High-density Surface Electromyography Research Articles

Decreased neural drive affects the early rate of force development after repeated burst-like isometric contractions.

The neural drive to the muscle is the primary determinant of the rate of force development (RFD) in the first 50 ms of a rapid contraction. It is still unproven if repetitive rapid contractions specifically impair the net neural drive to the muscles. To isolate the fatiguing effect of contraction rapidity, 17 male adult volunteers performed 100 burst-like (i.e., brief force pulses) isometric contractions of the knee extensors. The response to electrically-evoked single and octet femoral nerve stimulation was measured with high-density surface electromyography (HD-sEMG) from the vastus lateralis and medialis muscles. Root mean square (RMS) of each channel of HD-sEMG was normalized to the corresponding M-wave peak-to-peak amplitude, while muscle fiber conduction velocity (MFCV) was normalized to M-wave conduction velocity to compensate for changes in sarcolemma properties. Voluntary RFD 0-50 ms decreased (d = -0.56, p < 0.001) while time to peak force (d = 0.90, p < 0.001) and time to RFDpeak increased (d = 0.56, p = 0.034). Relative RMS (d = -1.10, p = 0.006) and MFCV (d = -0.53, p = 0.007) also decreased in the first 50 ms of voluntary contractions. Evoked octet RFD 0-50 ms (d = 0.60, p = 0.020), M-wave amplitude (d = 0.77, p = 0.009) and conduction velocity (d = 1.75, p < 0.001) all increased. Neural efficacy, i.e., voluntary/octet force ratio, largely decreased (d = -1.50, p < 0.001). We isolated the fatiguing impact of contraction rapidity and found that the decrement in RFD, particularly when calculated in the first 50 ms of muscle contraction, can mainly be explained by a decrease in the net neural drive.

Caffeine does not influence persistent inward current contribution to motoneuron firing.

The purpose of this study was to investigate whether caffeine consumption would change persistent inward current (PIC) contribution to motoneuron firing at increased contraction intensities and after repetitive sustained maximal contractions. Before and after the consumption of 6 mg·kg-1 of caffeine or placebo, 16 individuals performed isometric triangular-shaped ramp dorsiflexion contractions (to 20% and 40% of peak torque), followed by four maximal contractions sustained until torque production dropped to 60% of maximum, and consecutive 20% triangular-shaped contractions. Tibialis anterior motor unit firing frequencies were analyzed from high-density surface electromyograms. PIC contribution to motor unit firing was estimated by calculating the delta frequency (ΔF) using the paired motor unit technique. Motoneuron peak firing frequencies at 20% and 40% contractions and total torque-time integral during the repetitive sustained maximal contractions were also assessed. ΔF increased 0.69 peaks per second (pps) (95% CI = -0.98, -0.405; d = -0.87) from 20% to 40% contraction intensities and reduced 0.85 pps (95% CI = 0.66, 1.05; d = 0.99) after the repetitive sustained maximal contractions, regardless of caffeine consumption. Participants produced 337 Nm·s (95% CI = 49.9, 624; d = 0.63) more torque integral during the repetitive sustained maximal contractions after caffeine consumption. A strong repeated-measures correlation (r = 0.61; 95% CI = 0.49, 0.69) was observed between reductions of ΔF and peak firing frequencies after the repetitive sustained maximal contractions. PIC contribution to motoneuron firing increases from 20% to 40% contraction intensities, with no effect of caffeine (on rested tibialis anterior). Repetitive sustained maximal contractions reduced PIC contribution to motoneuron firing, regardless of caffeine or placebo consumption, evidencing that changes in intrinsic motoneuron properties contributed to performance loss. Caffeine-attenuated reduction of torque production capacity was unlikely mediated by PICs.NEW & NOTEWORTHY Persistent inward current (PIC) contribution to motoneuron firing increases with contraction intensities and is reduced after repetitive sustained maximal contractions, regardless of caffeine consumption. Reductions of PIC contribution to motoneuron firing and peak firing frequencies were largely associated, evidencing a novel mechanism underpinning decrements in maximal torque production capacity following repetitive sustained maximal contractions. Caffeine consumption attenuated neuromuscular performance reductions-allowing higher time-torque integral production during repetitive sustained maximal contractions. This was unlikely mediated by PIC.

Regional recruitment and differential behavior of motor units during postural control in older adults.

Aging is associated with neuromuscular system changes that may have implications for the recruitment and firing behaviors of motor units (MUs). In previous studies, we observed that young adults recruit subpopulations of triceps surae MUs during tasks that involved leaning in five directions: common units that were active during different leaning directions and unique units that were active in only one leaning direction. Furthermore, the MU subpopulation firing behaviors [average firing rate (AFR), coefficient of variation (CoVISI), and intermittent firing] modulated with leaning direction. The purpose of this study was to examine whether older adults exhibited this regional recruitment of MUs and firing behaviors. Seventeen older adults (aged 74.8 ± 5.3 yr) stood on a force platform and maintained their center of pressure leaning in five directions. High-density surface electromyography recordings from the triceps surae were decomposed into single MU action potentials. A MU tracking analysis identified groups of MUs as being common or unique across the leaning directions. Although leaning in different directions did not affect the AFR and CoVISI of common units (P > 0.05), the unique units responded to the leaning directions by increasing AFR and CoVISI, albeit modestly (F = 18.51, P < 0.001). The unique units increased their intermittency with forward leaning (F = 9.22, P = 0.003). The mediolateral barycenter positions of MU activity in both subpopulations were found in similar locations for all leaning directions (P > 0.05). These neuromuscular changes may contribute to the reduced balance performance seen in older adults.NEW & NOTEWORTHY In this study, we observed differences in motor unit recruitment and firing behaviors of distinct subpopulations of motor units in the older adult triceps surae muscle from those observed in the young adult. Our results suggest that the older adult central nervous system may partially lose the ability to regionally recruit and differentially control motor units. This finding may be an underlying cause of balance difficulties in older adults during directionally challenging leaning tasks.

Trunk extensor muscle endurance and its relationship to action potential conduction velocity and spectral parameters estimated using high-density electromyography

Trunk extensor muscle fatigue typically manifests as a decline in spectral content of surface electromyography. However, previous research on the relationship of this decline with trunk extensor muscle endurance have shown inconsistent results. The decline of spectral content mainly reflects the decrease in average motor unit action potential conduction velocity (CV). We evaluated whether the rate of change in CV, as well as two approaches employing the change in spectral content, are related to trunk extensor muscle endurance. Fourteen healthy male participants without a low-back pain history performed a non-strictly controlled static forward trunk bending trial until exhaustion while standing. For 13 participants, physiologically plausible CV estimates were obtained from high-density surface electromyography bilaterally from T6 to L5. Laterally between L1 and L2, the linear rate of CV change was strongly correlated to endurance time (R2 = 0.79), whereas analyses involving the linear rate of change in spectral measures showed a lower (R2 = 0.38) or no correlation. For medial electrode locations, estimating CV and its relationship with endurance time was less successful, while the linear rate of change in spectral measures correlated moderately to endurance time (R2 = 0.44; R2 = 0.56). This study provides guidance on monitoring trunk extensor muscle fatigue development using electromyography.

Low Back Pain-Induced Dynamic Trunk Muscle Control Impairments Are Associated with Altered Spatial EMG-Torque Relationships.

We quantified the relationship between high-density surface electromyographic (HDsEMG) oscillations (in both time and frequency domains) and torque steadiness during submaximal concentric/eccentric trunk extension/flexion contractions, in individuals with and without chronic low back pain (CLBP). Comparisons were made between regional differences in HDsEMG amplitude and HDsEMG-torque cross-correlation and coherence of the thoracolumbar erector spinae (ES), rectus abdominis (RA), and external oblique (EO) muscles between the two groups. HDsEMG signals were recorded from the thoracolumbar ES with two 64-electrode grids and from the RA and EO muscles with a single 64-electrode grid placed over each muscle. Torque signals were recorded with an isokinetic dynamometer. Coherence (δ band (0-5 Hz)) and cross-correlation analyses were used to examine the relationship between HDsEMG and torque signals. For this purpose, we used principal component analysis to reduce data dimensionality and improve HDsEMG-based torque estimation. We found that people with CLBP had poorer control during both concentric and eccentric trunk flexion and extension. Specifically, during trunk extension, they exhibited a higher HDsEMG-torque coherence in more cranial regions of the thoracolumbar ES and a higher HDsEMG cross-correlation compared with asymptomatic controls. During trunk flexion movements, they demonstrated higher HDsEMG amplitude of the abdominal muscles, with the center of activation being more cranial and a higher contribution of this musculature to the resultant torque (particularly the EO muscle). Our findings underscore the importance of evaluating torque steadiness in individuals with CLBP. Future research should consider the value of torque steadiness training and HDsEMG-based biofeedback for modifying trunk muscle recruitment strategies and improving torque steadiness performance in individuals with CLBP.

The dose-response relationship of quercetin on the motor unit firing patterns and contractile properties of muscle in men and women.

Quercetin is one type of ergogenic aid and its effects on the neuromuscular system have recently attracted interest, but its dose-effect is not yet fully understood. The aim of this study was to examine the effect of different doses of quercetin ingestion on motor unit firing patterns and muscle contractile properties in humans. Thirteen young males and females conducted neuromuscular performance tests before (PRE) and 60 min after (POST) ingestions of 500 or 200 mg of quercetin glycosides (Qg500/Qg200, respectively) or placebo (PLA) on three different days. At PRE and POST, motor unit firing rates were calculated from high-density surface electromyography of the vastus lateralis muscle during 120-s isometric contraction of knee extension at 10% of maximal voluntary contraction. Electrically elicited forces in knee extensor muscles were also measured. After 60 s of voluntary contraction, motor unit firing rates, normalized by the exerted muscle force at POST, were significantly lower at POST than PRE with Qg500 and Qg200 (p < 0.05), but not with PLA (p > 0.05). Changes in motor unit firing rates normalized by the exerted force from PRE to POST were significantly greater with Qg500 than Qg200 at the end of contraction (p < 0.05). Under all three conditions, the electrically elicited force did not significantly change from PRE to POST (p > 0.05). These results suggest that both 500 and 200-mg quercetin ingestions alter motor unit firing patterns, and that quercetin's effect is at least partially dose-dependent.

Evaluation of a passive upper limb exoskeleton using high-density surface electromyography

Evaluation of a passive upper limb exoskeleton using high-density surface electromyography

Tutorial. Surface electromyogram (sEMG) amplitude estimation: Best practices.

This tutorial intends to provide insight, instructions and "best practices" for those who are novices-including clinicians, engineers and non-engineers-in extracting electromyogram (EMG) amplitude from the bipolar surface EMG (sEMG) signal of voluntary contractions. A brief discussion of sEMG amplitude extraction from high density sEMG (HDsEMG) arrays and feature extraction from electrically elicited contractions is also provided. This tutorial attempts to present its main concepts in a straightforward manner that is accessible to novices in the field not possessing a wide range of technical background (if any) in this area. Surface EMG amplitude, also referred to as the sEMG envelope [often implemented as root mean square (RMS) sEMG or average rectified value (ARV) sEMG], quantifies the voltage variation of the sEMG signal and is grossly related to the overall neural excitation of the muscle and to peripheral parameters. The tutorial briefly reviews the physiological origin of the voluntary sEMG signal and sEMG recording, including electrode configurations, sEMG signal transduction, electronic conditioning and conversion by an analog-to-digital converter. These topics have been covered in greater detail in prior tutorials in this series. In depth descriptions of state-of-the-art methods for computing sEMG amplitude are then provided, including guidance on signal pre-conditioning, absolute value vs. square-law detection, selection of appropriate sEMG amplitude smoothing filters and attenuation of measurement noise. The tutorial provides a detailed list of best practices for sEMG amplitude estimation.

Simulating progressive motor neuron degeneration and collateral reinnervation in motor neuron diseases using a dynamic muscle model based on human single motor unit recordings

Objective. To simulate progressive motor neuron loss and collateral reinnervation in motor neuron diseases (MNDs) by developing a dynamic muscle model based on human single motor unit (MU) surface-electromyography (EMG) recordings. Approach. Single MU potentials recorded with high-density surface-EMG from thenar muscles formed the basic building blocks of the model. From the baseline MU pool innervating a muscle, progressive MU loss was simulated by removal of MUs, one-by-one. These removed MUs underwent collateral reinnervation with scenarios varying from 0% to 100%. These scenarios were based on a geometric variable, reflecting the overlap in MU territories using the spatiotemporal profiles of single MUs and a variable reflecting the efficacy of the reinnervation process. For validation, we tailored the model to generate compound muscle action potential (CMAP) scans, which is a promising surface-EMG method for monitoring MND patients. Selected scenarios for reinnervation that matched observed MU enlargements were used to validate the model by comparing markers (including the maximum CMAP and a motor unit number estimate (MUNE)) derived from simulated and recorded CMAP scans in a cohort of 49 MND patients and 22 age-matched healthy controls. Main results. The maximum CMAP at baseline was 8.3 mV (5th–95th percentile: 4.6 mV–11.8 mV). Phase cancellation caused an amplitude drop of 38.9% (5th–95th percentile, 33.0%–45.7%). To match observations, the geometric variable had to be set at 40% and the efficacy variable at 60%–70%. The Δ maximum CMAP between recorded and simulated CMAP scans as a function of fitted MUNE was −0.4 mV (5th–95th percentile = −4.0 – +2.4 mV). Significance. The dynamic muscle model could be used as a platform to train personnel in applying surface-EMG methods prior to their use in clinical care and trials. Moreover, the model may pave the way to compare biomarkers more efficiently, without directly posing unnecessary burden on patients.

Extracting Individual Muscle Drive and Activity From High-Density Surface Electromyography Signals Based on the Center of Gravity of Motor Unit.

Neural interfacing has played an essential role in advancing our understanding of fundamental movement neurophysiology and the development of human-machine interface. However, direct neural interfaces from brain and nerve recording are currently limited in clinical areas for their invasiveness and high selectivity. Here, we applied the surface electromyogram (EMG) in studying the neural control of movement and proposed a new non-invasive way of extracting neural drive to individual muscles. Sixteen subjects performed isometric contractions to complete six hand tasks. High-density surface EMG signals (256 channels in total) recorded from the forearm muscles were decomposed into motor unit firing trains. The location of each decomposed motor unit was represented by its center of gravity and was put into clustering for distinct muscle regions. All the motor units in the same cluster served as a muscle-specific motor pool from which individual muscle drive could be extracted directly. Moreover, we cross-validated the self-clustered muscle regions by magnetic resonance imaging (MRI) recorded from the subjects' forearms. All motor units that fall within the MRI region are considered correctly clustered. We achieved a clustering accuracy of 95.72% ± 4.01% for all subjects. We provided a new framework for collecting experimental muscle-specific drives and generalized the way of surface electrode placement without prior knowledge of the targeting muscle architecture.

Functionally Adaptive Myosite Selection Using High-Density sEMG for Upper Limb Myoelectric Prostheses.

Our study defines a novel electrode placement method called Functionally Adaptive Myosite Selection (FAMS), as a tool for rapid and effective electrode placement during prosthesis fitting. We demonstrate a method for determining electrode placement that is adaptable towards individual patient anatomy and desired functional outcomes, agnostic to the type of classification model used, and provides insight into expected classifier performance without training multiple models. FAMS relies on a separability metric to rapidly predict classifier performance during prosthesis fitting. The results show a predictable relationship between the FAMS metric and classifier accuracy (3.45%SE), allowing estimation of control performance with any given set of electrodes. Electrode configurations selected using the FAMS metric show improved control performance ( ) for target electrode counts compared to established methods when using an ANN classifier, and equivalent performance ( R2 ≥ .96) to previous top-performing methods on an LDA classifier, with faster convergence ( ). We used the FAMS method to determine electrode placement for two amputee subjects by using the heuristic to search through possible sets, and checking for saturation in performance vs electrode count. The resulting configurations that averaged 95.8% of the highest possible classification performance using a mean 25 number of electrodes (19.5% of the available sites). FAMS can be used to rapidly approximate the tradeoffs between increased electrode count and classifier performance, a useful tool during prosthesis fitting.

Optimizing the feature set and electrode configuration of high-density electromyogram via interpretable deep forest

Hand gesture recognition via high-density surface electromyogram (HDsEMG) has received increasing attentions in human–machine interactions. However, with an extremely large number of electrodes in the electrode arrays, the computational burden substantially increased in real world mobile computing applications. Additionally, with diverse features extracted from all electrodes, large information redundancy reduced the learning efficiency of machine learning models. Furthermore, most machine learning models were employed as black-box modules, increasing concerns on both ethics and the user trust on machine learning techniques, especially in human-centered applications. In this work, we applied deep forest in HDsEMG-based hand gesture recognition. Deep forest is a new deep learning architecture where information could be processed layer-by-layer via cascaded random forest modules. Built on highly interpretable decision trees, deep forest models retain the interpretability of decision trees. With the inherent interpretability of deep forest, the most important features and electrodes that deep forest focused on in the decision making process can also be used to optimize the feature set and electrode configuration. We also compared the model interpretation results with the anatomy structure of forearm muscles, providing relations between the interpretations of deep forest and the physiological basis of neuromuscular systems. HDsEMG (256 electrodes) from 20 subjects were used in our analyses. To simulate realistic application scenarios, all analyses and evaluations were performed in inter-subject validations. Results showed that with optimized feature set and electrode configuration, the number of features/electrodes was substantially reduced, with the overall classification accuracy improved. Data and codes are available online.

Stretchable and durable HD-sEMG electrodes for accurate recognition of swallowing activities on complex epidermal surfaces

Surface electromyography (sEMG) is widely used in monitoring human health. Nonetheless, it is challenging to capture high-fidelity sEMG recordings in regions with intricate curved surfaces such as the larynx, because regular sEMG electrodes have stiff structures. In this study, we developed a stretchable, high-density sEMG electrode array via layer-by-layer printing and lamination. The electrode offered a series of excellent human‒machine interface features, including conformal adhesion to the skin, high electron-to-ion conductivity (and thus lower contact impedance), prolonged environmental adaptability to resist water evaporation, and epidermal biocompatibility. This made the electrode more appropriate than commercial electrodes for long-term wearable, high-fidelity sEMG recording devices at complicated skin interfaces. Systematic in vivo studies were used to investigate its ability to classify swallowing activities, which was accomplished with high accuracy by decoding the sEMG signals from the chin via integration with an ear-mounted wearable system and machine learning algorithms. The results demonstrated the clinical feasibility of the system for noninvasive and comfortable recognition of swallowing motions for comfortable dysphagia rehabilitation.

Electromyographic signature of isometric squat in the highest refuge in Europe.

Reports of electromyography during hypoxic exercise are contrasting, due to protocol and muscle diversity. This work aimed to investigate alterations in muscle activation and myoelectrical fatigue during exercise at high-altitude in those muscles primarily involved in trekking. Twelve young adults balanced by gender and age were tested at low (1,667 m) and high (4,554 m, "Capanna Margherita", Italy) altitude, during an isometric squat lasting 60 seconds. High-density surface electromyography was performed from the quadriceps of right limb. The root mean square (RMS), median frequency with its slope, and muscle fiber conduction velocity (MFCV) were computed. Neither males nor females showed changes in median frequency (Med: 36.13 vs 35.63 Hz) and its slope (Med: -9 vs -12 degree) in response to high-altitude trekking, despite a great inter-individual heterogeneity, nor differences were found for MFCV. RMS was not significantly equivalent, with greater values at low altitude (0.385 ± 0.104 mV) than high altitude (0.346 ± 0.090 mV). Unexpected results can be due either to a postural compensation of the whole body compensating for a relatively greater effort or to the inability to support muscle activation after repeated physical efforts. Interesting results may emerge by measuring simultaneously electromyography, muscle oxygenation and kinematics comparing trekking at normoxia vs hypoxia.

Sensorimotor integration is affected by acute whole-body vibration: a coherence study.

Introduction: Several whole-body vibration (WBV) effects on performance have been related to potential changes in the neural drive, motor unit firing rate, and sensorimotor integration. In the present paper, motor unit coherence analysis was performed to detect the source of neural modulation based on the frequency domain. Methods: Thirteen men [25 ± 2.1years; Body Mass Index (BMI) = 23.9 ± 1.3kgm2; maximal voluntary force (MVF): 324.36 ± 41.26N] performed sustained contractions of the Tibialis Anterior (TA) at 10%MVF before and after acute WBV. The vibrating stimulus was applied barefoot through a platform to target the TA. High-Density surface Electromyography (HDsEMG) was used to record the myoelectrical activity of TA to evaluate coherence from motor unit cumulative spike-trains (CSTs). Results: Mean coherence showed a significant decrease in the alpha and low-beta bandwidths (alpha: from 0.143 ± 0.129 to 0.132 ± 0.129, p = 0.035; low-beta: from 0.117 ± 0.039 to 0.086 ± 0.03, p = 0.0001), whereas no significant changes were found in the other ones (p > 0.05). The discharge rate (DR) and the Force Covariance (CovF%) were not significantly affected by acute WBV exposure (p > 0.05). Discussion: According to the significant effects found in alpha and low-beta bandwidths, which reflect sensorimotor integration parameters, accompanied by no differences in the DR and CovF%, the present results underlined that possible neural mechanisms at the base of the previously reported performance enhancements following acute WBV are likely based on sensorimotor integration rather than direct neural drive modulation.

P2E5-12 A method of simultaneous record and analysis of voice and high-density surface electromyography of laryngeal muscle

P2E5-12 A method of simultaneous record and analysis of voice and high-density surface electromyography of laryngeal muscle

A Novel Event-Driven Spiking Convolutional Neural Network for Electromyography Pattern Recognition.

Electromyography (EMG) pattern recognition is an important technology for prosthesis control and human-computer interaction etc. However, the practical application of EMG pattern recognition is hampered by poor accuracy and robustness due to electrode shift caused by repeated wearing of the signal acquisition device. Moreover, the user's acceptability is low due to the heavy training burden, which is caused by the need for a large amount of training data by traditional methods. In order to explore the advantage of spiking neural network (SNN) in solving the poor robustness and heavy training burden problems in EMG pattern recognition, a spiking convolutional neural network (SCNN) composed of cyclic convolutional neural network (CNN) and fully connected modules is proposed and implemented in this study. High density surface electromyography (HD-sEMG) signals collected from 6 gestures of 10 subjects at 6 electrode positions are taken as the research object. Compared to CNN with the same structure, CNN-Long Short Term Memory (CNN-LSTM), linear kernel linear discriminant analysis classifier (LDA) and spiking multilayer perceptron (Spiking MLP), the accuracy of SCNN is 50.69%, 33.92%, 32.94% and 9.41% higher in the small sample training experiment, 6.50%, 4.23%, 28.73%, and 2.57% higher in the electrode shifts experiment respectively. In addition, the power consumption of SCNN is about 1/93 of CNN. The advantages of the proposed framework in alleviating user training burden, mitigating the adverse effect of electrode shifts and reducing power consumption make it very meaningful for promoting the development of user-friendly real-time myoelectric control system.

Neuromuscular assessment of force development, postural, and gait performance under cognitive-motor dual-tasking in healthy older adults and early Parkinson's disease patients: Study protocol for a cross-sectional Mobile Brain/Body Imaging (MoBI) study

Background: Neuromuscular dysfunction is common in older adults and more pronounced in neurodegenerative diseases. In Parkinson's disease (PD), a complex set of factors often prevents the effective performance of activities of daily living that require intact and simultaneous performance of the motor and cognitive tasks. Methods: The cross-sectional study includes a multifactorial mixed-measure design. Between-subject factor grouping the sample will be Parkinson’s Disease (early PD vs. healthy). The within-subject factors will be the task complexity (single- vs. dual-task) in each motor activity, i.e., overground walking, semi-tandem stance, and isometric knee extension, and a walking condition (wide vs. narrow lane) will be implemented for the overground walking activity only. To study dual-task (DT) effects, in each motor activity participants will be given a secondary cognitive task, i.e., a visual discrimination task for the overground walking, an attention task for the semi-tandem, and mental arithmetic for the isometric extension. Analyses of DT effects and underlying neuronal correlates will focus on both gait and cognitive performance where applicable. Based on an a priori sample size calculation, a total N = 42 older adults (55-75 years) will be recruited. Disease-specific changes such as laterality in motor unit behavior and cortical control of movement will be studied with high-density surface electromyography and electroencephalography during static and dynamic motor activities, together with whole-body kinematics. Discussion: This study will be one of the first to holistically address early PD neurophysiological and neuromuscular patterns in an ecologically valid environment under cognitive-motor DT conditions of different complexities. The outcomes of the study aim to identify the biomarker for early PD either at the electrophysiological, muscular or kinematic level or in the communication between these systems. Clinical Trial Registration: ClinicalTrials.Gov, NCT05477654. This study was approved by the Medical Ethical Committee (106/2021).

Cross-user gesture recognition from sEMG signals using an optimal transport assisted student-teacher framework

The cross-user gesture recognition is a puzzle in the myoelectric control system, owing to great variability in muscle activities across different users. To address this problem, a novel optimal transport (OT) assisted student-teacher (ST) framework (termed OT-ST) was proposed in this paper to facilitate transfer across user domains in an unsupervised domain adaptation (UDA) manner. In this framework, the initial parameters of the ST models were trained with the labeled data from users in the source domain. In the model transfer stage for a new user in the target domain, the teacher model was utilized to generate pseudo labels for unlabeled testing samples, providing guidance to the adaptation of the student model. The OT algorithm was employed to optimize the pseudo labels generated from the teacher model, avoiding the model bias and further improving the effect of domain adaptation. The performance of the proposed OT-ST framework was evaluated via experiments of classifying seven hand gestures using high-density surface electromyogram (HD-sEMG) recordings from extensor digitorum muscles of eight intact-limbed subjects. The OT-ST framework yielded a high accuracy of 96.50 ± 2.88% for new users, and outperformed other common machine learning and UDA methods significantly (p < 0.01), demonstrating its effectiveness. The OT-ST framework does not require special repetitive training or any labeled data for calibration. In addition, it can incrementally learn from new testing samples and improve the recognition ability. This study provides a promising method for developing user-generic myoelectric pattern recognition, with wide applications in human-computer interaction, consumer electronics and prosthesis control.

Task-related differences in peroneus longus muscle fiber conduction velocity.

It has been identified that the peroneus longus presents a regional activity. Specifically, a greater activation of the anterior and posterior compartments has been observed during eversion, whereas a lower activation of the posterior compartment has been reported during plantarflexion. In addition to myoelectrical amplitude, motor unit recruitment can be inferred indirectly from muscle fiber conduction velocity (MFCV). However, there are few reports of MFCV of the regions that make up a muscle, and even less, MFCV of the peroneus longus compartments. This study aimed to analyze the MFCV of peroneus longus compartments during eversion and plantarflexion. Twenty-one healthy individuals were assessed. High-density surface electromyography was recorded from the peroneus longus during eversion and plantarflexion at 10%, 30%, 50%, and 70% of maximal voluntary isometric contraction. The posterior compartment presented a lower MFCV than the anterior compartment during plantarflexion, and both compartments did not show differences in MFCV during eversion; however, the posterior compartment showed an increase in MFCV during eversion compared to plantarflexion. Differences observed in the MFCV of the peroneus longus compartments could support a regional activation strategy and, to some extent, explain different motor unit recruitment strategies of the peroneus longus during ankle movements.