Activity Number:
|
449
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 2, 2016 : 2:00 PM to 2:45 PM
|
Sponsor:
|
Section on Statistics in Imaging
|
Abstract #321695
|
|
Title:
|
Model-Free Estimation of Task-Based Dynamic Functional Connectivity and Its Confidence Intervals
|
Author(s):
|
Maria Kudela* and Mario Dzemidzic and Brandon G. Oberlin and Joaquín Goñi and David A. Kareken and Jaroslaw Harezlak
|
Companies:
|
Indiana University and Indiana University School of Medicine and Indiana University School of Medicine and Purdue University and Indiana University School of Medicine and Indiana University Fairbanks School of Public Health
|
Keywords:
|
time series ;
dynamic functional connectivity ;
time-varying correlation ;
task-based experiments ;
multivariate time series bootstrap ;
inference for functional data
|
Abstract:
|
Dynamic functional connectivity (dFC) is an emerging field of interest in functional magnetic resonance imaging (fMRI) studies. dFC measures time-dependent associations between brain regions estimated from the blood oxygenation level dependent (BOLD) time series. dFC changes are most commonly quantified by pairwise correlation coefficients between the time series within a short time sliding window of 30 to 50 measurements. In our work, we apply and expand a recently developed bootstrap-based technique (Kudela et al. 2015) to robustly estimate dFC and its confidence intervals in a task-based gustatory fMRI study. We summarize the dFC by reporting the percentage of the time points when the dynamic correlation is either significantly positive or negative and provide a group-level estimate of the dFC. We apply our method to a gustatory fMRI task and find strong connectivity between homologous and reward-related regions. Our initial investigation is being expanded to characterize whole brain network associations, and to relate dFC patterns with recent alcohol consumption, familial alcoholism and flavor types.
|
Authors who are presenting talks have a * after their name.