Optimizing shoulder elevation assist rate in exoskeletal rehabilitation based on muscular activity indices: a clinical feasibility study.

BACKGROUND: Restoring shoulder function is critical for upper-extremity rehabilitation following a stroke. The complex musculoskeletal anatomy of the shoulder presents a challenge for safely assisting elevation movements through robotic interventions. The level of shoulder elevation assistance in rehabilitation is often based on clinical judgment. There is no standardized method for deriving an optimal level of assistance, underscoring the importance of addressing abnormal movements during shoulder elevation, such as abnormal synergies and compensatory actions. This study aimed to investigate the effectiveness and safety of a newly developed shoulder elevation exoskeleton robot by applying a novel optimization technique derived from the muscle synergy index. METHODS: Twelve chronic stroke participants underwent an intervention consisting of 100 robot-assisted shoulder elevation exercises (10 × 10 times, approximately 40 min) for 10 days (4-5 times/week). The optimal robot assist rate was derived by detecting the change points using the co-contraction index, calculated from electromyogram (EMG) data obtained from the anterior deltoid and biceps brachii muscles during shoulder elevation at the initial evaluation. The primary outcomes were the Fugl-Meyer assessment-upper extremity (FMA-UE) shoulder/elbow/forearm score, kinematic outcomes (maximum angle of voluntary shoulder flexion and elbow flexion ratio during shoulder elevation), and shoulder pain outcomes (pain-free passive shoulder flexion range of motion [ROM] and visual analogue scale for pain severity during shoulder flexion). The effectiveness and safety of robotic therapy were examined using the Wilcoxon signed-rank sum test. RESULTS: All 12 patients completed the procedure without any adverse events. Two participants were excluded from the analysis because the EMG of the biceps brachii was not obtained. Ten participants (five men and five women; mean age: 57.0 [5.5] years; mean FMA-UE total score: 18.7 [10.5] points) showed significant improvement in the FMA-UE shoulder/elbow/forearm score, kinematic outcomes, and pain-free passive shoulder flexion ROM (P < 0.05). The shoulder pain outcomes remained unchanged or improved in all patients. CONCLUSIONS: The study presents a method for deriving the optimal robotic assist rate. Rehabilitation using a shoulder robot based on this derived optimal assist rate showed the possibility of safely improving the upper-extremity function in patients with severe stroke in the chronic phase.

Exploration of a machine learning approach for diagnosing sarcopenia among Chinese community-dwelling older adults using sEMG-based data.

BACKGROUND: In the practical application of sarcopenia screening, there is a need for faster, time-saving, and community-friendly detection methods. The primary purpose of this study was to perform sarcopenia screening in community-dwelling older adults and investigate whether surface electromyogram (sEMG) from hand grip could potentially be used to detect sarcopenia using machine learning (ML) methods with reasonable features extracted from sEMG signals. The secondary aim was to provide the interpretability of the obtained ML models using a novel feature importance estimation method. METHODS: A total of 158 community-dwelling older residents (≥ 60 years old) were recruited. After screening through the diagnostic criteria of the Asian Working Group for Sarcopenia in 2019 (AWGS 2019) and data quality check, participants were assigned to the healthy group (n = 45) and the sarcopenic group (n = 48). sEMG signals from six forearm muscles were recorded during the hand grip task at 20% maximal voluntary contraction (MVC) and 50% MVC. After filtering recorded signals, nine representative features were extracted, including six time-domain features plus three time-frequency domain features. Then, a voting classifier ensembled by a support vector machine (SVM), a random forest (RF), and a gradient boosting machine (GBM) was implemented to classify healthy versus sarcopenic participants. Finally, the SHapley Additive exPlanations (SHAP) method was utilized to investigate feature importance during classification. RESULTS: Seven out of the nine features exhibited statistically significant differences between healthy and sarcopenic participants in both 20% and 50% MVC tests. Using these features, the voting classifier achieved 80% sensitivity and 73% accuracy through a five-fold cross-validation. Such performance was better than each of the SVM, RF, and GBM models alone. Lastly, SHAP results revealed that the wavelength (WL) and the kurtosis of continuous wavelet transform coefficients (CWT_kurtosis) had the highest feature impact scores. CONCLUSION: This study proposed a method for community-based sarcopenia screening using sEMG signals of forearm muscles. Using a voting classifier with nine representative features, the accuracy exceeds 70% and the sensitivity exceeds 75%, indicating moderate classification performance. Interpretable results obtained from the SHAP model suggest that motor unit (MU) activation mode may be a key factor affecting sarcopenia.

The role of electrodiagnosis in focal neuropathies.

Electrodiagnostic (EDX) testing plays an important role in confirming a mononeuropathy, localizing the site of nerve injury, defining the pathophysiology, and assessing the severity and prognosis. The combination of nerve conduction studies (NCS) and needle electromyography findings provides the necessary information to fully assess a nerve. The pattern of NCS abnormalities reflects the underlying pathophysiology, with focal slowing or conduction block in neuropraxic injuries and reduced amplitudes in axonotmetic injuries. Needle electromyography findings, including spontaneous activity and voluntary motor unit potential changes, complement the NCS findings and further characterize chronicity and degree of axon loss and reinnervation. EDX is used as an objective marker to follow the progression of a mononeuropathy over time.

Effects of static exercises on hip muscle fatigue and knee wobble assessed by surface electromyography and inertial measurement unit data.

Hip muscle weakness can be a precursor to or a result of lower limb injuries. Assessment of hip muscle strength and muscle motor fatigue in the clinic is important for diagnosing and treating hip-related impairments. Muscle motor fatigue can be assessed with surface electromyography (sEMG), however sEMG requires specialized equipment and training. Inertial measurement units (IMUs) are wearable devices used to measure human motion, yet it remains unclear if they can be used as a low-cost alternative method to measure hip muscle fatigue. The goals of this work were to (1) identify which of five pre-selected exercises most consistently and effectively elicited muscle fatigue in the gluteus maximus, gluteus medius, and rectus femoris muscles and (2) determine the relationship between muscle fatigue using sEMG sensors and knee wobble using an IMU device. This work suggests that a wall sit and single leg knee raise activity fatigue the gluteus medius, gluteus maximus, and rectus femoris muscles most reliably (p < 0.05) and that the gluteus medius and gluteus maximus muscles were fatigued to a greater extent than the rectus femoris (p = 0.031 and p = 0.0023, respectively). Additionally, while acceleration data from a single IMU placed on the knee suggested that more knee wobble may be an indicator of muscle fatigue, this single IMU is not capable of reliably assessing fatigue level. These results suggest the wall sit activity could be used as simple, static exercise to elicit hip muscle fatigue in the clinic, and that assessment of knee wobble in addition to other IMU measures could potentially be used to infer muscle fatigue under controlled conditions. Future work examining the relationship between IMU data, muscle fatigue, and multi-limb dynamics should be explored to develop an accessible, low-cost, fast and standardized method to measure fatiguability of the hip muscles in the clinic.

Combining enhanced spectral resolution of EMG and a deep learning approach for knee pathology diagnosis.

Knee osteoarthritis (OA) is a prevalent, debilitating joint condition primarily affecting the elderly. This investigation aims to develop an electromyography (EMG)-based method for diagnosing knee pathologies. EMG signals of the muscles surrounding the knee joint were examined and recorded. The principal components of the proposed method were preprocessing, high-order spectral analysis (HOSA), and diagnosis/recognition through deep learning. EMG signals from individuals with normal and OA knees while walking were extracted from a publicly available database. This examination focused on the quadriceps femoris, the medial gastrocnemius, the rectus femoris, the semitendinosus, and the vastus medialis. Filtration and rectification were utilized beforehand to eradicate noise and smooth EMG signals. Signals' higher-order spectra were analyzed with HOSA to obtain information about nonlinear interactions and phase coupling. Initially, the bicoherence representation of EMG signals was devised. The resulting images were fed into a deep-learning system for identification and analysis. A deep learning algorithm using adapted ResNet101 CNN model examined the images to determine whether the EMG signals were conventional or indicative of knee osteoarthritis. The validated test results demonstrated high accuracy and robust metrics, indicating that the proposed method is effective. The medial gastrocnemius (MG) muscle was able to distinguish Knee osteoarthritis (KOA) patients from normal with 96.3±1.7% accuracy and 0.994±0.008 AUC. MG has the highest prediction accuracy of KOA and can be used as the muscle of interest in future analysis. Despite the proposed method's superiority, some limitations still require special consideration and will be addressed in future research.

Highly stretchable and customizable microneedle electrode arrays for intramuscular electromyography.

Stretchable three-dimensional (3D) penetrating microelectrode arrays have potential utility in various fields, including neuroscience, tissue engineering, and wearable bioelectronics. These 3D microelectrode arrays can penetrate and conform to dynamically deforming tissues, thereby facilitating targeted sensing and stimulation of interior regions in a minimally invasive manner. However, fabricating custom stretchable 3D microelectrode arrays presents material integration and patterning challenges. In this study, we present the design, fabrication, and applications of stretchable microneedle electrode arrays (SMNEAs) for sensing local intramuscular electromyography signals ex vivo. We use a unique hybrid fabrication scheme based on laser micromachining, microfabrication, and transfer printing to enable scalable fabrication of individually addressable SMNEA with high device stretchability (60 to 90%). The electrode geometries and recording regions, impedance, array layout, and length distribution are highly customizable. We demonstrate the use of SMNEAs as bioelectronic interfaces in recording intramuscular electromyography from various muscle groups in the buccal mass of Aplysia.

Study on Gesture Recognition Method with Two-Stream Residual Network Fusing sEMG Signals and Acceleration Signals.

Currently, surface EMG signals have a wide range of applications in human-computer interaction systems. However, selecting features for gesture recognition models based on traditional machine learning can be challenging and may not yield satisfactory results. Considering the strong nonlinear generalization ability of neural networks, this paper proposes a two-stream residual network model with an attention mechanism for gesture recognition. One branch processes surface EMG signals, while the other processes hand acceleration signals. Segmented networks are utilized to fully extract the physiological and kinematic features of the hand. To enhance the model's capacity to learn crucial information, we introduce an attention mechanism after global average pooling. This mechanism strengthens relevant features and weakens irrelevant ones. Finally, the deep features obtained from the two branches of learning are fused to further improve the accuracy of multi-gesture recognition. The experiments conducted on the NinaPro DB2 public dataset resulted in a recognition accuracy of 88.25% for 49 gestures. This demonstrates that our network model can effectively capture gesture features, enhancing accuracy and robustness across various gestures. This approach to multi-source information fusion is expected to provide more accurate and real-time commands for exoskeleton robots and myoelectric prosthetic control systems, thereby enhancing the user experience and the naturalness of robot operation.

Measuring Surface Electromyography with Textile Electrodes in a Smart Leg Sleeve.

This paper presents the design, development, and validation of a novel e-textile leg sleeve for non-invasive Surface Electromyography (sEMG) monitoring. This wearable device incorporates e-textile sensors for sEMG signal acquisition from the lower limb muscles, specifically the anterior tibialis and lateral gastrocnemius. Validation was conducted by performing a comparative study with eleven healthy volunteers to evaluate the performance of the e-textile sleeve in acquiring sEMG signals compared to traditional Ag/AgCl electrodes. The results demonstrated strong agreement between the e-textile and conventional methods in measuring descriptive metrics of the signals, including area, power, mean, and root mean square. The paired data t-test did not reveal any statistically significant differences, and the Bland-Altman analysis indicated negligible bias between the measures recorded using the two methods. In addition, this study evaluated the wearability and comfort of the e-textile sleeve using the Comfort Rating Scale (CRS). Overall, the scores confirmed that the proposed device is highly wearable and comfortable, highlighting its suitability for everyday use in patient care.

Rectified Latent Variable Model-Based EMG Factorization of Inhibitory Muscle Synergy Components Related to Aging, Expertise and Force-Tempo Variations.

Muscle synergy has been widely acknowledged as a possible strategy of neuromotor control, but current research has ignored the potential inhibitory components in muscle synergies. Our study aims to identify and characterize the inhibitory components within motor modules derived from electromyography (EMG), investigate the impact of aging and motor expertise on these components, and better understand the nervous system's adaptions to varying task demands. We utilized a rectified latent variable model (RLVM) to factorize motor modules with inhibitory components from EMG signals recorded from ten expert pianists when they played scales and pieces at different tempo-force combinations. We found that older participants showed a higher proportion of inhibitory components compared with the younger group. Senior experts had a higher proportion of inhibitory components on the left hand, and most inhibitory components became less negative with increased tempo or decreased force. Our results demonstrated that the inhibitory components in muscle synergies could be shaped by aging and expertise, and also took part in motor control for adapting to different conditions in complex tasks.

Mapping Method of Human Arm Motion Based on Surface Electromyography Signals.

This paper investigates a method for precise mapping of human arm movements using sEMG signals. A multi-channel approach captures the sEMG signals, which, combined with the accurately calculated joint angles from an Inertial Measurement Unit, allows for action recognition and mapping through deep learning algorithms. Firstly, signal acquisition and processing were carried out, which involved acquiring data from various movements (hand gestures, single-degree-of-freedom joint movements, and continuous joint actions) and sensor placement. Then, interference signals were filtered out through filters, and the signals were preprocessed using normalization and moving averages to obtain sEMG signals with obvious features. Additionally, this paper constructs a hybrid network model, combining Convolutional Neural Networks and Artificial Neural Networks, and employs a multi-feature fusion algorithm to enhance the accuracy of gesture recognition. Furthermore, a nonlinear fitting between sEMG signals and joint angles was established based on a backpropagation neural network, incorporating momentum term and adaptive learning rate adjustments. Finally, based on the gesture recognition and joint angle prediction model, prosthetic arm control experiments were conducted, achieving highly accurate arm movement prediction and execution. This paper not only validates the potential application of sEMG signals in the precise control of robotic arms but also lays a solid foundation for the development of more intuitive and responsive prostheses and assistive devices.

Research on a Calculation Model of Ankle-Joint-Torque-Based sEMG.

The purpose of this article is to establish a prediction model of joint movements and realize the prediction of joint movemenst, and the research results are of reference value for the development of the rehabilitation equipment. This will be carried out by analyzing the impact of surface electromyography (sEMG) on ankle movements and using the Hill model as a framework for calculating ankle joint torque. The table and scheme used in the experiments were based on physiological parameters obtained through the model. Data analysis was performed on ankle joint angle signal, movement signal, and sEMG data from nine subjects during dorsiflexion/flexion, varus, and internal/external rotation. The Hill model was employed to determine 16 physiological parameters which were optimized using a genetic algorithm. Three experiments were carried out to identify the optimal model to calculate torque and root mean square error. The optimized model precisely calculated torque and had a root mean square error of under 1.4 in comparison to the measured torque. Ankle movement models predict torque patterns with accuracy, thereby providing a solid theoretical basis for ankle rehabilitation control. The optimized model provides a theoretical foundation for precise ankle torque forecasts, thereby improving the efficacy of rehabilitation robots for the ankle.

Effects of prolonged vibration to the flexor carpi radialis muscle on intracortical excitability.

Prolonged local vibration (LV) can induce neurophysiological adaptations thought to be related to long-term potentiation or depression. Yet, how changes in intracortical excitability may be involved remains to be further investigated as previous studies reported equivocal results. We therefore investigated the effects of 30 min of LV applied to the right flexor carpi radialis muscle (FCR) on both short-interval intracortical inhibition (SICI) and intracortical facilitation (ICF). SICI and ICF were measured through transcranial magnetic stimulation before and immediately after 30 min of FCR LV (vibration condition) or 30 min of rest (control condition). Measurements were performed during a low-intensity contraction (n = 17) or at rest (n = 7). No significant SICI nor ICF modulations were observed, whether measured during isometric contractions or at rest (p = 0.2). Yet, we observed an increase in inter-individual variability for post measurements after LV. In conclusion, while intracortical excitability was not significantly modulated after LV, increased inter-variability observed after LV may suggest the possibility of divergent responses to prolonged LV exposure.

Comparison of Metabolic Power and Energy Cost of Submaximal and Sprint Running Efforts Using Different Methods in Elite Youth Soccer Players: A Novel Energetic Approach.

Sprinting is a decisive action in soccer that is considerably taxing from a neuromuscular and energetic perspective. This study compared different calculation methods for the metabolic power (MP) and energy cost (EC) of sprinting using global positioning system (GPS) metrics and electromyography (EMG), with the aim of identifying potential differences in performance markers. Sixteen elite U17 male soccer players (age: 16.4 ± 0.5 years; body mass: 64.6 ± 4.4 kg; and height: 177.4 ± 4.3 cm) participated in the study and completed four different submaximal constant running efforts followed by sprinting actions while using portable GPS-IMU units and surface EMG. GPS-derived MP was determined based on GPS velocity, and the EMG-MP and EC were calculated based on individual profiles plotting the MP of the GPS and all EMG signals acquired. The goodness of fit of the linear regressions was assessed by the coefficient of determination (R2), and a repeated measures ANOVA was used to detect changes. A linear trend was found in EMG activity during submaximal speed runs (R2 = 1), but when the sprint effort was considered, the trend became exponential (R2 = 0.89). The EMG/force ratio displayed two different trends: linear up to a 30 m sprint (R2 = 0.99) and polynomial up to a 50 m sprint (R2 = 0.96). Statistically significant differences between the GPS and EMG were observed for MP splits at 0-5 m, 5-10 m, 25-30 m, 30-35 m, and 35-40 m and for EC splits at 5-10 m, 25-30 m, 30-35 m, and 35-40 m (p ≤ 0.05). Therefore, the determination of the MP and EC based on GPS technology underestimated the neuromuscular and metabolic engagement during the sprinting efforts. Thus, the EMG-derived method seems to be more accurate for calculating the MP and EC in this type of action.

Identification of the Biomechanical Response of the Muscles That Contract the Most during Disfluencies in Stuttered Speech.

Stuttering, affecting approximately 1% of the global population, is a complex speech disorder significantly impacting individuals' quality of life. Prior studies using electromyography (EMG) to examine orofacial muscle activity in stuttering have presented mixed results, highlighting the variability in neuromuscular responses during stuttering episodes. Fifty-five participants with stuttering and 30 individuals without stuttering, aged between 18 and 40, participated in the study. EMG signals from five facial and cervical muscles were recorded during speech tasks and analyzed for mean amplitude and frequency activity in the 5-15 Hz range to identify significant differences. Upon analysis of the 5-15 Hz frequency range, a higher average amplitude was observed in the zygomaticus major muscle for participants while stuttering (p < 0.05). Additionally, when assessing the overall EMG signal amplitude, a higher average amplitude was observed in samples obtained from disfluencies in participants who did not stutter, particularly in the depressor anguli oris muscle (p < 0.05). Significant differences in muscle activity were observed between the two groups, particularly in the depressor anguli oris and zygomaticus major muscles. These results suggest that the underlying neuromuscular mechanisms of stuttering might involve subtle aspects of timing and coordination in muscle activation. Therefore, these findings may contribute to the field of biosensors by providing valuable perspectives on neuromuscular mechanisms and the relevance of electromyography in stuttering research. Further research in this area has the potential to advance the development of biosensor technology for language-related applications and therapeutic interventions in stuttering.

Exploring Human-Exoskeleton Interaction Dynamics: An In-Depth Analysis of Knee Flexion-Extension Performance across Varied Robot Assistance-Resistance Configurations.

Knee rehabilitation therapy after trauma or neuromotor diseases is fundamental to restore the joint functions as best as possible, exoskeleton robots being an important resource in this context, since they optimize therapy by applying tailored forces to assist or resist movements, contributing to improved patient outcomes and treatment efficiency. One of the points that must be taken into account when using robots in rehabilitation is their interaction with the patient, which must be safe for both and guarantee the effectiveness of the treatment. Therefore, the objective of this study was to assess the interaction between humans and an exoskeleton during the execution of knee flexion-extension movements under various configurations of robot assistance and resistance. The evaluation encompassed considerations of myoelectric activity, muscle recruitment, robot torque, and performed movement. To achieve this, an experimental protocol was implemented, involving an individual wearing the exoskeleton and executing knee flexion-extension motions while seated, with the robot configured in five distinct modes: passive (P), assistance on flexion (FA), assistance on extension (EA), assistance on flexion and extension (CA), and resistance on flexion and extension (CR). Results revealed distinctive patterns of movement and muscle recruitment for each mode, highlighting the complex interplay between human and robot; for example, the largest RMS tracking errors were for the EA mode (13.72 degrees) while the smallest for the CR mode (4.47 degrees), a non-obvious result; in addition, myoelectric activity was demonstrated to be greater for the completely assisted mode than without the robot (the maximum activation levels for the vastus medialis and vastus lateralis muscles were more than double those when the user had assistance from the robot). Tracking errors, muscle activations, and torque values varied across modes, emphasizing the need for careful consideration in configuring exoskeleton assistance and resistance to ensure effective and safe rehabilitation. Understanding these human-robot interactions is essential for developing precise rehabilitation programs, optimizing treatment effectiveness, and enhancing patient safety.

Accuracy of Repetitive Ocular Vestibular-Evoked Myogenic Potentials to Diagnose Myasthenia Gravis in Patients With Ptosis or Diplopia.

BACKGROUND AND OBJECTIVES: We developed repetitive ocular vestibular-evoked myogenic potentials (roVEMP) as an electrophysiologic test that allows us to elicit the characteristic decrement of extraocular muscles in patients with ocular myasthenia gravis (OMG). Case-control studies demonstrated that roVEMP reliably differentiates patients with OMG from healthy controls. We now aimed to evaluate the diagnostic accuracy of roVEMP for OMG diagnosis in patients with ptosis and/or diplopia. METHODS: In this blinded prospective diagnostic accuracy trial, we compared roVEMP in 89 consecutive patients presenting with ptosis and/or diplopia suspicious of OMG with a multimodal diagnostic approach, including clinical examination, antibodies, edrophonium testing, repetitive nerve stimulation of accessory and facial nerves, and single-fiber EMG (SFEMG). We calculated the roVEMP decrement as the ratio between the mean of the first 2 responses compared with the mean of the sixth-ninth responses in the train and used cutoff of >9% (unilateral decrement) in a 30 Hz stimulation paradigm. RESULTS: Following a complete diagnostic work-up, 39 patients (44%) were diagnosed with ocular MG, while 50 patients (56%) had various other neuro-ophthalmologic conditions, but not MG (non-MG). roVEMP yielded 88.2% sensitivity, 30.2% specificity, 50% positive predictive value (PPV), and 76.5% negative predictive value (NPV). For comparison, SFEMG resulted in 75% sensitivity, 56% specificity, 55.1% PPV, and 75.7% NPV. All other diagnostic tests (except for the ice pack test) also yielded significantly higher positive results in patients with MG compared with non-MG. DISCUSSION: The study revealed a high sensitivity of 88.2% for roVEMP in OMG, but specificity and PPV were too low to allow for the OMG diagnosis as a single test. Thus, differentiating ocular MG from other neuro-ophthalmologic conditions remains challenging, and the highest diagnostic accuracy is still obtained by a multimodal approach. In this study, roVEMP can complement the diagnostic armamentarium for the diagnosis of MG. CLASSIFICATION OF EVIDENCE: This study provides Class I evidence that in patients with diplopia and ptosis, roVEMP alone does not accurately distinguish MG from non-MG disorders. TRIAL REGISTRATION INFORMATION: ClinicalTrials.gov: NCT03049956.

Experimental evaluation of the impact of sEMG interfaces in enhancing embodiment of virtual myoelectric prostheses.

INTRODUCTION: Despite recent technological advances that have led to sophisticated bionic prostheses, attaining embodied solutions still remains a challenge. Recently, the investigation of prosthetic embodiment has become a topic of interest in the research community, which deals with enhancing the perception of artificial limbs as part of users' own body. Surface electromyography (sEMG) interfaces have emerged as a promising technology for enhancing upper-limb prosthetic control. However, little is known about the impact of these sEMG interfaces on users' experience regarding embodiment and their interaction with different functional levels. METHODS: To investigate this aspect, a comparison is conducted among sEMG configurations with different number of sensors (4 and 16 channels) and different time delay. We used a regression algorithm to simultaneously control hand closing/opening and forearm pronation/supination in an immersive virtual reality environment. The experimental evaluation includes 24 able-bodied subjects and one prosthesis user. We assess functionality with the Target Achievement Control test, and the sense of embodiment with a metric for the users perception of self-location, together with a standard survey. RESULTS: Among the four tested conditions, results proved a higher subjective embodiment when participants used sEMG interfaces employing an increased number of sensors. Regarding functionality, significant improvement over time is observed in the same conditions, independently of the time delay implemented. CONCLUSIONS: Our work indicates that a sufficient number of sEMG sensors improves both, functional and subjective embodiment outcomes. This prompts discussion regarding the potential relationship between these two aspects present in bionic integration. Similar embodiment outcomes are observed in the prosthesis user, showing also differences due to the time delay, and demonstrating the influence of sEMG interfaces on the sense of agency.

Estimation of electrical muscle activity during gait using inertial measurement units with convolution attention neural network and small-scale dataset.

In general, muscle activity can be directly measured using Electromyography (EMG) or calculated with musculoskeletal models. However, both methods are not suitable for non-technical users and unstructured environments. It is desired to establish more portable and easy-to-use muscle activity estimation methods. Deep learning (DL) models combined with inertial measurement units (IMUs) have shown great potential to estimate muscle activity. However, it frequently occurs in clinical scenarios that a very small amount of data is available and leads to limited performance of the DL models, while the augmentation techniques to efficiently expand a small sample size for DL model training are rarely used. The primary aim of the present study was to develop a novel DL model to estimate the EMG envelope during gait using IMUs with high accuracy. A secondary aim was to develop a novel model-based data augmentation method to improve the performance of the estimation model with small-scale dataset. Therefore, in the present study, a time convolutional network-based generative adversarial network, namely MuscleGAN, was proposed for data augmentation. Moreover, a subject-independent regression DL model was developed to estimate EMG envelope. Results suggested that the proposed two-stage method has better generalization and estimation performance than the commonly used existing methods. Pearson correlation coefficient and normalized root-mean-square errors derived from the proposed method reached up to 0.72 and 0.13, respectively. It was indicated that the MuscleGAN indeed improved the estimation accuracy of lower limb EMG envelope from 70% to 72%. Thus, even using only two IMUs and a very small-scale dataset, the proposed model is still capable of accurately estimating lower limb EMG envelope, demonstrating considerable potential for its application in clinical and daily life scenarios.

Monitoring Changes in TMS-Evoked EEG and EMG Activity During 1†Hz rTMS of the Healthy Motor Cortex.

Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive brain stimulation technique capable of inducing neuroplasticity as measured by changes in peripheral muscle electromyography (EMG) or electroencephalography (EEG) from pre-to-post stimulation. However, temporal courses of neuromodulation during ongoing rTMS are unclear. Monitoring cortical dynamics via TMS-evoked responses using EMG (motor-evoked potentials; MEPs) and EEG (transcranial-evoked potentials; TEPs) during rTMS might provide further essential insights into its mode of action - temporal course of potential modulations. The objective of this study was to first evaluate the validity of online rTMS-EEG and rTMS-EMG analyses, and second to scrutinize the temporal changes of TEPs and MEPs during rTMS. As rTMS is subject to high inter-individual effect variability, we aimed for single-subject analyses of EEG changes during rTMS. Ten healthy human participants were stimulated with 1,000 pulses of 1 Hz rTMS over the motor cortex, while EEG and EMG were recorded continuously. Validity of MEPs and TEPs measured during rTMS was assessed in sensor and source space. Electrophysiological changes during rTMS were evaluated with model fitting approaches on a group- and single-subject level. TEPs and MEPs appearance during rTMS was consistent with past findings of single pulse experiments. Heterogeneous temporal progressions, fluctuations or saturation effects of brain activity were observed during rTMS depending on the TEP component. Overall, global brain activity increased over the course of stimulation. Single-subject analysis revealed inter-individual temporal courses of global brain activity. The present findings are in favor of dose-response considerations and attempts in personalization of rTMS protocols.