Bayesian Analysis of Spatial Transformation Models with Applications in Neuroimaging Data

• Activity Number: 38
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Michelle Miranda*+ and Hongtu Zhu and Joseph G. Ibrahim
• Companies: and UNC-Chapel Hill and UNC
• Keywords: Big data ; MCMC ; Neuroimaging data ; Gaussian Markov random field ; Box-cox transformation ; Bayesian Analysis

Abstract

The aim of this paper is to develop a class of spatial transformation models (STM) to spatially model the varying association between imaging measures in a three-dimensional (3D) volume (or 2D surface) and a set of covariates. Our STMs include a varying Box-Cox transformation model for dealing with the issue of heterogeneous non-Gaussian distributed imaging data and a Gaussian Markov Random Field model for incorporating spatial smoothness of imaging data. Posterior computation proceeds via an efficient Markov chain Monte Carlo algorithm. Simulations and real data analysis demonstrate that the STM significantly outperforms the voxel-wise linear model with Gaussian noise in recovering meaningful geometric patterns. Our STM is able to reveal important brain regions with morphological changes in children with attention deficit hyperactivity disorder (ADHD).