Department of Biostatistics


Prospective Students



News & Events



Employment Opportunities

Resource Quicklinks

Computing Environment




Structure/Function Relationships in the Analysis of Anatomical and Functional Neuroimaging Data

 Shu-Chih Su, PhD Candidate, Johns Hopkins Department of Biostatistics

The major goal of this dissertation is to develop statistical methodology for analyzing and interpreting brain images in order to improve the effectiveness of the neoroimaging research. Neuroimaging data is complex, due to the high-throughput nature, inherent spatial associations, and in some cases, temporal associations. There are three major components within this dissertation; each aims to promote statistical principles in the analysis of neuroimaging data and to enhance our knowledge of how brain structure and function relate to our health and disease. In the first component, we proposed a modified test statistic using inter-voxel variance shrinkage with an application to functional Magnetic Resonance Imaging (fMRI) data. Due to the typically limited number of subjects in an fMRI study, accurate estimation of variances at each voxel is difficult. Thus combining information across voxels in the statistical analysis of fMRI data is desirable, in order to improve efficiency. We construct a hierarchical model and apply a Bayesian framework for the group analysis of fMRI contrast maps. The performance was evaluated through a series of simulation studies and on an auditory word-pair-associates experiment. The experimental motivation of the second and the third components involve investigating volumetric mediation of a toxicant's influence on cognitive function and understanding the anatomic resolution of leadís impact on brain structure. The motivating data come from an ongoing prospective study of lead's impact on the central nervous system structure and function. Previous results from this study have showed that lifetime cumulative lead dose, measured as lead concentration in tibia bone by X-ray fluorescence, was associated with persistent and progressive declines in cognitive function and with decreases in MRI-based brain volumes in former lead workers. In the second manuscript, these findings motivated us to explore a novel application of path analysis to evaluate effect mediation, specifically, whether the association of lead dose with cognitive function is mediated through brain volumes, on a voxel-wise basis. In the third manuscript, hierarchical models built on partitions of the brain were used to understand the impact of lead on the brain structure at multiple resolutions. Throughout all of these works, we aim to understand the structural and function of the brain and its relationship with the environment, in the terms of its physiology, functional architecture and dynamics.

Return to
Biostatistics Calendar | Return to Home Page
© 2004, The Johns Hopkins University. All rights reserved.
web policies, 615 N. Wolfe Street, Batimore, MD 21205-2179, 410-955-5000