An isometric muscle force estimation framework based on a high-density surface EMG array and an NMF algorithm

Characterizing upper limb motor dysfunction with temporal and spatial distribution of muscle synergy extracted from high-density surface electromyography

Objective.To promote the development of objective and comprehensive motion function assessment for patients, based on high-density surface electromyography (HD-sEMG), this study investigates the temporal and spatial variations of neuromuscular activities related to upper limb motor dysfunction.Approach.Patients with unilateral upper limb motor dysfunction and healthy controls were enrolled in the study. HD-sEMG was collected from both arms while they were performing eight hand and wrist movements. Muscle synergies were extracted from the HD-sEMG. Symmetry of bilateral upper limb synergies and synergy differences between motions were proposed as spatial indicators to measure alterations in synergy spatial distribution. Additionally, as a temporal characteristic, the correlation of bilateral upper limb activation coefficient was proposed to describe the coordination control of the central nervous system. All temporal and spatial indicators were compared between patients and healthy subjects.Main results.The patients showed a significant decrease (p< 0.05) in the symmetry of bilateral upper limb synergy spatial distribution and correlation of bilateral upper limb activation coefficient. Patients with motor dysfunction also showed an increase in synergy similarity between motions, indicating altered spatial distribution of muscle synergies.Significance.These findings provide valuable insights into specific patterns associated with motor dysfunction, informing motor function assessment, and guiding targeted interventions and rehabilitation strategies for neurologically disordered patients.

Mapping of spastic muscle activity after stroke: difference between passive stretch and active contraction

BACKGROUND

Investigating the spatial distribution of muscle activity would facilitate understanding the underlying mechanism of spasticity. The purpose of this study is to investigate the characteristics of spastic muscles during passive stretch and active contraction by high-density surface electromyography (HD-sEMG).

METHODS

Fourteen spastic hemiparetic subjects and ten healthy subjects were recruited. The biceps brachii (BB) muscle activity of each subject was recorded by HD-sEMG during passive stretch at four stretch velocities (10, 60, 120, 180˚/s) and active contraction at three submaximal contraction levels (20, 50, 80%MVC). The intensity and spatial distribution of the BB activity were compared by the means of two-way analysis of variance, independent sample t-test, and paired sample t-test.

RESULTS

Compared with healthy subjects, spastic hemiparetic subjects showed significantly higher intensity with velocity-dependent heterogeneous activation during passive stretch and more lateral and proximal activation distribution during active contraction. In addition, spastic hemiparetic subjects displayed almost non-overlapping activation areas during passive stretch and active contraction. The activation distribution of passive stretch was more distal when compared with the active contraction.

CONCLUSIONS

These alterations of the BB activity could be the consequence of deficits in the descending central control after stroke. The complementary spatial distribution of spastic BB activity reflected their opposite motor units (MUs) recruitment patterns between passive stretch and active contraction. This HD-sEMG study provides new neurophysiological evidence for the spatial relationship of spastic BB activity between passive stretch and active contraction, advancing our knowledge on the mechanism of spasticity.

TRIAL REGISTRATION

ChiCTR2000032245.

Online prediction of sustained muscle force from individual motor unit activities using adaptive surface EMG decomposition

Decoding movement intentions from motor unit (MU) activities to represent neural drive information plays a central role in establishing neural interfaces, but there remains a great challenge for obtaining precise MU activities during sustained muscle contractions. In this paper, we presented an online muscle force prediction method driven by individual MU activities that were decomposed from prolonged surface electromyogram (SEMG) signals in real time. In the training stage of the proposed method, a set of separation vectors was initialized for decomposing MU activities. After transferring each decomposed MU activity into a twitch force train according to its action potential waveform, a neural network was designed and trained for predicting muscle force. In the subsequent online stage, a practical double-thread-parallel algorithm was developed. One frontend thread predicted the muscle force in real time utilizing the trained network and the other backend thread simultaneously updated the separation vectors. To assess the performance of the proposed method, SEMG signals were recorded from the abductor pollicis brevis muscles of eight subjects and the contraction force was simultaneously collected. With the update procedure in the backend thread, the force prediction performance of the proposed method was significantly improved in terms of lower root mean square deviation (RMSD) of around 10% and higher fitness (R2) of around 0.90, outperforming two conventional methods. This study provides a promising technique for real-time myoelectric applications in movement control and health.

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

OBJECTIVE

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.

METHODS

FAMS relies on a separability metric to rapidly predict classifier performance during prosthesis fitting.

RESULTS

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).

SIGNIFICANCE

FAMS can be used to rapidly approximate the tradeoffs between increased electrode count and classifier performance, a useful tool during prosthesis fitting.

A real-time and convex model for the estimation of muscle force from surface electromyographic signals in the upper and lower limbs

Surface electromyography (sEMG) is a signal consisting of different motor unit action potential trains and records from the surface of the muscles. One of the applications of sEMG is the estimation of muscle force. We proposed a new real-time convex and interpretable model for solving the sEMG-force estimation. We validated it on the upper limb during isometric voluntary flexions-extensions at 30%, 50%, and 70% Maximum Voluntary Contraction in five subjects, and lower limbs during standing tasks in thirty-three volunteers, without a history of neuromuscular disorders. Moreover, the performance of the proposed method was statistically compared with that of the state-of-the-art (13 methods, including linear-in-the-parameter models, Artificial Neural Networks and Supported Vector Machines, and non-linear models). The envelope of the sEMG signals was estimated, and the representative envelope of each muscle was used in our analysis. The convex form of an exponential EMG-force model was derived, and each muscle's coefficient was estimated using the Least Square method. The goodness-of-fit indices, the residual signal analysis (bias and Bland-Altman plot), and the running time analysis were provided. For the entire model, 30% of the data was used for estimation, while the remaining 20% and 50% were used for validation and testing, respectively. The average R-square (%) of the proposed method was 96.77 ± 1.67 [94.38, 98.06] for the test sets of the upper limb and 91.08 ± 6.84 [62.22, 96.62] for the lower-limb dataset (MEAN ± SD [min, max]). The proposed method was not significantly different from the recorded force signal (p-value = 0.610); that was not the case for the other tested models. The proposed method significantly outperformed the other methods (adj. p-value < 0.05). The average running time of each 250 ms signal of the training and testing of the proposed method was 25.7 ± 4.0 [22.3, 40.8] and 11.0 ± 2.9 [4.7, 17.8] in microseconds for the entire dataset. The proposed convex model is thus a promising method for estimating the force from the joints of the upper and lower limbs, with applications in load sharing, robotics, rehabilitation, and prosthesis control for the upper and lower limbs.

A robust HD-sEMG sensor suitable for convenient acquisition of muscle activity in clinical post-stroke dysphagia

Objective.A flexible high-density surface electromyography (HD-sEMG) sensor combined with an adaptive algorithm was used to collect and analyze the swallowing activities of patients with Post-stroke dysphagia.Approach.The electrode frame, modified electrode, and bonded substrate of the sensor were fabricated using a flexible printed circuit process, controlled drop coating, and molding, respectively. The adaptation algorithm was achieved by using Laplace and Teager-Kaiser energy operators to extract active segments, a cross-correlation coefficient matrix (CCCM) to evaluate synergy, and multi-frame real-time dynamic root mean square (RMS) to visualize spatiotemporal information to screen lesions and level of dysphagia. Finally, support vector machines (SVM) were adopted to explore the classification accuracy of sex, age, and lesion location with small sample sizes.Main results.The sensor not only has a basic low contact impedance (0.262 kΩ) and high signal-to-noise ratio (37.284 ± 1.088 dB) but also achieves other characteristics suitable for clinical applications, such as flexibility (747.67 kPa) and durability (1000 times) balance, simple operation (including initial, repeated, and replacement use), and low cost ($ 15.2). The three conclusions are as follows. CCCM can be used as a criterion for judging the unbalanced muscle region of the patient's neck and can accurately locate unbalanced muscles. The RMS cloud map provides the time consumption, swallowing times, and unbalanced areas. When the lesion location involves the left and right hemispheres simultaneously, it can be used as an evidence of relatively severely unbalanced areas. The classification accuracy of SVM in terms of sex, age, and lesion location was as high as 100%.Significance.The HD-sEMG sensor in this study and the adaptation algorithm will contribute to the establishment of a larger-scale database in the future to establish more detailed and accurate quantitative standards, which will be the basis for developing more optimized screening mechanisms and rehabilitation assessment methods.

Relationship between EMG features and force in orbicularis oris muscle

BACKGROUND

Lip incompetence resulting from mouth breathing is a common clinical manifestation, while there are no definite indicators of amplitude and intensity of muscle functional training in clinical practice, which leads to unsatisfactory training results.

OBJECTIVE

The aim was to quantify the relationship between electromyography (EMG) and force in orbicularis oris muscle, so that the indicators of muscle functional training can be evaluated using EMG signals, so as to improve the training effects.

METHODS

The EMG and the force signals of orbicularis oris muscle from 0% to 100% MVC within 5 s in twelve healthy subjects (six males and six females; age, 25 ± 2 years; mass, 60 ± 15 kg) were recorded simultaneously for three trials. Four EMG features consisting of RMS, WAMP, SampEn and FuzzyEn were analyzed. The regression analyses were performed using first-order and third-order polynomial model.

RESULTS

There were high correlations between the four EMG features and muscle force with the two models. The third-order model yielded a higher coefficient of determination (R2) than the linear model (p< 0.001) and the result of FuzzyEn (R2: 0.884 ± 0.059) was the highest in the four features.

CONCLUSION

The third-order model with FuzzyEn of EMG signals may be used to guide the muscle functional training.

Study on Flexible sEMG Acquisition System and Its Application in Muscle Strength Evaluation and Hand Rehabilitation

Wearable devices based on surface electromyography (sEMG) to detect muscle activity can be used to assess muscle strength with the development of hand rehabilitation applications. However, conventional acquisition devices are usually complicated to operate and poorly comfortable for more medical and scientific application scenarios. Here, we report a flexible sEMG acquisition system that combines a graphene-based flexible electrode with a signal acquisition flexible printed circuit (FPC) board. Our system utilizes a polydimethylsiloxane (PDMS) substrate combined with graphene transfer technology to develop a flexible sEMG sensor. The single-lead sEMG acquisition system was designed and the FPC board was fabricated considering the requirements of flexible bending and twisting. We demonstrate the above design approach and extend this flexible sEMG acquisition system to applications for assessing muscle strength and hand rehabilitation training using a long- and short-term memory network training model trained to predict muscle strength, with 98.81% accuracy in the test set. The device exhibited good flexion and comfort characteristics. In general, the ability to accurately and imperceptibly monitor surface electromyography (EMG) signals is critical for medical professionals and patients.

A comparison of contributions of individual muscle and combination muscles to interaction force prediction using KPCA-DRSN model

Rapid and accurate prediction of interaction force is an effective way to enhance the compliant control performance. However, whether individual muscles or a combination of muscles is more suitable for interaction force prediction under different contraction tasks is of great importance in the compliant control of the wearable assisted robot. In this article, a novel algorithm that is based on sEMG and KPCA-DRSN is proposed to explore the relationship between interaction force prediction and sEMG signals. Furthermore, the contribution of each muscle to the interaction force is assessed based on the predicted results. First of all, the experimental platform for obtaining the sEMG is described. Then, the raw sEMG signal of different muscles is collected from the upper arm during different contractions. Meanwhile, the output force is collected by the force sensor. The Kernel Principal Component Analysis (KPCA) method is adopted to remove the invalid components of the raw sEMG signal. After that, the processed sequence is fed into the Deep Residual Shrinkage Network (DRSN) to predict the interaction force. Finally, based on the prediction results, the contribution of each sEMG signal from different muscles to the interaction force is evaluated by the mean impact value (MIV) indicator. The experimental results demonstrate that our methods can automatically extract the valid features of sEMG signal and provided fast and efficient prediction. In addition, the single muscle with the largest MIV index could predict the interaction force faster and more accurately than the muscle combination in different contraction tasks. The finding of our research provides a solid evidence base for the compliant control of the wearable robot.

A Simultaneous Gesture Classification and Force Estimation Strategy Based on Wearable A-Mode Ultrasound and Cascade Model

The existing Human-Machine Interfaces (HMI) based on gesture recognition using surface electromyography (sEMG) have made significant progress. However, the sEMG has inherent limitations as well as the gesture classification and force estimation have not been effectively combined. There are limitations in applications such as prosthetic control and clinical rehabilitation, etc. In this paper, a grasping gesture and force recognition strategy based on wearable A-mode ultrasound and two-stage cascade model is proposed, which can simultaneously estimate the force while classifying the grasping gesture. This paper experiments five grasping gestures and four force levels (5-50%MVC). The results demonstrate that the performance of the proposed model is significantly better than that of the traditional model both in classification and regression (p < 0.001). Additionally, the two-stage cascade regression model (TSCRM) used the Gaussian Process regression model (GPR) with the mean and standard deviation (MSD) feature obtains excellent results, with normalized root-mean-square error (nRMSE) and correlation coefficient (CC) of 0.10490.0374 and 0.94610.0354, respectively. Besides, the latency of the model meets the requirement of real-time recognition (T < 15ms). Therefore, the research outcomes prove the feasibility of the proposed recognition strategy and provide a reference for the field of prosthetic control, etc.

Synergy Analysis of Back Muscle Activities in Patients with Adolescent Idiopathic Scoliosis Based on High-Density Electromyogram

Adolescent idiopathic scoliosis (AIS) is a common structural spinal deformity and is typically associated with altered muscle properties. However, it is still unclear how muscle activities and the underlying neuromuscular control are changed in the entire scoliotic zone, restricting the corresponding pathology investigation and treatment enhancements. In this study, high-density electromyogram (HD-EMG) was utilized to explore the neuromuscular synergy of back muscle activities. For each of ten AIS patients and ten healthy subjects for comparison, a high-density electrode array was placed on their back from T8 to L4 to record EMG signals when performing five spinal motions (flexion/extension, lateral bending, axial rotation, siting, and standing). From the HD-EMG recordings, muscle synergies were extracted using the non-negative matrix factorization method and the topographical maps of EMG root-mean-square were constructed. For both the AIS and healthy subjects, the experimental results indicated that two muscle synergy groups could explain over 90% of recorded muscle activities for all five motions. During flexion/extension, the patients presented statistically significant higher activations on the convex side in the entire root-mean-square maps and synergy vector maps (p <0.05). During lateral bending and axial rotation, the patients exhibited less activated muscles on the dominant actuating side relative to the contralateral side and their synergy vector maps showed a less homogenous and more diffuse distribution of muscle contraction with statistically different centers of gravity. The findings suggest that a scoliotic spine might adopt an altered modular muscular coordination strategy to actuate different dominant muscles as adapted compensations for the deformation.

Decoding Muscle Force From Motor Unit Firings Using Encoder-Decoder Networks

Appropriate interpretation of motor unit (MU) activities after surface EMG (sEMG) decomposition is a key factor to decode motor intentions in a noninvasive and physiologically meaningful way. However, there are great challenges due to the difficulty in cross-trial MU tracking and unavoidable loss of partial MU information resulting from incomplete decomposition. In light of these issues, this study presents a novel framework for interpreting MU activities and applies it to decode muscle force. The resulting MUs were clustered and classified into different categories by characterizing their spatially distributed firing waveforms. The process served as a general MU tracking method. On this basis, after transferring the MU firing trains to twitch force trains by a twitch force model, a deep network was designed to predict the normalized force. In addition, MU category distribution was examined to calibrate the actual force level, while functions of some unavailable MUs were compensated. To investigate the effectiveness of this framework, high-density sEMG signals were recorded using an 8×8 electrode array from the abductor pollicis brevis muscles of eight subjects, while thumb abduction force was measured. The proposed method outperformed three common methods ( ) yielding the lowest root mean square deviation of 6.68% ± 1.29% and the highest fitness ( R² ) of 0.94 ± 0.04 between the predicted force and the actual force. This study offers a valuable, computational solution for interpreting individual MU activities, and its effectiveness was confirmed in muscle force estimation.

A novel sEMG-based force estimation method using deep-learning algorithm

This paper discusses the problem of force estimation represented by surface electromyography (sEMG) signals collected from an armband-like collection device. The scheme is proposed for the sake of two dimensions of sEMG signals: spatial and temporal information. From the point of space, first, appropriate channel number across all subjects is investigated. During this progress, an electrode channel selection method based on Spearman’s rank order correlation coefficient is utilized to detect signals from active muscle. Then, to reduce the computation and highlight the channel information, linear regression (LR) algorithm is conducted to weight each channel. Besides, the recurrent neural network (RNN) is used to capture the temporal information and model the relation between sEMG and output force. Experiments conducted on four subjects demonstrate that six channels are enough to characterize the muscle activity. By combining the selected channels with different weight coefficients, LR algorithm can fit the output force better than simply averaging them. Furthermore, RNN with long short-term memory cell shows the superiority in time series modeling, which can improve our results to a greater degree. Experimental results prove the feasibility of the proposed method.

Optimal strategy of sEMG feature and measurement position for grasp force estimation

Grasp force estimation based on surface electromyography (sEMG) is essential for the dexterous control of a prosthetic hand. Nowadays, although increasing the number of sEMG measurement positions and extracting more features are common methods to increase the accuracy of grasp force estimation, it will increase the computational burden. In this paper, an approach based on analysis of variance (ANOVA) and generalized regression neural network (GRNN) for optimal measurement positions and features is proposed, with the purpose of using fewer measurement positions or features to achieve higher estimation accuracy. Firstly, we captured six channels of sEMG from subjects’ forearm and grasp force synchronously. Then, four kinds of features in time domain are extracted from each channel of sEMG. By combining different measurement position sets (MPSs) and feature set (FSs), we construct 945 data sets. These data sets are fed to GRNN to realize grasp force estimation. Normalized root mean square error (NRMS), normalized mean of absolute error (NMAE), and correlation coefficient (CC) between estimated grasp force and actual force are introduced to evaluate the performance of grasp force estimation. Finally, ANOVA and Tukey HSD testing are introduced to analyze grasp force estimation results so as to obtain the optimal measurement positions and features. We obtain the optimal MPSs for grasp force estimation when different FSs are employed, and the optimal FSs when different MPSs are utilized.

Muscle Force Estimation Based on Neural Drive Information From Individual Motor Units

Estimation of muscle contraction force based on the macroscopic feature of surface electromyography (SEMG) has been widely reported, but the use of microscopic neural drive information has not been thoroughly investigated. In this study, a novel method is proposed to process individual motor unit (MU) activities (firing sequences and action potential waveforms) derived from the decomposition of high density SEMG (HD-SEMG), and it is applied to muscle force estimation. In the proposed method, a supervised machine learning approach was conducted to determine the twitch force of each MU according to its action potential waveforms, which enables separate calculation of every MU's contribution to force. Thus, the muscle force was predicted through a physiologically meaningful muscle force model. In the experiment, HD-SEMG data were recorded from the abductor pollicis brevis muscles of eight healthy subjects during their performance of thumb abduction with the force increasing gradually from zero to four force levels (10%, 20%, 30%, 40% of the maximal voluntary contraction), while the true muscle force was measured simultaneously. When the proposed method was used, the root mean square difference (RMSD) of the error of the estimated force with respect to the measured force was reported to be 8.3% ± 2.8%. The proposed method also significantly outperformed the other four common methods for force estimation (RMSD: from 11.7% to 20%, ), demonstrating its effectiveness. This study offers a useful tool for exploiting the neural drive information towards muscle force estimation with improved precision. The proposed method has wide applications in precise motor control, sport and rehabilitation medicine.

Spatial filtering for enhanced high-density surface electromyographic examination of neuromuscular changes and its application to spinal cord injury

Background

Spatial filtering of multi-channel signals is considered to be an effective pre-processing approach for improving signal-to-noise ratio. The use of spatial filtering for preprocessing high-density (HD) surface electromyogram (sEMG) helps to extract critical spatial information, but its application to non-invasive examination of neuromuscular changes have not been well investigated.

Methods

Aimed at evaluating how spatial filtering can facilitate examination of muscle paralysis, three different spatial filtering methods are presented using principle component analysis (PCA) algorithm, non-negative matrix factorization (NMF) algorithm, and both combination, respectively. Their performance was evaluated in terms of diagnostic power, through HD-sEMG clustering index (CI) analysis of neuromuscular changes in paralyzed muscles following spinal cord injury (SCI).

Results

The experimental results showed that: (1) The CI analysis of conventional single-channel sEMG can reveal complex neuromuscular changes in paralyzed muscles following SCI, and its diagnostic power has been confirmed to be characterized by the variance of Z scores; (2) the diagnostic power was highly dependent on the location of sEMG recording channel. Directly averaging the CI diagnostic indicators over channels just reached a medium level of the diagnostic power; (3) the use of either PCA-based or NMF-based filtering method yielded a greater diagnostic power, and their combination could even enhance the diagnostic power significantly.

Conclusions

This study not only presents an essential preprocessing approach for improving diagnostic power of HD-sEMG, but also helps to develop a standard sEMG preprocessing pipeline, thus promoting its widespread application.

Upper Limb End-Effector Force Estimation During Multi-Muscle Isometric Contraction Tasks Using HD-sEMG and Deep Belief Network

In this study, research was carried out on the end-effector force estimation of two representative multi-muscle contraction tasks: elbow flexion and palm-pressing. The aim was to ascertain whether an individual muscle or a combination of muscles is more suitable for the end-effector force estimation. High-density surface electromyography (HD-sEMG) signals were collected from four primary muscle areas of the upper arm and forearm: the biceps brachii (BB), brachialis (BR), triceps brachii (TB), brachioradialis (BRD), and extensor digitorum communis (EDC). The wrist pulling and palm-pressing forces were measured in elbow flexion and palm-pressing tasks, respectively. The deep belief network (DBN) was adopted to establish the relation between HD-sEMG and the measured force. The representative signals of the four primary areas, which were considered as the input signal of the force estimation model, were extracted by HD-sEMG using the principle component analysis (PCA) algorithm, and fed separately or together into the DBN. An index termed mean impact value (MIV) was proposed to describe the priority of different muscle groups for estimating the end-effector force. The experimental results demonstrated that, in multi-muscle isometric contraction tasks, the dominant muscles with the highest activation degree could track variations in the end-effector force more effectively, and are more suitable than a combination of muscles. The main contributions of this research are as follows: (1) To fuse the activation information from different muscles effectively, DBN was adopted to establish the relationship between HD-sEMG and the generated force, and achieved highly accurate force estimation. (2) Based on the well-trained DBN force estimation model, an index termed MIV was presented to evaluate the priority of muscles for estimating the generated force.

Investigation on the Contributions of Different Muscles to the Generated Force based on HD-sEMG and DBN

In this study, we proposed a multi-muscle contraction force estimation framework and implemented it on the elbow flexion task to explore the contributions of different muscles to integrated force at different force levels. High density surface electromyography (HD-sEMG) signals were collected from four muscle groups and the wrist pulling force was measured synchronously. The deep belief network (DBN) was adopted to establish the relationship between HD-sEMG and force. The representative signals of four primary areas, which were considered as the input signal of the force estimation model, were extracted from HD-sEMG by principle component analysis (PCA) algorithm, then fed separately or in common to the DBN to estimate the generated force. And the contributions of different muscle groups to the generated force was analyzed with an index called mean impact value (MIV). The experimental result demonstrates that in multi-muscle contraction task, not all muscles are suitable for force estimation, the force estimation accuracy obtained using only one muscle approximates even exceeds that obtained using multiple muscles, and the relative contributions of different muscle groups to the force can be obtained according to the ranking of MIVs. This scheme provides an effective method for muscle force estimation in multi-muscle contraction tasks, and can be further applied to biomechanics, sports and rehabilitation medicine.

A database of multi-channel intramuscular electromyogram signals during isometric hand muscles contractions

Hand movement is controlled by a large number of muscles acting on multiple joints in the hand and forearm. In a forearm amputee the control of a hand prosthesis is traditionally depending on electromyography from the remaining forearm muscles. Technical improvements have made it possible to safely and routinely implant electrodes inside the muscles and record high-quality signals from individual muscles. In this study, we present a database of intramuscular EMG signals recorded with fine-wire electrodes alongside recordings of hand forces in an isometric setup and with the addition of spike-sorted metadata. Six forearm muscles were recorded from twelve able-bodied subjects and nine forearm muscles from two subjects. The fully automated recording protocol, based on command cues, comprised a variety of hand movements, including some requiring slowly increasing/decreasing force. The recorded data can be used to develop and test algorithms for control of a prosthetic hand. Assessment of the signals was done in both quantitative and qualitative manners.

A Novel HD-sEMG Preprocessing Method Integrating Muscle Activation Heterogeneity Analysis and Kurtosis-Guided Filtering for High-Accuracy Joint Force Estimation

This study proposes a novel preprocessing method integrating muscle activation heterogeneity analysis and kurtosis-guided filtering to realize high-accuracy surface electromyogr-aphy (sEMG)-based force estimation. A total of 10 subjects were recruited. Each subject performed isometric elbow flexion tasks at 20%, 40%, and 60% maximum voluntary contraction (MVC) target force levels, and the joint force and high-density sEMG (HD-sEMG) signals from biceps brachii and brachialis were collected synchronously. The force estimation model was built using three-order polynomial fitting technique. The input signal extraction of the force model, also named as the preprocessing of HD-sEMG signal, was carried out in the following procedures: first, HD-sEMG signals were decomposed by principal component analysis into principal components and weight vectors; second, the first several weight maps were segmented to obtain heterogeneity information by the Otsu and Moore-Neighbor tracing methods, and the principal component covering the most activated areas was selected; and last, a kurtosis-guided filter was designed to process the selected principal component to obtain the input signal. For the sake of comparison, the joint force estimation experiments based ON five preprocessing methods were conducted. The experimental results demonstrated that the proposed method obtained 52%, 53%, and 59% reduction in the mean root mean square difference at 20% MVC, 40% MVC, and 60% MVC force-level tasks, respectively, compared to the preprocessing method with the first principal component plus fixed parameter filtering. This proposed HD-sEMG pre-processing method has reliable neuromuscular electro-physiological foundation, and has good application value for realizing high-accuracy muscle/joint force estimation in the fields of rehabilitation engineering, sports biomechanics, and muscle disease diagnosis etc.

Heterogeneity Counts More than Power for HD-sEMG-Based Joint Force Estimation

The aim of the proposed work is to utilize the heterogeneity information, in input signal extraction, to improve the joint force estimation from high-density surface electromyography (HD-sEMG). For this purpose, joint force and HD-sEMG signals from biceps brachii and brachialis were collected synchronously during isometric elbow flexion. The input signal of the force model was obtained after the following procedures: first, HD-sEMG signals were decomposed by principal component analysis into principal components and weight vectors; second, the first several weight maps were segmented to obtain heterogeneity information by the Otsu and Moore-Neighbor tracing methods, and the principal component covering the most activated areas (maximum heterogeneity) was selected; and last, the selected principal component was low-pass filtered to obtain the input signal. The force model was built using a polynomial fitting technique. The conventional power-based input signal was compared with our obtained input signal. According to the obtained results, the proposed heterogeneity-based input signal can reduce the force estimation error significantly than the power-based input signal. The proposed heterogeneity-based input signal extraction methods had more neuromuscular control information and will be tested in more muscles and force tasks in future works.

Spatial Reorganization of Myoelectric Activities in Extensor Digitorum for Sustained Finger Force Production

The study aims to explore the spatial distribution of multi-tendinous muscle modulated by central nervous system (CNS) during sustained contraction. Nine subjects were recruited to trace constant target forces with right index finger extension. Surface electromyography (sEMG) of extensor digitorum (ED) were recorded with a 32-channel electrode array. Nine successive topographic maps (TM) were obtained. Pixel wise analysis was utilized to extract subtracted topographic maps (STM), which exhibited inhomogeneous distribution. STMs were characterized into hot, warm, and cool regions corresponding to higher, moderate, and lower change ranges, respectively. The relative normalized area (normalized to the first phase) of these regions demonstrated different changing trends as rising, plateauing, and falling over time, respectively. Moreover, the duration of these trends were found to be affected by force level. The rising/falling periods were longer at lower force levels, while the plateau can be achieved from the initial phase for higher force output (45% maximal voluntary contraction). The results suggested muscle activity reorganization in ED plays a role to maintain sustained contraction. Furthermore, the decreased dynamical regulation ability to spatial reorganization may be prone to induce fatigue. This finding implied that spatial reorganization of muscle activity as a regulation mechanism contribute to maintain constant force production.

A Fatigue Involved Modification Framework for Force Estimation in Fatiguing Contraction

To alleviate the negative impacts of muscle fatigue on a force estimation model, a modification framework taking use of fatigue index was put forward in this paper. Muscle force and surface electromyography were first collected using high-density electrode grid and dynamometer. Then, multi-step signal pre-processing and a nonnegative matrix factorization-based signal optimization were conducted, with fatigue indices being extracted in the same time. Next, a degree 4 polynomial fitting model was employed to undertake the training process, and the relationship between the generated model parameters and fatigue indices was built up. In the end, the parameter-index relationship was applied on different testing sets to complete fatigue-modified force estimation. Significant improvement was found in most testing cases across different sexes and ages. Relative decreases of 36.5%, 20.7%, and 20.4% in the percentage root mean square error were achieved by young males, young females, and elderly males. The proposed method can boost the performances of force estimation models, thereby contributing to the development of a variety of fields including biomechanical study, rehabilitation treatment, and prosthesis research.

Feasibility Study of Advanced Neural Networks Applied to sEMG-Based Force Estimation

To find out the feasibility of different neural networks in sEMG-based force estimation, in this paper, three types of networks, namely convolutional neural network (CNN), long short-term memory (LSTM) network and their combination (C-LSTM) were applied to predict muscle force generated in static isometric elbow flexion across three different circumstances (multi-subject, subject-dependent and subject-independent). Eight healthy men were recruited for the experiments, and the results demonstrated that all the three models were applicable for force estimation, and LSTM and C-LSTM achieved better performances. Even under subject-independent situation, they maintained mean RMSE% of as low as 9.07 ± 1.29 and 8.67 ± 1.14. CNN turned out to be a worse choice, yielding a mean RMSE% of 12.13 ± 1.98. To our knowledge, this work was the first to employ CNN, LSTM and C-LSTM in sEMG-based force estimation, and the results not only prove the strength of the proposed networks, but also pointed out a potential way of achieving high accuracy in real-time, subject-independent force estimation.

A Novel Interpretation of Sample Entropy in Surface Electromyographic Examination of Complex Neuromuscular Alternations in Subacute and Chronic Stroke

The objective of this paper was to develop sample entropy (SampEn) as a novel surface electromyogram (EMG) biomarker to quantitatively examine post-stroke neuromuscular alternations. The SampEn method was performed on surface EMG interference patterns recorded from biceps brachii muscles of nine healthy control subjects, fourteen subjects with subacute stroke, and eleven subjects with chronic stroke, respectively. Measurements were collected during isometric contractions of elbow flexion at different constant force levels. By producing diagnostic decisions for individual muscles, two categories of abnormalities in some paretic muscles were discriminated in terms of abnormally increased and decreased SampEn. The efficiency of the SampEn was demonstrated by its comparable performance with a previously reported clustering index (CI) method. Mixed SampEn (or CI) patterns were observed in paretic muscles of subjects with stroke indicating complex neuromuscular changes at work as a result of a hemispheric brain lesion. Although both categories of SampEn (or CI) abnormalities were observed in both subacute and chronic stages of stroke, the underlying processes contributing to the SampEn abnormalities might vary a lot in stroke stage. The SampEn abnormalities were also found in contralateral muscles of subjects with chronic stroke indicating the necessity of applying interventions to contralateral muscles during stroke rehabilitation. Our work not only presents a novel method for quantitative examination of neuromuscular changes, but also explains the neuropathological mechanisms of motor impairments and offers guidelines for a better design of effective rehabilitation protocols toward improved motor recovery.

HD-sEMG-based research on activation heterogeneity of skeletal muscles and the joint force estimation during elbow flexion

OBJECTIVE

To investigate the activation heterogeneity of skeletal muscles and realize the joint force estimation during the elbow flexion task.

APPROACH

When an isometric elbow flexion task was performed, high-density surface electromyography (HD-sEMG) signals from a [Formula: see text] grid covering the front and inside of the upper arm and the generated joint force were recorded synchronously. HD-sEMG signals were preprocessed and then decomposed into source signals corresponding to biceps brachhi (BB) and brachialis (BR) and their contribution vectors using a fast, independent component analysis (FastICA) algorithm. The activation heterogeneity of BB and BR was investigated from the activation level and activation region, initially. Then, the contribution combinations of two sources were classified into several major clusters using the K-means clustering method. Afterwards, input signals for force estimation were extracted from the major clusters corresponding to different combinations, and the polynomial fitting technique was adopted as the force estimation model. Finally, the force estimation results were obtained and the analysis around the force estimation performance using different input signals was conducted.

MAIN RESULTS

Ten subjects were recruited in this research. The experimental results demonstrated that it is feasible to analyze the activation heterogeneity of muscles from the activation level and activation region, and to select the appropriate region of the HD-sEMG grid for high performance force estimation. For the isometric elbow flexion task, joint force estimation accuracy could be improved when the input signal was extracted from the specific area where the contribution difference of BB and BR to the HD-sEMG signals were relatively small.

SIGNIFICANCE

The proposed framework provided a novel way to explore the relationship between muscle activation and the generating joint force, and could be extended to multiple noteworthy research fields such as myoelectric prostheses, sports biomechanics, and muscle disease diagnosis.

A SEMG-Force Estimation Framework Based on a Fast Orthogonal Search Method Coupled with Factorization Algorithms

A novel framework based on the fast orthogonal search (FOS) method coupled with factorization algorithms was proposed and implemented to realize high-accuracy muscle force estimation via surface electromyogram (SEMG). During static isometric elbow flexion, high-density SEMG (HD-SEMG) signals were recorded from upper arm muscles, and the generated elbow force was measured at the wrist. HD-SEMG signals were decomposed into time-invariant activation patterns and time-varying activation curves using three typical factorization algorithms including principal component analysis (PCA), independent component analysis (ICA), and nonnegative matrix factorization (NMF). The activation signal of the target muscle was obtained by summing the activation curves, and the FOS algorithm was used to create basis functions with activation signals and establish the force estimation model. Static isometric elbow flexion experiments at three target levels were performed on seven male subjects, and the force estimation performances were compared among three typical factorization algorithms as well as a conventional method for extracting the average signal envelope of all HD-SEMG channels (AVG-ENVLP method). The overall root mean square difference (RMSD) values between the measured forces and the estimated forces obtained by different methods were 11.79 ± 4.29% for AVG-ENVLP, 9.74 ± 3.77% for PCA, 9.59 ± 3.81% for ICA, and 9.51 ± 4.82% for NMF. The results demonstrated that, compared to the conventional AVG-ENVLP method, factorization algorithms could substantially improve the performance of force estimation. The FOS method coupled with factorization algorithms provides an effective way to estimate the combined force of multiple muscles and has potential value in the fields of sports biomechanics, gait analysis, prosthesis control strategy, and exoskeleton devices for assisted rehabilitation.