EEG-Based BCI System to Detect Fingers Movements

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Restoring mobility for individuals with paralyzed or amputated limbs is a significant challenge in assistive technology.
• Existing prosthetic control systems often have limitations in dexterity and simultaneous movement capabilities.

Purpose of the Study:

• To develop and validate a brain-computer interface (BCI) system for controlling prosthetic fingers using electroencephalogram (EEG) signals.
• To enable simultaneous, multi-finger movements in prosthetic devices, enhancing functional restoration.

Main Methods:

• Utilized complex electroencephalogram (EEG) signal processing algorithms for outlier removal and feature extraction.
• Implemented a machine learning strategy based on an ensemble of one-class classifiers for multi-class classification of finger movement intentions.
• Identified and located brain regions sensitive to finger movements.

Main Results:

• Achieved an average accuracy of 81% across five subjects for predicting finger movement intentions.
• Demonstrated the capability for simultaneous control of multiple prosthetic fingers, surpassing limitations of existing prototypes.
• The multi-class classification approach yielded high accuracy compared to binary classification systems.

Conclusions:

• The proposed BCI system offers a novel prosthetic solution for individuals with severe disabilities.
• The system enhances the ability to perform daily tasks by enabling intuitive and simultaneous control of prosthetic fingers.
• This advancement in assistive technology holds significant potential for improving the quality of life for users.