Using brain connectivity metrics from synchrostates to perform motor imagery classification in EEG-based BCI systems

Area of Science:

• Neuroscience
• Cognitive Science
• Signal Processing

Background:

• Neural synchronisation is key to understanding brain cognition.
• Electroencephalogram (EEG) signals capture brain activity patterns.

Purpose of the Study:

• Investigate phase synchronisation patterns in EEG signals during motor imagery (MI) tasks.
• Develop a classification method for distinguishing cognitive tasks based on neural synchronisation.

Main Methods:

• Recorded EEG signals from participants performing MI tasks with emotional face stimuli.
• Identified specific phase-synchronised states ('synchrostates') for each task.
• Applied graph theory to extract network measures from synchrostates.
• Utilized classification algorithms to differentiate between MI tasks.

Main Results:

• Two MI tasks were classified with 85% accuracy, sensitivity, and specificity.
• Identified optimal feature subsets for improved discrimination.
• Demonstrated the specificity of synchrostates for cognitive tasks.

Conclusions:

• Phase synchronisation patterns are effective for classifying cognitive tasks.
• The developed method shows promise for online brain-computer interface (BCI) systems.
• Robust classification of MI tasks is achievable using EEG-derived neural connectivity.