Edge-centric functional network representations of human cerebral cortex reveal overlapping system-level architecture

Area of Science:

• Neuroscience
• Network Science
• Computational Biology

Background:

• Traditional network neuroscience uses node-centric models, limiting the analysis of connection interactions.
• Existing models cannot fully capture the dynamic interplay between different brain regions.

Purpose of the Study:

• To develop and validate an edge-centric network model for analyzing brain connectivity.
• To explore the properties and potential applications of edge functional connectivity (eFC).

Main Methods:

• Developed an edge-centric network model generating 'edge time series' and 'edge functional connectivity' (eFC).
• Applied network analysis to resting-state fMRI data to assess eFC consistency and reproducibility.
• Clustered eFC to identify communities of edges and overlapping brain regions.

Main Results:

• Demonstrated that eFC is consistent across datasets and reproducible within individuals over time.
• Showed that clustering eFC naturally divides the brain into overlapping communities, particularly in sensorimotor and attentional networks.
• Confirmed that eFC is systematically modulated by sensory input variations.

Conclusions:

• The edge-centric approach provides a novel framework for understanding brain network architecture beyond node-centric limitations.
• Identified reproducible eFC patterns and their modulation by sensory input.
• Proposed the edge-centric model as a tool for future research in disease biomarker identification, individual variation characterization, and high-resolution neural circuit mapping.