An automated 3D algorithm for neo-cortical thickness measurement

Abstract

A new method is presented for fully automated regional f quantification of human neo-cortical thickness from 3-D MR images. The method does not require explicit extraction of the cortical surface, but instead is based on a geodesic distance transform of the binarized 3-D volumetric gray matter map obtained by automated segmentation. Regional cortical thickness distributions are derived by atlas-based partitioning of the brain. Results for 64 normal controls are presented that show very good agreement with published reference values.