Classification of Diseases) were included. The first intervention protocol (Int1) consisted of the application of PO to the spastic muscle tendon and the second intervention protocol (Int2) to the muscle belly ofthe spastic antagonist muscle. For evaluation, the Modified Ashworth Scale (MAS) was used, while simultaneously capturing the mechanomyography (MMG) signals. Data were collected pre-intervention and 1 (Post1), 15 (Post15), 30 (Post30), 45 (Post45), and60 (Post60) minutes after the interventions. The MAS values (median ± interquartile range) post intervention were statistically lower when compared to the pre values in the 2 protocols studied; in Int1between Pre (2 ± 0) andPost15 (0 ± 1.75), Post30 (0 ± 1), Post45 (1 ± 1),and Post60 (1 ± 1), and in Int2only between Pre (2 ± 1) and Post1 (0 ± 1).The values found in the MMG in both its temporal and spectral domains did not follow a pattern (p&gt;0.05). The comparison between the protocols did not demonstrate statistical differences in any characteristics (MAS, MMGMF, and MMGRMS). However, PO was shown to be a therapeutic resource that modulated spasticity for up to 60 minutes after its application, and PO could contribute as a tool to aid the treatment of spasticity.

Control of prosthetic hand by using mechanomyography signals based on support-vector machine classifier

&lt;div&gt;Prosthetic devices are necessary to help amputees achieve their daily activity in the natural way possible. The prosthetic hand has controlled by type of signals such as electromyography (EMG) and mechanomyography (MMG). The MMG signals have represented mechanical signals that generate during muscle contraction. These signals can be detected by accelerometers or microphones and any kind of sensors that can detect muscle vibrations. The contribution of the current paper is classifying hand gestures and control prosthetic hands depends on pattern recognition through accelerometer and microphone are to detect MMG signals. In addition to the cost of prosthetic hand less than other designs. Six subjects are involved. In this present work is the devices. In this study, two of them are amputee subjects. Each subject performs seven classes of movements. Pattern recognition (PR) is used to classify hand gestures. The wavelet packet transform (WPT) and root mean square (RMS) as features extracted from the signals and support vector machine (SVM) as a classifier. The average accuracy is 88.94% for offline tests and 84.45% for online tests. 3D printing technology is used in this study to build prosthetic hands.&lt;/div&gt;

Influence of sensor mass and adipose tissue on the mechanomyography signal of elbow flexor muscles

Mechanomyography (MMG) is a non-invasive technique that records muscle contraction using sensors positioned on the skin's surface. Therefore, it can have its signal attenuated due to the adipose tissue, directly influencing the results. This study evaluates the influence of different mass added to a sensor's assembly and the adipose tissue on MMG signals of elbow flexor muscles. Test protocol consisted of skinfold thickness measurement of 22 volunteers, followed by applying 2–3 s electrical stimulation for muscle contraction during the acquisition of MMG signals. MMG signals were processed in the time domain, using the average of the absolute amplitude, and expressed in gravity values (G), termed here as MMG(G). Tests occurred four times with different sensor masses. MMG data were processed and analyzed statistically using Friedman and Kruskal-Wallis tests to determine the differences between the MMG signals measured with different sensor masses. The Mann-Whitney analysis indicated differences in the MMG signals between groups with different skinfold thickness. MMG(G) signals suffered attenuation with increasing sensor mass (0.4416 G to 0.94 g; 0.3902 G to 2.64 g; 0.3762 G to 5.44 g; 0.3762 G to 7.14 g) and adipose tissue.

Wearable MMG-Plus-One Armband: Evaluation of Normal Force on Mechanomyography (MMG) to Enhance Human-Machine Interfacing.

In this paper, we introduce a new mode of mechanomyography (MMG) signal capture for enhancing the performance of human-machine interfaces (HMIs) through modulation of normal pressure at the sensor location. Utilizing this novel approach, increased MMG signal resolution is enabled by a tunable degree of freedom normal to the sensor-skin contact area. We detail the mechatronic design, experimental validation, and user study of an armband with embedded acoustic sensors demonstrating this capacity. The design is motivated by the nonlinear viscoelasticity of the tissue, which increases with the normal surface pressure. This, in theory, results in higher conductivity of mechanical waves and hypothetically allows to interface with deeper muscle; thus, enhancing the discriminative information context of the signal space. Ten subjects (seven able-bodied and three trans-radial amputees) participated in a study consisting of the classification of hand gestures through MMG while increasing levels of contact force were administered. Four MMG channels were positioned around the forearm and placed over the flexor carpi radialis, brachioradialis, extensor digitorum communis, and flexor carpi ulnaris muscles. A total of 852 spectrotemporal features were extracted (213 features per each channel) and passed through a Neighborhood Component Analysis (NCA) technique to select the most informative neurophysiological subspace of the features for classification. A linear support vector machine (SVM) then classified the intended motion of the user. The results indicate that increasing the normal force level between the MMG sensor and the skin can improve the discriminative power of the classifier, and the corresponding pattern can be user-specific. These results have significant implications enabling embedding MMG sensors in sockets for prosthetic limb control and HMI.

A long short-term memory neural network model for knee joint acceleration estimation using mechanomyography signals

With the growth of the number of elderly and disabled with motor dysfunction, the demand for assisted exercise is increasing. Wearable power assistance robots are developed to provide athletic ability of limbs for the elderly or the disabled who have weakened limbs to better self-care ability. Existing wearable power-assisted robots generally use surface electromyography (sEMG) to obtain effective human motion intentions. Due to the characteristics of sEMG signals, it is limited in many applications. To solve the above problems, we design a long short-term memory (LSTM) neural network model based on human mechanomyography (MMG) signals to estimate the motion acceleration of knee joint. The acceleration can be further calculated by the torque required for movement control of the wearable power assistance robots for the lower limb. We detect MMG signals on the clothed thigh, extract features of the MMG signals, and then, use principal component analysis to reduce the features’ dimensions. Finally, the dimension-reduced features are inputted into the LSTM neural network model in time series for estimating the acceleration. The experimental results show that the average correlation coefficient ( R) is 94.48 ± 1.91% for the estimation of acceleration in the process of continuously performing under approximately π/4 rad/s. This approach can be applied in the practical applications of wearable field.