Group-wise parcellation of the cortex through multi-scale spectral clustering

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Brain Mapping

Background:

• Brain region delineation is crucial for understanding function.
• Traditional cytoarchitecture methods lack scalability and in vivo application.
• Existing in vivo methods primarily focus on single-subject parcellation.

Purpose of the Study:

• To develop a novel group-wise brain parcellation method.
• To enable scalable and consistent brain mapping across subjects.
• To leverage diffusion Magnetic Resonance Imaging (dMRI) for connectivity-driven parcellation.

Main Methods:

• A group-wise connectivity-driven parcellation approach using spectral clustering.
• Capturing local connectivity information at multiple scales.
• Enforcing direct correspondences between subjects for group analysis.

Main Results:

• Successful application to dMRI data from the Human Connectome Project (50 subjects per group).
• Demonstrated promising quantitative and qualitative results.
• Showcased effectiveness in terms of information loss, modality comparisons, group consistency, and inter-group similarities.

Conclusions:

• The proposed method offers a scalable and robust approach for group-wise brain parcellation.
• It enhances the understanding of group-specific brain behaviors through consistent mapping.
• The method shows significant potential for advancing in vivo neuroimaging studies.