Activity Number:
|
532
- Making Big and Complex Imaging Data Count with New Statistical Tools
|
Type:
|
Invited
|
Date/Time:
|
Thursday, August 6, 2020 : 1:00 PM to 2:50 PM
|
Sponsor:
|
SSC (Statistical Society of Canada)
|
Abstract #309260
|
|
Title:
|
Adaptive Regularization in Complex Settings: Multimodal Brain Imaging
|
Author(s):
|
Jaroslaw Harezlak* and Damian Brzyski and Kewin Paczek and Joaquin Goni and Timothy Randolph and Beau Ances
|
Companies:
|
Indiana University and Wroclaw University of Science and Technology, Poland and Jagiellonian University, Krakow, Poland and Purdue University and Fred Hutchinson Cancer Research Center and Washington University School of Medicine
|
Keywords:
|
Regularization;
Brain imaging;
Structural Connectivity;
Functional Connectivity;
Longitudinal Data
|
Abstract:
|
A problem frequently occurring in brain imaging research is a principled incorporation of information from different imaging modalities in association studies. Oftentimes, data from each modality are studied separately resulting in a loss of information. We propose a novel regularization method incorporating information from structural imaging, structural connectivity and functional connectivity in the longitudinal setting. In our work, the penalty term is defined from the structural and functional connectivity modularity information. We address both theoretical and computational issues and show that our method adapts to the incomplete or mis-specified brain connectivity information. Our regularization method is evaluated via extensive simulation studies and it is applied in a study of HIV+ individuals’ longitudinal neurodegeneration.
|
Authors who are presenting talks have a * after their name.