Discovering Cortical Folding Patterns in Neonatal Cortical Surfaces Using Large-Scale Dataset.

Area of Science:

• Neuroimaging and Computational Neuroscience
• Human Brain Development and Genetics

Background:

• Human brain cortical folding is complex, variable, and genetically influenced, established early in development.
• Understanding cortical folding patterns is crucial for neuroimaging analysis and linking brain structure to cognition and disorders.
• Neonatal brains offer an ideal window into primary cortical folding due to minimal postnatal environmental influence.

Purpose of the Study:

• To introduce a novel computational method for discovering major cortical folding patterns.
• To analyze a large-scale dataset of neonatal brain Magnetic Resonance Imaging (MRI) scans.
• To characterize and group distinct sulcal folding patterns in key brain regions.

Main Methods:

• Cortical folding was characterized by analyzing the distribution of sulcal pits, representing deep points within cortical sulci.
• Similarity between sulcal pit distributions was measured using spatial, geometrical, and topological features.
• A similarity network fusion technique adaptively combined these measurements, followed by hierarchical affinity propagation for pattern grouping.

Main Results:

• The method was applied to 677 neonatal brain MRIs, the largest neonatal dataset to date for this type of analysis.
• Multiple distinct and meaningful cortical folding patterns were identified in the central sulcus, superior temporal sulcus, and cingulate sulcus.
• The findings highlight genetically influenced, consistent patterns in early brain development.

Conclusions:

• The proposed novel method effectively discovers and groups major cortical folding patterns in large neonatal brain datasets.
• Sulcal pit distribution analysis provides a robust approach for characterizing genetically influenced cortical folding.
• This research advances neuroimaging techniques and deepens the understanding of early brain structural variations.