Sample entropy analysis of surface EMG for improved muscle activity onset detection against spurious background spikes

Clustering index analysis on EMG-Torque relation-based representation of complex neuromuscular changes after spinal cord injury

Spinal cord injury (SCI) resulting in complex neuromuscular pathology is not sufficiently well understood. To better quantify neuromuscular changes after SCI, this study uses a clustering index (CI) method for surface electromyography (sEMG) clustering representation to investigate the relation between sEMG and torque in SCI survivors. The sEMG signals were recorded from 13 subjects with SCI and 13 gender-age matched able-bodied subjects during isometric contraction of the biceps brachii muscle at different torque levels using a linear electrode array. Two torque representations, maximum voluntary contraction (MVC%) and absolute torque, were used. CI values were calculated for sEMG. Regression analyses were performed on CI values and torque levels of elbow flexion, revealing a strong linear relationship. The slopes of regressions between SCI survivors and control subjects were compared. The findings indicated that the range of distribution of CI values and slopes was greater in subjects with SCI than in control subjects (p < 0.05). The increase or decrease in slope was also observed at the individual level. This suggests that the CI and its sEMG clustering-torque relation may serve as valuable quantitative indicators for determining neuromuscular lesions after SCI, contributing to the development of effective rehabilitation strategies for improving motor performance.

Machine learning for hand pose classification from phasic and tonic EMG signals during bimanual activities in virtual reality

Myoelectric prostheses have recently shown significant promise for restoring hand function in individuals with upper limb loss or deficiencies, driven by advances in machine learning and increasingly accessible bioelectrical signal acquisition devices. Here, we first introduce and validate a novel experimental paradigm using a virtual reality headset equipped with hand-tracking capabilities to facilitate the recordings of synchronized EMG signals and hand pose estimation. Using both the phasic and tonic EMG components of data acquired through the proposed paradigm, we compare hand gesture classification pipelines based on standard signal processing features, convolutional neural networks, and covariance matrices with Riemannian geometry computed from raw or xDAWN-filtered EMG signals. We demonstrate the performance of the latter for gesture classification using EMG signals. We further hypothesize that introducing physiological knowledge in machine learning models will enhance their performances, leading to better myoelectric prosthesis control. We demonstrate the potential of this approach by using the neurophysiological integration of the "move command" to better separate the phasic and tonic components of the EMG signals, significantly improving the performance of sustained posture recognition. These results pave the way for the development of new cutting-edge machine learning techniques, likely refined by neurophysiology, that will further improve the decoding of real-time natural gestures and, ultimately, the control of myoelectric prostheses.

A systematic investigation of detectors for low signal-to-noise ratio EMG signals

Background

Active participation of stroke survivors during robot-assisted movement therapy is essential for sensorimotor recovery. Robot-assisted therapy contingent on movement intention is an effective way to encourage patients' active engagement. For severely impaired stroke patients with no residual movements, a surface electromyogram (EMG) has been shown to be a viable option for detecting movement intention. Although numerous algorithms for EMG detection exist, the detector with the highest accuracy and lowest latency for low signal-to-noise ratio (SNR) remains unknown.

Methods

This study, therefore, investigates the performance of 13 existing EMG detection algorithms on simulated low SNR (0dB and -3dB) EMG signals generated using three different EMG signal models: Gaussian, Laplacian, and biophysical model. The detector performance was quantified using the false positive rate (FPR), false negative rate (FNR), and detection latency. Any detector that consistently showed FPR and FNR of no more than 20%, and latency of no more than 50ms, was considered an appropriate detector for use in robot-assisted therapy.

Results

The results indicate that the Modified Hodges detector - a simplified version of the threshold-based Hodges detector introduced in the current study - was the most consistent detector across the different signal models and SNRs. It consistently performed for ~90% and ~40% of the tested trials for 0dB and -3dB SNR, respectively. The two statistical detectors (Gaussian and Laplacian Approximate Generalized Likelihood Ratio) and the Fuzzy Entropy detectors have a slightly lower performance than Modified Hodges.

Conclusions

Overall, the Modified Hodges, Gaussian and Laplacian Approximate Generalized Likelihood Ratio, and the Fuzzy Entropy detectors were identified as the potential candidates that warrant further investigation with real surface EMG data since they had consistent detection performance on low SNR EMG data.

Effect of muscular fatigue on the cumulative lumbar damage during repetitive lifting task: a comparative study of damage calculation methods

Several methods have been put forward to quantify cumulative loads; however, limited evidence exists as to the subsequent damages and the role of muscular fatigue. The present study assessed whether muscular fatigue could affect cumulative damage imposed on the L5-S1 joint. Trunk muscle electromyographic (EMG) activities and kinematics/kinetics of 18 healthy male individuals were evaluated during a simulated repetitive lifting task. A traditional EMG-assisted model of the lumbar spine was modified to account for the effect of erector spinae fatigue. L5-S1 compressive loads for each lifting cycle were estimated based on varying (i.e. actual), fatigue-modified, and constant Gain factors. The corresponding damages were integrated to calculate the cumulative damage. Moreover, the damage calculated for one lifting cycle was multiplied by the lifting frequency, as the traditional approach. Compressive loads and the damages obtained through the fatigue-modified model were predicted in close agreement with the actual values. Similarly, the difference between actual damages and those driven by the traditional approach was not statistically significant (p = 0.219). However, damages based on a constant Gain factor were significantly greater than those based on the actual (p = 0.012), fatigue-modified (p = 0.017), and traditional (p = 0.007) approaches.Practitioner summary: In this study, we managed to include the effect of muscular fatigue on cumulative lumbar damage calculations. Including the effect of muscular fatigue leads to an accurate estimation of cumulative damages while eliminating computational complexity. However, using the traditional approach also appears to provide acceptable estimates for ergonomic assessments.

Classification of hand and wrist movements via surface electromyogram using the random convolutional kernels transform

Prosthetic devices are vital for enhancing personal autonomy and the quality of life for amputees. However, the rejection rate for electric upper-limb prostheses remains high at around 30%, often due to issues like functionality, control, reliability, and cost. Thus, developing reliable, robust, and cost-effective human-machine interfaces is crucial for user acceptance. Machine learning algorithms using Surface Electromyography (sEMG) signal classification hold promise for natural prosthetic control. This study aims to enhance hand and wrist movement classification using sEMG signals, treated as time series data. A novel approach is employed, combining a variation of the Random Convolutional Kernel Transform (ROCKET) for feature extraction with a cross-validation ridge classifier. Traditionally, achieving high accuracy in time series classification required complex, computationally intensive methods. However, recent advances show that simple linear classifiers combined with ROCKET can achieve state-of-the-art accuracy with reduced computational complexity. The algorithm was tested on the UCI sEMG hand movement dataset, as well as on the Ninapro DB5 and DB7 datasets. We demonstrate how the proposed approach delivers high discrimination accuracy with minimal parameter tuning requirements, offering a promising solution to improve prosthetic control and user satisfaction.

Comparison of sEMG Onset Detection Methods for Occupational Exoskeletons on Extensive Close-to-Application Data

The design of human-machine interfaces of occupational exoskeletons is essential for their successful application, but at the same time demanding. In terms of information gain, biosensoric methods such as surface electromyography (sEMG) can help to achieve intuitive control of the device, for example by reduction of the inherent time latencies of a conventional, non-biosensoric, control scheme. To assess the reliability of sEMG onset detection under close to real-life circumstances, shoulder sEMG of 55 healthy test subjects was recorded during seated free arm lifting movements based on assembly tasks. Known algorithms for sEMG onset detection are reviewed and evaluated regarding application demands. A constant false alarm rate (CFAR) double-threshold detection algorithm was implemented and tested with different features. Feature selection was done by evaluation of signal-to-noise-ratio (SNR), onset sensitivity and precision, as well as timing error and deviation. Results of visual signal inspection by sEMG experts and kinematic signals were used as references. Overall, a CFAR algorithm with Teager-Kaiser-Energy-Operator (TKEO) as feature showed the best results with feature SNR = 14.48 dB, 91% sensitivity, 93% precision. In average, sEMG analysis hinted towards impending movements 215 ms before measurable kinematic changes.

Estimation of Gait Parameters for Adults with Surface Electromyogram Based on Machine Learning Models

Gait analysis has been studied over the last few decades as the best way to objectively assess the technical outcome of a procedure designed to improve gait. The treating physician can understand the type of gait problem, gain insight into the etiology, and find the best treatment with gait analysis. The gait parameters are the kinematics, including the temporal and spatial parameters, and lack the activity information of skeletal muscles. Thus, the gait analysis measures not only the three-dimensional temporal and spatial graphs of kinematics but also the surface electromyograms (sEMGs) of the lower limbs. Now, the shoe-worn GaitUp Physilog® wearable inertial sensors can easily measure the gait parameters when subjects are walking on the general ground. However, it cannot measure muscle activity. The aim of this study is to measure the gait parameters using the sEMGs of the lower limbs. A self-made wireless device was used to measure the sEMGs from the vastus lateralis and gastrocnemius muscles of the left and right feet. Twenty young female subjects with a skeletal muscle index (SMI) below 5.7 kg/m2 were recruited for this study and examined by the InBody 270 instrument. Four parameters of sEMG were used to estimate 23 gait parameters. They were measured using the GaitUp Physilog® wearable inertial sensors with three machine learning models, including random forest (RF), decision tree (DT), and XGBoost. The results show that 14 gait parameters could be well-estimated, and their correlation coefficients are above 0.800. This study signifies a step towards a more comprehensive analysis of gait with only sEMGs.

A new fractional fuzzy dispersion entropy and its application in muscle fatigue detection

Recently, fuzzy dispersion entropy (DispEn) has attracted much attention as a new nonlinear dynamics method that combines the advantages of both DispEn and fuzzy entropy. However, it suffers from limitation of insensitivity to dynamic changes. To solve this limitation, we proposed fractional fuzzy dispersion entropy (FFDispEn) based on DispEn, a novel fuzzy membership function and fractional calculus. The fuzzy membership function was defined based on the Euclidean distance between the embedding vector and dispersion pattern. Simulated signals generated by the one-dimensional (1D) logistic map were used to test the sensitivity of the proposed method to dynamic changes. Moreover, 29 subjects were recruited for an upper limb muscle fatigue experiment, during which surface electromyography (sEMG) signals of the biceps brachii muscle were recorded. Both simulated signals and sEMG signals were processed using a sliding window approach. Sample entropy (SampEn), DispEn and FFDispEn were separately used to calculate the complexity of each frame. The sensitivity of different algorithms to the muscle fatigue process was analyzed using fitting parameters through linear fitting of the complexity of each frame signal. The results showed that for simulated signals, the larger the fractional order q, the higher the sensitivity to dynamic changes. Moreover, DispEn performed poorly in the sensitivity to dynamic changes compared with FFDispEn. As for muscle fatigue detection, the FFDispEn value showed a clear declining tendency with a mean slope of -1.658 × 10-3 as muscle fatigue progresses; additionally, it was more sensitive to muscle fatigue compared with SampEn (slope: -0.4156 × 10-3) and DispEn (slope: -0.1675 × 10-3). The highest accuracy of 97.5% was achieved with the FFDispEn and support vector machine (SVM). This study provided a new useful nonlinear dynamic indicator for sEMG signal processing and muscle fatigue analysis. The proposed method may be useful for physiological and biomedical signal analysis.

Review of electromyography onset detection methods for real-time control of robotic exoskeletons

BACKGROUND

Electromyography (EMG) is a classical technique used to record electrical activity associated with muscle contraction and is widely applied in Biomechanics, Biomedical Engineering, Neuroscience and Rehabilitation Robotics. Determining muscle activation onset timing, which can be used to infer movement intention and trigger prostheses and robotic exoskeletons, is still a big challenge. The main goal of this paper was to perform a review of the state-of-the-art of EMG onset detection methods. Moreover, we compared the performance of the most commonly used methods on experimental EMG data.

METHODS

A total of 156 papers published until March 2022 were included in the review. The papers were analyzed in terms of application domain, pre-processing method and EMG onset detection method. The three most commonly used methods [Single (ST), Double (DT) and Adaptive Threshold (AT)] were applied offline on experimental intramuscular and surface EMG signals obtained during contractions of ankle and knee joint muscles.

RESULTS

Threshold-based methods are still the most commonly used to detect EMG onset. Compared to ST and AT, DT required more processing time and, therefore, increased onset timing detection, when applied on experimental data. The accuracy of these three methods was high (maximum error detection rate of 7.3%), demonstrating their ability to automatically detect the onset of muscle activity. Recently, other studies have tested different methods (especially Machine Learning based) to determine muscle activation onset offline, reporting promising results.

CONCLUSIONS

This study organized and classified the existing EMG onset detection methods to create consensus towards a possible standardized method for EMG onset detection, which would also allow more reproducibility across studies. The three most commonly used methods (ST, DT and AT) proved to be accurate, while ST and AT were faster in terms of EMG onset detection time, especially when applied on intramuscular EMG data. These are important features towards movement intention identification, especially in real-time applications. Machine Learning methods have received increased attention as an alternative to detect muscle activation onset. However, although several methods have shown their capability offline, more research is required to address their full potential towards real-time applications, namely to infer movement intention.

Tradeoffs of estimating reaction time with absolute and relative thresholds

Measuring the duration of cognitive processing with reaction time is fundamental to several subfields of psychology. Many methods exist for estimating movement initiation when measuring reaction time, but there is an incomplete understanding of their relative performance. The purpose of the present study was to identify and compare the tradeoffs of 19 estimates of movement initiation across two experiments. We focused our investigation on estimating movement initiation on each trial with filtered kinematic and kinetic data. Nine of the estimates involved absolute thresholds (e.g., acceleration 1000 back to 200 mm/s2, micro push-button switch), and the remaining ten estimates used relative thresholds (e.g., force extrapolation, 5% of maximum velocity). The criteria were the duration of reaction time, immunity to the movement amplitude, responsiveness to visual feedback during movement execution, reliability, and the number of manually corrected trials (efficacy). The three best overall estimates, in descending order, were yank extrapolation, force extrapolation, and acceleration 1000 to 200 mm/s2. The sensitive micro push-button switch, which was the simplest estimate, had a decent overall score, but it was a late estimate of movement initiation. The relative thresholds based on kinematics had the six worst overall scores. An issue with the relative kinematic thresholds was that they were biased by the movement amplitude. In summary, we recommend measuring reaction time on each trial with one of the three best overall estimates of movement initiation. Future research should continue to refine existing estimates while also exploring new ones.

The interaction between changes of muscle activation and cortical network dynamics during isometric elbow contraction: a sEMG and fNIRS study

Objective: The relationship between muscle activation during motor tasks and cerebral cortical activity remains poorly understood. The aim of this study was to investigate the correlation between brain network connectivity and the non-linear characteristics of muscle activation changes during different levels of isometric contractions. Methods: Twenty-one healthy subjects were recruited and were asked to perform isometric elbow contractions in both dominant and non-dominant sides. Blood oxygen concentrations in brain from functional Near-infrared Spectroscopy (fNIRS) and surface electromyography (sEMG) signals in the biceps brachii (BIC) and triceps brachii (TRI) muscles were recorded simultaneously and compared during 80% and 20% of maximum voluntary contraction (MVC). Functional connectivity, effective connectivity, and graph theory indicators were used to measure information interaction in brain activity during motor tasks. The non-linear characteristics of sEMG signals, fuzzy approximate entropy (fApEn), were used to evaluate the signal complexity changes in motor tasks. Pearson correlation analysis was used to examine the correlation between brain network characteristic values and sEMG parameters under different task conditions. Results: The effective connectivity between brain regions in motor tasks in dominant side was significantly higher than that in non-dominant side under different contractions (p < 0.05). The results of graph theory analysis showed that the clustering coefficient and node-local efficiency of the contralateral motor cortex were significantly varied under different contractions (p < 0.01). fApEn and co-contraction index (CCI) of sEMG under 80% MVC condition were significantly higher than that under 20% MVC condition (p < 0.05). There was a significant positive correlation between the fApEn and the blood oxygen value in the contralateral brain regions in both dominant or non-dominant sides (p < 0.001). The node-local efficiency of the contralateral motor cortex in the dominant side was positively correlated with the fApEn of the EMG signals (p < 0.05). Conclusion: In this study, the mapping relationship between brain network related indicators and non-linear characteristic of sEMG in different motor tasks was verified. These findings provide evidence for further exploration of the interaction between the brain activity and the execution of motor tasks, and the parameters might be useful in evaluation of rehabilitation intervention.

Effects of Law Enforcement Load Carriage Systems on Muscle Activity and Coordination during Walking: An Exploratory Study

Law enforcement officers (LEOs) commonly wear a duty belt (DB) or tactical vest (TV) and from prior findings, these forms of load carriage (LC) likely alter muscular activity. However, studies on the effects of LEO LC on muscular activity and coordination are limited in the current literature. The present study examined the effects of LEO load carriage on muscular activity and coordination. Twenty-four volunteers participated in the study (male = 13, age = 24.5 ± 6.0 years). Surface electromyography (sEMG) sensors were placed on the vastus lateralis, biceps femoris, multifidus, and lower rectus abdominus. Participants completed treadmill walking for two load carriage conditions (duty belt and tactical vest) and a control condition. Mean activity, sample entropy and Pearson correlation coefficients were computed for each muscle pair during the trials. The duty belt and tactical vest resulted in an increase in muscle activity in several muscles; however, no differences between the duty belt and tactical vest were found. Consistently across the conditions, the largest correlations were observed between the left and right multifidus (r = 0.33-0.68) and rectus abdominus muscles (0.34-0.55). There were statistically small effects (p < 0.05, η² = 0.031 to 0.076) of the LC on intermuscular coordination. No effect (p > 0.05) of the LC on sample entropy was found for any muscle. The findings indicate that LEO LC causes small differences in muscular activity and coordination during walking. Future research should incorporate heavier loads and longer durations.

Assessing Muscle Fatigue in Multiple Sclerosis using the Sample Entropy of Electromyographic Signals: A Proof of Concept

Background

Multiple sclerosis (MS) is a progressive and neurodegenerative disease of the central nervous system. Its symptoms vary greatly, which makes its diagnosis complex, expensive, and time-consuming. One of its most prevalent symptoms is muscle fatigue. It occurs in about 92% of patients with MS (PwMS) and is defined as a decrease in maximal strength or energy production in response to contractile activity. This article aims to compare the behavior of a healthy control (HC) with that of a patient with MS before and after muscle fatigue.

Methods

For this purpose, a static baropodometric test and a dynamic electromyographic analysis are performed to calculate the area of the stabilometric ellipse, the remitting MS (RMS) value, and the sample entropy (SampEn) of the signals, as a proof of concept to explore the feasibility of this test in the muscle fatigue quantitative analysis; in addition, the statistical analysis was realized to verify the results.

Results

According to the results, the ellipse area increased in the presence of muscle fatigue, indicating a decrease in postural stability. Likewise, the RMS value increased in the MS patient and decreased in the HC subject and the opposite behavior in the SampEn was observed in the presence of muscle fatigue.

Conclusion

Thus, this study demonstrates that SampEn is a viable parameter to estimate muscle fatigue in PwMS and other neuromuscular diseases.

Considerations for Applying Entropy Methods to Temporally Correlated Stochastic Datasets

The goal of this paper is to highlight considerations and provide recommendations for analytical issues that arise when applying entropy methods, specifically Sample Entropy (SampEn), to temporally correlated stochastic datasets, which are representative of a broad range of biomechanical and physiological variables. To simulate a variety of processes encountered in biomechanical applications, autoregressive fractionally integrated moving averaged (ARFIMA) models were used to produce temporally correlated data spanning the fractional Gaussian noise/fractional Brownian motion model. We then applied ARFIMA modeling and SampEn to the datasets to quantify the temporal correlations and regularity of the simulated datasets. We demonstrate the use of ARFIMA modeling for estimating temporal correlation properties and classifying stochastic datasets as stationary or nonstationary. We then leverage ARFIMA modeling to improve the effectiveness of data cleaning procedures and mitigate the influence of outliers on SampEn estimates. We also emphasize the limitations of SampEn to distinguish among stochastic datasets and suggest the use of complementary measures to better characterize the dynamics of biomechanical variables. Finally, we demonstrate that parameter normalization is not an effective procedure for increasing the interoperability of SampEn estimates, at least not for entirely stochastic datasets.

A Comprehensive Evaluation of Spine Kinematics, Kinetics, and Trunk Muscle Activities During Fatigue-Induced Repetitive Lifting

OBJECTIVE

Spine kinematics, kinetics, and trunk muscle activities were evaluated during different stages of a fatigue-induced symmetric lifting task over time.

BACKGROUND

Due to neuromuscular adaptations, postural behaviors of workers during lifting tasks are affected by fatigue. Comprehensive aspects of these adaptations remain to be investigated.

METHOD

Eighteen volunteers repeatedly lifted a box until perceived exhaustion. Body center of mass (CoM), trunk and box kinematics, and feet center of pressure (CoP) were estimated by a motion capture system and force-plate. Electromyographic (EMG) signals of trunk/abdominal muscles were assessed using linear and nonlinear approaches. The L5-S1 compressive force (Fc) was predicted via a biomechanical model. A two-way multivariate analysis of variance (MANOVA) was performed to examine the effects of five blocks of lifting cycle (C1 to C5) and lifting trial (T1 to T5), as independent variables, on kinematic, kinetic, and EMG-related measures.

RESULTS

Significant effects of lifting trial blocks were found for CoM and CoP shift in the anterior-posterior direction (respectively p < .001 and p = .014), trunk angle (p = .004), vertical box displacement (p < .001), and Fc (p = .005). EMG parameters indicated muscular fatigue with the extent of changes being muscle-specific.

CONCLUSION

Results emphasized variations in most kinematics/kinetics, and EMG-based indices, which further provided insight into the lifting behavior adaptations under dynamic fatiguing conditions.

APPLICATION

Movement and muscle-related variables, to a large extent, determine the magnitude of spinal loading, which is associated with low back pain.

Evaluation of Fast Sample Entropy Algorithms on FPGAs: From Performance to Energy Efficiency

Entropy is one of the most fundamental notions for understanding complexity. Among all the methods to calculate the entropy, sample entropy (SampEn) is a practical and common method to estimate time-series complexity. Unfortunately, SampEn is a time-consuming method growing in quadratic times with the number of elements, which makes this method unviable when processing large data series. In this work, we evaluate hardware SampEn architectures to offload computation weight, using improved SampEn algorithms and exploiting reconfigurable technologies, such as field-programmable gate arrays (FPGAs), a reconfigurable technology well-known for its high performance and power efficiency. In addition to the fundamental disclosed straightforward SampEn (SF) calculation method, this study evaluates optimized strategies, such as bucket-assist (BA) SampEn and lightweight SampEn based on BubbleSort (BS-LW) and MergeSort (MS-LW) on an embedded CPU, a high-performance CPU and on an FPGA using simulated data and real-world electrocardiograms (ECG) as input data. Irregular storage space and memory access of enhanced algorithms is also studied and estimated in this work. These fast SampEn algorithms are evaluated and profiled using metrics such as execution time, resource use, power and energy consumption based on input data length. Finally, although the implementation of fast SampEn is not significantly faster than versions running on a high-performance CPU, FPGA implementations consume one or two orders of magnitude less energy than a high-performance CPU.

Application of Surface Electromyography in Exercise Fatigue: A Review

Exercise fatigue is a common physiological phenomenon in human activities. The occurrence of exercise fatigue can reduce human power output and exercise performance, and increased the risk of sports injuries. As physiological signals that are closely related to human activities, surface electromyography (sEMG) signals have been widely used in exercise fatigue assessment. Great advances have been made in the measurement and interpretation of electromyographic signals recorded on surfaces. It is a practical way to assess exercise fatigue with the use of electromyographic features. With the development of machine learning, the application of sEMG signals in human evaluation has been developed. In this article, we focused on sEMG signal processing, feature extraction, and classification in exercise fatigue. sEMG based multisource information fusion for exercise fatigue was also introduced. Finally, the development trend of exercise fatigue detection is prospected.

Comparison of different algorithms based on TKEO for EMG change point detection

OBJECTIVE

A significant challenge in surface electromyography (EMG) is the accurate identification of onset and offset of muscle activation while maintaining high real-time performance. Teager-Kaiser energy operator (TKEO) is widely used in muscle activity monitoring systems because of its computational simplicity and strong real-time performance. However, in contrast to TKEO ontology, few studies have examined how well the energy operator variants from multiple fields perform in conditioning EMG signals. This paper aims to investigate the role of the energy operator and its variants in EMG change point detection by a threshold detector.

APPROACH

To compare the stability and accuracy of TKEO and its variants for EMG change point detection, the EMG data of extensor carpi radialis longus and flexor carpi radialis were acquired from twenty participants operating a controller under normal and disturbed conditions, and EMG change point detection was performed by four energy operators and their rectified versions.

MAIN RESULTS

Based on the "standard" change points collected by the controller, the detection results were evaluated by three evaluation indexes: detection rate, F1 Score, and accuracy. The experimental results show that the multiresolution energy operator (MTEO) and the TKEO with rectified (abs-TKEO) are more suitable for EMG change point detection.

SIGNIFICANCE

This paper compared the effect of the energy operator and its variants on a threshold-based EMG change point detector. The experimental results in this paper can provide a reference for the selection of EMG signal conditioning methods to improve the detection performance of the EMG change point detector.

Corticomuscular integrated representation of voluntary motor effort in robotic control for wrist-hand rehabilitation after stroke

OBJECTIVE

The central-to-peripheral voluntary motor effort (VME) in physical practice of the paretic limb is a dominant force for driving functional neuroplasticity on motor restoration post-stroke. However, current rehabilitation robots isolated the central and peripheral involvements in the control design, resulting in limited rehabilitation effectiveness. The purpose of this study was to design a corticomuscular coherence (CMC) and electromyography (EMG)-driven (CMC-EMG-driven) system with central-and-peripheral integrated representation of VME for wrist-hand rehabilitation after stroke.

APPROACH

The CMC-EMG-driven control was developed in a neuromuscular electrical stimulation (NMES)-robot system, i.e., CMC-EMG-driven NMES-robot system, to instruct and assist the wrist-hand extension and flexion in persons after stroke. A pilot single-group trial of 20 training sessions was conducted with the developed system to assess the feasibility for wrist-hand practice on the chronic strokes (n=16). The rehabilitation effectiveness was evaluated through clinical assessments, CMC, and EMG activation levels.

MAIN RESULTS

The trigger success rate and laterality index (LI) of CMC were significantly increased in wrist-hand extension across training sessions (p<0.05). After the training, significant improvements in the target wrist-hand joints and suppressed compensation from the proximal shoulder-elbow joints were observed through the clinical scores and EMG activation levels (p<0.05). The central-to-peripheral VME distribution across upper extremity (UE) muscles was also significantly improved, as revealed by the CMC values (p<0.05).

SIGNIFICANCE

Precise wrist-hand rehabilitation was achieved by the developed system, presenting suppressed cortical and muscular compensation from the contralesional hemisphere and the proximal UE, and improved distribution of the central-and-peripheral VME on UE muscles.

Pattern Classification of Hand Movement Tremor in MS Patients with DBS ON and OFF

Background:

Hand tremor is one of the consequences of MS disease degrading quality of patient’s life. Recently DBS is used as a prominent treatment to reduce this effect. Evaluation of this approach has significant importance because of the prevalence rate of disease.

Objective:

The purpose of this study was the nonlinear analysis of tremor signal in order to evaluate the quantitative effect of DBS on reducing MS tremor and differentiating between them using pattern recognition algorithms.

Material and Methods:

In this analytical study, nine features were extracted from the tremor signal. Through statistical analysis, the significance level of each feature was examined. Finally, tremor signals were categorized by SVM, weighted KNN and NN classifiers. The performance of methods was compared with an ROC graph.

Results:

The results have demonstrated that dominant frequency, maximum amplitude and energy of the first IMF, deviation of the direct path, sample entropy and fuzzy entropy have the potential to create a significant difference between the tremor signals. The classification accuracy rate of tremor signals in three groups for Weighted KNN, NN and SVM with Gaussian and Quadratic kernels resulted in 95.1%, 93.2%, 91.3% and 88.3%, respectively.

Conclusion:

Generally, nonlinear and nonstationary analyses have a high potential for a quantitative and objective measure of MS tremor. Weighted KNN has shown the best performance of classification with the accuracy of more than 95%. It has been indicated that DBS has a positive influence on reducing the MS tremor. Therefore, DBS can be used in the objective improvement of tremor in MS patients.

Onset and Offset Detection of Respiratory EMG Data Based on Energy Operator Signal

Onset and offset detection of electromyography (EMG) data is an important step in respiratory muscle coordination assessment. Impaired respiratory coordination can indicate breathing disorders and lung diseases. In this paper, we present an algorithm for onset and offset timing detection of real-world EMG signals from respiratory muscles, which are contaminated with electrocardiogram (ECG) artifacts. The algorithm is based on the Energy Operator signal, has a low computational cost, and includes a filtering procedure to remove ECG artifacts from EMG. Analysis of EMG signals from 2 respiratory muscles of 5 participants' data shows high agreement between the algorithm and manual method with a mean difference between two methods of 0.0407 seconds.

Cardinality and Short-Term Memory Concepts based Novel Feature Extraction for Myoelectric Pattern Recognition

The quality of the extracted traditional hand-crafted Electromyogram (EMG) features has been recently identified in the literature as a limiting factor prohibiting the translation from laboratory to clinical settings. To address this limitation, a shift of focus from traditional feature extraction methods to deep learning models was witnessed, as the latter can learn the best feature representation for the task at hand. However, while deep learning models achieve promising results based on raw EMG data, their clinical implementation is often challenged due to their significantly high computational costs (significantly large number of generated models' parameters and a huge amount of data needed for training). This paper is focused on combining the simplicity and low computational characteristics of traditional feature extraction with the memory concepts from Long Short-Term Memory (LSTM) models to efficiently extract the spatial-temporal dynamics of the EMG signals. The novelty of the proposed method can be summarized in a) the memory concept leveraged from deep learning structures, capturing short-term temporal dependencies of the EMG signals, b) the use of cardinality to generate logical combinations of spatially distinct EMG signals and as a feature extraction method and 3) low computational costs and the enhanced classification performance. The performance of the proposed method is validated using three EMG databases collected with 1) laboratory hardware (9 transradial amputees and 17 intact-limbed), and 2) wearables (22 intact-limed using two wearable consumer armbands). In comparison to several other well-known methods from the literature, the proposed method shows significantly enhanced myoelectric pattern recognition performance, with accuracies reaching up to 99%.

Long short-term memory (LSTM) recurrent neural network for muscle activity detection

BACKGROUND

The accurate temporal analysis of muscle activation is of great interest in many research areas, spanning from neurorobotic systems to the assessment of altered locomotion patterns in orthopedic and neurological patients and the monitoring of their motor rehabilitation. The performance of the existing muscle activity detectors is strongly affected by both the SNR of the surface electromyography (sEMG) signals and the set of features used to detect the activation intervals. This work aims at introducing and validating a powerful approach to detect muscle activation intervals from sEMG signals, based on long short-term memory (LSTM) recurrent neural networks.

METHODS

First, the applicability of the proposed LSTM-based muscle activity detector (LSTM-MAD) is studied through simulated sEMG signals, comparing the LSTM-MAD performance against other two widely used approaches, i.e., the standard approach based on Teager-Kaiser Energy Operator (TKEO) and the traditional approach, used in clinical gait analysis, based on a double-threshold statistical detector (Stat). Second, the effect of the Signal-to-Noise Ratio (SNR) on the performance of the LSTM-MAD is assessed considering simulated signals with nine different SNR values. Finally, the newly introduced approach is validated on real sEMG signals, acquired during both physiological and pathological gait. Electromyography recordings from a total of 20 subjects (8 healthy individuals, 6 orthopedic patients, and 6 neurological patients) were included in the analysis.

RESULTS

The proposed algorithm overcomes the main limitations of the other tested approaches and it works directly on sEMG signals, without the need for background-noise and SNR estimation (as in Stat). Results demonstrate that LSTM-MAD outperforms the other approaches, revealing higher values of F1-score (F1-score > 0.91) and Jaccard similarity index (Jaccard > 0.85), and lower values of onset/offset bias (average absolute bias < 6 ms), both on simulated and real sEMG signals. Moreover, the advantages of using the LSTM-MAD algorithm are particularly evident for signals featuring a low to medium SNR.

CONCLUSIONS

The presented approach LSTM-MAD revealed excellent performances against TKEO and Stat. The validation carried out both on simulated and real signals, considering normal as well as pathological motor function during locomotion, demonstrated that it can be considered a powerful tool in the accurate and effective recognition/distinction of muscle activity from background noise in sEMG signals.

Information-based analysis of the relation between human muscle reaction and walking path

BACKGROUND

Walking is one of the important actions of the human body. For this purpose, the human brain communicates with leg muscles through the nervous system. Based on the walking path, leg muscles act differently. Therefore, there should be a relation between the activity of leg muscles and the path of movement.

OBJECTIVE

In order to address this issue, we analyzed how leg muscle activity is related to the variations of the path of movement.

METHOD

Since the electromyography (EMG) signal is a feature of muscle activity and the movement path has complex structures, we used entropy analysis in order to link their structures. The Shannon entropy of EMG signal and walking path are computed to relate their information content.

RESULTS

Based on the obtained results, walking on a path with greater information content causes greater information content in the EMG signal which is supported by statistical analysis results. This allowed us to analyze the relation between muscle activity and walking path.

CONCLUSION

The method of analysis employed in this research can be applied to investigate the relation between brain or heart reactions and walking path.

Human motor function estimation based on EMG signal fractal dimension standard deviation

Wearable robots must adjust the assist mode/intensity according to human motion during the motion assistance process. By decoding the surface electromyography (sEMG) signal, the standard deviation of the fractal dimension is used as a characteristic index of muscle contraction-relaxation ability, and explore the feasibility of using the standard deviation of the fractal dimension to estimate the human motor function and thus provide a basis for decision-making for the flexible control of wearable robots. First, the sEMG signals of several subjects with different motor functions were collected and their time-domain and frequency-domain features were extracted. The experimental results for one hour of walking showed that the time-domain and frequency-domain feature values increased with muscle fatigue. The trend has little to do with the inherent motor function of the human body; Second, due to the strong nonlinearity, time-varying, and strong complexity of the sEMG signal, the fractal dimension nonlinear method is used to characterize the complexity of the EMG signal that is closely related to muscle function. Besides, theoretical and experimental studies have been conducted to clarify the feasibility of the complexity of fractal dimension representation and to provide theoretical support for the further use of the standard deviation of fractal dimension to estimate human motor function. The experimental results of continuous walking for one hour show that, macroscopically, the fractal dimension of each muscle of the individual subject does not change significantly with walking time, which shows that the fractal dimension has nothing to do with exercise time and muscle fatigue; On the microscopic level, the value of the fractal dimension changes when the subject’s muscles contract and relax. Subjects with strong motor function have smaller fractal dimensions when their muscles contract than subjects with weaker motor function, and the opposite happens when their muscles relax, and it can be seen that there is a positive correlation between the difference in the fractal dimension during muscle contraction and relaxation and the muscle contraction-relaxation ability and the human body’s inherent motor function. The test results verify the feasibility of using the standard deviation of fractal dimension to estimate the intrinsic motor function of the human body.

Information-based analysis of the relationship between brain and facial muscle activities in response to static visual stimuli

BACKGROUND: Human facial muscles react differently to different visual stimuli. It is known that the human brain controls and regulates the activity of the muscles. OBJECTIVE: In this research, for the first time, we investigate how facial muscle reaction is related to the reaction of the human brain. METHODS: Since both electromyography (EMG) and electroencephalography (EEG) signals, as the features of muscle and brain activities, contain information, we benefited from the information theory and computed the Shannon entropy of EMG and EEG signals when subjects were exposed to different static visual stimuli with different Shannon entropies (information content). RESULTS: Based on the obtained results, the variations of the information content of the EMG signal are related to the variations of the information content of the EEG signal and the visual stimuli. Statistical analysis also supported the results indicating that the visual stimuli with greater information content have a greater effect on the variation of the information content of both EEG and EMG signals. CONCLUSION: This investigation can be further continued to analyze the relationship between facial muscle and brain reactions in case of other types of stimuli.

Alterations in Quadriceps Neurologic Complexity After Anterior Cruciate Ligament Reconstruction

CONTEXT

Traditionally, quadriceps activation failure after anterior cruciate ligament reconstruction (ACLR) is estimated using discrete isometric torque values, providing only a snapshot of neuromuscular function. Sample entropy (SampEn) is a mathematical technique that can measure neurologic complexity during the entirety of contraction, elucidating qualities of neuromuscular control not previously captured.

OBJECTIVE

To apply SampEn analyses to quadriceps electromyographic activity in order to more comprehensively characterize neuromuscular deficits after ACLR.

DESIGN

Cross-sectional.

SETTING

Laboratory.

PARTICIPANTS

ACLR: n = 18; controls: n = 24.

INTERVENTIONS

All participants underwent synchronized unilateral quadriceps isometric strength, activation, and electromyography testing during a superimposed electrical stimulus.

MAIN OUTCOME MEASURES

Group differences in strength, activation, and SampEn were evaluated with t tests. Associations between SampEn and quadriceps function were evaluated with Pearson product-moment correlations and hierarchical linear regressions.

RESULTS

Vastus medialis SampEn was significantly reduced after ACLR compared with controls (P = .032). Vastus medialis and vastus lateralis SampEn predicted significant variance in activation after ACLR (r2 = .444; P = .003).

CONCLUSIONS

Loss of neurologic complexity correlates with worse activation after ACLR, particularly in the vastus medialis. Electromyographic SampEn is capable of detecting underlying patterns of variability that are associated with the loss of complexity between key neurophysiologic events after ACLR.

Effects of Tau and Sampling Frequency on the Regularity Analysis of ECG and EEG Signals Using ApEn and SampEn Entropy Estimators

Electrocardiography (ECG) and electroencephalography (EEG) signals provide clinical information relevant to determine a patient’s health status. The nonlinear analysis of ECG and EEG signals allows for discovering characteristics that could not be found with traditional methods based on amplitude and frequency. Approximate entropy (ApEn) and sampling entropy (SampEn) are nonlinear data analysis algorithms that measure the data’s regularity, and these are used to classify different electrophysiological signals as normal or pathological. Entropy calculation requires setting the parameters r (tolerance threshold), m (immersion dimension), and τ (time delay), with the last one being related to how the time series is downsampled. In this study, we showed the dependence of ApEn and SampEn on different values of τ, for ECG and EEG signals with different sampling frequencies (F_s), extracted from a digital repository. We considered four values of F_s (128, 256, 384, and 512 Hz for the ECG signals, and 160, 320, 480, and 640 Hz for the EEG signals) and five values of τ (from 1 to 5). We performed parametric and nonparametric statistical tests to confirm that the groups of normal and pathological ECG and EEG signals were significantly different (p < 0.05) for each F and τ value. The separation between the entropy values of regular and irregular signals was variable, demonstrating the dependence of ApEn and SampEn with F_s and τ. For ECG signals, the separation between the conditions was more robust when using SampEn, the lowest value of F_s, and τ larger than 1. For EEG signals, the separation between the conditions was more robust when using SampEn with large values of F_s and τ larger than 1. Therefore, adjusting τ may be convenient for signals that were acquired with different F_s to ensure a reliable clinical classification. Furthermore, it is useful to set τ to values larger than 1 to reduce the computational cost.

A Statistical Approach for the Assessment of Muscle Activation Patterns during Gait in Parkinson’s Disease

Recently, the statistical analysis of muscle activation patterns highlighted that not only one, but several activation patterns can be identified in the gait of healthy adults, with different occurrence. Although its potential, the application of this approach in pathological populations is still limited and specific implementation issues need to be addressed. This study aims at applying a statistical approach to analyze muscle activation patterns of gait in Parkinson’s Disease, integrating gait symmetry and co-activation. Surface electromyographic signal of tibialis anterior and gastrocnemius medialis were recorded during a 6-min walking test in 20 patients. Symmetry between right and left stride time series was verified, different activation patterns identified, and their occurrence (number and timing) quantified, as well as the co-activation of antagonist muscles. Gastrocnemius medialis presented five activation patterns (mean occurrence ranging from 2% to 43%) showing, with respect to healthy adults, the presence of a first shorted and delayed activation (between flat foot contact and push off, and in the final swing) and highlighting a new second region of anticipated activation (during early/mid swing). Tibialis anterior presented five activation patterns (mean occurrence ranging from 3% to 40%) highlighting absent or delayed activity at the beginning of the gait cycle, and generally shorter and anticipated activations during the swing phase with respect to healthy adults. Three regions of co-contraction were identified: from heel strike to mid-stance, from the pre- to initial swing, and during late swing. This study provided a novel insight in the analysis of muscle activation patterns in Parkinson’s Disease patients with respect to the literature, where unique, at times conflicting, average patterns were reported. The proposed integrated methodology is meant to be generalized for the analysis of muscle activation patterns in pathologic subjects.

Beta-Range Corticomuscular Coupling Reflects Asymmetries in Hand Movement

Hand movement in humans is verified as asymmetries and lateralization, and two hemispheres make some distinct but complementary contributions in the control of hand movement. However, little research has been done on whether the information transfer of the motor system is different between left and right hand movement. Considering the importance of functional corticomuscular coupling (FCMC) between the motor cortex and contralateral muscle in movement assessment, this study aimed to explore the differences between left and right hand by investigating the interaction between muscle and brain activity. Here, we applied the transfer spectral entropy (TSE) algorithm to quantize the connection between electroencephalogram (EEG) over the brain scalp and electromyogram (EMG) from extensor digitorum (ED) and flexor digitorum superficialis (FDS) muscles recorded simultaneously during a gripping task. Eight healthy subjects were enrolled in this study. Results showed that left hand yielded narrower and lower beta synchronization compared to the right. Further analysis indicated coupling strength in EEG-EMG(FDS) combination was higher at beta band than that in EEG-EMG(ED) combination, and exhibited distinct differences between descending (EEG to EMG direction) and ascending (EMG to EEG direction) direction. This study presents the distinctions of beta-range FCMC between left and right hand, and confirms the importance of beta synchronization in understanding the mechanism of motor stability control. The cortex-muscle FCMC might be used as an evaluation approach to explore the difference between left and right movement system.

Analysis of Influence of Segmentation, Features, and Classification in sEMG Processing: A Case Study of Recognition of Brazilian Sign Language Alphabet

Sign Language recognition systems aid communication among deaf people, hearing impaired people, and speakers. One of the types of signals that has seen increased studies and that can be used as input for these systems is surface electromyography (sEMG). This work presents the recognition of a set of alphabet gestures from Brazilian Sign Language (Libras) using sEMG acquired from an armband. Only sEMG signals were used as input. Signals from 12 subjects were acquired using a MyoTM armband for the 26 signs of the Libras alphabet. Additionally, as the sEMG has several signal processing parameters, the influence of segmentation, feature extraction, and classification was considered at each step of the pattern recognition. In segmentation, window length and the presence of four levels of overlap rates were analyzed, as well as the contribution of each feature, the literature feature sets, and new feature sets proposed for different classifiers. We found that the overlap rate had a high influence on this task. Accuracies in the order of 99% were achieved for the following factors: segments of 1.75 s with a 12.5% overlap rate; the proposed set of four features; and random forest (RF) classifiers.

Investigating the performance of an amplitude-independent algorithm for detecting the hand muscle activity of stroke survivors

To make robotic hand devices controlled by surface electromyography (sEMG) signals feasible and practical tools for assisting patients with hand impairments, the problems that prevent these devices from being widely used have to be overcome. The most significant problem is the involuntary amplitude variation of the sEMG signals due to the movement of electrodes during forearm motion. Moreover, for patients who have had a stroke or another neurological disease, the muscle activity of the impaired hand is weak and has a low signal-to-noise ratio (SNR). Thus, muscle activity detection methods intended for controlling robotic hand devices should not depend mainly on the amplitude characteristics of the sEMG signal in the detection process, and they need to be more reliable for sEMG signals that have a low SNR. Since amplitude-independent muscle activity detection methods meet these requirements, this paper investigates the performance of such a method on people who have had a stroke in terms of the detection of weak muscle activity and resistance to false alarms caused by the involuntary amplitude variation of sEMG signals; these two parameters are very important for achieving the reliable control of robotic hand devices intended for people with disabilities. A comparison between the performance of an amplitude-independent muscle activity detection algorithm and three amplitude-dependent algorithms was conducted by using sEMG signals recorded from six hemiparesis stroke survivors and from six healthy subjects. The results showed that the amplitude-independent algorithm performed better in terms of detecting weak muscle activity and resisting false alarms.

Complexity Analysis of Surface Electromyography for Assessing the Myoelectric Manifestation of Muscle Fatigue: A Review

The surface electromyography (sEMG) records the electrical activity of muscle fibers during contraction: one of its uses is to assess changes taking place within muscles in the course of a fatiguing contraction to provide insights into our understanding of muscle fatigue in training protocols and rehabilitation medicine. Until recently, these myoelectric manifestations of muscle fatigue (MMF) have been assessed essentially by linear sEMG analyses. However, sEMG shows a complex behavior, due to many concurrent factors. Therefore, in the last years, complexity-based methods have been tentatively applied to the sEMG signal to better individuate the MMF onset during sustained contractions. In this review, after describing concisely the traditional linear methods employed to assess MMF we present the complexity methods used for sEMG analysis based on an extensive literature search. We show that some of these indices, like those derived from recurrence plots, from entropy or fractal analysis, can detect MMF efficiently. However, we also show that more work remains to be done to compare the complexity indices in terms of reliability and sensibility; to optimize the choice of embedding dimension, time delay and threshold distance in reconstructing the phase space; and to elucidate the relationship between complexity estimators and the physiologic phenomena underlying the onset of MMF in exercising muscles.

Meditation Increases the Entropy of Brain Oscillatory Activity

We address the hypothesis that the entropy of neural dynamics indexes the intensity and quality of conscious content. Previous work established that serotonergic psychedelics can have a dysregulating effect on brain activity, leading to subjective effects that present a considerable overlap with the phenomenology of certain meditative states. Here we propose that the prolonged practice of meditation results in endogenous increased entropy of brain oscillatory activity. We estimated the entropy of band-specific oscillations during the meditative state of traditions classified as 'focused attention' (Himalayan Yoga), 'open monitoring' (Vipassana), and 'open awareness' (Isha Shoonya Yoga). Among all traditions, Vipassana resulted in the highest entropy increases, predominantly in the alpha and low/high gamma bands. In agreement with previous studies, all meditation traditions increased the global coherence in the gamma band, but also stabilized gamma-range dynamics by lowering the metastability. Finally, machine learning classifiers could successfully generalize between certain pairs of meditation traditions based on the scalp distribution of gamma band entropies. Our results extend previous findings on the spectral changes observed during meditation, showing how long-term practice can lead to the capacity for achieving brain states of high entropy. This constitutes an example of an endogenous, self-induced high entropy state.

Hand Gesture Recognition Using Compact CNN Via Surface Electromyography Signals

By training the deep neural network model, the hidden features in Surface Electromyography(sEMG) signals can be extracted. The motion intention of the human can be predicted by analysis of sEMG. However, the models recently proposed by researchers often have a large number of parameters. Therefore, we designed a compact Convolution Neural Network (CNN) model, which not only improves the classification accuracy but also reduces the number of parameters in the model. Our proposed model was validated on the Ninapro DB5 Dataset and the Myo Dataset. The classification accuracy of gesture recognition achieved good results.

A new detection method for EMG activity monitoring

This paper introduces a new approach for electromyography (EMG) activity monitoring based on an improved version of the adaptive linear energy detector (ALED), a widely used technique in voice activity detection. More precisely, we propose a modified ALED technique (named M-ALED) to improve the method's robustness with respect to noise. To achieve this objective, M-ALED relies on the Teager-Kaiser operator for signal pre-conditioning to increase the SNR and uses the order statistics to gain robustness against the signal's impulsiveness. We propose again to exploit the order statistics for the initial signal baseline estimation to deal with the cases where such information is unavailable. Finally, since M-ALED detects the signal's activity at the frame level, we propose in a second stage to refine this detection (at the sample level) by using a constant false alarm rate (CFAR) approach leading to the fine M-ALED (FM-ALED) solution. The performance of FM-ALED is assessed via real and synthetic EMG signal recordings and the obtained results highlight its effectiveness as compared with the state-of-the-art methods (it reduces the mean error probability by a factor close to 2).

Exploration of Feature Extraction Methods and Dimension for sEMG Signal Classification

It is necessary to complete the two parts of gesture recognition and wireless remote control to realize the gesture control of the automatic pruning machine. To realize gesture recognition, in this paper, we have carried out the research of gesture recognition technology based on surface electromyography signal, and discussed the influence of different numbers and different gesture combinations on the optimal size. We have calculated the 630-dimensional eigenvector from the benchmark scientific database of sEMG signals and extracted the features using principal component analysis (PCA). Discriminant analysis (DA) has been used to compare the processing effects of each feature extraction method. The experimental results have shown that the recognition rate of four gestures can reach 100.0%, the recognition rate of six gestures can reach 98.29%, and the optimal size is 516~523 dimensions. This study lays a foundation for the follow-up work of the pruning machine gesture control, and p rovides a compelling new way to promote the creative and human computer interaction process of forestry machinery.

sEMG-Based Trunk Compensation Detection in Rehabilitation Training

Stroke patients often use trunk to compensate for impaired upper limb motor function during upper limb rehabilitation training, which results in a reduced rehabilitation training effect. Detecting trunk compensations can improve the effect of rehabilitation training. This study investigates the feasibility of a surface electromyography-based trunk compensation detection (sEMG-bTCD) method. Five healthy subjects and nine stroke subjects with cognitive and comprehension skills were recruited to participate in the experiments. The sEMG signals from nine superficial trunk muscles were collected during three rehabilitation training tasks (reach-forward-back, reach-side-to-side, and reach-up-to-down motions) without compensation and with three common trunk compensations [lean-forward (LF), trunk rotation (TR), and shoulder elevation (SE)]. Preprocessing like filtering, active segment detection was performed and five time domain features (root mean square, variance, mean absolute value (MAV), waveform length, and the fourth order autoregressive model coefficient) were extracted from the collected sEMG signals. Excellent TCD performance was achieved in healthy participants by using support vector machine (SVM) classifier (LF: accuracy = 94.0%, AUC = 0.97, F1 = 0.94; TR: accuracy = 95.8%, AUC = 0.99, F1 = 0.96; SE: accuracy = 100.0%, AUC = 1.00, F1 = 1.00). By using SVM classifier, TCD performance in stroke participants was also obtained (LF: accuracy = 74.8%, AUC = 0.90, F1 = 0.73; TR: accuracy = 67.1%, AUC = 0.85, F1 = 0.71; SE: accuracy = 91.3%, AUC = 0.98, F1 = 0.90). Compared with the methods based on cameras or inertial sensors, better detection performance was obtained in both healthy and stroke participants. The results demonstrated the feasibility of the sEMG-bTCD method, and it helps to prompt the stroke patients to correct their incorrect posture, thereby improving the effectiveness of rehabilitation training.

Detection of stretch reflex onset based on empirical mode decomposition and modified sample entropy

Background

Accurate spasticity assessment provides an objective evaluation index for the rehabilitation treatment of patients with spasticity, and the key is detecting stretch reflex onset. The surface electromyogram of patients with spasticity is prone to false peaks, and its data length is unstable. These conditions decrease signal differences before and after stretch reflex onset. Therefore, a method for detecting stretch reflex onset based on empirical mode decomposition denoising and modified sample entropy recognition is proposed in this study.

Results

The empirical mode decomposition algorithm is better than the wavelet threshold algorithm in denoising surface electromyogram signal. Without adding Gaussian white noise to the electromyogram signal, the stretch reflex onset recognition rate of the electromyogram signal before and after empirical mode decomposition denoising was increased by 56%. In particular, the recognition rate of stretch reflex onset under the optimal parameter of the modified sample entropy can reach up to 100% and the average recognition rate is 93%.

Conclusions

The empirical mode decomposition algorithm can eliminate the baseline activity of the surface electromyogram signal before stretch reflex onset and effectively remove noise from the signal. The identification of stretch reflex onset using combined empirical mode decomposition and modified sample entropy is better than that via modified sample entropy alone, and stretch reflex onset can be accurately determined.

Exploration of Chinese Sign Language Recognition Using Wearable Sensors Based on Deep Belief Net

In this paper, deep belief net (DBN) was applied into the field of wearable-sensor based Chinese sign language (CSL) recognition. Eight subjects were involved in the study, and all of the subjects finished a five-day experiment performing CSL on a target word set consisting of 150 CSL subwords. During the experiment, surface electromyography (sEMG), accelerometer (ACC), and gyroscope (GYRO) signals were collected from the participants. In order to obtain the optimal structure of the network, three different sensor fusion strategies, including data-level fusion, feature-level fusion, and decision-level fusion, were explored. In addition, for the feature-level fusion strategy, two different feature sources, which are hand-crafted features and network generated features, and two different network structures, which are fully-connected net and DBN, were also compared. The result showed that feature level fusion could achieve the best recognition accuracy among the three fusion strategies, and feature-level fusion with network generated features and DBN could achieve the best recognition accuracy. The best recognition accuracy realized in this study was 95.1% for the user-dependent test and 88.2% for the user-independent test. The significance of the study is that it applied the deep learning method into the field of wearable sensors-based CSL recognition, and according to our knowledge it's the first study comparing human engineered features with the network generated features in the correspondent field. The results from the study shed lights on the method of using network-generated features during sensor fusion and CSL recognition.

Interpretation of Entropy Algorithms in the Context of Biomedical Signal Analysis and Their Application to EEG Analysis in Epilepsy

Biomedical signals are measurable time series that describe a physiological state of a biological system. Entropy algorithms have been previously used to quantify the complexity of biomedical signals, but there is a need to understand the relationship of entropy to signal processing concepts. In this study, ten synthetic signals that represent widely encountered signal structures in the field of signal processing were created to interpret permutation, modified permutation, sample, quadratic sample and fuzzy entropies. Subsequently, the entropy algorithms were applied to two different databases containing electroencephalogram (EEG) signals from epilepsy studies. Transitions from randomness to periodicity were successfully detected in the synthetic signals, while significant differences in EEG signals were observed based on different regions and states of the brain. In addition, using results from one entropy algorithm as features and the k-nearest neighbours algorithm, maximum classification accuracies in the first EEG database ranged from 63% to 73.5%, while these values increased by approximately 20% when using two different entropies as features. For the second database, maximum classification accuracy reached 62.5% using one entropy algorithm, while using two algorithms as features further increased that by 10%. Embedding entropies (sample, quadratic sample and fuzzy entropies) are found to outperform the rest of the algorithms in terms of sensitivity and show greater potential by considering the fine-tuning possibilities they offer. On the other hand, permutation and modified permutation entropies are more consistent across different input parameter values and considerably faster to calculate.