Functional Electrical Stimulation Controlled by Motor Imagery Brain-Computer Interface for Rehabilitation.

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Biomedical Engineering

Background:

• Sensorimotor rhythm (SMR)-based brain-computer interfaces (BCI) show promise for motor deficit rehabilitation.
• Challenges include unstructured motor imagery (MI) training, classifying single-hand MI tasks, and decoding voluntary MI-evoked SMRs.
• Functional Electrical Stimulation (FES) is often used in conjunction with BCI for enhanced motor recovery.

Purpose of the Study:

• To develop and validate an SMR-based BCI-FES system for 2-class motor imagery (MI) tasks in a single hand.
• To investigate the feasibility of this system for stroke and traumatic brain injury (TBI) patients.
• To evaluate the impact of FES and adaptive learning on task performance in patients.

Main Methods:

• Phase 1: Developed and validated an SMR-based BCI-FES system for 2-class MI tasks.
• Phase 2: Assessed system feasibility in stroke and TBI patients using goal-oriented tasks in a semi-asynchronous mode.
• Evaluated the effects of FES presence and adaptive learning on patient performance.

Main Results:

• Phase 1: Achieved significantly higher accuracy (approx. 71.25%) for 2-class MI classification compared to chance level.
• Distinguishing voluntary and passive SMRs did not reach significance.
• Phase 2: Adaptive learning significantly improved accuracy; accuracy under No-FES with adaptive learning was 61.9%, significantly above chance.
• Adaptive learning enhanced task performance in patients with motor deficits.

Conclusions:

• The developed SMR-based BCI-FES system shows potential for motor rehabilitation in stroke and TBI patients.
• Adaptive learning is a crucial component for improving BCI performance in this population.
• Future research should focus on larger sample sizes and kinesthetic MI for further advancements.