Microstate functional connectivity in EEG cognitive tasks revealed by a multivariate Gaussian hidden Markov model...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Signal Processing

Background:

• Tracking fast neural activity dynamics and transient networks is challenging due to limited understanding and methodologies.
• Existing methods struggle to capture the complex, dynamic nature of brain networks at high temporal resolutions.

Purpose of the Study:

• Introduce a novel approach to simultaneously identify distinct EEG microstates and their corresponding microstate functional connectivity (µFC) networks.
• Develop a sophisticated methodology for analyzing dynamic brain networks with improved accuracy and clinical applicability.

Main Methods:

• Utilized a multivariate Gaussian hidden Markov model (MGHMM) to decompose event-related potentials (ERPs) into quasi-stable EEG microstates.
• Concatenated trial segments belonging to specific microstates to measure µFC using the time-averaged phase-locking value.
• Validated the MGHMM approach with synthetic data and applied it to publicly available EEG data from visual cognitive tasks, comparing it with conventional dynamic FC methods.

Main Results:

• Successfully identified distinct EEG microstates and their associated µFC networks.
• Revealed dynamic modulations in the associations between EEG microstate networks and their corresponding µFC networks during cognitive tasks.
• Demonstrated superior performance of the MGHMM approach over conventional dynamic FC methods in terms of microstate correlations and functional connectivity distinction.

Conclusions:

• The MGHMM approach provides a robust and effective method for identifying and analyzing dynamic brain networks through EEG microstates and µFC.
• The proposed methodology shows significant improvements over existing techniques, suggesting its potential for future clinical applications in understanding brain function and dysfunction.