Exploring Self-Paced Embodiable Neurofeedback for Post-stroke Motor Rehabilitation

Area of Science:

• Neuroscience
• Rehabilitation Engineering
• Human-Computer Interaction

Background:

• Upper limb motor impairment after stroke significantly impacts quality of life.
• Neurofeedback-guided motor-imagery training (NF-MIT) shows promise but typically uses abstract feedback.
• Existing NF-MIT lacks embodiment, potentially limiting engagement and effectiveness.

Purpose of the Study:

• To develop and evaluate a self-paced NF-MIT paradigm with an embodiable feedback signal (EFS).
• To assess the feasibility of using an anthropomorphic robotic hand (RH) for real-time feedback in NF-MIT.
• To investigate electrophysiological, subjective, and performance outcomes in stroke patients and healthy controls.

Main Methods:

• Developed a closed-looped EEG-based brain-computer interface (BCI) integrated with an anthropomorphic RH.
• Participants (9 stroke patients, 9 healthy controls) used mental activity (motor imagery) to control RH movements.
• Evaluated feasibility through electrophysiological measures (ERD), subjective integration of the RH, and RH control performance.

Main Results:

• Participants exhibited expected event-related desynchronization (ERD) patterns over sensorimotor areas.
• Most participants subjectively integrated the robotic hand into their body schema.
• While high control was not achieved, most participants demonstrated some degree of RH control, with similar performance between patients and controls.

Conclusions:

• Self-paced NF-MIT with an embodiable feedback signal via a robotic hand is feasible for stroke patients.
• This novel approach can potentially complement traditional NF-MIT methods.
• The findings support the use of embodied feedback for enhancing motor imagery training in neurorehabilitation.