Dimensionless, Scale Invariant, Edge Weight Metric for the Study of Complex Structural Networks

Area of Science:

• Neuroimaging
• Network Neuroscience
• Computational Neuroscience

Background:

• Diffusion weighted imaging (DWI) and network analysis offer a powerful framework for in vivo brain structure studies.
• Graph theory, utilizing edge weights, is crucial for quantifying white matter connections between gray matter nodes.

Purpose of the Study:

• Introduce a novel dimensionless, scale-invariant edge weight metric for quantifying node connectivity.
• Assess the metric's robustness and consistency across different scales (rodents to humans) and resolutions.

Main Methods:

• Simulations to evaluate tractography seed density and orientation errors.
• Analysis of repeated DWI measures in human subjects.
• Connectivity characterization in excised rat brains at varying spatial resolutions.

Main Results:

• Edge weight estimates improve with increased seed density, given sufficient signal-to-noise ratio (SNR).
• Consistent and low-variability mean edge weight values were observed in human brain regions (cingulum, corpus callosum).
• Adequate resolution and SNR allow for robust characterization of network connections using the proposed metric.

Conclusions:

• The developed edge weight metric is a robust and scale-invariant measure of network connectivity.
• This metric can be reliably applied to quantify brain connectivity across various species and resolutions.
• The findings support the use of this metric for advanced DWI-derived brain network analysis.