Brain imaging has the potential to advance our understanding of the human brain and to improve diagnosis and treatment of many neurological diseases. Intrigued by key questions in neuroscience and medicine, numerous computational methods for analyzing brain structure and function have been developed, among which, statistical approaches and shape modeling play increasingly prominent roles.

In this talk, related issues in analyzing brain images from magnetic resonance imaging (MRI) will be first introduced. I will then present a unified framework for structural MRI image segmentation (boundary finding), where a Bayesian formulation, based on prior knowledge and the edge information of the input image (likelihood), is employed. The prior knowledge in the framework is based on principle component analysis of four different covariance matrices corresponding to independence, smoothness, statistical shape, and combined models, respectively. In addition, I will demonstrate that this principle component based statistical shape modeling can also be applied to characterize brain function by integrating and classifying the shape feature parameters extracted from the functional MRI data.