This work introduces a real-time intention decoding algorithm grounded in muscle synergies (Syn-ID). The algorithm detects the electromyographic (EMG) onset and infers the direction of the movement during reaching tasks to control a powered shoulder–elbow exoskeleton. Features related to muscle synergies are used in a Gaussian Mixture Model and probability accumulation-based logic to infer the user’s movement direction. The performance of the algorithm was verified by a feasibility study including eight healthy participants. The experiments comprised a transparent session, during which the exoskeleton did not provide any assistance, and an assistive session in which the Syn-ID strategy was employed. Participants were asked to reach eight targets equally spaced on a circumference of 25 cm radius (adjusted chance level: 18.1%). The results showed an average accuracy of 48.7% after 0.6 s from the EMG onset. Most of the confusion of the estimate was found along directions adjacent to the actual one (type 1 error: 33.4%). Effects of the assistance were observed in a statistically significant reduction in the activation of Posterior Deltoid and Triceps Brachii. The final positions of the movements during the assistive session were on average 1.42 cm far from the expected ones, both when the directions were estimated correctly and when type 1 errors occurred. Therefore, combining accurate estimates with type 1 errors, we computed a modified accuracy of 82.10±6.34%. Results were benchmarked with respect to a purely kinematics-based approach. The Syn-ID showed better performance in the first portion of the movement (0.14 s after EMG onset).

Commonly, quantitative gait analysis post-stroke is performed in fully equipped laboratories housing costly technologies for quantitative evaluation of a patient’s movement capacity. Combining such technologies with an electromyography (EMG)-driven musculoskeletal model can estimate muscle force properties non-invasively, offering clinicians insights into motor impairment mechanisms. However, lab-constrained areas and time-demanding sensor setup and data processing limit the practicality of these technologies in routine clinical care. We presented wearable technology featuring a multi-channel EMG-sensorized garment and an automated muscle localization technique. This allows unsupervised computation of muscle-specific activations, combined with five inertial measurement units (IMUs) for assessing joint kinematics and kinetics during various walking speeds. Finally, the wearable system was combined with a person-specific EMG-driven musculoskeletal model (referred to as human digital twins), enabling the quantitative assessment of movement capacity at a muscle-tendon level. This human digital twin facilitates the estimation of ankle dorsi-plantar flexion torque resulting from individual muscle-tendon forces. Results demonstrate the wearable technology’s capability to extract joint kinematics and kinetics. When combined with EMG signals to drive a musculoskeletal model, it yields reasonable estimates of ankle dorsi-plantar flexion torques (R2 = 0.65 ± 0.21) across different walking speeds for post-stroke individuals. Notably, EMG signals revealing an individual’s control strategy compensate for inaccuracies in IMU-derived kinetics and kinematics when input into a musculoskeletal model. Our proposed wearable technology holds promise for estimating muscle kinetics and resulting joint torque in time-limited and space-constrained environments. It represents a crucial step toward translating human movement biomechanics outside of controlled lab environments for effective motor impairment monitoring.

Stroke causes neurological and physical impairment in millions of people around the world every year. To better comprehend the upper-limb needs and challenges stroke survivors face and the issues associated with existing technology and formulate ideas for a technological solution, the authors conversed with 153 members of the ecosystem (60 neuro patients, 30 caregivers, and 63 medical providers). Patients fell into two populations depending on their upper-limb impairment: spastic (stiff, clenched hands) and flaccid (limp hands). For this work, the authors chose to focus on the second category and developed a set of design constraints based on the information collected through customer discovery. With these in mind, they designed and prototyped a 3D-printed powered wrist–hand grasping orthosis (exoskeleton) to aid in recovery. The orthosis is easily custom-sized based on two parameters and derived anatomical relationships. The researchers tested the prototype on a survivor of stroke and modeled the kinematic behavior of the orthosis with and without load. The prototype neared or exceeded the target design constraints and was able to grasp objects consistently and stably, as well as exercise the patients’ hands. In particular, donning time was only 42 s, as compared to the next fastest time of 3 min reported in literature. This device has the potential for effective neurorehabilitation in a home setting, and it lays the foundation for clinical trials and further device development.

Inertial measurement units (IMUs) have proven to be valuable tools in measuring the range of motion (RoM) of human upper limb joints. Although several studies have reported on the validity of IMUs compared to the gold standard (optical motion capture system, OMC), a quantitative summary of the accuracy of IMUs in measuring RoM of upper limb joints is still lacking. Thus, the primary objective of this systematic review and meta-analysis was to determine the concurrent validity of IMUs for measuring RoM of the upper extremity in adults. Fifty-one articles were included in the systematic review, and data from 16 were pooled for meta-analysis. Concurrent validity is excellent for shoulder flexion–extension (Pearson’s r = 0.969 [0.935, 0.986], ICC = 0.935 [0.749, 0.984], mean difference = −3.19 (p = 0.55)), elbow flexion–extension (Pearson’s r = 0.954 [0.929, 0.970], ICC = 0.929 [0.814, 0.974], mean difference = 10.61 (p = 0.36)), wrist flexion–extension (Pearson’s r = 0.974 [0.945, 0.988], mean difference = −4.20 (p = 0.58)), good to excellent for shoulder abduction–adduction (Pearson’s r = 0.919 [0.848, 0.957], ICC = 0.840 [0.430, 0.963], mean difference = −7.10 (p = 0.50)), and elbow pronation–supination (Pearson’s r = 0.966 [0.939, 0.981], ICC = 0.821 [0.696, 0.900]). There are some inconsistent results for shoulder internal–external rotation (Pearson’s r = 0.939 [0.894, 0.965], mean difference = −9.13 (p < 0.0001)). In conclusion, the results support IMU as a viable instrument for measuring RoM of upper extremity, but for some specific joint movements, such as shoulder rotation and wrist ulnar-radial deviation, IMU measurements need to be used with caution.

Body-worn sensor data is used in monitoring patient activity during rehabilitation and also can be extended to controlling rehabilitation devices based on the activity of the person. The primary focus of research has been on effectively capturing the spatiotemporal dependencies in the data collected by these sensors and efficiently classifying human activities. With the increasing complexity and size of models, there is a growing emphasis on optimizing their efficiency in terms of memory usage and inference time for real-time usage and mobile computers. While hybrid models combining convolutional and recurrent neural networks have shown strong performance compared to traditional approaches, self-attention-based networks have demonstrated even superior results. However, instead of relying on the same transformer architecture, there is an opportunity to develop a novel framework that incorporates recent advancements to enhance speed and memory efficiency, specifically tailored for human activity recognition (HAR) tasks. In line with this approach, we present GLULA, a unique architecture for HAR. GLULA combines gated convolutional networks, branched convolutions, and linear self-attention to achieve efficient and powerful solutions. To enhance the performance of our proposed architecture, we employed manifold mixup as an augmentation variant which proved beneficial in limited data settings. Extensive experiments were conducted on five benchmark datasets: PAMAP2, SKODA, OPPORTUNITY, DAPHNET, and USC-HAD. Our findings demonstrate that GLULA outperforms recent models in the literature on the latter four datasets but also exhibits the lowest parameter count and close to the fastest inference time among state-of-the-art models.

Lower limb spasm and spasticity may develop following spinal cord injury (SCI), causing hyper-excitability and increased tone, which can impact function and quality of life. Pharmaceutical interventions for spasticity may cause unwanted side effects such as drowsiness and weakness. Invasive and non-invasive electrical stimulation has been shown to reduce spasticity without these side effects. The aim of this study was to investigate the effect of sacral afferent stimulation (SAS), through surface electrical stimulation of the dorsal genital nerve (N = 7), and through implanted electrodes on the sacral afferent nerve roots, on lower limb spasm and spasticity (N = 2). Provoked spasms were interrupted with conditional SAS, where stimulation commenced following a provoked spasm, or unconditional stimulation, which was applied continuously. Conditionally and unconditionally applied SAS was shown to suppress acute provoked spasms in people with SCI. There was a statistically significant reduction in area under the curve of quadriceps electromyography during acute spasm with SAS compared to a control spasm. These results show that SAS may provide a safe, low-cost method of reducing acute spasm and spasticity in people living with SCI. SAS through implanted electrodes may also provide an additional function to sacral nerve stimulation devices.

Low-back pain is a common occupational hazard for industrial workers. Several studies show the advantages of using rigid and soft back-support passive exoskeletons and exosuits (exos) to reduce the low-back loading and risk of injury. However, benefits of using these exos have been shown to be task-specific. Therefore, in this study, we developed a benchmarking approach to assess exos for an industrial workplace at Hankamp Gears B.V. We assessed two rigid (Laevo Flex, Paexo back) and two soft (Auxivo Liftsuit 1.0, and Darwing Hakobelude) exos for tasks resembling the workplace. We measured the assistive moment provided by each exo and their respective influence on muscle activity as well as the user’s perception of comfort and exertion. Ten participants performed four lifting tasks (Static hold, Asymmetric, Squat, and Stoop), while their electromyography and subjective measures were collected. The two rigid exos provided the largest assistance during the Dynamic tasks. Reductions in erector spinae activity were seen to be task-specific, with larger reductions for the two rigid exos. Overall, Laevo Flex offered a good balance between assistive moments, reductions in muscle activity, as well as user comfort and reductions in perceived exertion. Thus, we recommend benchmarking exos for intended use in the industrial workplace. This will hopefully result in a better adoption of the back-support exoskeletons in the workplace and help reduce low-back pain.

This paper presents the design and validation of a wearable shoulder exoskeleton robot intended to serve as a platform for assistive controllers that can mitigate the risk of musculoskeletal disorders seen in workers. The design features a four-bar mechanism that moves the exoskeleton’s center of mass from the upper shoulders to the user’s torso, dual-purpose gravity compensation mechanism located inside the four-bar’s linkages that supports the full gravitational loading from the exoskeleton with partial user’s arm weight compensation, and a novel 6 degree-of-freedom (DoF) compliant misalignment compensation mechanism located between the end effector and the user’s arm to allow shoulder translation while maintaining control of the arm’s direction. Simulations show the four-bar design lowers the center of mass by $ 11 $ cm and the kinematic chain can follow the motion of common upper arm trajectories. Experimental tests show the gravity compensation mechanism compensates gravitational loading within $ \pm 0.5 $ Nm over the range of shoulder motion and the misalignment compensation mechanism has the desired 6 DoF stiffness characteristics and range of motion to adjust for shoulder center translation. Finally, a workspace admittance controller was implemented and evaluated showing the system is capable of accurately reproducing simulated impedance behavior with transparent low-impedance human operation.

Transcutaneous spinal cord stimulation (TSCS) is gaining popularity as a noninvasive alternative to epidural stimulation. However, there is still much to learn about its effects and utility in assisting recovery of motor control. In this study, we applied TSCS to healthy subjects concurrently performing a functional training task to study its effects during a training intervention. We first carried out neurophysiological tests to characterize the H-reflex, H-reflex recovery, and posterior root muscle reflex thresholds, and then conducted balance tests, first without TSCS and then with TSCS. Balance tests included trunk perturbations in forward, backward, left, and right directions, and subjects’ balance was characterized by their response to force perturbations. A balance training task involved the subjects playing a catch-and-throw game in virtual reality (VR) while receiving trunk perturbations and TSCS. Balance tests with and without TSCS were conducted after the VR training to measure subjects’ post-training balance characteristics and then neurophysiological tests were carried out again. Statistical comparisons using t-tests between the balance and neurophysiological data collected before and after the VR training intervention found that the immediate effect of TSCS was to increase muscle activity during forward perturbations and to reduce balance performance in that direction. Muscle activity decreased after training and even more once TSCS was turned off. We thus observed an interaction of effects where TSCS increased muscle activity while the physical training decreased it.

Orthotic wrist supports will be beneficial for people with muscular weakness to keep their hand in a neutral rest position and prevent potential wrist contractures. Compensating the weight of the hands is complex since the level of support depends on both wrist and forearm orientations. To explore simplified approaches, two different weight compensation strategies (constant and linear) were compared to the theoretical ideal sinusoidal profile and no compensation in eight healthy subjects using a mechanical wrist support system. All three compensation strategies showed a significant reduction of 47–53% surface electromyography activity in the anti-gravity m. extensor carpi radialis. However, for the higher palmar flexion region, a significant increase of 44–61% in the m. flexor carpi radialis was found for all compensation strategies. No significant differences were observed between the various compensation strategies. Two conclusions can be drawn: (1) a simplified torque profile (e.g., constant or linear) for weight compensation can be considered as equally effective as the theoretically ideal sinusoidal profile and (2) even the theoretically ideal profile provides no perfect support as other factors than weight, such as passive joint impedance, most likely influence the required compensation torque for the wrist joint.

Telerehabilitation and robotics, either traditional rigid or soft, have been extensively studied and used to improve hand functionality after a stroke. However, a limited number of devices combined these two technologies to such a level of maturity that was possible to use them at the patients’ home, unsupervised. Here we present a novel investigation that demonstrates the feasibility of a system that integrates a soft inflatable robotic glove, a cloud-connected software interface, and a telerehabilitation therapy. Ten chronic moderate-to-severe stroke survivors independently used the system at their home for 4 weeks, following a software-led therapy and being in touch with occupational therapists. Data from the therapy, including automatic assessments by the robot, were available to the occupational therapists in real-time, thanks to the cloud-connected capability of the system. The participants used the system intensively (about five times more movements per session than the standard care) for a total of more than 8 hr of therapy on average. We were able to observe improvements in standard clinical metrics (FMA +3.9 ± 4.0, p < .05, COPM-P + 2.5 ± 1.3, p < .05, COPM-S + 2.6 ± 1.9, p < .05, MAL-AOU +6.6 ± 6.5, p < .05) and range of motion (+88%) at the end of the intervention. Despite being small, these improvements sustained at follow-up, 2 weeks after the end of the therapy. These promising results pave the way toward further investigation for the deployment of combined soft robotic/telerehabilitive systems at-home for autonomous usage for stroke rehabilitation.