Related Experiment Video
Updated: Nov 12, 2025
08:51
Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
Published on: November 1, 2019
5.8K
A Comprehensive Analysis of Multilayer Community Detection Algorithms for Application to EEG-Based Brain Networks
Maria Grazia Puxeddu1,2, Manuela Petti1,2, Laura Astolfi1,2
1Department of Computer, Control and Management Engineering "Antonio Ruberti", University of Rome Sapienza, Rome, Italy.
Frontiers in Systems Neuroscience
|March 18, 2021
Summary
This study compares algorithms for detecting communities in multilayer brain networks from electroencephalographic (EEG) data. It provides guidelines for selecting the best algorithm for analyzing dynamic brain activity and modular structures.
Area of Science:
- Neuroscience
- Network Science
- Computational Biology
Background:
- Brain networks exhibit modular organization crucial for cognitive functions.
- Multilayer brain networks, derived from electroencephalographic (EEG) signals, capture dynamic neural activity.
- Current methods lack consensus on optimal algorithms for community detection in these complex networks.
Purpose of the Study:
- To comprehensively analyze and compare state-of-the-art algorithms for multilayer community detection in EEG-based brain networks.
- To evaluate algorithm performance in identifying both static and dynamic modular structures.
- To provide guidelines for selecting appropriate algorithms based on network properties.
Main Methods:
- Compared three multilayer community detection algorithms (genLouvain, DynMoga, FacetNet) against a single-layer approach.
- Utilized custom benchmark graphs with varying densities, cluster numbers, noise levels, and layers for statistical evaluation.
- Applied selected algorithms to real EEG data from resting-state conditions (eyes open/closed).
Main Results:
- The simulation study established performance benchmarks for each algorithm under diverse conditions.
- Identified algorithm suitability based on specific network characteristics like density and dynamics.
- Demonstrated the feasibility of multilayer EEG network analysis for capturing brain states.
Conclusions:
- Algorithm choice significantly impacts the accuracy of modular structure detection in multilayer EEG networks.
- The study offers practical guidance for researchers in selecting optimal community detection methods.
- Multilayer analysis of EEG data is a viable approach for understanding dynamic brain organization.
