Bayesian Optimized Spectral Filters Coupled With Ternary ECOC for Single-Trial EEG Classification

Area of Science:

• Biomedical Engineering
• Neuroscience
• Machine Learning

Background:

• Brain-computer interfaces (BCIs) show promise in assistive and rehabilitative systems.
• Optimizing signal processing is crucial for enhancing BCI performance.
• Existing methods may not fully leverage subject-specific spectral and spatial information.

Purpose of the Study:

• To propose a novel Bayesian framework for simultaneously optimizing subject-specific spectral and spatial filters for BCIs.
• To enhance feature discrimination for improved EEG-based BCI accuracy.
• To adapt information theory principles for robust EEG signal classification.

Main Methods:

• Developed a Bayesian framework for optimizing filter banks and spatial filters.
• Utilized common spatial patterns (CSP) coupled with error-correcting output coding (ECOC) classifiers.
• Proposed a modified ECOC approach with ternary class codewords for increased robustness.

Main Results:

• The framework successfully constructs optimized subject-specific spectral filters, creating significantly discriminant features.
• The adapted ECOC approach enhances robustness to misclassification errors in EEG data.
• Evaluated on BCI Competition datasets (BCIC-III and BCIC-IV), the proposed approach outperformed existing methods.

Conclusions:

• Optimized spectral filters play a critical role in improving overall classification accuracy for BCIs.
• The proposed Bayesian framework offers a significant advancement in BCI signal processing.
• The modified ECOC strategy provides a more robust classification for EEG data.