AI-Driven Precision Neurorehabilitation: Integrating Digital Biomarkers, Neuroplasticity, and Personalized Therapy

Abstract

Background: Traditional neurorehabilitation often relies on standardized protocols that fail to account for individual variability in neuroplastic potential. The integration of artificial intelligence (AI) with digital biomarkers offers a paradigm shift toward precision neurorehabilitation, enabling continuous, objective monitoring and real-time personalization of therapy. Objective: To evaluate the clinical efficacy, feasibility, and patient adherence of an AI-driven precision neurorehabilitation system utilizing wearable digital biomarkers in patients recovering from stroke and traumatic brain injury (TBI) in Islamabad, Pakistan. Methods: A prospective, single-blind randomized controlled trial was conducted at two tertiary hospitals in Islamabad (PIMS and Shifa International Hospital). A total of 120 adult patients with stroke or mild-to-moderate TBI were randomized 1:1 to an experimental group (AI-driven, biomarker-guided personalized therapy via wearable inertial sensors and a bilingual gamified tablet application) or a control group (standard protocol-driven physiotherapy) for 8 weeks. Primary clinical outcomes included the Fugl-Meyer Assessment for Upper Extremity (FMA-UE), Berg Balance Scale (BBS), and Modified Rankin Scale (mRS). Results: Of the 120 participants, 114 completed the 8-week intervention. Result: The AI-driven group demonstrated significantly superior improvements in motor function (FMA-UE: 54.1 ± 6.8 vs. 44.5 ± 7.6, p < 0.001), balance (BBS: 50.8 ± 4.3 vs. 45.1 ± 5.4, p < 0.001), and functional independence (mRS: 1.0 vs. 2.0, p < 0.001) compared to the control group. Digital biomarker data revealed high device compliance (89.4%), strong utilization of the localized Urdu interface (72%), and a significant positive correlation between daily active therapy time and FMA-UE gains (r = 0.68, p < 0.01). Conclusion: AI-driven precision neurorehabilitation significantly accelerates functional recovery by continuously optimizing neuroplasticity through personalized therapeutic dosing. The high feasibility and acceptability of this system in Islamabad demonstrate that advanced digital health technologies can be successfully adapted for Low- and Middle-Income Country (LMIC) settings, offering a scalable model for global neurorehabilitation.