Cortical sulcal atlas construction using a diffeomorphic mapping approach.

Area of Science:

• Neuroscience
• Computational Geometry
• Medical Imaging

Background:

• Sulcal and gyral patterns on the human cortex are unique and contain important anatomical information.
• Previous methods for analyzing sulcal shapes have limitations in capturing complex geometric properties and establishing correspondences.

Purpose of the Study:

• To present a novel geometric approach for constructing shape atlases of human cortical sulcal curves.
• To facilitate computational analyses and correspondences between diverse sulcal shapes.

Main Methods:

• Representing sulci and gyri as continuous open curves in 3D space (R3).
• Studying sulcal shapes as elements of an infinite-dimensional sphere with a Riemannian L2 metric.
• Utilizing geodesics in a quotient space (modulo rigid rotations and reparameterizations) for sulcal mapping.
• Integrating an elastic shape model into a surface registration framework.

Main Results:

• The developed shape manifold possesses properties that facilitate computational analyses and correspondences.
• The resulting sulcal shape atlas effectively preserves local geometry inherent in the population.
• Experimental results demonstrate the method's efficacy on deep brain sulci in a cohort of 69 healthy adults.

Conclusions:

• The geometric approach provides a robust framework for analyzing and mapping human cortical sulcal shapes.
• The shape atlas construction method preserves crucial geometric details, aiding in comparative neuroanatomy.
• This technique offers a powerful tool for studying variations in cortical morphology within populations.