Summary: UCLA researchers in the Department of Neurology have developed a novel AI-driven mixed reality headset for precision stroke rehabilitation.
Background: Stroke is a leading cause of long-term motor disability, often requiring prolonged and intensive rehabilitation to restore motor function. Current methods for evaluating motor impairments rely heavily on subjective clinical assessments, introducing inter-operator variability and human error. Consequently, therapeutic interventions are typically generalized instead of tailored to a patient’s individual deficits, which may reduce efficacy of clinical outcome. In addition, long-term intensive therapy is frequently confined to specialized clinical environments, reducing accessibility, long-term adherence, and recovery potential. As a result, there remains a significant unmet need for a scalable, data-driven rehabilitation approach capable of objective motor assessment and personalized therapy across diverse healthcare settings.
Innovation: Researchers at UCLA have developed an AI-driven mixed reality (MR) platform designed to facilitate stroke rehabilitation. Deployed via an MR headset, the system utilizes markerless 3D kinematic analysis and real-time motion capture to quantify motor impairments through embedded cameras. The system removes the need for physical markers through its real-time analysis of movement quality. Unlike existing rehabilitation methods that utilize scripted exercises, this technology analyzes patient movement deficits, quantifying impairments such as weakness, synergies, loss of dexterity, and compensatory movements, through AI algorithms trained on clinical kinematic datasets. The system’s algorithms are trained on extensive stroke and healthy control datasets, enabling swift classification and generation of individualized rehabilitation plans. Based on real-time patient performance, the MR environment provides immediate visual, auditory, and haptic cues to reinforce correct movement patterns and promote neuroplasticity. By replacing subjective assessments with data-driven analysis, this technology can reduce inter-rater variability while improving treatment personalization. Crucially, the system’s portability enables deployment across clinical, home, and tele-rehabilitation settings, significantly enhancing accessibility to intensive stroke rehabilitation. By integrating adaptive AI with real-time kinematic data, this platform represents a scalable approach to modernizing stroke rehabilitation.
Potential Applications: ● Stroke rehabilitation ● Broader neurological rehabilitation ● Movement disorder intervention ● Tele-health & remote monitoring ● Orthopedic medicine ● Physical & occupational therapy ● Sports medicine ● Clinical trials
Advantages: ● Data-driven ● Adaptive AI trained on extensive stroke and healthy datasets ● Closed-loop system ● Personalized therapy ● Accessibility ● Enhanced neuroplasticity ● No physical markers ● Closed-loop system Development-To-Date: Initial conception; currently pitching to VCs
Reference: UCLA Case No. 2025-99L
Lead Inventor: Ahmet Arac, Faculty in the Department of Neurology