AI-guided personalized rehabilitation to improve upper-limb recovery after stroke

Personalized Rehabilitation Pathways to Maximal Motor Functional Return Through an AI Recovery Prediction System for Diverse Stroke Survivors

This project tests whether an explainable AI system can predict recovery and recommend tailored rehab (usual care alone or plus acupuncture, robotic training, or NMES) for people aged 65–80 who had a first-time unilateral stroke 1–6 months earlier.

Quick facts

• Phase: Not applicable
• Study type: Interventional
• Enrollment: 400 (estimated)
• Ages: 65 Years to 80 Years
• Sex: All
• Sponsor: Chinese University of Hong Kong (Academic / other)
• Locations: 1 site (Hong Kong, Sha Tin)
• Trial ID: NCT06828679 on ClinicalTrials.gov

What this trial studies

Four hundred subacute stroke survivors will be randomly assigned to one of four groups receiving usual rehabilitation alone or usual care plus acupuncture, robotic training, or neuromuscular electrical stimulation, with treatments lasting at least one month and at least 20 sessions. Detailed baseline assessments (clinical exams, MRI, EEG, EMG, TMS, and blood tests) will be collected before rehab, with clinical scores and EMG measured immediately after and at six months to quantify short- and longer-term recovery. Midpoint clinical scores will be recorded during therapy to monitor progress. All multimodal data will be analyzed offline using machine learning to develop PRAISE-HK, an explainable AI that predicts individual recovery potential and the treatment likely to maximize functional return.

Who should consider this trial

Good fit: Ideal candidates are people aged 65–80, 1–6 months after a first-time unilateral supratentorial stroke with moderate-to-severe upper-limb weakness (Fugl-Meyer upper extremity 10–50) who can give consent and have detectable EMG in target forearm muscle groups.

Not a fit: Patients who are unconscious or bed-bound, have uncontrollable diabetes, severe comorbid organ failure, a cardiac pacemaker, anticipated non-adherence, bilateral or brainstem/cerebellar strokes, or who fall outside the age/onset windows are unlikely to benefit from or be eligible for this protocol.

Why it matters

Potential benefit: If successful, the system could help match each person to the rehabilitation approach most likely to boost arm function and independence while making rehab delivery more efficient.

How similar studies have performed: Robotics, NMES, and acupuncture have produced mixed but sometimes positive functional gains in prior trials, while multimodal predictive and explainable AI approaches for individualizing rehab are promising but remain relatively novel and not yet widely validated.

Eligibility criteria

IInclusion Criteria:

* Age 65-80
* 1-6 months after onset of a first-time unilateral stroke rostral to midbrain
* Moderate-to-severe motor impairment of one upper limb (Fugl-Meyer Assessment for Upper Extremity of 10-50 out of 66);
* Able to provide written informed consent;
* Detectable electromyographic (EMG) activities in flexor digitorum-flexor carpi radialis and extensor digitorum-extensor carpi ulnaris muscle groups, with EMG from each muscle group exceeding 3 standard deviations above baseline mean. This last criterion is essential for successful NMES training

Exclusion Criteria:

* Unconscious or bed-bound;
* Uncontrollable diabetes;
* Anticipated non-adherence to treatment schedule;
* On cardiac pacemaker;
* Other severe comorbidities (heart/kidney failure, deranged liver function).

Where this trial is running

Hong Kong, Sha Tin

• The Chinese University of Hong Kong — Hong Kong, Sha Tin, Hong Kong (Recruiting)

Study contacts

• Study coordinator: Yat Sing Kelvin Lau, MSc

Conditions Acute Stroke Intervention; Acute Stroke; Acute Ischemic Stroke; stroke; motor neuroscience; MRI; rehabilitation; muscle coordination