A Framework for Adaptation of Training Task, Assistance and Feedback for Optimizing Motor (Re)-Learning With a Robotic Exoskeleton

Abstract

Motor (re)-learning using a robotic exoskeleton can be improved by understanding what constitutes the optimal training environment for a subject that can maximize motor learning. We present a framework for motor (re)-learning using a robotic exoskeleton that provides subject-specific training by adapting one or more factors that constitute the training environment (training task, assistance, and feedback) based on the performance of the subject on the task. We present a torque-based task that requires subjects to dynamically regulate their joint torques for dexterous manipulation training and adapt the difficulty of the task based on their performance. We also present an adaptive impedance control for the robotic exoskeleton that transitions between error augmentation and haptic guidance training based on performance. Furthermore, we present an adaptive visual feedback approach that changes the transparency of the feedback based on performance. We conduct a study with ten healthy human subjects (five per training group) using a hand exoskeleton to evaluate if adaptations in task, assistance (assistive torque from the exoskeleton), and feedback (visual depiction of the torque trajectories) can modulate challenge and affect motor learning. Half of the participants were trained using an adaptive training paradigm (task, assistance, and feedback all adapted based on performance) and the other half with a nonadaptive training paradigm. A multiple comparison test showed that performance scores after adaptive training were significantly lower than nonadaptive training (MD: 62.6, 95% CI: 42.65-82.55, p <; 0.0001) demonstrating that adaptations in task, assistance, and feedback can modulate challenge and significantly affect motor learning.