Brain mapping using topology graphs obtained by surface segmentation

Abstract

Brain mapping is a technique used to alleviate the tedious and time-consuming process of annotating brains by mapping existing annotations from brain atlases to individual brains. We introduce an automated surface-based brain mapping approach. After reconstructing a volume data set (trivariate scalar field) from raw imaging data, an isosurface is extracted approximating the brain cortex. The cortical surface can be segmented into gyral and sulcal regions by exploiting geometrical properties. Our surface segmentation is executed at a coarse level of resolution, such that discrete curvature estimates can be used to detect cortical regions. The topological information obtained from the surface segmentation is stored in a topology graph. A topology graph contains a high-level representation of the geometrical regions of a brain cortex. By deriving topology graphs for both atlas brain and individual brains, a graph node matching defines a mapping of brain cortex regions and their annotations.