Functionally driven brain networks using multi-layer graph clustering.

Area of Science:

• Neuroscience
• Graph Theory
• Computational Biology

Background:

• Resting-state brain connectivity analysis is crucial for understanding neurodevelopmental disorders.
• Analyzing high-dimensional functional brain networks requires methods to extract meaningful connectivity information at both individual and population levels.

Purpose of the Study:

• To develop a technique for extracting common network modules (hubs) from population connectivity data.
• To create a functional brain network atlas.
• To characterize individual variations in brain connectivity for statistical analysis.

Main Methods:

• Proposed a multi-layer graph clustering technique.
• Fused information from a collection of population connectivity networks.
• Extracted common network modules serving as population hubs.
• Developed subject-specific factors for intra- and inter-connectivity analysis.

Main Results:

• Created a population network atlas of functional connectivity using MEG data from typically developing children.
• Demonstrated the technique's utility in describing individualized variations in brain connectivity in children with autism spectrum disorder.

Conclusions:

• The multi-layer graph clustering technique effectively identifies common brain network hubs and characterizes individual connectivity variations.
• This approach provides a valuable tool for studying neurodevelopmental disorders by creating a functional network atlas and enabling group-level statistical analyses.