Methods Used to Estimate EEG Source-Space Networks: A Comparative Simulation-Based Study.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Electrophysiology

Background:

• Spontaneous brain activity during rest is a crucial area of neuroscience research.
• Resting-state functional connectivity analysis using Electroencephalography (EEG) is a growing field.
• Standardized pipelines for EEG source connectivity are lacking, hindering consistent results.

Purpose of the Study:

• To investigate the impact of methodological choices on the accuracy of resting-state network reconstruction from EEG.
• To evaluate the influence of channel density, inverse solutions, and functional connectivity measures.
• To optimize EEG-based estimation of resting-state cortical networks, specifically the default mode network (DMN).

Main Methods:

• Simulated EEG data using neural mass models for the default mode network (DMN).
• Tested varying channel densities, two inverse solutions (e.g., eLORETA), and two functional connectivity measures (e.g., PLV).
• Assessed the correspondence between reconstructed and reference connectivity matrices.

Main Results:

• Increased electrode density significantly enhanced the accuracy of network reconstruction.
• The combination of eLORETA (inverse solution) and PLV (functional connectivity) demonstrated superior accuracy.
• This pairing showed a higher correlation between reconstructed and reference connectivity matrices compared to other combinations.

Conclusions:

• Higher channel density in EEG recordings improves the reliability of resting-state network analysis.
• The eLORETA/PLV method offers a more accurate approach for estimating resting-state cortical networks from EEG.
• This research contributes to the standardization of electrophysiology connectomics methods.