Lateralization in scalp EEG brain connectivity during hand motor imagery can improve task classification for...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Motor imagery (MI) tasks involve imagining movement and are crucial for brain-computer interfaces (BCIs).
• Traditional MI decoders often rely on µ-band power spectrum features from electroencephalography (EEG).
• The reorganization of brain connectivity during MI and its predictive value for classification remain areas for exploration.

Purpose of the Study:

• To evaluate brain connectivity reorganization during motor imagery (MI) tasks.
• To assess the predictive value of EEG-based functional connectivity measures for MI classification.
• To compare connectivity measures against traditional µ-band power spectrum features.

Main Methods:

• Analyzed EEG data from BCI Competition IV 2a and PhysioNet Motor Imagery datasets for left- and right-hand MI.
• Evaluated functional connectivity measures: Phase Locking Value (PLV), cross-correlation (CC), weighted Phase Lag Index (wPLI), and Granger causality (GC).
• Compared decoding performance using Random Forest classifiers against µ-band power features; employed graph-theoretical metrics for network analysis.

Main Results:

• Phase Locking Value (PLV) demonstrated the most reliable MI decoding performance, comparable to µ-band power features across both datasets.
• A moderate correlation was found between network centrality differences (PLV-based) and the importance of single-channel power values.
• While network topology was stable, a small set of contralateral connections significantly enhanced classification accuracy.

Conclusions:

• Motor imagery primarily modulates a limited number of task-specific functional connections.
• Functional connectivity measures offer complementary, network-level insights beyond traditional power-based approaches.
• These findings can inform interpretable feature selection and the design of future brain-computer interface models.