Autonomous modeling of repetitive movement for rehabilitation exercise monitoring.

BACKGROUND: Insightful feedback generation for daily home-based stroke rehabilitation is currently unavailable due to the inefficiency of exercise inspection done by therapists. We aim to produce a compact anomaly representation that allows a therapist to pay attention to only a few specific sections in a long exercise session record and boost their efficiency in feedback generation. METHODS: This study proposes a data-driven technique to model a repetitive exercise using unsupervised phase learning on an artificial neural network and statistical learning on principal component analysis (PCA). After a model is built on a set of normal healthy movements, the model can be used to extract a sequence of anomaly scores from a movement of the same prescription. RESULTS: The method not only works on a standard marker-based motion capture system but also performs well on a more compact and affordable motion capture system based-on Kinect V2 and wrist-worn inertial measurement units that can be used at home. An evaluation of four different exercises shows its potential in separating anomalous movements from normal ones with an average area under the curve (AUC) of 0.9872 even on the compact motion capture system. CONCLUSIONS: The proposed processing technique has the potential to help clinicians in providing high-quality feedback for telerehabilitation in a more scalable way.

Regenerative rehabilitation: exploring the synergistic effects of rehabilitation and implantation of a bio-functional scaffold in enhancing nerve regeneration.

Clinically, rehabilitation is one of the most common treatment options for traumatic injuries. Despite that, recovery remains suboptimal and recent breakthroughs in regenerative approaches may potentially improve clinical outcomes. To date, there have been numerous studies on the utilization of either rehabilitative or regenerative strategies for traumatic injury treatment. However, studies that document the combined effects of rehabilitation and regenerative tissue engineering options remain scarce. Here, in the context of traumatic nerve injury treatment, we use a rat spinal cord injury (SCI) model as a proof of concept to evaluate the synergistic effects of regenerative tissue engineering and rehabilitation. Specifically, we implanted a pro-regenerative hybrid fiber-hydrogel scaffold and subjected SCI rats to intensive rehabilitation. Of note, the rehabilitation session was augmented by a novel customized training device that imparts normal hindlimb gait movements to rats. Morphologically, more regenerated axons were observed when rats received rehabilitation (∼2.5 times and ∼2 times enhancement after 4 and 12 weeks of recovery, respectively, p < 0.05). Besides that, we also observed a higher percentage of anti-inflammatory cells (36.1 ± 12.9% in rehab rats vs. 3.31 ± 1.48% in non-rehab rats, p < 0.05) and perineuronal net formation in rehab rats at Week 4. Physically, rehab animals were also able to exert higher ankle flexion force (∼0.779 N vs. ∼0.495 N at Week 4 and ∼1.36 N vs. ∼0.647 N at Week 12 for rehab vs. non-rehab rats, p < 0.001) and performed better than non-rehab rats in the open field test. Taken together, we conclude that coupling rehabilitation with regenerative scaffold implantation strategies can further promote functional recovery after traumatic nerve injuries.

Automatic Inference of Rat's Hindlimb Trajectory to Synchronize with Forelimb Gait Through Phase.

Synchronous forelimb-hindlimb gait pattern is important to facilitate natural walking behavior of an injured rat with total transection. Since our ultimate research goal is to build a rehabilitation robotic system to simulate the natural walking pattern for spinalized rats, this research aims to address an immediate goal of automating the inference of the rat's hindlimb trajectory from its own forelimb movement. Our proposed method uses unsupervised learning to extract independent forelimb and hinblimb phases. From the phase information, a relationship between forelimb and hindlimb trajectory can then be calculated. Results show that the proposed method has the potential to be used in a rehabilitation robotic system.

Personalized Markerless Upper-Body Tracking with a Depth Camera and Wrist-Worn Inertial Measurement Units.

A markerless motion capture technique is proposed based on a fusion between a depth camera (Kinect V2) and a pair of wrist-worn inertial measurement units (IMU). The method creates a personalized articulated human mesh model from one depth image frame and uses that model to improve the accuracy of the upper-body joint tracking. The IMUs are useful as an additional clue for the arm tracking, especially during an occlusion. An evaluation of the method against a marker-based system as a gold standard using data from 6 subjects is done. The result shows over 20% reduction in upper-limb joint position errors when compared to Kinect's skeleton tracking. All the collected data are calibrated, synchronized, and made publicly available for research purposes.

Unsupervised Phase Learning and Extraction from Repetitive Movements.

Phase extraction from repetitive movements is one crucial part in various applications such as interactive robotics, physical rehabilitation, or gait analysis. However, pre-existing automatic phase extraction techniques are specific to a target movement due to some handcrafted-features. To make it more universal, a novel unsupervised-learning-based phase extraction technique is proposed. A neural network architecture and a cost function are designed to learn the concept of phase from records of a repetitive movement without any given phase label. The method is tested on a rat's gait cycle and a human's upper limb movement. The phases are successfully extracted at the sample level despite the variations in movement speed, trajectory, or subject's anthropometric features.