IEEE Transactions on Neural Systems and Rehabilitation Engineering

Mobility declines with age to the extent that walking speed is often considered a vital sign. Identifying neurological mechanisms behind this decline would aid with early identification and intervention. Electroencephalography (EEG) metrics may provide insight into neural factors contributing to mobility decline with aging. Recent research has shown a differentiation in aperiodic EEG metrics (exponent and offset) across age groups, cognitive abilities, and populations with neurological injury. However, it is unknown if aperiodic EEG differs between mobility tasks or brain regions. The purpose of this study was to 1) compare aperiodic EEG in older and younger adults at rest and while walking and 2) determine if oscillatory and aperiodic EEG in sensorimotor brain regions are predictors of a lower walking speed, regardless of age and other demographic factors. We analyzed EEG collected while participants were sitting at rest and walking on a treadmill in 31 younger adults (age: 24 {+/-} 4, mean {+/-} s.d.) and 59 older adults (age: 74 {+/-} 6). We found that older adults had lower aperiodic exponent and offset at rest and during walking, but only a subset of brain regions showed age group differences. Using machine learning methods, we found that right sensorimotor alpha and left sensorimotor offset and beta had the largest effect on individualized walking speed, after the demographics of age, waist circumference, and sex. These results suggest aperiodic and oscillatory EEG in specific brain areas may give additional insights into brain health.

This paper, for the first time, compares the behaviors of nonlinear versus linear muscle networks in decoding hidden peripheral synergistic neural patterns during dynamic functional tasks. In this paper, we report a case study during which one healthy subject conducts a series of four lower limb repetitive tasks. Specifically, the paper focuses on tasks that involve the right knee joint, including walking, sit-tostand, stepping, and drop-jump. Twelve muscles were recorded using the Delsys Trigno system. The linear muscle network was generated using coherence analysis, and the nonlinear network was generated using Spearmans correlation. The results show that the degree, clustering coefficient, and global efficiency of the muscle network have the highest value among tasks in the linear domain for the walking task, while a low linear synergistic network behavior for the sit-to-stand is observed. On the other hand, the results show that the nonlinear functional muscle network decodes high connectivity (degree) and clustering coefficient and efficiency for the sit-tostand when compared with other tasks. We have also developed a two-dimensional functional connectivity plane composed of linear and nonlinear features and shown that it can span the lower-limb dynamic task space. The results of this paper for the first time highlight the importance of observing both linear and nonlinear connectivity patterns, especially for complex dynamic tasks. It should also be noted that through a simultaneous EEG recording (using BrainVision System), we have shown that, indeed, cortical activity may indirectly explain highly-connected nonlinear muscle network for the sit-to-stand task, highlighting the importance of nonlinear muscle network as a neurophysiological window of observation beyond the periphery.

As an emerging neurostimulation for improving motor rehabilitation, applying vagus nerve stimulation (VNS) after successful movement during training facilitates motor recovery in animals with neuromotor impairment. To translate this procedure to human rehabilitation in a non-invasive, objective, and automated manner, real-time classification of movement quality on a trial-by-trial basis in a minimally constrained state is required. In this work, we developed an integrated closed-loop system using video-based real-time movement classification that can automatically trigger transcutaneous VNS (tVNS) wirelessly as soon as successful movement is detected. We also created a film-like conformable tVNS electrode to be attached over the outer ear. For movement training, we focused on the use case of dance therapy (backward walking), which is widely used for people with Parkinsons disease and older adults. Our markerless video analysis model could detect steps with 0.91 precision and 0.72 recall and classify successful backward steps with a 0.93 F1 score. The classification triggers tVNS through Bluetooth Low Energy communications with a trigger relay device we created. The integrated system enabled real-time automated classification and stimulation, triggering tVNS with 71.3% of the successful movements and taking 2.24 s from video capture to tVNS. We consider our work to be an important step toward patient-driven rehabilitation at home showcasing non-invasive, low-cost, and automated closed-loop neurostimulation technologies.

Recent advancements in data-driven methods have enabled real-time estimation of biomechanical states for exoskeleton control. While biological joint moments can be directly used to scale exoskeleton assistance, this approach is often suboptimal. An optimized mapping between biological joint moments and exoskeleton assistance could enhance end-to-end controllers based on the users physiological state. We introduce a flexible parametrization of biological moment-based control using delay, scaling, and shaping terms to transform joint moment estimates into commanded torque. We performed human-in-the-loop optimization, using metabolic cost to evaluate each iterations controller parameters, for 9 subjects across three ambulation modes: level walking at 1.1 m/s, 1.5 m/s, and 5{degrees} inclined walking. We evaluated three methods of exoskeleton control: 1. Personalized/Task Dependent, 2. Task Dependent/Non-personalized, and 3. Task Agnostic/Non-personalized. On average, our personalized approach provided the greatest benefit of 18.3% reduction in metabolic cost compared to walking without the exoskeleton, with the task dependent and task agnostic controllers producing similar reductions of 8.6% and 8.4%, respectively. Our results show that while generalizable, task agnostic control parameters can improve user energetics across cyclic tasks, fully personalized exoskeleton control parameters yield larger metabolic reductions, highlighting the value of personalizing exoskeleton assistance to users across many diverse tasks.

Robotic-Assisted Gait training (RAGT) offers an innovative therapeutic option for restoration of functional gait in stroke survivors, complementing existing physical rehabilitation strategies. However, there is a limited understanding of the neurophysiological response induced by this training in end-users. Neural desynchronization and Cortico-Muscular Coherence (CMC) are two biomarkers that define the level of muscle-cortex association during gait phases and can be used to estimate induced users adaptation during RAGT. In this study, we measure Event-Related Spectral Perturbation (ERSP) and CMC from three healthy individuals and three stroke survivors during overground-gait with and without an exoskeleton. Results show that (1) the use of the exoskeleton in healthy individuals is associated with a different and more refined motor-control represented in a high{theta} -desynchronization, (2) altered and noisy ERSP and lower and non-focal {beta}-CMC patterns are observed in Stroke patients when performing overground-gait both with and without the Exoskeleton, and (3) Exoskeleton use in stroke survivors is associated with a reduction in swing-time during gait-cycle, but this effect is not correlated with an increment of{theta} -desynchronization and/or {beta}-CMC. ERSP and CMC demonstrated evidence of neural modulation in able-bodied users during RAGT, which could not be detected in subacute stroke survivors during RAGT. These results suggest that the gait-parameters changes observed during exoskeleton use in subacute stroke survivors are unlikely to be neurally driven.

Human-in-the-loop optimization (HILO) is an established method for identifying subject-specific optimal controllers for performance augmentation. For HILO algorithms to be useful in rehabilitation, however, the optimization algorithm may need to account for how the human response changes over time in response to assistance. In this study, we tested a modified version of Bayesian optimization (BO), dynamic Bayesian optimization (DBO), in a three-parameter optimization problem that sought to identify participant-specific optimal solutions for increasing walking speed. As opposed to BO, DBO accounts for the non-stationarity of human responses. Sixteen healthy participants received bilateral hip torque pulses delivered by a hip exoskeleton. The exoskeleton torque parameters were determined using HILO with either DBO or BO. Validation iterations were introduced to objectively compare performance across optimizers at different time points of HILO. The results showed that both DBO and BO significantly increased walking speed compared to baseline. When comparing performance between DBO and BO, DBO emerged as an improvement over BO both in terms of efficacy, modeling accuracy, and personalization. DBO induced changes in walking speed relative to baseline that exceeded those induced by BO at three of the four validation iterations. DBO outperformed BO in modeling accuracy in later validation iterations. DBO personalization induced changes in walking speed that were significantly greater than those induced by previously identified assistive solutions, while this was not the case of BO. Overall, our results indicate that DBO outperformed BO due to its greater ability to account for non-stationary aspects of the human response.

Wearable inertial measurement units (IMUs) provide a practical and objective approach for gait assessment in clinical populations. Although several handcrafted gait features have been proposed, these features may not fully capture the multidimensional signal characteristics associated with different pathological gait patterns. This study proposes a digital biomarker called Embedding-Distance Gait Biomarker (EDGB) based on supervised contrastive representation learning of wearable IMU signals. A compact multi-input convolutional neural network is developed to encode raw acceleration, angular velocity, and their temporal derivatives into a 32-dimensional latent representation. Class-specific prototypes are computed from the training embeddings of healthy, neurological, and orthopedic participants. The proposed EDGB is then derived from the distances between each trial embedding and the learned group prototypes. The proposed architecture is evaluated on the publicly available Voisard clinical gait dataset using a subject-level split, with 20% of participants held out for testing to prevent leakage across repeated trials. On unseen test subjects, the proposed biomarker distinguished healthy from neurological, healthy from orthopedic, and neurological from orthopedic gait patterns with AUCs of 90.59%, 88.47%, and 99.50%, respectively. The biomarker also demonstrated a large group effect, with clinical category explaining 71% of its variance. Reliability analysis showed significant consistency across repeated trials, with an ICC (2,1) of 0.82, indicating that most variability reflected between-subject differences rather than within-subject trial-to-trial fluctuations.

Neural activities within the beta frequency range (13-30 Hz) are not stationary, but occur in transient packets known as beta bursts. Parkinsons disease (PD) is characterized by the occurrence of beta bursts of increased duration and amplitude within the cortico-basal ganglia network. The pathophysiological importance of beta bursts is exemplified by the fact that they serve as a clinically useful feedback signal in beta amplitude triggered adaptive Deep Brain Stimulation (aDBS). Prolonged duration beta bursts are closely associated with motor impairments in PD, whilst bursts of shorter duration may have a physiological role. Consequently, we aimed to develop a deep learning-based pipeline capable of predicting long (> 150ms) and short (< 150ms) duration beta bursts from subthalamic nucleus local field potential (LFP) recordings. Our approach achieved promising accuracy values of 87% and 85.2% in two patients implanted with a DBS device that was capable of long-term wireless LFP sensing. Our findings highlight the feasibility of prolonged beta burst prediction and could inform the development of a new type of intelligent DBS approach with the capability of delivering stimulation only during the occurrence of prolonged bursts.

The velocity vector field (flow) controller is a well-established control strategy for lower limb exoskeletons. In this paper, we analyze this controller and propose modifications to improve its performance. We demonstrate that flow control acts as a variable proportional-derivative error regulator, where the parameter {Gamma} represents the desired norm of the hip-knee joint velocity vector (path speed). Based on this, we introduce two modifications to {Gamma}: (1) a constant {Gamma} set to the mean desired path speed, and (2) a variable {Gamma} that mimics natural path speed during unassisted walking. We compared the modified flow controllers with a slow-{Gamma} version in experiments involving seven participants walking on a treadmill at 0.6m/s, 0.8m/s, and 1.0m/s. Compared to the slow-{Gamma} controller, the RMS tracking error decreased by 30.7{+/-} 11.3% and the range of motion of the knee increased by 48.2 {+/-} 5.5% for the mean-{Gamma} controller, while the variable-{Gamma} controller had 32.4 {+/-} 14.7% smaller RMS error and 50.5 {+/-} 6.5% larger range of motion of the knee. Additionally, the slow-{Gamma} controller consistently applied resistive power, whereas participants reported more comfortable and natural gait with the modified controllers. We also compared them with the original tuning of flow controller, with results indicating superior performance from the proposed modifications. These findings demonstrate effectiveness across different walking speeds and offer a tuning strategy for future flow controller use.

Falls are a leading cause of injury and growing healthcare care costs, particularly in aging populations, yet subtle balance impairments often go undetected until severe problems arise. Accurately quantifying human balance has remained a critical challenge. Here, we introduce perturbation recovery time, a novel balance metric inspired by nonlinear dynamic system theory that quantifies the duration required for post-perturbation deviations to return to steady-state gait consistently. Unlike traditional assessments that rely on regular walking patterns, we used external perturbations to uncover hidden balance mechanisms that may not appear during steady-state walking. We identified key kinematic balance features, including the anterior-posterior distance between the center of mass and the center of pressure, whole-body angular momentum in the frontal and sagittal planes, and vertical center of mass position and acceleration, which encode significant balance information and enhance the sensitivity of the metric. These features reinforce the importance of foot placement, inverted pendulum dynamics, push-off control, and trunk motion to maintain human balance. The perturbation recovery time metric effectively identified subtle balance changes caused by controlled artificial impairments of sensing, actuation, and consistency. These findings demonstrate that the recovery process following external walking perturbations encodes critical information for quantifying human balance, offering a promising approach for early detection of balance impairments.

The purpose of this study is to develop a deep neural network model for predicting soleus muscle activation patterns from time-series lower-limb joint angles collected during level ground walking at different speeds and to evaluate the prediction performance. An open dataset was used to obtain full lower-limb kinematics, including pelvis, hip, knee, and ankle joints, and soleus muscle activation patterns from 20 control adults. Long short-term memory (LSTM)-based deep learning models were developed and then evaluated for prediction performance by conducting both the random split cross-validation (CVM1) and the leave-one-subject-out cross-validation (CVM2). For both cross-validation methods, the developed models yielded promising error and regression metrics such as the root mean square error and coefficient of determination. However, the CVM2 demonstrated that the prediction performance can be sensitive to individual datasets. The subject factors, such as age, sex, and walking speed, appear to have a negligible effect on the prediction performance for the CVM1. This study demonstrated the feasibility of the developed models to be a template for a potential tool that quantifies muscle activation patterns from joint angles during level ground walking at different speeds.

Robotic exoskeletons require human control and decision making to switch between different locomotion modes, which can be inconvenient and cognitively demanding. To support the development of automated locomotion mode recognition systems (i.e., high-level controllers), we designed an environment recognition system using computer vision and deep learning. We collected over 5.6 million images of indoor and outdoor real-world walking environments using a wearable camera system, of which ~923,000 images were annotated using a 12-class hierarchical labelling architecture (called the ExoNet database). We then trained and tested the EfficientNetB0 convolutional neural network, designed for efficiency using neural architecture search, to predict the different walking environments. Our environment recognition system achieved ~73% image classification accuracy. While these preliminary results benchmark Efficient-NetB0 on the ExoNet database, further research is needed to compare different image classification algorithms to develop an accurate and real-time environment-adaptive locomotion mode recognition system for robotic exoskeleton control.

The movement-related cortical potential (MRCP) is a low-frequency component of the electroencephalography (EEG) signal recorded from the motor cortex and its neighboring cortical areas. Since the MRCP encodes motor intention and execution, it may be utilized as an interface between patients and neurorehabilitation technologies. This study investigates the EEG signal recorded from the Cz electrode to discriminate between four levels of rate of force development (RFD) of the tibialis anterior muscle. For classification, three feature sets were evaluated to describe the EEG traces. These were (i) MRCP morphological characteristics in the{delta} -band such as amplitude and timing, (ii) MRCP statistical characteristics in the{delta} -band such as mean, standard deviation, and kurtosis, and (iii) wideband time-frequency features in the 0.5-90 Hz range. Using a support vector machine for classification, the four levels of RFD were classified with a mean (SD) accuracy of 82% (7%) accuracy when using the time-frequency feature space, and with an accuracy of 75% (12%) when using the MRCP statistical characteristics. It was also observed that some of the key features from the statistical and morphological sets responded monotonically to the intensity of the RFD. Examples are slope and standard deviation in the (0, 1)s window for the statistical, and min1 and minn for the morphological sets. This monotonical response of features explains the observed performance of the{delta} -band MRCP and corresponding high discriminative power. Results from temporal analysis considering the pre-movement phase ((-3, 0)s) and three windows of the post-movement phase ((0, 1)s, (1, 2)s, and (2, 3)s)) suggest that the complete MRCP waveform represents high information content regarding the planning, execution, duration, and ending of the isometric dorsiflexion task using the tibialis anterior muscle. Results shed light on the role of{delta} -band in translating to motor command, with potential applications in neural engineering systems.

After stroke, paretic arm muscles are constantly exposed to abnormal neural drive from the injured brain. As such, hypertonia, broadly defined as an increase in muscle tone, is prevalent especially in distal muscles, which impairs daily function or in long-term leads to a flexed resting posture in the wrist and fingers. However, there currently is no quantitative measure that can reliably track how hypertonia is expressed on daily basis. In this study, we propose a novel time-based surface electromyography (sEMG) measure that can overcome the limitations of the coarse clinical scales often measured in functionally irrelevant context and the magnitude-based sEMG measures that suffer from signal non-stationarity. We postulated that the key to robust quantification of hypertonia is to capture the "true" baseline in sEMG for each measurement session, by which we can define the relative duration of activity over a short time segment continuously tracked in a sliding window fashion. We validate that the proposed measure of sEMG active duration is robust across parameter choices (e.g., sampling rate, window length, threshold criteria), robust against typical noise sources present in paretic muscles (e.g., low signal-to-noise ratio, sporadic motor unit action potentials), and reliable across measurements (e.g., sensors, trials, and days), while providing a continuum of scale over the full magnitude range for each session. Furthermore, sEMG active duration could well characterize the clinically observed differences in hypertonia expressed across different muscles and impairment levels. The proposed measure can be used for continuous and quantitative monitoring of hypertonia during activities of daily living while at home, which will allow for the study of the practical effect of pharmacological and/or physical interventions that try to combat its presence.

Real-time brain-computer interfaces (BCIs) that decode electroencephalograms (EEG) during motor imagery (MI) are a powerful adjunct to rehabilitation therapy after neurotrauma. Immersive virtual reality (VR) could complement BCIs by delivering multisensory feedback congruent to the users MI, enabling therapies that engage users in task-oriented scenarios. Yet, therapeutic outcomes rely on the users proficiency in evoking MI to attain volitional BCI-commanded VR interaction. While previous studies suggested that users could improve BCI-evoked MI within a single session, the effects of multiple training sessions on sensorimotor neuromodulation remain unknown. Here, we present a longitudinal study assessing the impact of VR-mediated BCI training on lower-limb sensorimotor neuromodulation, wherein an EEG-based BCI was coupled with congruent real-time multisensory feedback in immersive VR. We show that unimpaired individuals could learn to modulate their sensorimotor activations during MI virtual walking over multiple training sessions, also resulting in increased BCI control accuracy. Additionally, when extending the system to immersive VR cycling, four individuals with chronic complete spinal cord injury (SCI) showed similar improvements. This is the first study demonstrating that individuals could learn modulating sensorimotor activity associated with MI using BCI integrated with immersive VR over multiple training sessions, even after SCI-induced motor and sensory decline. These results suggest that VR-BCI training may facilitate neuroplasticity, potentially strengthening sensorimotor pathways and functional connectivity relevant to motor control and recovery.

BackgroundThe gait pattern results from a complex interaction of several body parts, orchestrated by the (central) nervous system that controls the active and passive systems of the body. An impairment of gait due to a stroke results in a decline in quality of life and independence. Setting up efficient gait training requires an objective and wholesome assessment of the patients movement pattern to target individual gait alterations. However, current assessment tools are limited in their ability to capture the complexity of the movement and the amount of data acquired during gait analysis. AimsIn this study, we explore the potential of variational autoencoders (VAE) to learn and recognise different gait patterns within both, pathologic and healthy gait. MethodsFor this purpose, the lower-limb joint angles of 71 participants (29 stroke survivors, 42 healthy controls) were used to train and test a VAE. ResultsThe good reconstruction results (range r = 0.52 - 0.91, average normalized RMSE 23.36 % {+/-} 4.13) indicate that VAEs extract meaningful information from the gait pattern. Furthermore, the extracted latent features are sensitive enough to distinguish between the gait patterns of stroke survivors and a healthy cohort (p<0.001). ConclusionsThe presented approach allows the assessment of gait data in an objective and wholesome manner, thereby integrating the individual characteristics of each persons gait, making it a suitable tool for monitoring the progress of rehabilitation efforts.

Despite the common focus of gait in rehabilitation, there are few tools that allow quantitatively characterizing gait in the clinic. We recently described an algorithm, trained on a large dataset from our clinical gait analysis laboratory, which produces accurate cycle-by-cycle estimates of spatiotemporal gait parameters including step timing and walking velocity. Here, we demonstrate this system generalizes well to clinical care with a validation study on prosthetic users seen in therapy and outpatient clinic. Specifically, estimated walking velocity was similar to annotated 10-meter walking velocities, and cadence and foot contact times closely mirrored our wearable sensor measurements. Additionally, we found that a 2D keypoint detector pre-trained on largely able-bodied individuals struggles to localize prosthetic joints, particularly for those individuals with more proximal or bilateral amputations, but it is possible to train a prosthetic-specific joint detector. Further work is required to validate the other outputs from our algorithm including sagittal plane joint angles and step length. Code and trained weights will be released upon publication.

For exoskeletons to be successful in real-world settings, they will need to be effective across a variety of terrains, including on inclines. While some single-joint exoskeletons have assisted incline walking, recent successes in level-ground assistance suggest that greater improvements may be possible by optimizing assistance of the whole leg. To understand how exoskeleton assistance should change with incline, we used human-in-the-loop optimization to find whole-leg exoskeleton assistance torques that minimized metabolic cost on a range of grades. We optimized assistance for three expert, able-bodied participants on 5 degree, 10 degree and 15 degree inclines using a hip-knee-ankle exoskeleton emulator. For all assisted conditions, the cost of transport was reduced by at least 50% relative to walking in the device with no assistance, a large improvement to walking that is comparable to the benefits of whole-leg assistance on level-ground. This corresponds to large absolute reductions in metabolic cost, with the most strenuous conditions reduced by 4.9 W/kg, more than twice the entire energy cost of level walking. Optimized extension torque magnitudes and exoskeleton power increased with incline, with hip extension, knee extension and ankle plantarflexion often growing as large as allowed by comfort-based limits. Applied powers on steep inclines were double the powers applied during level-ground walking, indicating that larger exoskeleton power may be optimal in scenarios where biological powers and costs are higher. Future exoskeleton devices can be expected to deliver large improvements in walking performance across a range of inclines, if they have sufficient torque and power capabilities.

A detailed comprehension of the central and peripheral processes underlying walking is essential to develop effective therapeutic interventions to slow down gait decline with age, and rehabilitation strategies to maximize motor recovery for patients with damages at the central nervous system. The combined use of electromyography (EMG) and electroencephalography (EEG), in the framework of coherence analysis, has recently established for neuromotor integrity/impairment assessment. In this study, we propose corticomuscular (EEG-EMG) and inter/intramuscular (EMG-EMG) coherences as measures of the cognitive reserve, i.e., the process whereby a wider repertoire of cognitive strategies, as well as more flexible and efficient strategies, can moderate the manifestation of brain disease/damage. We recorded EEG signals from the main brain source locations and superficial EMG signals from the main leg muscles involved in gait in 16 healthy young adults (age [&le;]30 years) and 13 healthy elderly (age [&ge;]65 years) during three different overground walking conditions (i.e., spontaneous walking, walking with cognitive dual-task, and walking with targets drawn on the floor). In all conditions, we calculated corticomuscular and inter/intramuscular coherences. We observed higher corticomuscular and inter/intramuscular coherences during targeted walking compared to spontaneous walking in both groups, even if the increase was greater in young people. Considering dual-task walking compared to spontaneous walking, only corticomuscular coherence in the elderly increased. These results suggest age-related differences in cognitive reserve that reflect different abilities to perform complex cognitive or motor tasks during gait. This study demonstrates the feasibility, repeatability, and effectiveness of the proposed method to investigate brain-to-muscle connectivity during different gait conditions, to study the related changes with age, and to quantify the cognitive reserve.

For a seamless integration of electroencephalography (EEG)-based motor imagery brain-computer interfaces (MI-BCIs), it is vital to be able to classify movements of the same joint. However, a fundamental challenge in classifying the same joint movements arises from the close spatial proximity of the corresponding brain regions. To address this challenge, we explore the feasibility of distinguishing up and down movements specific to the right upper limb using multiple frequency bands combined with a channel averaging method. Six electrodes positioned in close proximity to the motor cortex and two distinct frequency bands: mu (8-12Hz) and beta (12-30Hz) were selected. This isolates and enhances electromagnetic activity in the brain commonly associated with motor and cognitive processing. The results of our study revealed promising outcomes across two classification methods. Utilizing a support vector machine (SVM) classifier, our proposed approach achieved an average accuracy of 59.3% and a k-nearest neighbor(KNN) classifier approach yielded an average accuracy of 61.63% in distinguishing between upward and downward movements of the right arm. These results demonstrate the potential of combining spatially focused EEG acquisition with frequency-specific analysis for improved MI-BCI performance.