Brain Connectivity-Informed Adaptive Regularization for Generalized Outcomes

• Activity Number: 212 - Scientifically and Clinically Motivated Statistical Methods for Human Brain Data Analysis
• Type: Invited
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Jaroslaw Harezlak* and Damian Brzyski and Marta Karas and Beau Ances and Joaquin Goni and Mario Dzemidzic and Timothy Randolph
• Companies: Indiana University School of Public Health and Wroclaw Technological University and Johns Hopkins School of Public Health and Washington University School of Medicine and Purdue University and Indiana University School of Medicine and Fred Hutchinson Cancer Research Center
• Keywords: regularization; generalized outcomes; brain imaging; connectivity; HIV infection

Abstract

One of the challenging problems in brain imaging research is a principled incorporation of information from different imaging modalities in regression models. Frequently, data from each modality is analyzed separately using, for instance, dimensionality reduction techniques, which result in a possible loss of information. We propose a novel regularization method, griPEER (generalized ridgified Partially Empirical Eigenvectors for Regression) to estimate the association between the brain structure features and a scalar outcome within the generalized linear regression framework. griPEER provides a principled approach to use external information from the structural brain connectivity to improve the regression coefficient estimation. Our proposal incorporates a penalty term, derived from the structural connectivity Laplacian matrix, in the penalized generalized linear regression framework. We address both theoretical and computational issues and show that our method is robust to the incomplete structural brain connectivity information. griPEER is evaluated via extensive simulation studies and it is applied in classification of the HIV+ and HIV- individuals.