Multiscale Community Detection in Functional Brain Networks Constructed Using Dynamic Time Warping.

Area of Science:

• Neuroscience
• Network Science
• Data Analysis

Background:

• Resting-state functional magnetic resonance imaging (fMRI) is crucial for brain network analysis.
• Traditional Pearson correlation methods often overlook the temporal dynamics of fMRI time series.
• Accurate detection of community structures in brain networks is essential for understanding brain function.

Purpose of the Study:

• To develop a novel method for analyzing functional brain network community structures using fMRI data.
• To address the limitations of traditional correlation methods by incorporating time series dynamics.
• To identify robust community structures and functional relationships within individual brain networks.

Main Methods:

• Applied the Dynamic Time Warp (DTW) algorithm to compare fMRI time series synchronization and asynchrony between brain regions.
• Utilized Modularity, Variation of Information (VI), and Normalized Mutual Information (NMI) to detect communities across multiple network scales.
• Determined the optimal community structure for each subject based on these metrics.

Main Results:

• Successfully identified key components of seven fundamental functional subnetworks.
• Discovered novel associations between the putamen, Heschl's gyrus, and the inferior parietal network.
• Validated the method's ability to determine individual functional brain network density, consistent with existing research.

Conclusions:

• The DTW algorithm offers a more accurate approach to analyzing functional brain network dynamics compared to traditional methods.
• This method enhances the detection of community structures and reveals previously unknown functional relationships.
• The findings confirm the utility of this approach for studying real-world brain network organization and density.