Functional Connectivity Analysis of Cerebellum Using Spatially Constrained Spectral Clustering.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Brain Imaging

Background:

• The cerebellum, despite its small size, contains a disproportionately large number of neurons.
• Understanding the cerebellum's functional network is crucial for comprehending brain function.
• Previous studies have suggested specific regional importance within the cerebellum.

Purpose of the Study:

• To derive a functional network atlas of the human cerebellum during resting-state.
• To compare cerebellar functional networks between males and females.
• To introduce and validate a novel spatially constrained spectral clustering algorithm.

Main Methods:

• Development and application of a spatially constrained spectral clustering algorithm.
• Comparison with conventional spectral graph theory methods (spectral clustering, N-cut).
• Utilized resting-state functional MRI (fMRI) data from the Human Connectome Project (HCP).
• Combined functional atlas with anatomical atlas to create a 46-region atlas.
• Performed gender-based network analysis using minimum spanning trees.

Main Results:

• The spatially constrained spectral clustering method demonstrated superior performance in noisy conditions.
• Statistically significant differences were found in cerebellar network organization between males and females.
• The left lobule VI was identified as dominant in both genders, consistent with prior research.
• The study generated the first resting-state cerebellar atlas based on HCP data.

Conclusions:

• The novel algorithm effectively maps cerebellar functional networks.
• Significant sex-based differences exist in cerebellar functional organization.
• The findings contribute to a deeper understanding of cerebellar structure-function relationships and sexual dimorphism in the brain.