Publications
789 results found
Guo W, Ma C, Wang Z, et al., 2021, Long exposure convolutional memory network for accurate estimation of finger kinematics from surface electromyographic signals., Journal of Neural Engineering, Vol: 18, Pages: 1-12, ISSN: 1741-2552
OBJECTIVE: Estimation of finger kinematics is an important function of an intuitive human-machine interface, such as gesture recognition. Here, we propose a novel deep learning method, named Long Exposure Convolutional Memory Network (LE-ConvMN), and use it to proportionally estimate finger joint angles through surface electromyographic (sEMG) signals. APPROACH: We use a convolution structure to replace the neuron structure of traditional Long Short-Term Memory (LSTM) networks, and use the long exposure data structure which retains the spatial and temporal information of the electrodes as input. The Ninapro database, which contains continuous finger gestures and corresponding sEMG signals was used to verify the efficiency of the proposed deep learning method. The proposed method was compared with LSTM and Sparse Pseudo-input Gaussian Process (SPGP) on this database to predict the 10 main joint angles on the hand based on sEMG. The correlation coefficient (CC) was evaluated using the three methods on eight healthy subjects, and all the methods adopted the root mean square (RMS) features. MAIN RESULTS: The experimental results showed that the average CC, RMSE, NRMSE of the proposed LE-ConvMN method (0.82±0.03,11.54±1.89,0.12±0.013) was significantly higher than SPGP (0.65±0.05, p＜0.001; 15.51±2.82, p＜0.001; 0.16±0.01, p＜0.001) and LSTM (0.64±0.06, p＜0.001; 14.77±3.21, p＜0.001; 0.15±0.02, p=＜0.001). Furthermore, the proposed real-time-estimation method has a computation cost of only approximately 82 ms to output one state of ten joints (average value of 10 tests on TitanV GPU). SIGNIFICANCE: The proposed LE-ConvMN method could efficiently estimate the continuous movement of fingers with sEMG, and its performance is significantly superior to two established deep learning methods.
Enoka RM, Farina D, 2021, Force Steadiness: From Motor Units to Voluntary Actions., Physiology (Bethesda), Vol: 36, Pages: 114-130
Voluntary actions are controlled by the synaptic inputs that are shared by pools of spinal motor neurons. The slow common oscillations in the discharge times of motor units due to these synaptic inputs are strongly correlated with the fluctuations in force during submaximal isometric contractions (force steadiness) and moderately associated with performance scores on some tests of motor function. However, there are key gaps in knowledge that limit the interpretation of differences in force steadiness.
Ting JE, Vecchio AD, Sarma D, et al., 2021, Sensing and decoding the neural drive to paralyzed muscles during attempted movements of a person with tetraplegia using a sleeve array
<jats:title>Abstract</jats:title><jats:p>Motor neurons in the brain and spinal cord convey information about motor intent that can be extracted and interpreted to control assistive devices, such as computers, wheelchairs, and robotic manipulators. However, most methods for measuring the firing activity of single neurons rely on implanted microelectrodes. Although intracortical brain-computer interfaces (BCIs) have been shown to be safe and effective, the requirement for surgery poses a barrier to widespread use. Here, we demonstrate that a wearable sensor array can detect residual motor unit activity in paralyzed muscles after severe cervical spinal cord injury (SCI). Despite generating no observable hand movement, volitional recruitment of motor neurons below the level of injury was observed across attempted movements of individual fingers and overt wrist and elbow movements. Subgroups of motor units were coactive during flexion or extension phases of the task. Single digit movement intentions were classified offline from the EMG power (RMS) or motor unit firing rates with median classification accuracies >75% in both cases. Simulated online control of a virtual hand was performed with a binary classifier to test feasibility of real time extraction and decoding of motor units. The online decomposition algorithm extracted motor units in 1.2 ms, and the firing rates predicted the correct digit motion 88 ± 24% of the time. This study provides the first demonstration of a wearable interface for recording and decoding firing rates of motor neurons below the level of injury in a person with tetraplegia after motor complete SCI.</jats:p><jats:sec><jats:title>Significance Statement</jats:title><jats:p>A wearable electrode array and machine learning methods were used to record and decode myoelectric signals and motor unit firing in paralyzed muscles of a person with motor complete tetraplegia. Motor unit action p
Hug F, Del Vecchio A, Avrillon S, et al., 2021, Muscles from the same muscle group do not necessarily share common drive: evidence from the human triceps surae., Journal of applied physiology (Bethesda, Md. : 1985), Vol: 130, Pages: 342-354, ISSN: 1522-1601
It has been proposed that movements are produced through groups of muscles, or motor modules, activated by common neural commands. However, the neural origin of motor modules is still debated. Here, we used complementary approaches to determine: 1) whether three muscles of the same muscle group [soleus, gastrocnemius medialis (GM), and gastrocnemius lateralis (GL)] are activated by a common neural drive, and 2) whether the neural drive to GM and GL could be differentially modified by altering the mechanical requirements of the task. Eighteen human participants performed an isometric standing heel raise and submaximal isometric plantarflexions (10%, 30%, and 50% of maximal effort). High-density surface electromyography recordings were decomposed into motor unit action potentials and coherence analysis was applied on the motor unit spike trains. We identified strong common drive to each muscle but minimal common drive between the muscles. Further, large between-muscle differences were observed during the isometric plantarflexions, such as a delayed recruitment time of GL compared with GM and soleus motor units and opposite time-dependent changes in the estimates of neural drive to muscles during the torque plateau. Finally, the feet position adopted during the heel-raise task (neutral vs. internally rotated) affected only the GL neural drive with no change for GM. These results provide conclusive evidence that not all anatomically defined synergist muscles are controlled by strong common neural drive. Independent drive to some muscles from the same muscle group may allow for more flexible control to comply with secondary goals such as joint stabilization.NEW & NOTEWORTHY In this study, we demonstrated that the three muscles composing the human triceps surae share minimal common drive during isometric contractions. Our results suggest that reducing the number of effectively controlled degrees of freedom may not always be the strategy used by the central nervous sys
Clarke AK, Atashzar SF, Vecchio AD, et al., 2021, Deep learning for robust decomposition of high-density surface EMG signals, IEEE Transactions on Biomedical Engineering, Vol: 68, Pages: 526-534, ISSN: 0018-9294
Blind source separation (BSS) algorithms, such as gradient convolution kernel compensation (gCKC), can efficiently and accurately decompose high-density surface electromyography (HD-sEMG) signals into constituent motor unit (MU) action potential trains. Once the separation matrix is blindly estimated on a signal interval, it is also possible to apply the same matrix to subsequent signal segments. Nonetheless, the trained separation matrices are sub-optimal in noisy conditions and require that incoming data undergo computationally expensive whitening. One unexplored alternative is to instead use the paired HD-sEMG signal and BSS output to train a model to predict MU activations within a supervised learning framework. A gated recurrent unit (GRU) network was trained to decompose both simulated and experimental unwhitened HD-sEMG signal using the output of the gCKC algorithm. The results on the experimental data were validated by comparison with the decomposition of concurrently recorded intramuscular EMG signals. The GRU network outperformed gCKC at low signal-to-noise ratios, proving superior performance in generalising to new data. Using 12 seconds of experimental data per recording, the GRU performed similarly to gCKC, at rates of agreement of 92.5% (84.5%-97.5%) and 94.9% (88.8%-100.0%) respectively for GRU and gCKC against matched intramuscular sources.
Braecklein M, Ibanez J, Barsakcioglu DY, et al., 2021, Towards human motor augmentation by voluntary decoupling beta activity in the neural drive to muscle and force production, JOURNAL OF NEURAL ENGINEERING, Vol: 18, ISSN: 1741-2560
- Author Web Link
- Cite
- Citations: 11
Avrillon S, Del Vecchio A, Farina D, et al., 2021, Individual differences in the neural strategies to control the lateral and medial head of the quadriceps during a mechanically constrained task., Journal of applied physiology (Bethesda, Md. : 1985), Vol: 130, Pages: 269-281, ISSN: 1522-1601
The interindividual variability in the neural drive sent from the spinal cord to muscles is largely unknown, even during highly constrained motor tasks. Here, we investigated individual differences in the strength of neural drive received by the vastus lateralis (VL) and vastus medialis (VM) during an isometric task. We also assessed the proportion of common neural drive within and between these muscles. Twenty-two participants performed a series of submaximal isometric knee extensions at 25% of their peak torque. High-density surface electromyography recordings were decomposed into motor unit action potentials. Coherence analyses were applied on the motor unit spike trains to assess the degree of neural drive that was shared between motor neurons. Six participants were retested ∼20 mo after the first session. The distribution of the strength of neural drive between VL and VM varied between participants and was correlated with the distribution of normalized interference electromyography (EMG) signals (r > 0.56). The level of within- and between-muscle coherence varied across individuals, with a significant positive correlation between these two outcomes (VL: r = 0.48; VM: r = 0.58). We also observed a large interindividual variability in the proportion of muscle-specific drive, that is, the drive unique to each muscle (VL range: 6%-83%, VM range: 6%-86%). All the outcome measures were robust across sessions, providing evidence that the individual differences did not depend solely on the variability of the measures. Together, these results demonstrate that the neural strategies to control the VL and VM muscles widely vary across individuals, even during a constrained task.NEW & NOTEWORTHY We observed that the distribution of the strength of neural drive between the vastus lateralis and vastus medialis during a single-joint isometric task varied across participants. Also, we observed that the proportion of ne
Urh F, Strnad D, Clarke A, et al., 2021, On the Need for Spatial Whitening of High-Density Surface Electromyograms in Motor Unit Identification by Neural Networks, Pages: 915-922, ISSN: 1680-0737
In recent years, many new decomposition methods for identification of motor unit (MU) firings from high-density surface electromyograms (HDEMG) have been developed, with recent attempts focused on the use of different neural networks (NN). In this study, we evaluated the need for HDEMG signal whitening in NN-based MU identification. For this purpose, we analyzed the learning efficiency of two different types of NN, namely dense NN and long short-term memory (LSTM) NN, on the same HDEMG signals, with and without spatial whitening applied to them. All the HDEMG signals used were simulated with advanced HDEMG simulator, providing a full control of MU firing patterns and MU characteristics in our test environment. Spatial whitening of HDEMG signals significantly improved the precision of MU identification, regardless of the type of NN tested. For dense NN, precision of identified MU increased from 32.2 ± 20.2% to 93.1 ± 8.7%, whereas miss rate decreased from 48.4 ± 23.9% to 12.0 ± 13.3% when whitening of HDEMG signals was employed. For LSTM NN the precision of MU identification increased from 59.7 ± 19.7% to 99.4 ± 2.0% whereas miss rate decreased from 43.1 ± 22.3% to 12.7 ± 9.7% with whitening.
Rahimian E, Zabihi S, Asif A, et al., 2021, FS-HGR: Few-Shot Learning for Hand Gesture Recognition via Electromyography, IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, Vol: 29, Pages: 1004-1015, ISSN: 1534-4320
- Author Web Link
- Cite
- Citations: 39
Tottrup L, Atashzar SF, Farina D, et al., 2020, Nerve injury decreases hyperacute resting-state connectivity between the anterior cingulate and primary somatosensory cortex in anesthetized rats, IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol: 28, Pages: 2691-2698, ISSN: 1534-4320
A better understanding of neural pain processing and of the development of pain over time, is critical to identify objective measures of pain and to evaluate the effect of pain alleviation therapies. One issue is, that the brain areas known to be related to pain processing are not exclusively responding to painful stimuli, and the neuronal activity is also influenced by other brain areas. Functional connectivity reflects synchrony or covariation of activation between groups of neurons. Previous studies found changes in connectivity days or weeks after pain induction. However, less in known on the temporal development of pain. Our objective was therefore to investigate the interaction between the anterior cingulate cortex (ACC) and primary somatosensory cortex (SI) in the hyperacute (minute) and sustained (hours) response in an animal model of neuropathic pain. Intra-cortical local field potentials (LFP) were recorded in 18 rats. In 10 rats the spared nerve injury model was used as an intervention. The intra-cortical activity was recorded before, immediately after, and three hours after the intervention. The interaction was quantified as the calculated correlation and coherence. The results from the intervention group showed a decrease in correlation between ACC and SI activity, which was most pronounced in the hyperacute phase but a longer time frame may be required for plastic changes to occur. This indicated that both SI and ACC are involved in hyperacute pain processing.
Del Vecchio A, Sylos-Labini F, Mondi V, et al., 2020, Spinal motoneurons of the human newborn are highly synchronized during leg movements, SCIENCE ADVANCES, Vol: 6, ISSN: 2375-2548
- Author Web Link
- Cite
- Citations: 27
Gardner M, Mancero Castillo C, Wilson S, et al., 2020, A multimodal intention detection sensor suite for shared autonomy of upper-limb robotic prostheses, Sensors, Vol: 20, ISSN: 1424-8220
Neurorobotic augmentation (e.g., robotic assist) is now in regular use to support individuals suffering from impaired motor functions. A major unresolved challenge, however, is the excessive cognitive load necessary for the human–machine interface (HMI). Grasp control remains one of the most challenging HMI tasks, demanding simultaneous, agile, and precise control of multiple degrees-of-freedom (DoFs) while following a specific timing pattern in the joint and human–robot task spaces. Most commercially available systems use either an indirect mode-switching configuration or a limited sequential control strategy, limiting activation to one DoF at a time. To address this challenge, we introduce a shared autonomy framework centred around a low-cost multi-modal sensor suite fusing: (a) mechanomyography (MMG) to estimate the intended muscle activation, (b) camera-based visual information for integrated autonomous object recognition, and (c) inertial measurement to enhance intention prediction based on the grasping trajectory. The complete system predicts user intent for grasp based on measured dynamical features during natural motions. A total of 84 motion features were extracted from the sensor suite, and tests were conducted on 10 able-bodied and 1 amputee participants for grasping common household objects with a robotic hand. Real-time grasp classification accuracy using visual and motion features obtained 100%, 82.5%, and 88.9% across all participants for detecting and executing grasping actions for a bottle, lid, and box, respectively. The proposed multimodal sensor suite is a novel approach for predicting different grasp strategies and automating task performance using a commercial upper-limb prosthetic device. The system also shows potential to improve the usability of modern neurorobotic systems due to the intuitive control design.
Martinez-Valdes E, Negro F, Falla D, et al., 2020, Inability to increase the neural drive to muscle is associated with task failure during submaximal contractions., J Neurophysiol, Vol: 124, Pages: 1110-1121
We investigated changes in motor unit (MU) behavior and vasti-muscle contractile properties during sustained submaximal fatiguing contractions with a new time-domain tracking technique to understand the mechanisms responsible for task failure. Sixteen participants performed a nonfatiguing 15-s isometric knee extension at 50% of the maximum voluntary (MVC) torque, followed by a 30% MVC sustained contraction until exhaustion. Two grids of 64 surface electromyography electrodes were placed over vastus medialis and lateralis. Signals were decomposed into MU discharge times and the MUs from the 30% MVC sustained contraction were followed until task failure by overlapping decomposition intervals. These MUs were then tracked between 50% and 30% MVC. During the sustained fatiguing contraction, MUs of the two muscles decreased their discharge rate until ∼40% of the endurance time, referred to as the reversal time, and then increased their discharge rate until task failure. This reversal in firing behavior predicted total endurance time and was matched by opposite changes in twitch force (increase followed by a decrease). Despite the later increase in MU firing rates, peak discharge rates at task failure did not reach the frequency attained during a nonfatiguing 50% MVC contraction. These results show that changes in MU firing properties are influenced by adjustments in contractile properties during the course of the contraction, allowing the identification of two phases. Nevertheless, the contraction cannot be sustained, possibly because of progressive motoneuron inhibition/decreased excitability, as the later increase in firing rate saturates at a much lower frequency compared with a higher-force nonfatiguing contraction.NEW & NOTEWORTHY Motor unit firing and contractile properties during a submaximal contraction until failure were assessed with a new tracking technique. Two distinct phases in firing behavior were observed, which compensated for changes in twitch ar
Atashzar SF, Huang H-Y, Duca FD, et al., 2020, Energetic Passivity Decoding of Human Hip Joint for Physical Human-Robot Interaction, IEEE ROBOTICS AND AUTOMATION LETTERS, Vol: 5, Pages: 5953-5960, ISSN: 2377-3766
- Author Web Link
- Cite
- Citations: 6
Nuccio S, Del Vecchio A, Casolo A, et al., 2020, Muscle fiber conduction velocity in the vastus lateralis and medialis muscles of soccer players after ACL reconstruction, Scandinavian Journal of Medicine and Science in Sports, Vol: 30, Pages: 1976-1984, ISSN: 0905-7188
The neural factors underlying the persistency of quadriceps weakness after anterior cruciate ligament reconstruction (ACLR) have been only partially explained. This study examined muscle fiber conduction velocity (MFCV) as an indirect parameter of motor unit recruitment strategies in the vastus lateralis (VL) and medialis (VM) muscles of soccer players with ACLR. High‐density surface electromyography (HDsEMG) was acquired from VL and VM in nine soccer players (22.7 ± 2.9 years; BMI: 22.08 ± 1.72 kg·m−2; 7.7 ± 2.2 months post‐surgery). Voluntary muscle force and the relative myoelectrical activity from the reconstructed and contralateral sides were recorded during linearly increasing isometric knee extension contractions up to 70% of maximal voluntary isometric force (MVIF). The relation of MFCV and force was examined by linear regression analysis at the individual subject level. The initial (intercept), peak (MFCV70), and rate of change (slope) of MFCV related to force were compared between limbs and muscles. The MVIF was lower in the reconstructed side than in the contralateral side (−%20.5; P < .05). MFCV intercept was similar among limbs and muscles (P > .05). MFCV70 and MFCV slope were lower in the reconstructed side compared to the contralateral for both VL (−28.5% and −10.1%, respectively; P < .001) and VM (−22.6% and −8.1%, respectively; P < .001). The slope of MFCV was lower in the VL than VM, but only in the reconstructed side (−12.4%; P < .001). These results suggest possible impairments in recruitment strategies of high‐threshold motor units (HTMUs) as well as deficits in sarcolemmal excitability, fiber diameter, and discharge rate of knee extensor muscles following ACLR.
Pascual Valdunciel A, Gonzalez-Sanchez M, Muceli S, et al., 2020, Intramuscular stimulation of muscle afferents attains prolonged tremor reduction in essential tremor patients., IEEE Transactions on Biomedical Engineering, Vol: PP, ISSN: 0018-9294
This study proposes and clinically tests intramuscular electrical stimulation below motor threshold to achieve prolonged reduction of wrist flexion/extension tremor in Essential Tremor (ET) patients. The developed system consisted of an intramuscular thin-film electrode structure that included both stimulation and electromyography (EMG) recording electrodes, and a control algorithm for the timing of intramuscular stimulation based on EMG (closed-loop stimulation). Data were recorded from nine ET patients with wrist flexion/extension tremor recruited from the Gregorio Maran Hospital (Madrid, Spain). Patients participated in two experimental sessions comprising: 1) sensory stimulation of wrist flexors/extensors via thin-film multichannel intramuscular electrodes; and 2) surface stimulation of the nerves innervating the same target muscles. For each session, four of these patients underwent random 60-s trials of two stimulation strategies for each target muscle: 1) selective and adaptive timely stimulation (SATS) - based on EMG of the antagonist muscle; and 2) continuous stimulation (CON) of target muscles. Two patients underwent SATS stimulation trials alone while the other three underwent CON stimulation trials alone in each session. Kinematics of wrist, elbow, and shoulder, together with clinical scales, were used to assess tremor before, right after, and 24 h after each session. Intramuscular SATS achieved, on average, 32% acute (during stimulation) tremor reduction on each trial, while continuous stimulation augmented tremorgenic activity. Furthermore, tremor reduction was significantly higher using intramuscular than surface stimulation. Prolonged reduction of tremor amplitude (24 h after the experiment) was observed in four patients. These results showed acute and prolonged (24 h) tremor reduction using a minimally invasive neurostimulation technology based on SATS of primary sensory afferents of wrist muscles. This strategy might open the possibility of an alte
Tanzarella S, Muceli S, Del Vecchio A, et al., 2020, Non-invasive analysis of motor neurons controlling the intrinsic and extrinsic muscles of the hand, Journal of Neural Engineering, Vol: 17, ISSN: 1741-2552
OBJECTIVE: We present a non-invasive framework for investigating efferent commands to 14 extrinsic and intrinsic hand muscles. We extend previous studies (limited to a few muscles) on common synaptic input among pools of motor neurons in a large number of muscles. APPROACH: Seven subjects performed sinusoidal isometric contractions to complete seven types of grasps, with each finger and with three combinations of fingers in opposition with the thumb. High-density surface EMG (HD-sEMG) signals (384 channels in total) recorded from the 14 muscles were decomposed into the constituent motor unit action potentials. This provided a non-invasive framework for the investigation of motor neuron discharge patterns, muscle coordination and efferent commands of the hand muscles during grasping. Moreover, during grasping tasks, it was possible to identify common neural information among pools of motor neurons innervating the investigated muscles. For this purpose, principal component analysis (PCA) was applied to the smoothed discharge rates of the decoded motor units. MAIN RESULTS: We found that the first principal component (PC1) of the ensemble of decoded motor neuron spike trains explained a variance of (53.8 ± 10.5) % and was positively correlated with force (R=0.67 ± 0.01 across all subjects and tasks). By grouping the pools of motor neurons from extrinsic or intrinsic muscles, the PC1 explained a proportion of variance of (57.3 ± 10.8) % and (57.9 ± 11.8) %, respectively, and was correlated with force with R=0.61 ± 0.13 and 0.64 ± 0.12, respectively. SIGNIFICANCE: These observations demonstrate a low dimensional control of motor neurons across multiple muscles that can be exploited for extracting control signals in neural interfacing. The proposed framework was designed for hand rehabilitation perspectives, such as post-stroke rehabilitation and hand-exoskeleton control.
Del Vecchio A, Holobar A, Falla D, et al., 2020, Tutorial: Analysis of motor unit discharge characteristics from high-density surface EMG signals, Journal of Electromyography and Kinesiology, Vol: 53, Pages: 1-14, ISSN: 1050-6411
Recent work demonstrated that it is possible to identify motor unit discharge times from high-density surface EMG (HDEMG) decomposition. Since then, the number of studies that use HDEMG decomposition for motor unit investigations has increased considerably. Although HDEMG decomposition is a semi-automatic process, the analysis and interpretation of the motor unit pulse trains requires a thorough inspection of the output of the decomposition result. Here, we report guidelines to perform an accurate extraction of motor unit discharge times and interpretation of the signals. This tutorial includes a discussion of the differences between the extraction of global EMG signal features versus the identification of motor unit activity for physiological investigations followed by a comprehensive guide on how to acquire, inspect, and decompose HDEMG signals, and robust extraction of motor unit discharge characteristics.
Sagastegui Alva PG, Muceli S, Atashzar SF, et al., 2020, Wearable multichannel haptic device for encoding proprioception in the upper limb., Journal of Neural Engineering, Vol: 17, Pages: 1-11, ISSN: 1741-2552
OBJECTIVE: We present the design, implementation, and evaluation of a wearable multichannel haptic system. The device is a wireless closed-loop armband driven by surface electromyography and provides sensory feedback encoding proprioception. The study is motivated by restoring proprioception information in upper limb prostheses. APPROACH: The armband comprises eight vibrotactile actuators that generate distributed patterns of mechanical waves around the limb to stimulate perception and to transfer proportional information on the arm motion. An experimental study was conducted to assess: the sensory threshold in 8 locations around the forearm, the user adaptation to the sensation provided by the device, the user performance in discriminating multiple stimulation levels, and the device performance in coding proprioception using four spatial patterns of stimulation. Eight able-bodied individuals performed reaching tasks by controlling a cursor with an EMG interface in a virtual environment. Vibrotactile patterns were tested with and without visual information on the cursor position with the addition of a random rotation of the reference control system to disturb the natural control and proprioception. RESULTS: The sensation threshold depended on the actuator position and increased over time. The maximum resolution for stimuli discrimination was four. Using this resolution, four patterns of vibrotactile activation with different spatial and magnitude properties were generated to evaluate their performance in enhancing proprioception. The optimal vibration pattern varied among the participants. When the feedback was used in closed-loop control with the EMG interface, the task success rate, completion time, execution efficiency, and average target-cursor distance improved for the optimal stimulation pattern compared to the condition without visual or haptic information on the cursor position. SIGNIFICANCE: The results indicate that the vibrotactile device enhanced the par
Stachaczyk M, Atashzar SF, Farina D, 2020, Adaptive spatial filtering of high-density EMG for reducing the influence of noise and artefacts in myoelectric control, IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol: 28, Pages: 1511-1517, ISSN: 1534-4320
Electromyography (EMG) is a source of neural information for controlling neuroprosthetic devices. To enhance the information content of conventional bipolar EMG, high-density EMG systems include tens to hundreds of closely spaced electrodes that non-invasively record the electrical activity of muscles with high spatial resolution. Despite the advantages of relying on multiple signal sources, however, variations in electrode-skin contact impedance and noise remain challenging for multichannel myocontrol systems. These spatial and temporal non-stationarities negatively impact the control accuracy and therefore substantially limit the clinical viability of high-density EMG techniques. Here, we propose an adaptive algorithm for automatic artefact/noise detection and attenuation for high-density EMG control. The method infers the presence of noise in each EMG channel by spectro-temporal measures of signal similarity. These measures are then used for establishing a scoring system based on an adaptive weighting and reinforcement formulation. The method was experimentally tested as a pre-processing step for a multi-class discrimination problem of 4-digit activation. The approach was proven to enhance the discriminative information content of high-density EMG signals, as well as to attenuate non-stationary artefacts, with improvements in accuracy and robustness of the classification.
Konstantin A, Yu T, Le Carpentier E, et al., 2020, Simulation of motor unit action potential recordings from intramuscular multichannel scanning electrodes, IEEE Transactions on Biomedical Engineering, Vol: 67, Pages: 2005-2014, ISSN: 0018-9294
Multi-channel intramuscular EMG (iEMG) provides information on motor neuron behavior, muscle fiber (MF) innervation geometry and, recently, has been proposed as a means to establish a human-machine interface. Objective: to provide a reliable benchmark for computational methods applied to such recordings, we propose a simulation model for iEMG signals acquired by intramuscular multi-channel electrodes. Methods: we propose several modifications to the existing motor unit action potentials (MUAPs) simulation methods, such as farthest point sampling (FPS) for the distribution of motor unit territory centers in the muscle cross-section, accurate fiber-neuron assignment algorithm, modeling of motor neuron action potential propagation delay, and a model of multi-channel scanning electrode. Results: we provide representative applications of this model to the estimation of motor unit territories and the iEMG decomposition evaluation. Also, we extend it to a full multi-channel iEMG simulator using classic linear EMG modeling. Conclusions: altogether, the proposed models provide accurate MUAPs across the entire motor unit territories and for various electrode configurations. Significance: they can be used for the development and evaluation of mathematical methods for multi-channel iEMG processing and analysis.
Barsakcioglu DY, Bracklein M, Holobar A, et al., 2020, Control of spinal motoneurons by feedback from a non-invasive real-time interface, IEEE Transactions on Biomedical Engineering, Vol: 68, Pages: 926-935, ISSN: 0018-9294
Interfacing with human neural cells during natural tasks provides the means for investigating the working principles of the central nervous system and for developing human-machine interaction technologies. Here we present a computationally efficient non-invasive, real-time interface based on the decoding of the activity of spinal motoneurons from wearable high-density electromyogram (EMG) sensors. We validate this interface by comparing its decoding results with those obtained with invasive EMG sensors and offline decoding, as reference. Moreover, we test the interface in a series of studies involving real-time feedback on the behavior of a relatively large number of decoded motoneurons. The results on accuracy, intuitiveness, and stability of control demonstrate the possibility of establishing a direct non-invasive interface with the human spinal cord without the need for extensive training. Moreover, in a control task, we show that the accuracy in control of the proposed neural interface may approach that of the natural control of force. These results are the first that demonstrate the feasibility and validity of a non-invasive direct neural interface with the spinal cord, with wearable systems and matching the neural information flow of natural movements.
Yu T, Akhmadeev K, Le Carpentier E, et al., 2020, On-line recursive decomposition of intramuscular EMG signals using GPU-implemented bayesian filtering, IEEE Transactions on Biomedical Engineering, Vol: 67, Pages: 1806-1818, ISSN: 0018-9294
Objective: Real-time intramuscular electromyography (iEMG) decomposition, which is needed in biofeedback studies and interfacing applications, is a complex procedure that involves identifying the motor neuron spike trains from a streaming iEMG recording. Methods: We have previously proposed a sequential decomposition algorithm based on a Hidden Markov Model of EMG, which used Bayesian filter to estimate unknown parameters of motor unit (MU) spike trains, as well as their action potentials (MUAPs). Here, we present a modification of this original model in order to achieve a real-time performance of the algorithm as well as a parallel computation implementation of the algorithm on Graphics Processing Unit (GPU). Specifically, the Kalman filter previously used to estimate the MUAPs, is replaced by a least-mean-square filter. Additionally, we introduce a number of heuristics that help to omit the most improbable decomposition scenarios while searching for the best solution. Then, a GPU-implementation of the proposed algorithm is presented. Results: Simulated iEMG signals containing up to 10 active MUs, as well as five experimental fine-wire iEMG signals acquired from the tibialis anterior muscle, were decomposed in real time. The accuracy of decompositions depended on the level of muscle activation, but in all cases exceeded 85%. Conclusion: The proposed method and implementation provide an accurate, real-time interface with spinal motor neurons. Significance: The presented real time implementation of the decomposition algorithm substantially broadens the domain of its application.
Aliakbaryhosseinabadi S, Farina D, Mrachacz-Kersting N, 2020, Real-time neurofeedback is effective in reducing diversion of attention from a motor task in healthy individuals and patients with amyotrophic lateral sclerosis, JOURNAL OF NEURAL ENGINEERING, Vol: 17, ISSN: 1741-2560
- Author Web Link
- Cite
- Citations: 3
Martinez-Valdes E, Negro F, Farina D, et al., 2020, Divergent response of low- versus high-threshold motor units to experimental muscle pain., J Physiol, Vol: 598, Pages: 2093-2108
KEY POINTS: The neural strategies behind the control of force during muscle pain are not well understood as previous research has been limited in assessing pain responses only during low-force contractions. Here we compared, for the first time, the behaviour of motor units recruited at low and high forces in response to pain. The results showed that motor units activated at low forces were inhibited while those recruited at higher forces increased their activity in response to pain. When analysing lower- and higher-threshold motor unit behaviour at high forces we observed differential changes in discharge rate and recruitment threshold across the motor unit pool. These adjustments allow the exertion of high forces in acutely painful conditions but could eventually lead to greater fatigue and stress of the muscle tissue. ABSTRACT: During low-force contractions, motor unit discharge rates decrease when muscle pain is induced by injecting nociceptive substances into the muscle. Despite this consistent observation, it is currently unknown how the central nervous system regulates motor unit behaviour in the presence of muscle pain at high forces. For this reason, we analysed the tibialis anterior motor unit behaviour at low and high forces. Surface EMG signals were recorded from 15 healthy participants (mean age (SD) 26 (3) years, six females) using a 64-electrode grid while performing isometric ankle dorsiflexion contractions at 20% and 70% of the maximum voluntary force (MVC). Signals were decomposed and the same motor units were tracked across painful (intramuscular hypertonic saline injection) and non-painful (baseline, isotonic saline, post-pain) contractions. At 20% MVC, discharge rates decreased significantly in the painful condition (baseline vs. pain: 12.7 (1.1) Hz to 11.5 (0.9) Hz, P < 0.001). Conversely, at 70% MVC, discharge rates increased significantly during pain (baseline vs. pain: 19.7 (2.8) Hz to 21.3 (3.5) Hz, p = 0.029) and recr
Zuo S, Heidari H, Farina D, et al., 2020, Miniaturized magnetic sensors for implantable magnetomyography, Advanced Materials Technologies, Vol: 5, Pages: 1-15, ISSN: 2365-709X
Magnetism‐based systems are widely utilized for sensing and imaging biological phenomena, for example, the activity of the brain and the heart. Magnetomyography (MMG) is the study of muscle function through the inquiry of the magnetic signal that a muscle generates when contracted. Within the last few decades, extensive effort has been invested to identify, characterize and quantify the magnetomyogram signals. However, it is still far from a miniaturized, sensitive, inexpensive and low‐power MMG sensor. Herein, the state‐of‐the‐art magnetic sensing technologies that have the potential to realize a low‐profile implantable MMG sensor are described. The technical challenges associated with the detection of the MMG signals, including the magnetic field of the Earth and movement artifacts are also discussed. Then, the development of efficient magnetic technologies, which enable sensing pico‐Tesla signals, is advocated to revitalize the MMG technique. To conclude, spintronic‐based magnetoresistive sensing can be an appropriate technology for miniaturized wearable and implantable MMG systems.
Hahne JM, Wilke MA, Koppe M, et al., 2020, Longitudinal case study of regression-based hand prosthesis control in daily life, Frontiers in Neuroscience, Vol: 14, Pages: 1-8, ISSN: 1662-453X
Hand prostheses are usually controlled by electromyographic (EMG) signals from the remnant muscles of the residual limb. Most prostheses used today are controlled with very simple techniques using only two EMG electrodes that allow to control a single prosthetic function at a time only. Recently, modern prosthesis controllers based on EMG classification, have become clinically available, which allow to directly access more functions, but still in a sequential manner only. We have recently shown in laboratory tests that a regression-based mapping from EMG signals into prosthetic control commands allows for a simultaneous activation of two functions and an independent control of their velocities with high reliability. Here we aimed to study how such regression-based control performs in daily life in a two-month case study. The performance is evaluated in functional tests and with a questionnaire at the beginning and the end of this phase and compared with the participant’s own prosthesis, controlled with a classical approach. Already 1 day after training of the regression model, the participant with transradial amputation outperformed the performance achieved with his own Michelangelo hand in two out of three functional metrics. No retraining of the model was required during the entire study duration. During the use of the system at home, the performance improved further and outperformed the conventional control in all three metrics. This study demonstrates that the high fidelity of linear regression-based prosthesis control is not restricted to a laboratory environment, but can be transferred to daily use.
Chen C, Ma S, Sheng X, et al., 2020, Adaptive real-time identification of motor unit discharges from non-stationary high-density surface electromyographic signals, IEEE Transactions on Biomedical Engineering, Vol: 67, Pages: 3501-3509, ISSN: 0018-9294
Objective. Estimation of the discharge pattern of motor units by electromyography (EMG) decomposition has been applied for neurophysiologic investigations, clinical diagnosis, and human-machine interfacing. However, most of the methods for EMG decomposition are currently applied offline. Here, we propose an approach for high-density surface EMG decomposition in real-time. Methods. A real-time decomposition scheme including two sessions, offline training and online decomposition, is proposed based on the convolutional kernel compensation algorithm. The estimation parameters, separation vectors and the thresholds for spike extraction, are first computed during offline training, and then they are directly applied to estimate motor unit spike trains (MUSTs) during the online decomposition. The estimation parameters are updated with the identification of new discharges to adapt to non-stationary conditions. The decomposition accuracy was validated on simulated EMG signals by convolving synthetic MUSTs with motor unit action potentials (MUAPs). Moreover, the accuracy of the online decomposition was assessed from experimental signals recorded from forearm muscles using a signal-based performance metrics (pulse-to-noise ratio, PNR). Main results. The proposed algorithm yielded a high decomposition accuracy and robustness to non-stationary conditions. The accuracy of MUSTs identified from simulated EMG signals was > 80% for most conditions. From experimental EMG signals, on average, 12±2 MUSTs were identified from each electrode grid with PNR of 25.0±1.8 dB, corresponding to an estimated decomposition accuracy > 75%. Conclusion and Significance. These results indicate the feasibility of real-time identification of motor unit activities non-invasively during variable force contractions, extending the potential applications of high-density EMG as a neural interface.
Dimitrov H, Bull AMJ, Farina D, 2020, Real-time interface algorithm for ankle kinematics and stiffness from electromyographic signals, IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol: 28, Pages: 1416-1427, ISSN: 1534-4320
Shortcomings in capabilities of below-knee (transtibial) prostheses, compared to their biological counterparts, still cause medical complications and functional deficit to millions of amputees around the world. Although active (powered actuation) transtibial prostheses have the potential to bridge these gaps, the current control solutions limit their efficacy. Here we describe the development of a novel interface for two degrees-of-freedom position and stiffness control for below-knee amputees. The developed algorithm for the interface relies entirely on muscle electrical signals from the lower leg. The algorithm was tested for voluntary position and stiffness control in eight able-bodied and two transtibial amputees and for voluntary stiffness control with foot position estimation while walking in eight able-bodied and one transtibial amputee. The results of the voluntary control experiment demonstrated a promising target reaching success rate, higher for amputees compared to the able-bodied individuals (82.5% and 72.5% compared to 72.5% and 68.1% for the position and position and stiffness matching tasks respectively). Further, the algorithm could provide the means to control four stiffness levels during walking in both amputee and able-bodied individuals while providing estimates of foot kinematics (gait cycle cross-correlation >75% for the sagittal and >90% for the frontal plane and gait cycle root mean square error <7.5° in sagittal and <3° in frontal plane for able-bodied and amputee individuals across three walking speeds). The results from the two experiments demonstrate the feasibility of using this novel algorithm for online control of multiple degrees of freedom and of their stiffness in lower limb prostheses.
Kapelner T, Sartori M, Negro F, et al., 2020, Neuro-Musculoskeletal Mapping for Man-Machine Interfacing., Scientific Reports, Vol: 10, Pages: 5834-5834, ISSN: 2045-2322
We propose a myoelectric control method based on neural data regression and musculoskeletal modeling. This paradigm uses the timings of motor neuron discharges decoded by high-density surface electromyogram (HD-EMG) decomposition to estimate muscle excitations. The muscle excitations are then mapped into the kinematics of the wrist joint using forward dynamics. The offline tracking performance of the proposed method was superior to that of state-of-the-art myoelectric regression methods based on artificial neural networks in two amputees and in four out of six intact-bodied subjects. In addition to joint kinematics, the proposed data-driven model-based approach also estimated several biomechanical variables in a full feed-forward manner that could potentially be useful in supporting the rehabilitation and training process. These results indicate that using a full forward dynamics musculoskeletal model directly driven by motor neuron activity is a promising approach in rehabilitation and prosthetics to model the series of transformations from muscle excitation to resulting joint function.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.