Sparse Learning and Structure Identification for Ultrahigh-Dimensional Image-on-Scalar Regression

This article considers high-dimensional image-on-scalar regression, where the spatial heterogeneity of covariate effects on imaging responses is investigated via a flexible partially linear spatially varying coefficient model. To tackle the challenges of spatial smoothing over the imaging response's complex domain consisting of regions of interest, we approximate the spatially varying coefficient functions via bivariate spline functions over triangulation. We first study estimation when the active constant coefficients and varying coefficient functions are known in advance. We then further develop a unified approach for simultaneous sparse learning and model structure identification in the presence of ultrahigh-dimensional covariates. Our method can identify zero, nonzero constant, and spatially varying components correctly and efficiently. The estimators of constant coefficients and varying coefficient functions are consistent and asymptotically normal for constant coefficient estimators. The method is evaluated by Monte Carlo simulation studies and applied to a dataset provided by the Alzheimer's Disease Neuroimaging Initiative.for this article are available online.

Automated summary

This article discusses high-dimensional image-on-scalar regression and proposes a flexible partially linear spatially varying coefficient model to investigate the spatial heterogeneity of covariate effects on imaging responses. The spatially varying coefficient functions are approximated via bivariate spline functions over triangulation to address the challenges of spatial smoothing over the imaging response's complex domain. The method can simultaneously perform sparse learning and model structure identification in the presence of ultrahigh-dimensional covariates, accurately and efficiently identifying zero, nonzero constant, and spatially varying components.