High-density Electromyogram Research Articles

Effects of botulinum toxin on paraspinal muscle imbalance during spinal motions in adolescents with idiopathic scoliosis: Assessment using high-density electromyography

Background The pathogenesis of adolescent idiopathic scoliosis (AIS) remains unclear, with muscle imbalance being a widely discussed hypothesis. Objective This study examined the impact of muscle imbalance on AIS by injecting botulinum toxin A (BTX) into patients’ unilateral muscles and investigating altered back muscle synergies. Methods Three AIS patients received BTX injections in the concave-side paraspinal muscles. High-density electromyogram arrays placed from T8 to L4 recorded signals during five spinal motions at four timepoints (one pre-BTX and three post-BTX). Muscle synergies were extracted using non-negative matrix factorization and compared with data from ten healthy and ten AIS subjects from our previous studies. Results Post-BTX, muscle activity maps during flexion/extension, sitting, and standing exhibited reduced symmetry, with concave/convex ratios decreasing and being statistically lower than those of healthy subjects at post-2 and post-3 follow-ups ( p < 0.01). Muscles on the dominant side during lateral bending or axial rotation demonstrated decreased activation and differently distributed center of gravity positions on synergy maps compared to healthy subjects at all timepoints ( p < 0.05). Post-BTX changes were particularly notable for the patient with mild deformity. Conclusions BTX affected the activation of paraspinal muscles, providing insights into the role of muscle imbalance in AIS and informing future therapeutic strategies.

Synaptic inputs to motor neurons underlying muscle coactivation for functionally different tasks have different spectral characteristics.

The central nervous system (CNS) may produce the same endpoint trajectory or torque profile with different muscle activation patterns. What differentiates these patterns is the presence of cocontraction, which does not contribute to effective torque generation but allows to modulate joints' mechanical stiffness. Although it has been suggested that the generation of force and the modulation of stiffness rely on separate pathways, a characterization of the differences between the synaptic inputs to motor neurons (MNs) underlying these tasks is still missing. In this study, participants coactivated the same pair of upper-limb muscles, i.e., the biceps brachii and the triceps brachii, to perform two functionally different tasks: limb stiffness modulation or endpoint force generation. Spike trains of MNs were identified through decomposition of high-density electromyograms (EMGs) collected from the two muscles. Cross-correlogram showed a higher synchronization between MNs recruited to modulate stiffness, whereas cross-muscle coherence analysis revealed peaks in the β-band, which is commonly ascribed to a cortical origin. These peaks did not appear during the coactivation for force generation, thus suggesting separate cortical inputs for stiffness modulation. Moreover, a within-muscle coherence analysis identified two subsets of MNs that were selectively recruited to generate force or regulate stiffness. This study is the first to highlight different characteristics, and probable different neural origins, of the synaptic inputs driving a pair of muscles under different functional conditions. We suggest that stiffness modulation is driven by cortical inputs that project to a separate set of MNs, supporting the existence of a separate pathway underlying the control of stiffness.NEW & NOTEWORTHY The characterization of the pathways underlying force generation or stiffness modulation are still unknown. In this study, we demonstrated that the common input to motor neurons of antagonist muscles shows a high-frequency component when muscles are coactivated to modulate stiffness but not to generate force. Our results provide novel insights on the neural strategies for the recruitment of multiple muscles by identifying specific spectral characteristics of the synaptic inputs underlying functionally different tasks.

Longitudinal changes in intrinsic motoneuron excitability in amyotrophic lateral sclerosis are dependent on disease progression.

Increased amplitude of persistent inward currents (PICs) is observed in pre-symptomatic genetically modified SOD1 mice models of amyotrophic lateral sclerosis (ALS). However, at the symptomatic stage this reverses and there is a large reduction in PIC amplitude. It remains unclear whether these changes in PICs can be observed in humans, with cross-sectional studies in humans reporting contradictory findings. In people with ALS, we estimated the PIC contribution to self-sustained firing of motoneurons, using the paired-motor unit analysis to calculate the Δfrequency (ΔF), to compare the weaker and stronger muscles during the course of disease. We hypothesised that, with disease progression, ΔFs would relatively increase in the stronger muscles; and decline in the weaker muscles. Forty-three individuals with ALS were assessed in two occasions on average 17weeks apart. Tibialis anterior high-density electromyograms were recorded during dorsiflexion (40% of maximal capacity) ramped contractions, followed by clinical tests. ∆F increased from 3.14 (2.57, 3.71)peaks per second (pps) to 3.55 (2.94, 4.17)pps on the stronger muscles (0.41 (0.041, 0.781)pps, standardised difference (d)=0.287 (0.023, 0.552), P=0.030). ∆F reduced from 3.38 (95% CI 2.92, 3.84)pps to 2.88 (2.40, 3.36)pps on the weaker muscles (-0.50 (-0.80, -0.21)pps, d=0.353 (0.138, 0.567), P=0.001). The ALSFRS-R score reduced 3.9 (2.3, 5.5) points. These data indicate that the contribution of PICs to motoneuron self-sustained firing increases over time in early stages of the disease when there is little weakness before decreasing as the disease progresses and muscle weakness exacerbates, in alignment with the findings from studies using SOD1 mice. KEY POINTS: Research on mouse model of amyotrophic lateral sclerosis (ALS) suggests that the amplitude of persistent inward currents (PICs) is increased in early stages before decreasing as the disease progresses. Cross-sectional studies in humans have reported contradictory findings with both higher and lower PIC contributions to motoneuron self-sustained firing. In this longitudinal (∼17weeks) study we tracked changes in PIC contribution to motoneuron self-sustained firing, using the ΔF calculation (i.e. onset-offset hysteresis of motor unit pairs), in tibialis anterior muscles with normal strength and with clinical signs of weakness in people with ALS. ΔFs decreased over time in muscles with clinical signs of weakness. The PIC contribution to motoneuron self-sustained firing increases before the onset of muscle weakness, and subsequently decreases when muscle weakness progresses.

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

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

Concurrent and Continuous Prediction of Finger Kinetics and Kinematics via Motoneuron Activities.

Robust neural decoding of intended motor output is crucial to enable intuitive control of assistive devices, such as robotic hands, to perform daily tasks. Few existing neural decoders can predict kinetic and kinematic variables simultaneously. The current study developed a continuous neural decoding approach that can concurrently predict fingertip forces and joint angles of multiple fingers. We obtained motoneuron firing activities by decomposing high-density electromyogram (HD EMG) signals of the extrinsic finger muscles. The identified motoneurons were first grouped and then refined specific to each finger (index or middle) and task (finger force and dynamic movement) combination. The refined motoneuron groups (separate matrix) were then applied directly to new EMG data in real-time involving both finger force and dynamic movement tasks produced by both fingers. EMG-amplitude-based prediction was also performed as a comparison. We found that the newly developed decoding approach outperformed the EMG-amplitude method for both finger force and joint angle estimations with a lower prediction error (Force: 3.47±0.43 vs 6.64±0.69% MVC, Joint Angle: 5.40±0.50° vs 12.8±0.65°) and a higher correlation (Force: 0.75±0.02 vs 0.66±0.05, Joint Angle: 0.94±0.01 vs 0.5±0.05) between the estimated and recorded motor output. The performance was also consistent for both fingers. The developed neural decoding algorithm allowed us to accurately and concurrently predict finger forces and joint angles of multiple fingers in real-time. Our approach can enable intuitive interactions with assistive robotic hands, and allow the performance of dexterous hand skills involving both force control tasks and dynamic movement control tasks.

Motor Unit Identification in the M Waves Recorded by High-Density Electromyogram.

We describe and test the methodology supporting the identification of individual motor unit (MU) firings in the motor response (M wave) to percutaneous nerve stimulation recorded by surface high-density electromyography (HD-EMG) on synthetic and experimental data. A set of simulated voluntary contractions followed by 100 simulated M waves with a normal distribution (MU mean firing latency: 10 ms, Standard Deviation - SDLAT 0.1-1.3 ms) constituted the synthetic signals. In experimental condition, at least 52 progressively increasing M waves were elicited in the soleus muscle of 12 males, at rest (REST), and at 10% (C10) and 20% (C20) of maximal voluntary contraction (MVC). The MU decomposition filters were identified from 15-20 s long isometric plantar flexions performed at 10-70% of MVC and, afterwards, applied to M waves. Synthetic signal analysis demonstrated high accuracy of MU identification in M waves (precision ≥ 85%). In experimental conditions 42.6 ± 11.2 MUs per participant were identified from voluntary contractions. When the MU filters were applied to the M wave recordings, 28.4 ± 14.3, 23.7 ± 14.9 and 20.2 ± 13.5 MU firings were identified in the maximal M waves, with individual MU firing latencies of 10.0 ± 2.8 (SDLAT: 1.2 ± 1.2), 9.6 ± 3.0 (SDLAT: 1.5 ± 1.3) and 10.1 ± 3.7 (SDLAT: 1.7 ± 1.6) ms in REST, C10 and C20 conditions, respectively. We present evidence that supports the feasibility of identifying MU firings in M waves recorded by HD-EMG.

Motor Unit Discharge Characteristics and Conduction Velocity of the Vastii Muscles in Long-Term Resistance-Trained Men.

Adjustments in motor unit (MU) discharge properties have been shown after short-term resistance training; however, MU adaptations in long-term resistance-trained (RT) individuals are less clear. Here, we concurrently assessed MU discharge characteristics and MU conduction velocity in long-term RT and untrained (UT) men. Motor unit discharge characteristics (discharge rate, recruitment, and derecruitment threshold) and MU conduction velocity were assessed after the decomposition of high-density electromyograms recorded from vastus lateralis (VL) and vastus medialis (VM) of RT (>3 yr; n = 14) and UT ( n = 13) during submaximal and maximal isometric knee extension. Resistance-trained men were on average 42% stronger (maximal voluntary force [MVF], 976.7 ± 85.4 N vs 685.5 ± 123.1 N; P < 0.0001), but exhibited similar relative MU recruitment (VL, 21.3% ± 4.3% vs 21.0% ± 2.3% MVF; VM, 24.5% ± 4.2% vs 22.7% ± 5.3% MVF) and derecruitment thresholds (VL, 20.3% ± 4.3% vs 19.8% ± 2.9% MVF; VM, 24.2% ± 4.8% vs 22.9% ± 3.7% MVF; P ≥ 0.4543). There were also no differences between groups in MU discharge rate at recruitment and derecruitment or at the plateau phase of submaximal contractions (VL, 10.6 ± 1.2 pps vs 10.3 ± 1.5 pps; VM, 10.7 ± 1.6 pps vs 10.8 ± 1.7 pps; P ≥ 0.3028). During maximal contractions of a subsample population (10 RT, 9 UT), MU discharge rate was also similar in RT compared with UT (VL, 21.1 ± 4.1 pps vs 14.0 ± 4.5 pps; VM, 19.5 ± 5.0 pps vs 17.0 ± 6.3 pps; P = 0.7173). Motor unit conduction velocity was greater in RT compared with UT individuals in both VL (4.9 ± 0.5 m·s -1 vs 4.5 ± 0.3 m·s -1 ; P < 0.0013) and VM (4.8 ± 0.5 m·s -1 vs 4.4 ± 0.3 m·s -1 ; P < 0.0073). Resistance-trained and UT men display similar MU discharge characteristics in the knee extensor muscles during maximal and submaximal contractions. The between-group strength difference is likely explained by superior muscle morphology of RT as suggested by greater MU conduction velocity.

Myoelectric Control With Fixed Convolution-Based Time-Domain Feature Extraction: Exploring the Spatio–Temporal Interaction

The role of feature extraction in electromyogram (EMG) based pattern recognition has recently been emphasized with several publications promoting deep learning (DL) solutions that outperform traditional methods. It has been shown that the ability of DL models to extract temporal, spatial, and spatio–temporal information provides significant enhancements to the performance and generalizability of myoelectric control. Despite these advancements, it can be argued that DL models are computationally very expensive, requiring long training times, increased training data, and high computational resources, yielding solutions that may not yet be feasible for clinical translation given the available technology. The aim of this paper is, therefore, to leverage the benefits of spatio–temporal DL concepts into a computationally feasible and accurate traditional feature extraction method. Specifically, the proposed novel method extracts a set of well-known time-domain features into a matrix representation, convolves them with predetermined fixed filters, and temporally evolves the resulting features over a short and long-term basis to extract the EMG temporal dynamics. The proposed method, based on Fixed Spatio–Temporal Convolutions, offers significant reductions in the computational costs, while demonstrating a solution that can compete with, and even outperform, recent DL models. Experimental tests were performed on sparse-and high-density EMG (HD-EMG) signals databases, across a total of 44 subjects performing a maximum of 53 movements. Despite the simplification compared to deep approaches, our results show that the proposed solution significantly reduces the classification error rates by 3% to 10% in comparison to recent DL models, while being efficient for real-time implementations.

Neural data-driven model of spinal excitability changes induced by transcutaneous electrical stimulation in spinal cord injury subjects.

The efficacy of trans-spinal direct current stimulation (tsDCS) as neurorehabilitation technology remains sub-optimal, partly due to the variability introduced by subject-specific neurophysiological features and stimulation conditions (e.g. electrode placement, stimulating amplitude, polarity, etc.) Hence, current therapies apply tsDCS in an open-loop fashion, resulting in a lack of standardized protocols for controlling elicited neuronal adaptations in closed-loop. Through the combination of high-density electromyogram (HD-EMG) decomposition, biophysical neuronal modelling and metaheuristic optimization, this work presents a novel neural data-driven framework for estimating subject-specific features and quantifying acute neuronal adaptations elicited by tsDCS on incomplete spinal cord injury subjects. This approach consists of calibrating the anatomical parameters (e.g. soma diameter) of in silico $\alpha-$motoneuron (MN) models for firing similarly to in vivo MNs decoded from HD-EMG. Assuming that cathodal-tsDCS elicits excitability changes in the MN pool, while preserving their anatomical parameters, optimization of an excitability gain common to the entire pool was performed to minimize discrepancies in firing rate and recruitment time between in vivo and in silico MNs after cathodal-tsDCS. This quantification of excitability changes on MN models calibrated in a person specific way enables closing the loop with neuro-modulation devices to tailor neurorehabilitation therapies. Clinical Relevance - This framework addresses a key limitation in non-invasive neuro-modulative technologies via a novel model-assisted framework that enables quantifying acute excitability changes induced on a person-specific in silico MN pool calibrated using in vivo neural data. This will enable the development of advanced controllers for modulating targeted neuronal adaptations in closed-loop.

Startling stimuli increase maximal motor unit discharge rate and rate of force development in humans.

Maximal rate of force development in adult humans is determined by the maximal motor unit discharge rate, however, the origin of the underlying synaptic inputs remains unclear. Here, we tested a hypothesis that the maximal motor unit discharge rate will increase in response to a startling cue, a stimulus that purportedly activates the pontomedullary reticular formation neurons that make mono- and disynaptic connections to motoneurons via fast-conducting axons. Twenty-two men were required to produce isometric knee extensor forces "as fast and as hard" as possible from rest to 75% of maximal voluntary force, in response to visual (VC), visual-auditory (VAC; 80 dB), or visual-startling cue (VSC; 110 dB). Motoneuron activity was estimated via decomposition of high-density surface electromyogram recordings over the vastus lateralis and medialis muscles. Reaction time was significantly shorter in response to VSC compared with VAC and VC. The VSC further elicited faster neuromechanical responses including a greater number of discharges per motor unit per second and greater maximal rate of force development, with no differences between VAC and VC. We provide evidence, for the first time, that the synaptic input to motoneurons increases in response to a startling cue, suggesting a contribution of subcortical pathways to maximal motoneuron output in humans.NEW & NOTEWORTHY Motor unit discharge characteristics are a key determinant of rate of force development in humans, but the neural substrate(s) underpinning such output remains unknown. Using decomposition of high-density electromyogram, we show greater number of discharges per motor unit per second and greater rate of force development after a startling auditory stimulus. These observations suggest a possible subcortical contribution to maximal in vivo motor unit discharge rate in adult humans.

The Effectiveness of Narrowing the Window size for LD & HD EMG Channels based on Novel Deep Learning Wavelet Scattering Transform Feature Extraction Approach.

The use of the Electromyogram (EMG) signals as a source of control to command externally powered prostheses is often challenged by the signal complexity and non-stationary behavior. Mainly, two factors affect classification accuracy: selecting the optimum feature extraction methods and overlapping segmentation/window size. Nowadays, studies attempt to use deep learning (DL) methods to improve classification accuracy. However, DL models are frequently hampered by their requirements of a vast quantity of training data to attain decent performance and the high computing costs. Therefore, researchers tried to replace the deep learning models with other low computational cost methods like deep wavelet scattering transform (DWST) as a feature extraction technique. In terms of windows size, selecting a larger window size increases the classification accuracy, but at the same time, it increases the processing time, which makes the system unsuitable for real-time applications. Accordingly, researchers attempted to minimise the size of the overlapping windows as much as possible without impacting classification performance. This work suggests to utilise DWST transform to achieve two goals (a) extracting the features from EMG signal with low computational cost. Even though many studies have used DWST approaches to extract features from other biological signals, but not been examined before for EMG signals. (b) study the effect of extracting the features from high-density EMG datasets (HD EMG) and low-density EMG datasets (LD EMG), reducing the analysis window size by up to 32msec with minimal impact on classification performance. The outcomes of the proposed method are compared with other well-known feature extraction algorithms to validate these achievements. The proposed strategy exceeds other methods by more than 25% in accuracy.

Compression of EMG Signals Using Deep Convolutional Autoencoders.

Efficient storage and transmission of electromyogram (EMG) data are important for emerging applications such as telemedicine and big data, as a vital tool for further advancement of the field. However, due to limitations in internet speed and hardware resources, transmission and storage of EMG data are challenging. As a solution, this work proposes a new method for EMG data compression using deep convolutional autoencoders (CAE). Eight-channel EMG data from 10 subjects, and high-density EMG data from 18 subjects, were investigated for compression. The CAE architecture was designed to extract an abstract data representation that is heavily compressed, but from which the salient information for classification can be effectively reconstructed. The proposed method attained efficient compression; for CR = 1600, the average PRDN (percentage RMS difference normalized) was 31.5% and the wrist motions classification accuracy (CA) reduced roughly 5%. The CAE substantially outperformed the state-of-the-art high-efficiency video coding and a well-known wavelet-thresholding compression technique. Moreover, by reducing the bit-resolution of the CAE's compressed data from 24 bits to 6 bits, an additional 4-fold compression was achieved without significant degradation of the reconstruction performance. Furthermore, the CAE's inter-subject performance was promising; e.g., for CR = 1600, the PRDN for the inter-subject case was only 2.6% less than that of the within-subject performance. The powerful EMG compression performance with remarkable reconstruction results reflects the CAEs potential as an automatic end-to-end approach with the ability to learn the complete encoding and decoding process. Furthermore, the excellent inter-subject performance demonstrates the generalizability and usability of the proposed approach.

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

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

A generic neural network model to estimate populational neural activity for robust neural decoding

BackgroundRobust and continuous neural decoding is crucial for reliable and intuitive neural-machine interactions. This study developed a novel generic neural network model that can continuously predict finger forces based on decoded populational motoneuron firing activities. MethodWe implemented convolutional neural networks (CNNs) to learn the mapping from high-density electromyogram (HD-EMG) signals of forearm muscles to populational motoneuron firing frequency. We first extracted the spatiotemporal features of EMG energy and frequency maps to improve learning efficiency, given that EMG signals are intrinsically stochastic. We then established a generic neural network model by training on the populational neuron firing activities of multiple participants. Using a regression model, we continuously predicted individual finger forces in real-time. We compared the force prediction performance with two state-of-the-art approaches: a neuron-decomposition method and a classic EMG-amplitude method. ResultsOur results showed that the generic CNN model outperformed the subject-specific neuron-decomposition method and the EMG-amplitude method, as demonstrated by a higher correlation coefficient between the measured and predicted forces, and a lower force prediction error. In addition, the CNN model revealed more stable force prediction performance over time. ConclusionsOverall, our approach provides a generic and efficient continuous neural decoding approach for real-time and robust human-robot interactions.

Adaptive Real-Time Decomposition of Electromyogram During Sustained Muscle Activation: A Simulation Study.

Real-time decomposition of electromyogram (EMG) into constituent motor unit (MU) activity has shown promising applications in neurophysiology and human-machine interactions. Existing decomposition methods could not accommodate stochastic variations in EMG signals such as drifts of action potential amplitudes and MU recruitment-derecruitment (rotation) patterns during long-term recordings. The objective of this study was to develop an adaptive real-time decomposition approach suitable for prolonged muscle activation. We developed a parallel-double-thread computation algorithm. The backend thread initiated and periodically refined and updated the MU information (separation matrix) using independent component analysis and convolution kernel compensation. The frontend thread performed the real-time decomposition. We evaluated our algorithm on synthesized high-density EMG signals, in which MUs were recruited-derecruited sporadically and MU action potentials amplitude drifted over time. Different signal-to-noise levels were also simulated. Compared with the decomposition without the adaptive processes, periodically fine-tuned and updated separation matrix increased identifiable MU number by 3-4 fold over 30-minute of signals. The increased MU number was more prominent at higher signal-to-noise ratios. The decomposition accuracy also increased by up to 10% with greater improvement observed at higher muscle contraction levels. The adaptive algorithm can maintain the decomposition performance over time, allows us to continuously track the same MUs during sustained activation, and, at the same time, can add newly recruited MU information to existing separation matrix. Our approach showed robust performance over time, which has the potential to longitudinally evaluate MU firing and recruitment properties and improve neural decoding performance for neural-machine interactions.

Evaluation of Lumbar Muscle Activation Patterns during Trunk Movements using High-Density Electromyography: A Preliminary Report.

Lumbar paraspinal muscles are heavily involved in daily and work-related activities including trunk bending, trunk twisting, and lifting. Repetitive or inappropriate activation of the lumbar muscles while performing these activities can lead to low back pain. The aim of this preliminary study was to quantify the activation patterns of multiple lumbar muscles when participants performed three different trunk movement tasks, including sustained lumbar flexion posture, dynamic lumbar flexion and extension, and left-right twisting movements. Two 8×8 high-density electromyogram (HD-EMG) electrode arrays were used to record the lumbar muscle activity during these movements. We observed a symmetric and rapid increase in the amplitude of EMG in the erector spinae muscles during the sustained flexion or oscillation tasks. Asymmetric activation patterns were observed in bilateral lumbar muscles during the trunk twisting task. In addition, we observed substantial bilateral co-activation of the lumbar muscles for both twisting directions. These preliminary results demonstrated the potential feasibility of using HD-EMG as a tool to monitor spatial activation patterns of the lumbar muscles during different trunk movements. This approach can also be further developed to assess lumbar muscle function in individuals with low back pain.

HD-sEMG Signal Denoising Method for Improved Classification Performance in Transhumeral Amputees Pros thesis Control.

Surface myoelectric pattern recognition (sMPR) based control strategy is a popularly adopted scheme for multifunctional upper limb prostheses. Meanwhile, above-elbow amputees (transhumeral: TH) usually have limited residual arm muscles, that mostly hinder the provision of requisite signals necessary for physiologically appropriate sMPR control. Hence, the need to maximally explore the limited signals to realize adequate sMPR control scheme in practical settings. This study proposes an effective signal denoising method driven by Multi-scale Local Polynomial Transform (MLPT) concept that can improve the signal quality, thus allowing adequate decoding of TH amputees' motion intent from high-density electromyogram (HD-sEMG) signals. The proposed method's performance was systematically investigated with HD-sEMG signals obtained from TH amputees that performed multiple classes of targeted upper limb movement tasks, and compared with two common signal denoising methods based on wavelet transform. The obtained results show that the proposed MLPT method outperformed both existing methods for motion tasks decoding with over 13.0% increment in accuracy across subjects. The possibility of generating distinct and repeatable myoelectric contraction patterns using the MLPT based denoised HDs-EMG recordings was investigated. The obtained results proved that the MLPT method can better denoise and aid the reconstruction of myoelectric signal patterns of the amputees. Therefore, this suggest the potential of the MLPT method in characterizing high-level upper limb amputees' muscle activation patterns in the context of sMPR prostheses control scheme.

Decoding HD-EMG Signals for Myoelectric Control - How Small Can the Analysis Window Size be?

Recently the use of high-density electromyogram (HD-EMG) signal acquisition setups has been promoted for myoelectric control and several databases have been made open access for scientific research. Despite this accelerated growth in the literature, coupled with industry interest, some fundamental research questions are unanswered. For instance, can HD-EMG signals be decoded at smaller window sizes and, if yes, what is the additional value of sophisticated motor unit decomposition methods. We used three open access databases, with 128 (8 and 65 hand movements) to 256 electrodes (34 hand movements) to explore the impact of varying the number of electrodes and segmentation windows size on EMG decoding accuracy. The analysis considered varying windows sizes from 32 ms to 256 ms and electrodes from 8 to 128/256. Simple time-domain and auto-regressive model parameters were extracted to train a linear discriminant analysis classifier to identify the intended hand motions. Our analysis indicates that with HD-EMG, even with windows sizes as small as 32 ms, one can achieve considerably high decoding accuracy, e.g. <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\ge\!\!90\%$</tex-math></inline-formula> , in our selected databases. As such, HD-EMG setups allow significant reductions in the analysis windows lengths up to 32 ms without compromising decoding performance.

Novel Insights Into Biarticular Muscle Actions Gained From High-Density Electromyogram.

Biarticular muscles have traditionally been considered to exhibit homogeneous neuromuscular activation. The regional activation of biarticular muscles, as revealed from high-density surface electromyograms, seems however to discredit this notion. We thus hypothesize the regional activation of biarticular muscles may contribute to different actions about the joints they span. We then discuss the mechanistic basis and methodological implications underpinning our hypothesis.

On the Reuse of Motor Unit Filters in High Density Surface Electromyograms Recorded at Different Contraction Levels

We analyzed the efficiency of Motor Unit (MU) tracking across different experimentally recorded contraction levels of Biceps Brachii (BB), Tibialis Anterior (TA), First Dorsal Interosseous (FDI) and Abductor Digiti Minimi (ADM) muscles and in simulated conditions. We used the Convolution Kernel Compensation (CKC) algorithm to estimate the MU filters from high-density electromyograms (HDEMGs) at contraction levels ranging from 5 % to 70 % of Maximum Voluntary Contraction (MVC), and applied these MU filters to all the recorded contractions levels of the same muscle. For each MU filter estimation-application pair we assessed the number of identified MUs, Pulse-to-Noise Ratio (PNR), Mean Discharge Rate (MDR) and MUAP peak-to-peak (P2P) amplitude, normalized by HDEMG P2P amplitude. In simulated conditions, we also calculated the Precision ( $Pr$ ) and Sensitivity ( $Se$ ) of MU firing identification. We also reported the number of jointly identified MUs for all possible contraction level pairs. The results in simulated conditions confirmed the efficiency of MU filter transfer from low to high contraction levels. The large majority of MUs identified at lower contraction levels were tracked successfully at higher contraction levels, though many of them were not directly identified at higher contraction levels. Sensitivities of identified MU firings were above 97 %, decreasing only by about 1 % when transferring the MU filters from low to high contraction levels. In the experimental conditions the efficiency of the MU tracking depended significantly on the investigated muscle. The highest efficiency was demonstrated in the FDI muscle, where about 90 % and 70 % of MUs identified at the contraction level of MU filter estimation were also tracked at 10 % and 20 % higher contraction level, respectively. These figures decreased to 47 % and 20 % of MUs in TA, and to 21 % and 18 % of MUs in the BB muscle. The observed differences between the MU filter transfer efficiencies are likely caused by the differences in the size and anatomy of the investigated muscles.