Channel Selection for Optimal EEG Measurement in Motor Imagery-Based Brain-Computer Interfaces

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Motor imagery brain-computer interfaces (MI-BCI) often utilize numerous electroencephalographic (EEG) channels, leading to increased system complexity, noise, and reduced portability.
• Optimizing EEG channel selection is crucial for improving the online interoperability, portability, and user comfort of MI-BCI systems.

Purpose of the Study:

• To develop and validate a method for selecting a minimal yet effective subset of EEG channels for MI-BCI applications.
• To analyze the relationship between selected EEG channels and MI-BCI performance, aiming to reduce variability and noise.

Main Methods:

• A novel channel selection method was developed to identify common EEG channels across subjects that maintain high MI-BCI performance.
• The method was evaluated using the BCI Competition IV dataset 2a, a standard benchmark for motor imagery tasks.
• Performance was assessed for both two-class and four-class classification scenarios.

Main Results:

• The proposed method achieved classification accuracies comparable to state-of-the-art approaches using significantly fewer EEG channels.
• Binary classification accuracy reached 77-83% with as few as 6 EEG channels.
• Four-class classification accuracy exceeded 60% using 10 EEG channels.

Conclusions:

• The developed channel selection strategy effectively optimizes EEG signal acquisition for MI-BCI systems.
• This approach contributes to the development of more efficient, non-invasive, and wearable brain-computer interfaces.