Machine learning algorithms that use data streams captured from soft wearable sensors have the potential to automatically detect PD symptoms and inform clinicians about the progression of disease. However, these algorithms must be trained with annotated data from clinical experts who can recognize symptoms, and collecting such data are costly. Understanding how many sensors and how much labeled data are required is key to successfully deploying these models outside of the clinic. Here we recorded movement data using 6 flexible wearable sensors in 20 individuals with PD over the course of multiple clinical assessments conducted on 1 day and repeated 2 weeks later. Participants performed 13 common tasks, such as walking or typing, and a clinician rated the severity of symptoms (bradykinesia and tremor). We then trained convolutional neural networks and statistical ensembles to detect whether a segment of movement showed signs of bradykinesia or tremor based on data from tasks performed by other individuals. Our results show that a single wearable sensor on the back of the hand is sufficient for detecting bradykinesia and tremor in the upper extremities, whereas using sensors on both sides does not improve performance. Increasing the amount of training data by adding other individuals can lead to improved performance, but repeating assessments with the same individuals—even at different medication states—does not substantially improve detection across days. Our results suggest that PD symptoms can be detected during a variety of activities and are best modeled by a dataset incorporating many individuals.

Robots offer an alternative, potentially advantageous method of providing repetitive, high-dosage, and high-intensity training to address the gait impairments caused by stroke. In this study, we compared the effects of the Stride Management Assist (SMA®) System, a new wearable robotic device developed by Honda R&D Corporation, Japan, with functional task specific training (FTST) on spatiotemporal gait parameters in stroke survivors. A single blinded randomized control trial was performed to assess the effect of FTST and task-specific walking training with the SMA® device on spatiotemporal gait parameters. Participants (n = 50) were randomly assigned to FTST or SMA. Subjects in both groups received training 3 times per week for 6–8 weeks for a maximum of 18 training sessions. The GAITRite® system was used to collect data on subjects’ spatiotemporal gait characteristics before training (baseline), at mid-training, post-training, and at a 3-month follow-up. After training, significant improvements in gait parameters were observed in both training groups compared to baseline, including an increase in velocity and cadence, a decrease in swing time on the impaired side, a decrease in double support time, an increase in stride length on impaired and non-impaired sides, and an increase in step length on impaired and non-impaired sides. No significant differences were observed between training groups; except for SMA group, step length on the impaired side increased significantly during self-selected walking speed trials and spatial asymmetry decreased significantly during fast-velocity walking trials. SMA and FTST interventions provided similar, significant improvements in spatiotemporal gait parameters; however, the SMA group showed additional improvements across more parameters at various time points. These results indicate that the SMA® device could be a useful therapeutic tool to improve spatiotemporal parameters and contribute to improved functional mobility in stroke survivors. Further research is needed to determine the feasibility of using this device in a home setting vs a clinic setting, and whether such home use provides continued benefits. This study is registered under the title “Development of walk assist device to improve community ambulation” and can be located in clinicaltrials.gov with the study identifier: NCT01994395 .

Background: Atypical walking in the months and years after stroke constrain community reintegration and reduce mobility, health, and quality of life. The ReWalk ReStore™ is a soft robotic exosuit designed to assist the propulsion and ground clearance subtasks of post-stroke walking by actively assisting paretic ankle plantarflexion and dorsiflexion. Previous proof-of-concept evaluations of the technology demonstrated improved gait mechanics and energetics and faster and farther walking in users with post-stroke hemiparesis. We sought to determine the safety, reliability, and feasibility of using the ReStore™ during post-stroke rehabilitation. Methods: A multi-site clinical trial (NCT03499210) was conducted in preparation for an application to the United States Food and Drug Administration (FDA). The study included 44 users with post-stroke hemiparesis who completed up to 5 days of training with the ReStore™ on the treadmill and over ground. In addition to primary and secondary endpoints of safety and device reliability across all training activities, an exploratory evaluation of the effect of multiple exposures to using the device on users’ maximum walking speeds with and without the device was conducted prior to and following the five training visits. Results: All 44 study participants completed safety and reliability evaluations. Thirty-six study participants completed all five training days. No device-related falls or serious adverse events were reported. A low rate of device malfunctions was reported by clinician-operators. Regardless of their reliance on ancillary assistive devices, after only 5 days of walking practice with the device, study participants increased both their device-assisted (Δ: 0.10 ± 0.03 m/s) and unassisted (Δ: 0.07 ± 0.03 m/s) maximum walking speeds (P’s < 0.05). Conclusions: When used under the direction of a licensed physical therapist, the ReStore™ soft exosuit is safe and reliable for use during post-stroke gait rehabilitation to provide targeted assistance of both paretic ankle plantarflexion and dorsiflexion during treadmill and overground walking. Trial registration: NCT03499210. Prospectively registered on March 28, 2018.

Recent advancements in deep learning have produced significant progress in markerless human pose estimation, making it possible to estimate human kinematics from single camera videos without the need for reflective markers and specialized labs equipped with motion capture systems. Such algorithms have the potential to enable the quantification of clinical metrics from videos recorded with a handheld camera. Here we used DeepLabCut, an open-source framework for markerless pose estimation, to fine-tune a deep network to track 5 body keypoints (hip, knee, ankle, heel, and toe) in 82 below-waist videos of 8 patients with stroke performing overground walking during clinical assessments. We trained the pose estimation model by labeling the keypoints in 2 frames per video and then trained a convolutional neural network to estimate 5 clinically relevant gait parameters (cadence, double support time, swing time, stance time, and walking speed) from the trajectory of these keypoints. These results were then compared to those obtained from a clinical system for gait analysis (GAITRite®, CIR Systems). Absolute accuracy (mean error) and precision (standard deviation of error) for swing, stance, and double support time were within 0.04 ± 0.11 s; Pearson’s correlation with the reference system was moderate for swing times (r = 0.4–0.66), but stronger for stance and double support time (r = 0.93–0.95). Cadence mean error was −0.25 steps/min ± 3.9 steps/min (r = 0.97), while walking speed mean error was −0.02 ± 0.11 m/s (r = 0.92). These preliminary results suggest that single camera videos and pose estimation models based on deep networks could be used to quantify clinically relevant gait metrics in individuals poststroke, even while using assistive devices in uncontrolled environments. Such development opens the door to applications for gait analysis both inside and outside of clinical settings, without the need of sophisticated equipment.

Artificial intelligence (AI) tools are increasingly able to learn from larger and more complex data, thus allowing clinicians and scientists to gain new insights from the information they collect about their patients every day. In rehabilitation medicine, AI can be used to find patterns in huge amounts of healthcare data. These patterns can then be leveraged at the individual level, to design personalized care strategies and interventions to optimize each patient’s outcomes. However, building effective AI tools requires many careful considerations about how we collect and handle data, how we train the models, and how we interpret results. In this perspective, we discuss some of the current opportunities and challenges for AI in rehabilitation. We first review recent trends in AI for the screening, diagnosis, treatment, and continuous monitoring of disease or injury, with a special focus on the different types of healthcare data used for these applications. We then examine potential barriers to designing and integrating AI into the clinical workflow, and we propose an end-to-end framework to address these barriers and guide the development of effective AI for rehabilitation. Finally, we present ideas for future work to pave the way for AI implementation in real-world rehabilitation practices.

Wearable electronics with dual-sensing modalities enable artifact-free health monitoring through differential detection.