Enhancing prosthetic hand control: A synergistic multi-channel electroencephalogram.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Electromyogram (EMG) is standard for prosthetic hand control but faces limitations due to muscle function and fatigue.
• Classifying noninvasive electroencephalogram (EEG) signals for prosthetic control remains challenging due to complex temporal brain network shifts.

Purpose of the Study:

• To investigate brain synergy, defined as coordinated temporal patterns in brain networks, for decoding hand movements.
• To develop and validate an EEG-based prosthetic hand control system utilizing synergistic features.

Main Methods:

• Acquired 32-channel EEG data from 10 healthy participants during hand grasp and open tasks.
• Analyzed synergistic spatial distribution and power spectra using independent component analysis.
• Extracted time-domain and synergistic features from 15 selected EEG channels.
• Trained a Bayesian optimizer-based support vector machine (SVM) classifier.

Main Results:

• Achieved an average testing accuracy of 94.39 ± 0.84% using synergistic features with the optimized SVM.
• Synergistic features significantly outperformed time-domain features in classifying hand movements (p < 0.05).
• The classifier's output was successfully employed for prosthetic hand control.

Conclusions:

• Brain synergy analysis provides a promising, effective approach for prosthetic hand control, addressing EMG limitations.
• This synergistic method enhances the decoding of hand movements from EEG, contributing to advanced neuroprosthetics.
• The findings highlight the potential of synergy-based control for future prosthetic research and development.