A Computationally Efficient Method for Hybrid EEG-fNIRS BCI Based on the Pearson Correlation

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Hybrid brain-computer interfaces (BCI) integrate electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) for improved performance.
• Channel selection is critical for optimizing hybrid BCI system efficiency and accuracy.
• Existing methods face challenges with system complexity and computational time.

Purpose of the Study:

• To propose a novel channel selection method for hybrid EEG-fNIRS BCIs.
• To reduce computational burden and complexity in hybrid BCI systems.
• To enhance the reliability and accuracy of motor imagery classification.

Main Methods:

• Simultaneous recording of EEG and fNIRS signals.
• Novel channel selection using Pearson product-moment correlation coefficient to identify highly correlated channels per hemisphere.
• Extraction of four statistical features and their combinations for classification.
• Utilized K-Nearest Neighbors (KNN) and Tree classifiers.

Main Results:

• The proposed channel selection method significantly reduces computational burden.
• High reliability and classification accuracy comparable to existing literature were achieved.
• Demonstrated the effectiveness of Pearson correlation for hybrid EEG-fNIRS channel selection.

Conclusions:

• The novel channel selection approach effectively optimizes hybrid EEG-fNIRS BCIs.
• This method offers a computationally efficient solution without compromising classification accuracy.
• Paves the way for more practical and accessible hybrid BCI applications.