Adaptive Tensor Regression in Neuroimaging Data Analysis

• Activity Number: 243
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Yan Zhang*+
• Companies: North Carolina State University
• Keywords: Brain imaging ; Generalized linear model ; Tensor regression ; Statistical parametric map ; Multiple comparisons ; Nonparametric

Abstract

For brain imaging data, traditional voxel-based analysis is useful in producing statistical parametric maps for testing regional significance, but can also be insufficient due to the high dimensionality and complex structural correlations of images. Zhou et al. (2013) introduced a class of tensor regression models which takes into account the inherent structure of the image/tensor covariates. In this talk, we propose an adaptive tensor regression model, in which voxel-based analysis is combined with the tensor regression. We demonstrate that this new strategy results in much improved estimation and interpretation, through both simulations and an analysis of the Alzheimer's disease neuroimaging data.