Exploring the Effects of Offline Paradigms and Feature Extraction Techniques on Performance of Motor Imagery...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Motor Imagery (MI) Brain-Computer Interfaces (BCIs) enable communication via brain signals.
• Technical and human factors pose challenges to MI BCI classification performance.

Purpose of the Study:

• To investigate the impact of paradigm choice and phase synchronization features on MI BCI classification.
• To compare performance using primary datasets versus older supplemental datasets.

Main Methods:

• Utilized Area Under the Curve (AUC) Receiver Operating Characteristics (ROC) curves to evaluate classification performance.
• Compared advanced paradigms and phase synchronization-based features (TD-CSP-wPLI, S-CSP-wPLI) against baseline methods.
• Analyzed performance across different frequency bands (16-30Hz, 12-15Hz).

Main Results:

• Advanced paradigms and features significantly improved MI BCI classification performance.
• Phase synchronization features, particularly TD-CSP-wPLI (16-30Hz) and S-CSP-wPLI (12-15Hz), yielded the most substantial performance gains.
• Enhanced classification performance correlated with improved usability.

Conclusions:

• The selection of advanced paradigms and phase synchronization features is crucial for optimizing MI BCI systems.
• Specific frequency bands are more effective for feature extraction in MI BCI applications.
• Improvements in classification accuracy directly translate to better user experience and system usability.