|
Activity Number:
|
240
- Statistical Analysis of Complex Imaging Data
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2020 : 1:00 PM to 2:50 PM
|
|
Sponsor:
|
WNAR
|
|
Abstract #309360
|
|
|
Title:
|
A Source Separation Method for Investigating Brain Connectome Traits
|
|
Author(s):
|
Ying Guo* and Yikai Wang
|
|
Companies:
|
Emory University and Emory University
|
|
Keywords:
|
connectome;
neuroimaging;
source separation;
low-rank factorization;
brain connectivity;
sparsity regularization
|
|
Abstract:
|
In recent years, network-oriented study has become increasingly important in neuroscience research. Examining brain functional and structural connectome traits could potentially offer new insights on brain organization and how it is altered by neurological disorders. In this talk, we present a signal separation framework to conduct fully data-driven decomposition of multi-subject connectivity matrices to extract population-level latent connectivity traits. The findings could help uncover neural circuits that are differentially expressed in clinical subpopulations or get disrupted in brain disorders. The proposed method aims to achieve more efficient and accurate source separation for connectivity matrices by assuming a low-rank factorization structure that is motivated by the characteristics of the observed neural connectivity. We also propose a novel sparsity regularization method to reduce spurious findings. An efficient estimation method is proposed to optimize the objective function with the new sparsity regularization. We demonstrate the performance of the method via simulation studies and further illustrate brain connectome traits extracted from connectivity data.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2020 program
|
