Multi-scale Voxel-based Morphometry via Weighted Spherical Harmonic Representation

Abstract. Although the voxel-based morphometry (VBM) has been widely used in quantifying the amount of gray matter of the human brain, the optimal amount of registration that should be used in VBM has not been addressed. In this paper, we present a novel multi-scale VBM using the weighted spherical harmonic (SPHARM) representation to address the issue. The weighted-SPHARM provides the explicit smooth functional representation of true unknown cortical boundary. Based on this new representation, the gray matter tissue density is constructed using the Euclidean distance map from a voxel to the estimated smooth cortical boundary. The methodology is applied in localizing abnormal cortical regions in a group of autistic subjects.