From Coherence to Multivariate Causal Estimators of EEG Connectivity

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Electroencephalography (EEG) is crucial for studying brain activity.
• Estimating functional connectivity from EEG data is complex due to noise and signal properties.
• Current methods for EEG functional connectivity have limitations.

Purpose of the Study:

• To review and compare existing EEG functional connectivity estimation methods.
• To assess the robustness of different methods against noise and artifacts.
• To propose advanced methods for analyzing dynamic brain networks and causal connectivity.

Main Methods:

• Comparison of linear and non-linear, bivariate and multivariate EEG connectivity measures.
• Evaluation of method performance regarding robustness to noise, common drive, and volume conduction.
• Introduction of time-varying connectivity estimation.
• Application of advanced graph analysis for network community structure and hierarchy.

Main Results:

• Identified advantages and flaws of current EEG functional connectivity measures.
• Provided guidance on selecting robust connectivity metrics.
• Demonstrated the utility of time-varying analysis for dynamic brain processing.
• Showcased advanced graph analysis for richer network representation.

Conclusions:

• There is a need for robust EEG functional connectivity methods that account for noise and artifacts.
• Future research should focus on effective (causal) connectivity and dynamic network analysis.
• Advanced graph analysis offers a more comprehensive understanding of brain network organization.