The challenge

One-third of patients who had stroke suffered persistent disabilities, and upper limb (UL) motor impairment is one of the main disabilities. Recent clinical studies had been conducted using non-invasive EEG-based BCI via motor imagery, for post-stroke rehabilitation, yielded motor improvement of 7.2 on the Fugl-Meyer Motor Assessment (FMA-UE)score in chronic stroke patients that is significantly better than standard care. However, all the stroke patients underwent the same “one-size-fits-all” treatment option involving all six different activities of daily living (ADL)-oriented tasks regardless of their impairment or ability.
We hypothesize that precision personalized stroke rehabilitation intervention that is tailored to the patient hold more promise than a “one-size-fits-all” stroke rehabilitation strategy.

The proposed solution

1. To address the “one-size-fits-all” stroke rehabilitation strategy, RRIS will develop an Ability data-driven personalized stroke rehabilitation based on the stroke patient’s UL impairment and motor ability, by first matching 6 UL tasks in RRIS Ability Database with the 6 ADL tasks of the BCI-SR Intervention via similarity indices. A personalized subset of ADL tasks treatment options is then generated by a data-driven recommendation based on the patient’s ability, movement pattern of the treatment option and the normative data from the RRIS Ability Database. A multi-modal BCI is proposed to perform EEG subject-specific calibration using NIRS to ensure motor imagery compliance.  

2. 20 stroke subjects with UL impairments (score 11-45 on the FMA-UE) will be recruited to undergo the UL tasks assessment at RRIS. They will then undergo the personalized stroke rehabilitation using the Multimodal EEG and NIRS-based BCI with Soft Robotic therapy for 1 hour over 6 weeks, 3 times a week. The effectiveness of the personalized stroke rehabilitation can then be retrospectively compared to the use of “one-size-fits-all” ADL tasks in the previous clinical trial.