A1006
Title: Incorporation of spatial- and connectivity-based cortical brain distances in regularized regression
Authors: Jaroslaw Harezlak - Indiana University School of Public Health-Bloomington (United States) [presenting]
Timothy Randolph - Fred Hutchinson Cancer Research Center (United States)
Damian Brzyski - Wroclaw University of Science and Technology (Poland)
Joaquin Goni - Purdue University (United States)
Aleksandra Steiner - University of Wroclaw (Poland)
Abstract: The aim is to address the problem of adaptive incorporation of spatial information and connectivity-based information in brain imagining data in the multiple linear regression setting. In the example considered, we model scalar outcomes as functions of the brain cortical properties, e.g. cortical thickness and cortical area. We utilize both connectivity and spatial proximity information to build adaptive penalty terms in the regularized regression problem. The general idea of incorporating external information in the regularization approach via linear mixed model representation has been recently established in our prior work, specifically in the method called: ridgified Partially Empirical Eigenvectors for Regression (riPEER). We incorporate multiple sources of information, including structural connectivity network structure as well as the spatial distance between the cortical regions to estimate the regression parameters with multiple penalty terms via a riPEER extension called disPEER (distance-based Partially Empirical Eigenvectors for Regression). We present a simulation study testing various realistic scenarios and apply disPEER to data arising from the Human Connectome Project (HCP) study.