Functional Linear Regression Model with Scalar on Image Predictors

• Activity Number: 638
• Type: Topic Contributed
• Date/Time: Thursday, August 7, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Yihong Zhao*+ and R. Todd Ogden and Huaihou Chen
• Companies: New York University and Columbia University and New York University
• Keywords: image analysis ; wavelet-based LASSO ; penalized regression ; screening strategies ; permutation test ; bootstrap based confidence intervals

Abstract

Resting-state functional magnetic resonance imaging (fMRI) is sensitive to functional brain changes related to many psychiatric disorders and thus becomes increasingly important in medical research. One useful approach for fitting linear models with scalar outcomes and image predictors involves transforming the functional data to the wavelet domain and converting the data fitting problem to a variable selection problem. Applying the LASSO procedure in this situation has been shown to be efficient and powerful. In this study we explore possible directions for improvements to this method. The finite sample performance of the proposed methods will be compared through simulations and real data applications in mental health research. We believe applying these procedures can lead to improved estimation and prediction as well as better stability.