[Analysis of imagery motor effective networks based on dynamic partial directed coherence]

Area of Science:

• Neuroscience
• Brain-Computer Interfaces
• Network Science

Background:

• Understanding brain functional mechanisms and cognitive status is crucial.
• Electroencephalogram (EEG) signals provide valuable data for brain network analysis.
• Existing methods for measuring directional brain interactions have limitations.

Purpose of the Study:

• To propose and validate a dynamic partial directed coherence (dPDC) method for modeling brain networks during motor imagery.
• To analyze the network characteristics and significance of motor imagery using dPDC.
• To identify active brain regions associated with motor imagery.

Main Methods:

• Utilized scalp-recorded electroencephalogram (EEG) signals.
• Applied a time-frequency method, partial directed coherence (PDC), and introduced its dynamic version (dPDC).
• Calculated network parameters (out-degree, in-degree, clustering coefficient, eccentricity) for 9 subjects using data from BCI competitions IV (2008).

Main Results:

• The effective brain network during motor imagery exhibits small-world properties.
• Significant differences in out-degree were observed between left and right hand motor imagery in specific regions (ROI2, ROI3).
• Active regions for motor imagery were identified in fronto-central (ROI2, ROI3) and parieto-occipital (ROI5, ROI6) areas.

Conclusions:

• The dPDC algorithm effectively models brain networks for motor imagery.
• dPDC can accurately reflect changes in motor imagery.
• This method serves as a practical index for researching neural mechanisms of motor imagery.