Department of Mathematics and Statistics

Department of Mathematics and Statistics
Department of Mathematics and Statistics

Department Colloquium


Farouk Nathoo  (University of Victoria)

Friday, October 9th, 2020

Time: 2:30 p.m.  Place: Online (via Zoom)

Speaker: Farouk Nathoo (University of Victoria)

Title: Spatial Statistical Modeling for Neuroimaging Data.

Abstract: I will describe three projects involving the analysis of neuroimaging data and the development of hierarchical spatial Bayesian models for each. In the first, we develop an approach for determining the location and dynamics of brain activity from combined magnetoencephalography and electroencephalography data. The resulting inverse problem is ill-posed and we propose a distributed solution based on a Bayesian spatial finite mixture model that incorporates the Potts model to represent the spatial dependence in an allocation process that partitions the cortical surface into a small number of latent states. In the second project, we consider statistical modelling of functional magnetic resonance imaging (fMRI) data which is challenging in part as the data are both spatially and temporally correlated. Motivated by an event‐related fMRI experiment, we propose a novel hierarchical Bayesian model with automatic selection of the auto‐regressive orders of the noise process that vary spatially over the brain. In the third project, we develop a Bayesian bivariate spatial model for multivariate regression analysis applicable to studies examining the influence of genetic variation on brain structure. Our model is motivated by an imaging genetics study of the Alzheimer's Disease Neuroimaging Initiative, where the objective is to examine the association between images of volumetric and cortical thickness values summarizing the structure of the brain as measured by magnetic resonance imaging (MRI) and a set of 486 SNPs from 33 Alzheimer's Disease (AD) candidate genes obtained from 632 subjects. A bivariate spatial process model is developed to accommodate the correlation structures typically seen in structural brain imaging data and we develop a mean-field variational Bayes algorithm and a Gibbs sampling algorithm to fit the model. We compare the new spatial model to an existing non-spatial model in our motivating application.