A Novel Methodology for Classifying EMG Movements Based on SVM and Genetic Algorithms

Optimizing RNNs for EMG Signal Classification: A Novel Strategy Using Grey Wolf Optimization

Accurate classification of electromyographic (EMG) signals is vital in biomedical applications. This study evaluates different architectures of recurrent neural networks for the classification of EMG signals associated with five movements of the right upper extremity. A Butterworth filter was implemented for signal preprocessing, followed by segmentation into 250 ms windows, with an overlap of 190 ms. The resulting dataset was divided into training, validation, and testing subsets. The Grey Wolf Optimization algorithm was applied to the gated recurrent unit (GRU), long short-term memory (LSTM) architectures, and bidirectional recurrent neural networks. In parallel, a performance comparison with support vector machines (SVMs) was performed. The results obtained in the first experimental phase revealed that all the RNN networks evaluated reached a 100% accuracy, standing above the 93% achieved by the SVM. Regarding classification speed, LSTM ranked as the fastest architecture, recording a time of 0.12 ms, followed by GRU with 0.134 ms. Bidirectional recurrent neural networks showed a response time of 0.2 ms, while SVM had the longest time at 2.7 ms. In the second experimental phase, a slight decrease in the accuracy of the RNN models was observed, standing at 98.46% for LSTM, 96.38% for GRU, and 97.63% for the bidirectional network. The findings of this study highlight the effectiveness and speed of recurrent neural networks in the EMG signal classification task.

Synthesis of sEMG Signals for Hand Gestures Using a 1DDCGAN

The emergence of modern prosthetics controlled by bio-signals has been facilitated by AI and microchip technology innovations. AI algorithms are trained using sEMG produced by muscles during contractions. The data acquisition procedure may result in discomfort and fatigue, particularly for amputees. Furthermore, prosthetic companies restrict sEMG signal exchange, limiting data-driven research and reproducibility. GANs present a viable solution to the aforementioned concerns. GANs can generate high-quality sEMG, which can be utilised for data augmentation, decrease the training time required by prosthetic users, enhance classification accuracy and ensure research reproducibility. This research proposes the utilisation of a one-dimensional deep convolutional GAN (1DDCGAN) to generate the sEMG of hand gestures. This approach involves the incorporation of dynamic time wrapping, fast Fourier transform and wavelets as discriminator inputs. Two datasets were utilised to validate the methodology, where five windows and increments were utilised to extract features to evaluate the synthesised sEMG quality. In addition to the traditional classification and augmentation metrics, two novel metrics-the Mantel test and the classifier two-sample test-were used for evaluation. The 1DDCGAN preserved the inter-feature correlations and generated high-quality signals, which resembled the original data. Additionally, the classification accuracy improved by an average of 1.21-5%.

Electromyographic Validation of Spontaneous Facial Mimicry Detection Using Automated Facial Action Coding

Although electromyography (EMG) remains the standard, researchers have begun using automated facial action coding system (FACS) software to evaluate spontaneous facial mimicry despite the lack of evidence of its validity. Using the facial EMG of the zygomaticus major (ZM) as a standard, we confirmed the detection of spontaneous facial mimicry in action unit 12 (AU12, lip corner puller) via an automated FACS. Participants were alternately presented with real-time model performance and prerecorded videos of dynamic facial expressions, while simultaneous ZM signal and frontal facial videos were acquired. Facial videos were estimated for AU12 using FaceReader, Py-Feat, and OpenFace. The automated FACS is less sensitive and less accurate than facial EMG, but AU12 mimicking responses were significantly correlated with ZM responses. All three software programs detected enhanced facial mimicry by live performances. The AU12 time series showed a roughly 100 to 300 ms latency relative to the ZM. Our results suggested that while the automated FACS could not replace facial EMG in mimicry detection, it could serve a purpose for large effect sizes. Researchers should be cautious with the automated FACS outputs, especially when studying clinical populations. In addition, developers should consider the EMG validation of AU estimation as a benchmark.

Sign Language Recognition Using the Electromyographic Signal: A Systematic Literature Review

The analysis and recognition of sign languages are currently active fields of research focused on sign recognition. Various approaches differ in terms of analysis methods and the devices used for sign acquisition. Traditional methods rely on video analysis or spatial positioning data calculated using motion capture tools. In contrast to these conventional recognition and classification approaches, electromyogram (EMG) signals, which measure muscle electrical activity, offer potential technology for detecting gestures. These EMG-based approaches have recently gained attention due to their advantages. This prompted us to conduct a comprehensive study on the methods, approaches, and projects utilizing EMG sensors for sign language handshape recognition. In this paper, we provided an overview of the sign language recognition field through a literature review, with the objective of offering an in-depth review of the most significant techniques. These techniques were categorized in this article based on their respective methodologies. The survey discussed the progress and challenges in sign language recognition systems based on surface electromyography (sEMG) signals. These systems have shown promise but face issues like sEMG data variability and sensor placement. Multiple sensors enhance reliability and accuracy. Machine learning, including deep learning, is used to address these challenges. Common classifiers in sEMG-based sign language recognition include SVM, ANN, CNN, KNN, HMM, and LSTM. While SVM and ANN are widely used, random forest and KNN have shown better performance in some cases. A multilayer perceptron neural network achieved perfect accuracy in one study. CNN, often paired with LSTM, ranks as the third most popular classifier and can achieve exceptional accuracy, reaching up to 99.6% when utilizing both EMG and IMU data. LSTM is highly regarded for handling sequential dependencies in EMG signals, making it a critical component of sign language recognition systems. In summary, the survey highlights the prevalence of SVM and ANN classifiers but also suggests the effectiveness of alternative classifiers like random forests and KNNs. LSTM emerges as the most suitable algorithm for capturing sequential dependencies and improving gesture recognition in EMG-based sign language recognition systems.

Development of a Low-Cost Portable EMG for Measuring the Muscular Activity of Workers in the Field

This study explores the development and validation of a low-cost electromyography (EMG) device for monitoring muscle activity and muscle fatigue by monitoring the key features in EMG time and frequency domains. The device consists of a Raspberry Pico microcontroller interfacing a Myoware EMG module. The experiment involved 34 volunteers (14 women, 20 men) who performed isometric and isotonic contractions using a hand dynamometer. The low-cost EMG device was compared to a research-grade EMG device, recording EMG signals simultaneously. Key features including root mean square (RMS), median power frequency (MDF), and mean power frequency (MNF) were extracted to evaluate muscle fatigue. During isometric contraction, a strong congruence between the two devices, with similar readings and behavior of the extracted features, was observed, and the Wilcoxon signed rank test confirmed no significant difference in the ability to detect muscle fatigue between the devices. For isotonic contractions, the low-cost device demonstrated behavior similar to the professional EMG device in 70.58% of cases, despite some susceptibility to noise and movement. This suggests the potential viability of the low-cost EMG device as a portable tool for assessing muscle fatigue, enabling accessible and cost-effective management of muscle health in various work scenarios.

Altered Resting-State Brain Activity and Functional Connectivity in Post-Stroke Apathy: An fMRI Study

Apathy is a common neuropsychiatric disease after stroke and is linked to a lower quality of life while undergoing rehabilitation. However, it is still unknown what are the underlying neural mechanisms of apathy. This research aimed to explore differences in the cerebral activity and functional connectivity (FC) of subjects with post-stroke apathy and those without it. A total of 59 individuals with acute ischemic stroke and 29 healthy subjects with similar age, sex, and education were recruited. The Apathy Evaluation Scale (AES) was used to evaluate apathy at 3 months after stroke. Patients were split into two groups-PSA (n = 21) and nPSA (n = 38)-based on their diagnosis. The fractional amplitude of low-frequency fluctuation (fALFF) was used to measure cerebral activity, as well as region-of-interest to region-of-interest analysis to examine functional connectivity among apathy-related regions. Pearson correlation analysis between fALFF values and apathy severity was performed in this research. The values of fALFF in the left middle temporal regions, right anterior and middle cingulate regions, middle frontal region, and cuneus region differed significantly among groups. Pearson correlation analysis showed that the fALFF values in the left middle temporal region (p < 0.001, r = 0.66) and right cuneus (p < 0.001, r = 0.48) were positively correlated with AES scores in stroke patients, while fALFF values in the right anterior cingulate (p < 0.001, r = -0.61), right middle frontal gyrus (p < 0.001, r = -0.49), and middle cingulate gyrus (p = 0.04, r = -0.27) were negatively correlated with AES scores in stroke patients. These regions formed an apathy-related subnetwork, and functional connectivity analysis unveiled that altered connectivity was linked to PSA (p < 0.05). This research found that abnormalities in brain activity and FC in the left middle temporal region, right middle frontal region, right cuneate region, and right anterior and middle cingulate regions in stroke patients were associated with PSA, revealing a possible neural mechanism and providing new clues for the diagnosis and treatment of PSA.

Investigating the Effects of a Focal Muscle Vibration Protocol on Sensorimotor Integration in Healthy Subjects

Background: The ability to perceive two tactile stimuli as asynchronous can be measured using the somatosensory temporal discrimination threshold (STDT). In healthy humans, the execution of a voluntary movement determines an increase in STDT values, while the integration of STDT and movement execution is abnormal in patients with basal ganglia disorders. Sensorimotor integration can be modulated using focal muscle vibration (fMV), a neurophysiological approach that selectively activates proprioceptive afferents from the vibrated muscle. Method: In this study, we investigated whether fMV was able to modulate STDT or STDT-movement integration in healthy subjects by measuring them before, during and after fMV applied over the first dorsalis interosseous, abductor pollicis brevis and flexor radialis carpi muscles. Results: The results showed that fMV modulated STDT-movement integration only when applied over the first dorsalis interosseous, namely, the muscle performing the motor task involved in STDT-movement integration. These changes occurred during and up to 10 min after fMV. Differently, fMV did not influence STDT at rest. We suggest that that fMV interferes with the STDT-movement task processing, possibly disrupting the physiological processing of sensory information. Conclusions: This study showed that FMV is able to modulate STDT-movement integration when applied over the muscle involved in the motor task. This result provides further information on the mechanisms underlying fMV, and has potential future implications in basal ganglia disorders characterized by altered sensorimotor integration.

A Deep Learning Approach for Predicting Multiple Sclerosis

This paper proposes a deep learning model based on an artificial neural network with a single hidden layer for predicting the diagnosis of multiple sclerosis. The hidden layer includes a regularization term that prevents overfitting and reduces the model complexity. The purposed learning model achieved higher prediction accuracy and lower loss than four conventional machine learning techniques. A dimensionality reduction method was used to select the most relevant features from 74 gene expression profiles for training the learning models. The analysis of variance test was performed to identify the statistical difference between the mean of the proposed model and the compared classifiers. The experimental results show the effectiveness of the proposed artificial neural network.

Parkinson’s Disease Diagnosis Using Laplacian Score, Gaussian Process Regression and Self-Organizing Maps

Parkinson’s disease (PD) is a complex degenerative brain disease that affects nerve cells in the brain responsible for body movement. Machine learning is widely used to track the progression of PD in its early stages by predicting unified Parkinson’s disease rating scale (UPDRS) scores. In this paper, we aim to develop a new method for PD diagnosis with the aid of supervised and unsupervised learning techniques. Our method is developed using the Laplacian score, Gaussian process regression (GPR) and self-organizing maps (SOM). SOM is used to segment the data to handle large PD datasets. The models are then constructed using GPR for the prediction of the UPDRS scores. To select the important features in the PD dataset, we use the Laplacian score in the method. We evaluate the developed approach on a PD dataset including a set of speech signals. The method was evaluated through root-mean-square error (RMSE) and adjusted R-squared (adjusted R²). Our findings reveal that the proposed method is efficient in the prediction of UPDRS scores through a set of speech signals (dysphonia measures). The method evaluation showed that SOM combined with the Laplacian score and Gaussian process regression with the exponential kernel provides the best results for R-squared (Motor-UPDRS = 0.9489; Total-UPDRS = 0.9516) and RMSE (Motor-UPDRS = 0.5144; Total-UPDRS = 0.5105) in predicting UPDRS compared with the other kernels in Gaussian process regression.

Repeatability and Temporal Consistency of Lower Limb Biomechanical Variables Expressing Interlimb Coordination during the Double-Support Phase in People with and without Stroke Sequelae

Reliable biomechanical methods to assess interlimb coordination during the double-support phase in post-stroke subjects are needed for assessing movement dysfunction and related variability. The data obtained could provide a significant contribution for designing rehabilitation programs and for their monitorisation. The present study aimed to determine the minimum number of gait cycles needed to obtain adequate values of repeatability and temporal consistency of lower limb kinematic, kinetic, and electromyographic parameters during the double support of walking in people with and without stroke sequelae. Eleven post-stroke and thirteen healthy participants performed 20 gait trials at self-selected speed in two separate moments with an interval between 72 h and 7 days. The joint position, the external mechanical work on the centre of mass, and the surface electromyographic activity of the tibialis anterior, soleus, gastrocnemius medialis, rectus femoris, vastus medialis, biceps femoris, and gluteus maximus muscles were extracted for analysis. Both the contralesional and ipsilesional and dominant and non-dominant limbs of participants with and without stroke sequelae, respectively, were evaluated either in trailing or leading positions. The intraclass correlation coefficient was used for assessing intra-session and inter-session consistency analysis. For most of the kinematic and the kinetic variables studied in each session, two to three trials were required for both groups, limbs, and positions. The electromyographic variables presented higher variability, requiring, therefore, a number of trials ranging from 2 to >10. Globally, the number of trials required inter-session ranged from 1 to >10 for kinematic, from 1 to 9 for kinetic, and 1 to >10 for electromyographic variables. Thus, for the double support analysis, three gait trials were required in order to assess the kinematic and kinetic variables in cross-sectional studies, while for longitudinal studies, a higher number of trials (>10) were required for kinematic, kinetic, and electromyographic variables.

Use of Surface Electromyography to Estimate End-Point Force in Redundant Systems: Comparison between Linear Approaches

Estimation of the force exerted by muscles from their electromyographic (EMG) activity may be useful to control robotic devices. Approximating end-point forces as a linear combination of the activities of multiple muscles acting on a limb may lead to an inaccurate estimation because of the dependency between the EMG signals, i.e., multi-collinearity. This study compared the EMG-to-force mapping estimation performed with standard multiple linear regression and with three other algorithms designed to reduce different sources of the detrimental effects of multi-collinearity: Ridge Regression, which performs an L2 regularization through a penalty term; linear regression with constraints from foreknown anatomical boundaries, derived from a musculoskeletal model; linear regression of a reduced number of muscular degrees of freedom through the identification of muscle synergies. Two datasets, both collected during the exertion of submaximal isometric forces along multiple directions with the upper limb, were exploited. One included data collected across five sessions and the other during the simultaneous exertion of force and generation of different levels of co-contraction. The accuracy and consistency of the EMG-to-force mappings were assessed to determine the strengths and drawbacks of each algorithm. When applied to multiple sessions, Ridge Regression achieved higher accuracy (R² = 0.70) but estimations based on muscle synergies were more consistent (differences between the pulling vectors of mappings extracted from different sessions: 67%). In contrast, the implementation of anatomical constraints was the best solution, both in terms of consistency (R² = 0.64) and accuracy (74%), in the case of different co-contraction conditions. These results may be used for the selection of the mapping between EMG and force to be implemented in myoelectrically controlled robotic devices.