Project Summary The processes by which people with cerebral palsy (CP) adapt and learn new movement patterns are poorly understood. The Parent R01 for this NINDS Research Supplement to Promote Diversity in Health-Related Research seeks to quantify and understand motor adaptation in CP. Specifically, how individual differences in adaptation influence responses to rehabilitation that aims to improve walking function. Multimodal feedback training is a promising approach to support adaptation and motor learning, but few studies have evaluated adaptation with real-time feedback training during walking or its impacts on walking function. The Parent R01 seeks to fill these gaps by completing the systematic experimental analyses necessary to quantify walking adaption rates (Aim-1) and determine whether repeated exposure to multimodal feedback training can alter adaptation rates (Aim-2) to induce motor learning and improve walking function (Aim-3). We provide multimodal feedback using (1) sensorimotor feedback from adaptive ankle resistance delivered via a light- weight, wearable robot and (2) audiovisual feedback of plantar flexor activity from EMG recordings. To complement and extend this work, this Research Supplement will evaluate novel computational approaches bridging machine learning and causal modeling to evaluate individual responses to feedback training. The trainee will develop expertise bridging engineering, data science, and rehabilitation to (Supplement Aim-1) develop individual models of response to feedback training that can be used to track progression within and across sessions and (Supplement Aim-2) expand access to biomechanical monitoring and feedback training outside of research laboratories. Supplement Aim-1 uses Bayesian Additive Regression Trees to leverage the step-by-step data captured during the Parent R01 12 sessions of treadmill training to quantify changes in gait. Supplement Aim-2 tests the accuracy of Shallow Recurrent Decoder Networks to reconstruct key biomechanical variables from a sparse set of wearable sensors. This research will be complemented by extensive training in data science, rehabilitation, and biomechanics by a multidisciplinary team of mentors to support the trainee's future goals to pursue a tenure-track faculty position in engineering. Together, this research will introduce new methods to understand and optimize rehabilitation to individual needs and priorities.