Abstract
Shape analysis of neuroanatomical structures has proven useful in the study of neuropathology and neurodevelopment. Advances in medical imaging have made it possible to study this shape variation in vivo. In this paper, we propose the use of a spherical wavelet transformation to extract cortical surface shape features, as wavelets can characterize the underlying functions in a local fashion in both space and frequency. Our results demonstrate the utility of the wavelet approach in both detecting the spatial scale and pattern of shape variation in synthetic data, and for quantifying and visualizing shape variations of cortical surface models in subject populations.