The Effect of Modular Degeneracy on Neuroimaging Data

Area of Science:

• Neuroimaging
• Network Science
• Computational Neuroscience

Background:

• Community structure detection is crucial for systems-level analysis.
• Modularity maximization algorithms, while popular, suffer from degeneracy, producing multiple solutions.
• This degeneracy poses challenges in analyzing complex brain networks.

Purpose of the Study:

• To investigate the degeneracy effect of modularity on cingulate cortex parcellation using resting-state functional magnetic resonance imaging (rsfMRI) data.
• To propose and validate a novel iterative approach to overcome the limitations of existing modularity algorithms.
• To enhance the reliability of community structure detection in neuroimaging data.

Main Methods:

• Applied modularity-based community detection to Human Connectome Project rsfMRI data for cingulate cortex parcellation.
• Developed and implemented a new iterative algorithm to address the degeneracy problem inherent in modularity maximization.
• Compared the stability and consistency of the proposed method against current modularity approaches.

Main Results:

• Standard modularity algorithms produced variable numbers of cingulate cortex subdivisions across repeated analyses.
• The proposed iterative method demonstrated significantly improved stability, yielding consistent partitions.
• The new approach effectively mitigated the degeneracy problem, offering a more reliable parcellation.

Conclusions:

• The degeneracy of modularity algorithms limits reliable brain network parcellation.
• The novel iterative method provides a robust solution for stable community structure detection in rsfMRI data.
• This work offers a more dependable tool for applying modularity principles to neuroimaging analysis.