Homology and topology based metrics for evaluating cortical parcellations generated using diffusion MRI

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Brain Mapping

Background:

• Diffusion MRI is used for cerebral cortex segmentation, but processing choices significantly impact results.
• There is no established consensus on optimal diffusion MRI processing protocols or in vivo ground truth for parcellation.
• Current validation methods for cortical parcellations rely on indirect evidence, including cross-modality comparisons, reproducibility, and functional distinctiveness.

Purpose of the Study:

• To propose and validate a novel strategy for assessing diffusion MRI-based cortical parcellation outcomes.
• To evaluate parcellation results based on established principles of brain organization, specifically inter-hemispheric homology and topology.
• To provide subject-independent measures for guiding improvements in diffusion MRI acquisition, reconstruction, and clustering for parcellation.

Main Methods:

• Developed a new assessment strategy for diffusion MRI parcellation by evaluating results against known brain organizational principles.
• Focused on inter-hemispheric homology and topology for assessing parcellation accuracy within individual subjects.
• This method allows for evaluation independent of other imaging modalities or acquisitions.

Main Results:

• The proposed method enables direct, subject-specific assessment of parcellation quality based on neuroanatomical principles.
• This approach complements existing validation techniques by offering an intrinsic measure of parcellation fidelity.
• Results can guide optimization of diffusion MRI data processing pipelines.

Conclusions:

• Assessing diffusion MRI parcellations using inter-hemispheric homology and topology offers a robust, subject-specific validation method.
• This strategy can significantly improve the reliability and accuracy of cerebral cortex segmentation using diffusion MRI.
• The proposed measures are crucial for advancing diffusion MRI-based brain mapping and understanding cortical organization.