Minimum spanning tree analysis of the human connectome.

Area of Science:

• Neuroscience
• Network Science
• Medical Imaging

Background:

• Comparing brain network organization across individuals is challenging due to variations in connection number and strength.
• Existing methods often focus on cortical connections, potentially overlooking other significant pathways.

Purpose of the Study:

• To introduce Minimum Spanning Tree (MST) analysis as a method for creating a standardized, empirical reference network for human brain structure.
• To utilize this MST reference network as a null model for comparing individual brain network topology.

Main Methods:

• Analysis of three diffusion-weighted imaging datasets from healthy adults.
• Construction of a group-level MST from the mean connectivity matrix to serve as an empirical null model.
• Comparison of individual subject MSTs against the reference MST.

Main Results:

• Individual brain networks showed significant overlap (58%-88%) with the reference MST, indicating consistent structural backbones.
• Identified hub nodes in the MST corresponded to known hub regions, including rich club nodes.
• Cortical-subcortical connections were consistently found in the MST and enhanced network efficiency, suggesting their role in major neural communication.

Conclusions:

• MST analysis offers an elegant and straightforward approach to characterizing structural brain networks.
• The MST provides a robust empirical null model for comparing individual brain network topology.
• MST characteristics can effectively index age-related changes in brain structure.