A0476
Title: Biclustering multivariate longitudinal data: Application to white matter recovery trajectories after sport-related mTBI
Authors: Jaroslaw Harezlak - Indiana University School of Public Health-Bloomington (United States) [presenting]
Luo Xiao - North Carolina State University (United States)
Abstract: Biclustering is the task of simultaneously clustering the samples and features of a dataset. In doing so, subsets of samples that exhibit similar behaviours across subsets of features can be identified. Motivated by a longitudinal diffusion tensor imaging study of sport-related concussion (SRC), the problem of biclustering multivariate longitudinal data in which subjects and features are grouped simultaneously based on longitudinal patterns rather than a magnitude is presented. A penalized regression-based method is proposed for solving this problem by exploiting the heterogeneity in the longitudinal patterns within subjects and features. The performance of the proposed methods is evaluated via a simulation study, and an analysis of the motivating data set is performed. Subgroups of SRC cases that exhibit heterogeneous patterns of white-matter abnormalities are revealed.
