Anatomy-guided discovery of large-scale consistent connectivity-based cortical landmarks

Area of Science:

• Neuroscience
• Brain Mapping
• Computational Anatomy

Background:

• Establishing structural and functional correspondences across brains is crucial for human brain mapping.
• Previous multimodal diffusion tensor imaging (DTI)/functional magnetic resonance imaging (fMRI) studies used white matter connections for brain architecture analysis.
• Existing methods for discovering cortical landmarks often overlook rich anatomical information like gyral/sulcal patterns.

Purpose of the Study:

• To present a novel anatomy-guided framework for discovering dense cortical landmarks.
• To integrate anatomical, morphological, and fiber connectional information for landmark optimization.
• To establish a universal and individualized brain reference system for neuroimaging.

Main Methods:

• Developed an anatomy-guided discovery framework integrating anatomical, morphological, and fiber connectional data.
• Formulated landmark initialization, optimization, and prediction as an energy minimization problem.
• Utilized DTI and fMRI data for validation.

Main Results:

• Identified 555 cortical landmarks with group-wise consistent anatomical and fiber connectional profiles.
• Demonstrated that the identified landmarks are reproducible and predictable.
• Showcased accurate structural and functional correspondences across individuals and populations using fMRI data.

Conclusions:

• The novel framework successfully integrates diverse neuroimaging data for robust landmark discovery.
• The identified landmarks provide a reliable basis for understanding brain structure-function relationships.
• This approach offers a universal yet individualized brain reference system for advancing neuroimaging research.