A Bayesian Approach to the Study of Dynamic Functional Connectivity Networks in fMRI Data

• Activity Number: 512
• Type: Invited
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Michele Guindani* and Marina Vannucci and Ryan Warnick and Erik Erhardt and Elena Allen and Vince D. Calhoun
• Companies: MD Anderson Cancer Center and Rice University and Rice University and University of New Mexico and University of New Mexico and University of New Mexico
• Keywords: fMRI data ; Dynamic Connectivity ; Bayesian Modeling ; Bayesian Graphical Models

Abstract

fMRI studies have traditionally assumed stationarity of the connectivity patterns observed in a subject during a fMRI experiment. While the assumption has successfully allowed to study large-scale properties of brain functioning, it is generally recognized that functional connectivity varies with time and tasks performed. In this talk, we describe a novel Bayesian methodological framework for the analysis of temporal dynamics of functional networks in task-based fRMi data collected on a single subject. Our proposed formulation allows joint modeling of the task-related activations in addition to the dynamics of individual functional connectivity. Furthermore, we allow simultaneous learning of the common and differential edges (interactions) in the inferred time-varying functional networks. We illustrate the proposed approach by means of simulation and an analysis on a real fMRI dataset.