A Cerebellar Partitioning Method Using Spectral Clustering With Optimized Nonlinear Functional Connectivity

Area of Science:

• Neuroscience
• Brain Imaging
• Computational Biology

Background:

• The cerebellum exhibits high individual specificity in functional signals and is linked to neuropsychiatric disorders.
• Current cerebellar atlases hinder functional and structural understanding, data dimensionality reduction, and model applicability for functional imaging data.
• Progress in cerebellum-related research is impeded by the lack of suitable cerebellar partitioning methods.

Purpose of the Study:

• To develop a novel cerebellar partitioning algorithm for improved functional and structural analysis.
• To enhance the utility of cerebellar functional imaging data for research and clinical applications.
• To validate the reproducibility and comparative performance of the new partitioning method.

Main Methods:

• Utilized order-preserving variations with spatial constraints to optimize functional connectivity matrices.
• Employed spectral clustering and clustering ensemble techniques to create a cerebellar partitioning algorithm with a variable number of partitions.
• Validated the method using functional magnetic resonance imaging (fMRI) data and compared it against existing cerebellar atlases.

Main Results:

• The developed partitioning method demonstrated high reproducibility across individuals.
• The new partitions showed superior signal coherence and spatial congruence with cerebellar structural templates compared to existing atlases.
• Application to Parkinson's disease (PD) data significantly improved a classification model's usability, with optimal classification at 185 partitions.

Conclusions:

• The novel cerebellar partitioning algorithm offers a more effective tool for analyzing cerebellar functional imaging data.
• The method enhances the understanding of cerebellar involvement in neuropsychiatric disorders, exemplified by improved Parkinson's disease classification.
• The optimal number of cerebellar partitions may be task-dependent, suggesting flexibility in application for different research questions.