Weighting the structural connectome: Exploring its impact on network properties and predicting cognitive performance...

Area of Science:

• Neuroscience
• Network Science
• Computational Biology

Background:

• Brain function arises from complex neural network interactions, not isolated activity.
• The human connectome encompasses structural (anatomical) and functional (dynamic) aspects.
• Weighting methods for structural connections can influence study interpretations.

Purpose of the Study:

• To investigate how different structural connectome (SC) weighting methods affect network properties and cognitive performance prediction.
• To introduce and compare the Axon Diameter Distribution (ADD) weighting method with traditional methods.
• To explore the functional relevance of weighted SCs using Human Connectome Project (HCP) data.

Main Methods:

• Developed three SC weighting models: Number of Streamlines (NOS), Fractional Anisotropy (FA), and Axon Diameter Distribution (ADD).
• Utilized the AxSI method to extract ADD-weighted SCs.
• Employed graph theory properties and the NIH Toolbox to build predictive models for cognitive performance using HCP data.

Main Results:

• Different weighting models yield distinct network properties.
• The ADD-weighted SC, particularly when combined with a functional subnetwork model, demonstrated superior predictive power for cognitive performance.
• The ADD method, extracted via AxSI, showed unique insights compared to NOS and FA.

Conclusions:

• The choice of SC weighting method significantly influences network analysis and conclusions.
• ADD-weighted SCs offer a promising approach for understanding the relationship between brain structure and cognitive function.
• The ADD method, integrated with functional subnetworks, provides a more accurate model for estimating cognitive performance.