Abstract:
|
Motivated by a recent interest in data fusion methods in medical imaging, we discuss novel approaches for joint analysis of multiple neuroimaging datasets. In the first part of the talk, I discuss our novel Bayesian structured graphical learning approach for estimating brain functional connectivity guided by structural information. We illustrate that multimodal learning results in biologically reproducible results compared to analysis using a single platform. In the second part, we propose a novel approach for modeling the longitudinal progression of Alzheimer's disease using T1w-MRI data. Our approach differs from existing approaches in this area in terms of accounting for the spatial configuration of the voxels as well as the noise in the brain images, which leads to increased predictive accuracy and discovery of important brain regions associated with neurodegeneration that are simply missed by existing methods without noise correction due to the attenuation to null phenomenon.
|