Upper extremity (UE) recovery remains a huge rehabilitation challenge with largely incomplete recovery of the upper limb post stroke. This is due to heavier priorities placed on other stroke competencies such as mobility, activities of daily living training and home integration which results in suboptimal amounts of time spent in upper limb training post stroke.

In this study the investigators plan to pilot brain computer interface (BCI) integrated wearable hand robotic glove (HandyRehab) system for upper limb stroke rehabilitation.

Robot-aided therapy have shown promising results in rehabilitation recovery in stroke patients while reducing strain and effort for both the therapist and suitable patients.

HandyRehab (HR) is a portable, wearable hand robotic glove used for functional training. The integration of Brain Computer Interface (BCI) platform with HR provides an alternative way of communication or intended movement through end-effectors for stroke and potentially facilitate neuroplasticity, improving motor functions.

This study aims to validate the usability of intensive training with HandyRehab (HR) compared with conventional occupational therapy (CT). Feasibility and safety of the novel BCI-HR will be examined and objective measures of clinical efficacy will be compared across all 3 platforms.

Inclusion Criteria:

1. Age 21-85 years, males and females

2. First diagnosis of stroke (ischemic or haemorrhagic), confirmed by neurologist/neurosurgeon/CT/MRI imaging

3. > 16 weeks post stroke

4. Hemiplegic pattern of post-stroke weakness

5. MRC ≥ 2/5 motor power and above for shoulder abduction & elbow flexion

6. MRC 0 to 4/5 motor power and above for finger flexors &/or extensors of thumb, index, middle fingers

7. Screening Fugl-Meyer wrist hand sub score <18/24

8. Spasticity MAS <3 for thumb, index, and middle fingers

9. Able to discriminate thumb and index sensation to pain

10. Hand sizes within 170-200mm (length) &75-85mm (width), compatible with HandyRehab robotic glove

11. BCI compatible brain states using a standardised screening protocol

12. Able to understand simple commands with Mini Mental state examination scores MMSE > 21/30)

13. Able to give informed consent

    • Subjects who are unable to clear BCI screening 11. will be randomised to either CT or HR groups.

Exclusion Criteria:

Neurological

• Recurrent stroke
• Diagnosis of neurodegenerative disease; e.g. Parkinson's' disease, Dementia, ALS

Medical:

- unstable medical or neurological conditions, life expectancy <6 months, end-organ renal failure on dialysis, severe heart failure, postural hypotension, history of uncontrolled sepsis, epilepsy with seizure within 3 months of informed consent, skin conditions which could potentially be worsened by wearable robot (ulcers, open wounds, eczema, infections etc)

Postural:

• Unable to tolerate upright posture or sit unaided for < 90min with rest breaks
• Cognitive/behavioural/visual:
• Severe dysphasia/aphasia, severe visual neglect/blindness, untreated severe depression, psychiatric illness, unable to understand study requirements

Upper limb:

• Moderate to severe spasticity (Modified Ashworth scale MAS ≥2)
• Hand/arm related pain (VAS Pain ≥ 5/10),
• Presence of finger contractures, reduced functional range of motion, limiting functional wrist/finger movements, active fractures or arthritis with deformities
• Severe limb ataxia/apraxia
• Severe post stroke hemi-anaesthesia in affected UE

BCI incompatibility:

• Motor imagery EEG signals unable to be detected
• Presence of craniectomy skull defect (affecting BCI cap fit and electrode contact)
• Concomitant participation in other interventional research trials
• Resident of nursing home or overseas country which may compromise attendance at research site
• Pregnant or lactating females will not be allowed to participate

Assessors (occupational therapist performing measurements/outcome) not involved in participant training will be blinded to assignments of participants.