Title: Topological Correction of Brain Surface Meshes Using Spherical Harmonics

Surface reconstruction methods allow advanced analysis of brain data beyond what can be achieved using volumetric images alone. For instance, signals from distinct functional areas within the cortical sheet can be aligned, averaged and compared across subjects, while correctly modeling the folded geometry of the cortex. Even so, automated generation of cortical surface meshes from MRI data often leads to topological defects and geometrical artifacts that usually must be corrected to permit subsequent analysis. Topological defects prevent the surface from being homeomorphic with a sphere, while artifacts are topologically correct sharp peaks (usually due to noise) that do not represent the true cortical anatomy. Spherical harmonics (SPH) have been recently used for modeling brain surface meshes, usually in the realm of shape analysis (Brechbühler et al., 1995, Thompson and Toga, 1996). Here, we propose a method to repair cortical surface defects using a reconstruction based on SPH.