Clustering of resting state networks.

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Systems Neuroscience

Background:

• Resting state functional magnetic resonance imaging (fMRI) allows for the investigation of intrinsic brain activity.
• Understanding the organizational principles of the human brain at rest is crucial for neuroscience.

Purpose of the Study:

• To demonstrate a hierarchical structure of resting state brain activity in healthy individuals.
• To identify and characterize large-scale functional brain networks using a data-driven approach.

Main Methods:

• Applied the fuzzy-c-means clustering algorithm to resting-state fMRI data from two groups of healthy participants.
• Determined optimal cluster numbers using a cluster dispersion measure and assessed cluster similarity with an inner product metric.
• Investigated results with varying numbers of clusters (e.g., 2, 7, 11) and initialization conditions.

Main Results:

• The two-cluster solution differentiated task-positive and task-negative systems.
• Optimal clustering (7 and 11 clusters) identified established resting state networks, including the default mode, frontoparietal, attention, somatomotor, visual, and language networks.
• Language and ventral attention networks showed significant subcortical involvement, and the parcellation was robust across algorithm runs and initialization methods.

Conclusions:

• Clustering resting state activity revealed hierarchical organization of brain networks.
• The identified resting state networks align with previously established findings, reinforcing their structured nature.