PROJECT ABSTRACT Functional electrical stimulation (FES) is a common rehabilitation technology that incorporates electrical stimulation timed with a functional task to augment paretic muscle function in people with neuro-pathologies such as stroke and spinal cord injury. The rigor of previous research has established transcutaneous and implanted FES systems for standing, walking, reaching and grasping, with both neuro-prosthetic and therapeutic effects. Stroke is the leading cause of disability. Footdrop is a highly prevalent post-stroke gait deficit, leading to insufficient ankle dorsiflexion during the swing phase of gait, and contributing to reduced mobility. FES systems that compensate for footdrop to improve gait function and reduce falls risk are gaining increasing popularity, with commercial FES footdrop systems now available for clinical use. Despite their promising functional value, accessibility, and positive neuroplasticity effects, current FES systems have some fundamental limitations, which limit their clinical prescription. The overall goal of this R21 is to overcome 2 major limitations and technical gaps in FES – rapid onset of muscle fatigue and lack of closed-loop control of FES intensity. Most existing FES systems do not automatically modulate stimulation intensity in response to muscle fatigue, and may overstimulate the muscles. To address this limitation, this R21 facilitates a novel inter- institutional cross-disciplinary collaboration between scientists with complimentary expertise in clinical testing of FES for improving stroke gait (MPI Kesar) and engineering of data-driven FES control systems (MPI Sharma). We propose the first clinical testing of a new bioinspired FES controls approach developed by MPI Sharma that exploits fatigue predictions from a model-predictive controller (MPC) to control FES intensity, avoid overstimulation, delay fatigue, and maximize FES-induced gait performance. For the first time, we will systematically test FES controllers that incorporate MPI Sharma’s data-driven MPC muscle and fatigue behavior model in people with post-stroke hemiparesis, so that FES parameters can be computed optimally and proactively, via prediction (feedforward FES control), in addition to reacting to changes in ongoing gait performance (feedback FES control). For this MPC-based FES controller, we propose to use ultrasound to inform the model about fatigue onset, which is unprecedented in FES for stroke gait. Our project aims are to (1) develop and evaluate a novel data-driven, model predictive FES controller that utilizes ultrasound- derived feedback for footdrop correction during post-stroke gait; and 2) compare our novel FES control system with conventional FES during treadmill and overground gait in people post-stroke. This work will inform future clinical testing and development. Our immediate deliverable will be a leading-edge FES neuro-prosthetic technology that uses ultrasound-sensing and MPC closed-loop optimizati...