Motor imagery classification by means of source analysis for brain-computer interface applications.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) enable communication and control through brain activity.
• Motor imagery (MI) classification from electroencephalography (EEG) is crucial for BCIs.
• Current EEG analysis methods may lack precision in localizing neural sources.

Purpose of the Study:

• To evaluate the efficacy of source analysis techniques for classifying motor imagery tasks.
• To determine if source localization improves the accuracy of MI classification from scalp EEG.
• To explore the potential of advanced source analysis in BCIs.

Main Methods:

• Independent Component Analysis (ICA) was used for spatio-temporal filtering of EEG signals.
• Equivalent dipole analysis and cortical current density imaging were applied for neural source reconstruction.
• Classification accuracy was assessed based on the location of reconstructed neural sources in the motor cortex.

Main Results:

• An approximately 80% classification rate for motor imagery tasks was achieved in a human subject.
• Both equivalent dipole analysis and cortical current density imaging yielded comparable high accuracy.
• Correct classification correlated with the localization of equivalent neural sources in the correct motor cortex hemisphere.

Conclusions:

• Source analysis of EEG provides a clearer representation of cortical activity during motor imagery.
• This approach significantly enhances the classification accuracy of motor imagery tasks for BCI applications.
• The findings suggest source analysis is a promising method for improving BCI performance.