Title: Spatially varying coefficient model for functional image-on-scalar regression
Authors: Lily Wang - Iowa State University (United States) [presenting]
Guannan Wang - College of William \& Mary (United States)
Abstract: Motivated by recent work analyzing imaging data in the neuroimaging studies, a class of linear functional response regression models for imaging responses and scalar predictors is considered. To help define the possibly irregular domain of the active (non-null) region, our smoothing method is based on penalized bivariate splines over triangulations, which solves the problem of ``leakage'' across non-rectangular domains where many conventional smoothing tools suffer. Highly efficient and scalable estimation algorithm is developed. We conduct Monte Carlo simulation to examine the finite-sample performance of the proposed method. An Alzheimer's Disease Study example will be provided to illustrate the application of the proposed methods.