Project Summary/Abstract Cerebral Palsy (CP) is the most common motor disorder and cause of disability in children, with a prevalence of roughly half a million children in the United States. Nearly all children with CP undergo physical therapy to improve gait, strength, balance, or flexibility. This therapy depends on dosage far in excess of available clinical session time, making adherence to an at-home therapy regimen critical to success. As therapy may be difficult, repetitive, or just not fun, maintaining engagement, adherence, and exercise relevance and quality in pediatric rehabilitation patients are critical barriers to an effective at-home rehabilitative physical therapy regimen. To promote engagement and adherence, a solution must be attractive and fun, build rapport with the patient, and promote focus within a session as well as longitudinal motivation. Active video games have proven to be an improvement over typical practice but leave room for innovation; recent research suggests a promising complementary approach based on an artificially intelligent, physically embodied social agent. By activating humans’ natural social responses, socially assistive robots have been shown to enhance engagement and adherence in related settings. Charles River Analytics and our partners at Spaulding Rehabilitation Hospital propose a Phase I SBIR to prototype and demonstrate the feasibility of “Rehabilitation Engagement at Home with a socially Assistive roBot for Pediatric Adherence” (REHAB-PAL). REHAB-PAL is a socially assistive robot for children undergoing rehabilitative physical therapy that provides companionship and engagement during therapy. REHAB-PAL improves adherence to the home-care regimen by encouraging the prescribed therapeutic exercises with fun and supportive interaction. To rigorously isolate its effects, we will compare the robot to a behaviorally analogous screen-based virtual agent. We will prototype and demonstrate REHAB-PAL’s feasibility via three Specific Aims: (1) Design and develop socially assistive application software for a commercially available small humanoid robot and a screen-based virtual agent. (2) Integrate with a visuo-motor integration task. (3) Evaluate clinical suitability of REHAB-PAL for gait and balance training in children with CP. Ultimately, we aim to show that REHAB-PAL offers patients an engaging and encouraging companion throughout a potentially dreary treatment, while offering clinicians a superior tool to keep pediatric patients motivated and focused on their treatment and a valuable proxy to provide continuity, consistency and monitoring during home care. By enhancing the efficacy of therapeutic dosage outside the clinic, REHAB-PAL has the potential to extend and improve care, with less frequent clinic visits and improved adherence.