Exploring the distribution of connectivity weights in resting-state EEG networks

TL;DR

This study investigates the distribution patterns of functional connectivity weights in resting-state EEG networks through simulations and validation, revealing consistent right-skewed distributions unaffected by channel density or coupling measures.

Contribution

It is the first to systematically analyze how distribution patterns of connectivity weights in resting-state EEG are influenced by various factors, providing new insights into neural network properties.

Findings

Connectivity weights are right-skewed and unaffected by channel density.

Volume conduction influences the uniformity of connectivity weight distribution.

Coupling measures affected by volume conduction correlate with network distribution metrics.

Abstract

The resting-state brain networks (RSNs) reflects the functional connectivity patterns between brain modules, providing essential foundations for decoding intrinsic neural information within the brain. It serves as one of the primary tools for describing the spatial dynamics of the brain using various neuroimaging techniques, such as electroencephalography (EEG) and magnetoencephalography (MEG). However, the distribution rules or potential modes of functional connectivity weights in the resting state remain unclear. In this context, we first start from simulation, using forward solving model to generate scalp EEG with four channel densities (19, 32, 64, 128). Subsequently, we construct scalp brain networks using five coupling measures, aiming to explore whether different channel density or coupling measures affect the distribution pattern of functional connectivity weights. Next, we quantify the distribution pattern by calculating the skewness, kurtosis, and Shannon entropy of the functional connectivity network weights. Finally, the results of the simulation were validated in a normative database. We observed that: 1) The functional connection weights exhibit a right-skewed distribution, and are not influenced by channel density or coupling measures; 2) The functional connection weights exhibit a relatively uniform distribution, with the potential for volume conduction to affect the degree of uniformity in the distribution; 3) Networks constructed using coupling measures influenced by volume conduction exhibit significant correlations between the average connection weight and measures of skewness, kurtosis, and Shannon entropy. This study contributes to a deeper understanding of RSNs, providing valuable insights for research in the field of neuroscience, and holds promise for being associated with brain cognition and disease diagnosis.