The long term goal of this research is to advance personalization of walking rehabilitation for individuals post- stroke by developing therapeutic strategies targeting an individual’s specific motor control deficit. Stroke is an incredibly heterogeneous population, and while various therapeutic approaches have produced large effects in some individuals, group effects are often minimized by those who fail to respond to the intervention, leading to a paucity of efficacious randomized controlled trials. In addition, evidence supporting mechanisms by which walking is improved is limited, and we currently lack models predicting which individuals are likely to respond to an intervention and the mechanisms by which they improve. An urgent need exists to maximize treatment effect by targeting specific motor control impairments and improve predictive capability by developing theory-based clinical decision-making frameworks to translate interventions tailored to specific deficits for walking rehabilitation after stroke. Our overall goal for this project is to test a motor control deficit-based treatment approach that we developed, in order to provide information necessary for future translation of personalized interventions. We previously published the existence of distinct post-stroke motor control deficits based on the percentage of overall propulsive forces generated by the paretic leg termed paretic propulsion (Pp), a widely accepted biomechanical outcome measure that we developed. 1) Low Pp is associated with large and early paretic flexor EMG activity, lengthened paretic step length, and decreased paretic hip extension; and 2) High Pp pattern is characterized by decreased knee flexion during paretic swing, shortened paretic step length, and prolonged paretic hip extension. Our pilot data reveal that individuals with these walking patterns are most effectively rehabilitated by unique treatment strategies: 1) Low Pp by walking on an inclined treadmill requiring increased force production; and 2) High Pp by walking on a declined treadmill, promoting effective stance to swing transitions through normalization of joint kinetics and kinematics. The hypothesized ideal training strategy (INCLINE for low Pp and DECLINE for high Pp) is the personalized strategy and will be compared to non- personalized strategies (DECLINE for low Pp and INCLINE for high Pp). Both personalized strategies will be compared to a CONTROL group training on a flat treadmill at equivalent amounts of walking activity. Pilot training data demonstrate that personalized strategies demonstrate a larger effect on self-selected walking speed (SSWS) and symmetry of Pp. Thus, the purposes of this proposal are to compare clinical and biomechanical outcomes from personalized strategies to both non-personalized strategies and control interventions and to identify the variables that predict meaningful changes in SSWS. To accomplish these purposes, we will equally randomize 60 individuals (30 with ...