An Integrated Approach for Real-Time Monitoring of Knee Dynamics with IMUs and Multichannel EMG.

Measuring human joint dynamics is crucial for understanding how our bodies move and function, providing valuable insights into biomechanics and motor control. Cerebral palsy (CP) is a neurological disorder affecting motor control and posture, leading to diverse gait abnormalities, including altered knee angles. The accurate measurement and analysis of knee angles in individuals with CP are crucial for understanding their gait patterns, assessing treatment outcomes, and guiding interventions. This paper presents a novel multimodal approach that combines inertial measurement unit (IMU) sensors and electromyography (EMG) to measure knee angles in individuals with CP during gait and other daily activities. We discuss the performance of this integrated approach, highlighting the accuracy of IMU sensors in capturing knee joint movements when compared with an optical motion-tracking system and the complementary insights offered by EMG in assessing muscle activation patterns. Moreover, we delve into the technical aspects of the developed device. The presented results show that the angle measurement error falls within the reported values of the state-of-the-art IMU-based knee joint angle measurement devices while enabling a high-quality EMG recording over prolonged periods of time. While the device was designed and developed primarily for measuring knee activity in individuals with CP, its usability extends beyond this specific use-case scenario, making it suitable for applications that involve human joint evaluation.

Evaluation of Simple Algorithms for Proportional Control of Prosthetic Hands Using Intramuscular Electromyography.

Although seemingly effortless, the control of the human hand is backed by an elaborate neuro-muscular mechanism. The end result is typically a smooth action with the precise positioning of the joints of the hand and an exerted force that can be modulated to enable precise interaction with the surroundings. Unfortunately, even the most sophisticated technology cannot replace such a comprehensive role but can offer only basic hand functionalities. This issue arises from the drawbacks of the prosthetic hand control strategies that commonly rely on surface EMG signals that contain a high level of noise, thus limiting accurate and robust multi-joint movement estimation. The use of intramuscular EMG results in higher quality signals which, in turn, lead to an improvement in prosthetic control performance. Here, we present the evaluation of fourteen common/well-known algorithms (mean absolute value, variance, slope sign change, zero crossing, Willison amplitude, waveform length, signal envelope, total signal energy, Teager energy in the time domain, Teager energy in the frequency domain, modified Teager energy, mean of signal frequencies, median of signal frequencies, and firing rate) for the direct and proportional control of a prosthetic hand. The method involves the estimation of the forces generated in the hand by using different algorithms applied to iEMG signals from our recently published database, and comparing them to the measured forces (ground truth). The results presented in this paper are intended to be used as a baseline performance metric for more advanced algorithms that will be made and tested using the same database.

Learning regularized representations of categorically labelled surface EMG enables simultaneous and proportional myoelectric control.

BACKGROUND: Processing the surface electromyogram (sEMG) to decode movement intent is a promising approach for natural control of upper extremity prostheses. To this end, this paper introduces and evaluates a new framework which allows for simultaneous and proportional myoelectric control over multiple degrees of freedom (DoFs) in real-time. The framework uses multitask neural networks and domain-informed regularization in order to automatically find nonlinear mappings from the forearm sEMG envelope to multivariate and continuous encodings of concurrent hand- and wrist kinematics, despite only requiring categorical movement instruction stimuli signals for calibration. METHODS: Forearm sEMG with 8 channels was collected from healthy human subjects (N = 20) and used to calibrate two myoelectric control interfaces, each with two output DoFs. The interfaces were built from (I) the proposed framework, termed Myoelectric Representation Learning (MRL), and, to allow for comparisons, from (II) a standard pattern recognition framework based on Linear Discriminant Analysis (LDA). The online performances of both interfaces were assessed with a Fitts's law type test generating 5 quantitative performance metrics. The temporal stabilities of the interfaces were evaluated by conducting identical tests without recalibration 7 days after the initial experiment session. RESULTS: Metric-wise two-way repeated measures ANOVA with factors method (MRL vs LDA) and session (day 1 vs day 7) revealed a significant ([Formula: see text]) advantage for MRL over LDA in 5 out of 5 performance metrics, with metric-wise effect sizes (Cohen's [Formula: see text]) separating MRL from LDA ranging from [Formula: see text] to [Formula: see text]. No significant effect on any metric was detected for neither session nor interaction between method and session, indicating that none of the methods deteriorated significantly in control efficacy during one week of intermission. CONCLUSIONS: The results suggest that MRL is able to successfully generate stable mappings from EMG to kinematics, thereby enabling myoelectric control with real-time performance superior to that of the current commercial standard for pattern recognition (as represented by LDA). It is thus postulated that the presented MRL approach can be of practical utility for muscle-computer interfaces.

Electrotactile Feedback for the Discrimination of Different Surface Textures Using a Microphone.

Most commercial prosthetic hands lack closed-loop feedback, thus, a lot of research has been focusing on implementing sensory feedback systems to provide the user with sensory information during activities of daily living. This study evaluates the possibilities of using a microphone and electrotactile feedback to identify different textures. A condenser microphone was used as a sensor to detect the friction sound generated from the contact between different textures and the microphone. The generated signal was processed to provide a characteristic electrical stimulation presented to the participants. The main goal of the processing was to derive a continuous and intuitive transfer function between the microphone signal and stimulation frequency. Twelve able-bodied volunteers participated in the study, in which they were asked to identify the stroked texture (among four used in this study: Felt, sponge, silicone rubber, and string mesh) using only electrotactile feedback. The experiments were done in three phases: (1) Training, (2) with-feedback, (3) without-feedback. Each texture was stroked 20 times each during all three phases. The results show that the participants were able to differentiate between different textures, with a median accuracy of 85%, by using only electrotactile feedback with the stimulation frequency being the only variable parameter.

Contactless Real-Time Heartbeat Detection via 24 GHz Continuous-Wave Doppler Radar Using Artificial Neural Networks.

The measurement of human vital signs is a highly important task in a variety of environments and applications. Most notably, the electrocardiogram (ECG) is a versatile signal that could indicate various physical and psychological conditions, from signs of life to complex mental states. The measurement of the ECG relies on electrodes attached to the skin to acquire the electrical activity of the heart, which imposes certain limitations. Recently, due to the advancement of wireless technology, it has become possible to pick up heart activity in a contactless manner. Among the possible ways to wirelessly obtain information related to heart activity, methods based on mm-wave radars proved to be the most accurate in detecting the small mechanical oscillations of the human chest resulting from heartbeats. In this paper, we presented a method based on a continuous-wave Doppler radar coupled with an artificial neural network (ANN) to detect heartbeats as individual events. To keep the method computationally simple, the ANN took the raw radar signal as input, while the output was minimally processed, ensuring low latency operation (<1 s). The performance of the proposed method was evaluated with respect to an ECG reference ("ground truth") in an experiment involving 21 healthy volunteers, who were sitting on a cushioned seat and were refrained from making excessive body movements. The results indicated that the presented approach is viable for the fast detection of individual heartbeats without heavy signal preprocessing.

A decision support system for electrode shaping in multi-pad FES foot drop correction.

BACKGROUND: Functional electrical stimulation (FES) can be applied as an assistive and therapeutic aid in the rehabilitation of foot drop. Transcutaneous multi-pad electrodes can increase the selectivity of stimulation; however, shaping the stimulation electrode becomes increasingly complex with an increasing number of possible stimulation sites. We described and tested a novel decision support system (DSS) to facilitate the process of multi-pad stimulation electrode shaping. The DSS is part of a system for drop foot treatment that comprises a custom-designed multi-pad electrode, an electrical stimulator, and an inertial measurement unit. METHODS: The system was tested in ten stroke survivors (3-96 months post stroke) with foot drop over 20 daily sessions. The DSS output suggested stimulation pads and parameters based on muscle twitch responses to short stimulus trains. The DSS ranked combinations of pads and current amplitudes based on a novel measurement of the quality of the induced movement and classified them based on the movement direction (dorsiflexion, plantar flexion, eversion and inversion) of the paretic foot. The efficacy of the DSS in providing satisfactory pad-current amplitude choices for shaping the stimulation electrode was evaluated by trained clinicians. The range of paretic foot motion was used as a quality indicator for the chosen patterns. RESULTS: The results suggest that the DSS output was highly effective in creating optimized FES patterns. The position and number of pads included showed pronounced inter-patient and inter-session variability; however, zones for inducing dorsiflexion and plantar flexion within the multi-pad electrode were clearly separated. The range of motion achieved with FES was significantly greater than the corresponding active range of motion (p < 0.05) during the first three weeks of therapy. CONCLUSIONS: The proposed DSS in combination with a custom multi-pad electrode design covering the branches of peroneal and tibial nerves proved to be an effective tool for producing both the dorsiflexion and plantar flexion of a paretic foot. The results support the use of multi-pad electrode technology in combination with automatic electrode shaping algorithms for the rehabilitation of foot drop. TRIAL REGISTRATION: This study was registered at the Current Controlled Trials website with ClinicalTrials.gov ID NCT02729636 on March 29, 2016.

Feasibility of a hybrid brain-computer interface for advanced functional electrical therapy.

We present a feasibility study of a novel hybrid brain-computer interface (BCI) system for advanced functional electrical therapy (FET) of grasp. FET procedure is improved with both automated stimulation pattern selection and stimulation triggering. The proposed hybrid BCI comprises the two BCI control signals: steady-state visual evoked potentials (SSVEP) and event-related desynchronization (ERD). The sequence of the two stages, SSVEP-BCI and ERD-BCI, runs in a closed-loop architecture. The first stage, SSVEP-BCI, acts as a selector of electrical stimulation pattern that corresponds to one of the three basic types of grasp: palmar, lateral, or precision. In the second stage, ERD-BCI operates as a brain switch which activates the stimulation pattern selected in the previous stage. The system was tested in 6 healthy subjects who were all able to control the device with accuracy in a range of 0.64-0.96. The results provided the reference data needed for the planned clinical study. This novel BCI may promote further restoration of the impaired motor function by closing the loop between the "will to move" and contingent temporally synchronized sensory feedback.

Inferring position of motor units from high-density surface EMG.

The spatial distribution of muscle fibre activity is of interest in guiding therapy and assessing recovery of motor function following injuries of the peripheral or central nervous system. This paper presents a new method for stable estimation of motor unit territory centres from high-density surface electromyography (HDsEMG). This completely automatic process applies principal component compression and a rotatable Gaussian surface fit to motor unit action potential (MUAP) distributions to map the spatial distribution of motor unit activity. Each estimated position corresponds to the signal centre of the motor unit territory. Two subjects were used to test the method on forearm muscles, using two different approaches. With the first dataset, motor units were identified by decomposition of intramuscular EMG and the centre position of each motor unit territory was estimated from synchronized HDsEMG data. These positions were compared to the positions of the intramuscular fine wire electrodes with depth measured using ultrasound. With the second dataset, decomposition and motor unit localization was done directly on HDsEMG data, during specific muscle contractions. From the first dataset, the mean estimated depth of the motor unit centres were 8.7, 11.6, and 9.1 mm, with standard deviations 0.5, 0.1, and 1.3 mm, and the respective depths of the fine wire electrodes were 8.4, 15.8, and 9.1 mm. The second dataset generated distinct spatial distributions of motor unit activity which were used to identify the regions of different muscles of the forearm, in a 3-dimensional and projected 2-dimensional view. In conclusion, a method is presented which estimates motor unit centre positions from HDsEMG. The study demonstrates the shifting spatial distribution of muscle fibre activity between different efforts, which could be used to assess individual muscles on a motor unit level.

Analysis of modulations of mental fatigue on intra-individual variability from single-trial event related potentials.

BACKGROUND: Intra-individual variability (IIV), a measure of variance within an individual's performance, has been demonstrated as metrics of brain responses for neural functionality. However, how mental fatigue modulates IIV remains unclear. Consequently, the development of robust mental fatigue detection methods at the single-trial level is challenging. NEW METHODS: Based on a long-duration flanker task EEG dataset, the modulations of mental fatigue on IIV were explored in terms of response time (RT) and trial-to-trial latency variations of event-related potentials (ERPs). Specifically, latency variations were quantified using residue iteration decomposition (RIDE) to reconstruct latency-corrected ERPs. We compared reconstructed ERPs with raw ERPs by means of temporal principal component analysis (PCA). Furthermore, a single-trial classification pipeline was developed to detect the changes of mental fatigue levels. RESULTS: We found an increased IIV in the RT metric in the fatigue state compared to the alert state. The same sequence of ERPs (N1, P2, N2, P3a, P3b, and slow wave, or SW) was separated from both raw and reconstructed ERPs using PCA, whereas differences between raw and reconstructed ERPs in explained variances for separated ERPs were found owing to IIV. Particularly, a stronger N2 was detected in the fatigue than alert state after RIDE. The single-trial fatigue detection pipeline yielded an acceptable accuracy of 73.3%. COMPARISON WITH EXISTING METHODS: The IIV has been linked to aging and brain disorders, and as an extension, our finding demonstrates IIV as an efficient indicator of mental fatigue. CONCLUSIONS: This study reveals significant modulations of mental fatigue on IIV at the behavioral and neural levels and establishes a robust mental fatigue detection pipeline.

Surface-distributed low-frequency asynchronous stimulation delays fatigue of stimulated muscles.

INTRODUCTION: One important reason why functional electrical stimulation (FES) has not gained widespread clinical use is the limitation imposed by rapid muscle fatigue due to non-physiological activation of the stimulated muscles. We aimed to show that asynchronous low-pulse-rate (LPR) electrical stimulation applied by multipad surface electrodes greatly postpones the occurrence of muscle fatigue compared with conventional stimulation (high pulse rate, HPR). METHODS: We compared the produced force vs. time of the forearm muscles responsible for finger flexion in 2 stimulation protocols, LPR (fL = 10 Hz) and HPR (fH = 40 Hz). RESULTS: Surface-distributed low-frequency asynchronous stimulation (sDLFAS) doubles the time interval before the onset of fatigue (104 ± 80%) compared with conventional synchronous stimulation. CONCLUSIONS: Combining the performance of multipad electrodes (increased selectivity and facilitated positioning) with sDLFAS (decreased fatigue) can improve many FES applications in both the lower and upper extremities.

Long term sustained attention alters dynamic functional connectivity patterns.

Mental fatigue has attracted much attention from researchers as it plays a key role in performance efficiency and safety situations. Functional connectivity analysis using graph theory is an effective method for revealing changes in cognition resources influenced by mental fatigue. Previous studies have revealed that functional networks are dynamically reorganized. Therefore, it is critical to explore dynamic timescales of networks related to specific cognitive abilities. In this study, we used an open EEG dataset of twenty-one subjects recorded in a 60-minutes sustained attention task. After preprocessing, we constructed connectivity matrices using the weighted phase lag index (wPLI) in the theta band and characterized them with dynamic graph measures, namely characteristic path length (CPL) and clustering coefficient (CC). The results show that the frontal-parietal brain networks in theta band are involved in a sustaining attention task. When averaging from temporal and spatial activations, CPL and CC decreased with time-on-task. Our results indicate that mental fatigue results in deteriorations in sustaining attention, and graph theory analysis can provide support for mental fatigue analysis.Clinical Relevance- Identification of the effects of long term sustained attention on dynamic brain networks may be potential for mechanism study and detection of mental states and attentional deficits caused by mental diseases.

Compressed spike-triggered averaging in iterative decomposition of surface EMG.

BACKGROUND AND OBJECTIVE: Analysis of motor unit activity is important for assessing and treating diseases or injuries affecting natural movement. State-of-the-art decomposition translates high-density surface electromyography (HDsEMG) into motor unit activity. However, current decomposition methods offer far from complete separation of all motor units. METHODS: This paper proposes a peel-off approach to automatic decomposition of HDsEMG into motor unit action potential (MUAP) trains, based on the Fast Independent Component Analysis algorithm (FastICA). The novel steps include utilizing compression by means of Principal Component Analysis and spike-triggered averaging, to estimate surface MUAP distributions with less noise, which are iteratively subtracted from the HDsEMG dataset. Furthermore, motor unit spike trains are estimated by high-dimensional density-based clustering of peaks in the FastICA source output. And finally, a new reliability measure is used to discard poor motor unit estimates by comparing the variance of the FastICA source output before and after the peel-off step. The method was validated using reconstructed synthetic data at three different signal-to-noise levels and was compared to an established deflationary FastICA approach. RESULTS: Both algorithms had very high recall and precision, over 90%, for spikes from matching motor units, referred to as matched performance. However, the peel-off algorithm correctly identified more motor units for all noise levels. When accounting for unidentified motor units, total recall was up to 33 percentage points higher; and when accounting for duplicate estimates, total precision was up to 24 percentage points higher, compared to the state-of-the-art reference. In addition, a comparison was done using experimental data where the proposed algorithm had a matched recall of 97% and precision of 85% with respect to the reference algorithm. CONCLUSION: These results show a substantial performance increase for decomposition of simulated HDsEMG data and serve to validate the proposed approach. This performance increase is an important step towards complete decomposition and extraction of information of motor unit activity.

A fast blind source separation algorithm for decomposing ultrafast ultrasound images into spatiotemporal muscle unit kinematics.

Objective.Ultrasound can detect individual motor unit (MU) activity during voluntary isometric contractions based on their subtle axial displacements. The detection pipeline, currently performed offline, is based on displacement velocity images and identifying the subtle axial displacements. This identification can preferably be made through a blind source separation (BSS) algorithm with the feasibility of translating the pipeline fromofflinetoonline. However, the question remains how to reduce the computational time for the BSS algorithm, which includes demixing tissue velocities from many different sources, e.g. the active MU displacements, arterial pulsations, bones, connective tissue, and noise.Approach.This study proposes a fast velocity-based BSS (velBSS) algorithm suitable for online purposes that decomposes velocity images from low-force voluntary isometric contractions into spatiotemporal components associated with single MU activities. The proposed algorithm will be compared against spatiotemporal independent component analysis (stICA), i.e. the method used in previous papers, for various subjects, ultrasound- and EMG systems, where the latter acts as MU reference recordings.Main results. We found that the computational time for velBSS was at least 20 times less than for stICA, while the twitch responses and spatial maps extracted from stICA and velBSS for the same MU reference were highly correlated (0.96 ± 0.05 and 0.81 ± 0.13).Significance.The present algorithm (velBSS) is computationally much faster than the currently available method (stICA) while maintaining the same performance. It provides a promising translation towards an online pipeline and will be important in the continued development of this research field of functional neuromuscular imaging.

The rubber hand illusion evaluated using different stimulation modalities.

Tactile feedback plays a vital role in inducing ownership and improving motor control of prosthetic hands. However, commercially available prosthetic hands typically do not provide tactile feedback and because of that the prosthetic user must rely on visual input to adjust the grip. The classical rubber hand illusion (RHI) where a brush is stroking the rubber hand, and the user's hidden hand synchronously can induce ownership of a rubber hand. In the classic RHI the stimulation is modality-matched, meaning that the stimulus on the real hand matches the stimulus on the rubber hand. The RHI has also been used in previous studies with a prosthetic hand as the "rubber hand," suggesting that a hand prosthesis can be incorporated within the amputee's body scheme. Interestingly, previous studies have shown that stimulation with a mismatched modality, where the rubber hand was brushed, and vibrations were felt on the hidden hand also induced the RHI. The aim of this study was to compare how well mechanotactile, vibrotactile, and electrotactile feedback induced the RHI in able-bodied participants and forearm amputees. 27 participants with intact hands and three transradial amputees took part in a modified RHI experiment. The rubber hand was stroked with a brush, and the participant's hidden hand/residual limb received stimulation with either brush stroking, electricity, pressure, or vibration. The three latter stimulations were modality mismatched with regard to the brushstroke. Participants were tested for ten different combinations (stimulation blocks) where the stimulations were applied on the volar (glabrous skin), and dorsal (hairy skin) sides of the hand. Outcome was assessed using two standard tests (questionnaire and proprioceptive drift). All types of stimulation induced RHI but electrical and vibration stimulation induced a stronger RHI than pressure. After completing more stimulation blocks, the proprioceptive drift test showed that the difference between pre- and post-test was reduced. This indicates that the illusion was drifting toward the rubber hand further into the session.

Exploration of sensations evoked during electrical stimulation of the median nerve at the wrist level.

Objective.Nerve rehabilitation following nerve injury or surgery at the wrist level is a lengthy process during which not only peripheral nerves regrow towards receptors and muscles, but also the brain undergoes plastic changes. As a result, at the time when nerves reach their targets, the brain might have already allocated some of the areas within the somatosensory cortex that originally processed hand signals to some other regions of the body. The aim of this study is to show that it is possible to evoke a variety of somatotopic sensations related to the hand while stimulating proximally to the injury, therefore, providing the brain with the relevant inputs from the hand regions affected by the nerve damage.Approach.This study included electrical stimulation of 28 able-bodied participants where an electrode that acted as a cathode was placed above the Median nerve at the wrist level. The parameters of electrical stimulation, amplitude, frequency, and pulse shape, were modulated within predefined ranges to evaluate their influence on the evoked sensations.Main results.Using this methodology, the participants reported a wide variety of somatotopic sensations from the hand regions distal to the stimulation electrode.Significance.Furthermore, to propose an accelerated stimulation tuning procedure that could be implemented in a clinical protocol and/or standalone device for providing meaningful sensations to the somatosensory cortex during nerve regeneration, we trained machine-learning techniques using the gathered data to predict the location/area, naturalness, and sensation type of the evoked sensations following different stimulation patterns.

A multi-pad electrode based functional electrical stimulation system for restoration of grasp.

BACKGROUND: Functional electrical stimulation (FES) applied via transcutaneous electrodes is a common rehabilitation technique for assisting grasp in patients with central nervous system lesions. To improve the stimulation effectiveness of conventional FES, we introduce multi-pad electrodes and a new stimulation paradigm. METHODS: The new FES system comprises an electrode composed of small pads that can be activated individually. This electrode allows the targeting of motoneurons that activate synergistic muscles and produce a functional movement. The new stimulation paradigm allows asynchronous activation of motoneurons and provides controlled spatial distribution of the electrical charge that is delivered to the motoneurons. We developed an automated technique for the determination of the preferred electrode based on a cost function that considers the required movement of the fingers and the stabilization of the wrist joint. The data used within the cost function come from a sensorized garment that is easy to implement and does not require calibration. The design of the system also includes the possibility for fine-tuning and adaptation with a manually controllable interface. RESULTS: The device was tested on three stroke patients. The results show that the multi-pad electrodes provide the desired level of selectivity and can be used for generating a functional grasp. The results also show that the procedure, when performed on a specific user, results in the preferred electrode configuration characteristics for that patient. The findings from this study are of importance for the application of transcutaneous stimulation in the clinical and home environments.

Comparing algorithms for assessing upper limb use with inertial measurement units.

The various existing measures to quantify upper limb use from wrist-worn inertial measurement units can be grouped into three categories: 1) Thresholded activity counting, 2) Gross movement score and 3) machine learning. However, there is currently no direct comparison of all these measures on a single dataset. While machine learning is a promising approach to detecting upper limb use, there is currently no knowledge of the information used by machine learning measures and the data-related factors that influence their performance. The current study conducted a direct comparison of the 1) thresholded activity counting measures, 2) gross movement score,3) a hybrid activity counting and gross movement score measure (introduced in this study), and 4) machine learning measures for detecting upper-limb use, using previously collected data. Two additional analyses were also performed to understand the nature of the information used by machine learning measures and the influence of data on the performance of machine learning measures. The intra-subject random forest machine learning measure detected upper limb use more accurately than all other measures, confirming previous observations in the literature. Among the non-machine learning (or traditional) algorithms, the hybrid activity counting and gross movement score measure performed better than the other measures. Further analysis of the random forest measure revealed that this measure used information about the forearm's orientation and amount of movement to detect upper limb use. The performance of machine learning measures was influenced by the types of movements and the proportion of functional data in the training/testing datasets. The study outcomes show that machine learning measures perform better than traditional measures and shed some light on how these methods detect upper-limb use. However, in the absence of annotated data for training machine learning measures, the hybrid activity counting and gross movement score measure presents a reasonable alternative. We believe this paper presents a step towards understanding and optimizing measures for upper limb use assessment using wearable sensors.

Smart Protocols for Physical Therapy of Foot Drop Based on Functional Electrical Stimulation: A Case Study.

Functional electrical stimulation (FES) is used for treating foot drop by delivering electrical pulses to the anterior tibialis muscle during the swing phase of gait. This treatment requires that a patient can walk, which is mostly possible in the later phases of rehabilitation. In the early phase of recovery, the therapy conventionally consists of stretching exercises, and less commonly of FES delivered cyclically. Nevertheless, both approaches minimize patient engagement, which is inconsistent with recent findings that the full rehabilitation potential could be achieved by an active psycho-physical engagement of the patient during physical therapy. Following this notion, we proposed smart protocols whereby the patient sits and ankle movements are FES-induced by self-control. In six smart protocols, movements of the paretic ankle were governed by the non-paretic ankle with different control strategies, while in the seventh voluntary movements of the paretic ankle were used for stimulation triggering. One stroke survivor in the acute phase of recovery participated in the study. During the therapy, the patient's voluntary ankle range of motion increased and reached the value of normal gait after 15 sessions. Statistical analysis did not reveal the differences between the protocols in FES-induced movements.

End-to-End Estimation of Hand- and Wrist Forces From Raw Intramuscular EMG Signals Using LSTM Networks.

Processing myoelectrical activity in the forearm has for long been considered a promising framework to allow transradial amputees to control motorized prostheses. In spite of expectations, contemporary muscle-computer interfaces built for this purpose typically fail to satisfy one or more important desiderata, such as accuracy, robustness, and/or naturalness of control, in part due to difficulties in acquiring high-quality signals continuously outside laboratory conditions. In light of such problems, surgically implanted electrodes have been made a viable option that allows for long-term acquisition of intramuscular electromyography (iEMG) measurements of spatially precise origin. As it stands, the question of how information embedded in such signals is best extracted and combined across multiple channels remains open. This study presents and evaluates an approach to this end that uses deep neural networks based on the Long Short-Term Memory (LSTMs) architecture to regress forces exerted by multiple degrees of freedom (DoFs) from multichannel iEMG. Three deep learning models, representing three distinct regression strategies, were evaluated: (I) One-to-One, wherein each DoF is separately estimated by an LSTM model processing a single iEMG channel, (II) All-to-One, wherein each DoF is separately estimated by an LSTM model processing all iEMG channels, and (III) All-to-All, wherein a single LSTM model with access to all iEMG channels estimates all DoFs simultaneously. All models operate on raw iEMG, with no preliminary feature extraction required. When evaluated on a dataset comprising six iEMG channels with concurrent force measurements acquired from 14 subjects, all LSTM strategies were found to significantly outperform a baseline feature-based linear control regression method. This finding indicates that recurrent neural networks can learn to transform raw forearm iEMG signals directly into representations that correlate with forces exerted at the level of the hand to a greater degree than simple features do. Furthermore, the All-to-All and All-to-One strategies were found to exhibit better performance than the One-to-One strategy. This finding suggests that, in spite of the spatially local nature of signals, iEMG from muscles not directly actuating the relevant DoF can provide contextual information that aid in decoding motor intent.

Can Deep Synthesis of EMG Overcome the Geometric Growth of Training Data Required to Recognize Multiarticulate Motions?

By being predicated on supervised machine learning, pattern recognition approaches to myoelectric prosthesis control require electromyography (EMG) training data collected concurrently with every detectable motion. Within this framework, calibration protocols for simultaneous control of multifunctional prosthetic hands rapidly become prohibitively long-the number of unique motions grows geometrically with the number of controllable degrees of freedom (DoFs). This paper proposes a technique intended to circumvent this combinatorial explosion. Using EMG windows from 1-DoF motions as input and EMG windows from 2-DoF motions as targets, we train generative deep learning models to synthesize EMG windows appertaining to multi-DoF motions. Once trained, such models can be used to complete datasets consisting of only 1-DoF motions, enabling simple calibration protocols with durations that scale linearly with the number of DoFs. We evaluated synthetic EMG produced in this way via a classification task using a database of forearm surface EMG collected during 1-DoF and 2-DoF motions. Multi-output classifiers were trained on either (I) real data from 1-DoF and 2-DoF motions, (II) real data from only 1-DoF motions, or (III) real data from 1-DoF motions appended with synthetic EMG from 2-DoF motions. When tested on data containing all possible motions, classifiers trained on synthetic-appended data (III) significantly outperformed classifiers trained on 1-DoF real data (II), although significantly underperformed classifiers trained on both 1- and 2-DoF real data (I) (I < 0.05). These findings suggest that it is feasible to model EMG concurrent with multiarticulate motions as nonlinear combinations of EMG from constituent 1-DoF motions, and that such modelling can be harnessed to synthesize realistic training data.