Artificial Intelligence-Driven Virtual Reality Rehabilitation for Upper Limb Recovery in Acute and Subacute Stroke Patients

Background: Upper limb motor deficits are common after stroke, and early rehabilitation is crucial for functional recovery. Virtual reality (VR) combined with artificial intelligence (AI) has shown potential to enhance motor learning, but evidence from randomized controlled trials in acute and subacute hospitalized patients is limited.

Objective: To evaluate whether AI-driven VR rehabilitation in addition to standard care improves upper limb motor recovery compared with standard rehabilitation alone in acute and subacute stroke patients.

Design: Prospective, assessor-blinded, randomized controlled trial with stratified 1:1 allocation by baseline Fulg Meyer Assessment Score.S Participants: Adults (≥18 years) with confirmed acute or subacute stroke, admitted to the hospital, with upper limb motor déficits.

Intervention: Experimental group will receive standard rehabilitation plus AI-driven VR therapy during hospitalization. Control group will receive standard rehabilitation alone.

Primary Outcome: Change in FM-UE score from baseline to hospital discharge. Secondary Outcomes: VR-derived digital biomarkers, therapy adherence, NIHSS, FM-UE, and Modified Ashworth Scale at each session, and the following baseline and treatment variables: etiology, Modified Rankin Scale (mRS), handedness, stroke lateralization, aphasia, acute reperfusion, reperfusion treatment, days from stroke to start of rehabilitation, adverse events, patient satisfaction, and date of discharge. FM-UE at 1 month post-intervention.

Randomization and Blinding: Patients will be randomized in a 1:1 ratio to either group, stratified by baseline FM-UE (≤30 vs >30) to ensure balance between groups for stroke severity. Within each stratum, block randomization will be used. Outcome assessors will be blinded.

Sample Size: 50 patients (25 per group).

Ethics and Registration: Approved by hospital ethics committee. Significance: This trial will provide evidence on the effectiveness of AI-driven VR therapy for early upper limb rehabilitation in hospitalized stroke patients.