Brain Dynamic Functional Connectivity: Modularity Assessment in a Task-Based Gustatory Study

• Activity Number: 324 - Applications of Functional Data Analysis to Medical Studies
• Type: Topic Contributed
• Date/Time: Tuesday, August 1, 2017 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Maria Kudela* and Mario Dzemidzic and Joaquin Goni and Brandon G Oberlin and David Kareken and Jaroslaw Harezlak
• Companies: and Department of Neurology Indiana University School of Medicine and Purdue University and Department of Neurology Indiana University School of Medicine and Department of Neurology Indiana University School of Medicine and Indiana University School of Public Health
• Keywords: functional brain imaging ; dynamic functional connectivity ; modularity ; additive mixed models

Abstract

Times series data collected in the fMRI brain activation studies are frequently used to study the co-activation (functional connectivity - FC) of brain regions. In our work, we use the data from a gustatory task fMRI study. We propose a novel combination of statistical and graph theory methods to quantify the changes in the dynamic FC (dFC). Specifically, we utilize the subject-specific nonparametric estimates of dFC (Kudela et al. 2017) in the additive mixed model framework to obtain the group-level dFC estimates. Subsequently, Louvain algorithm is applied to assess the dFC modularity defined as the mutually exclusive division of brain regions into blocks with intra-connectivity greater than the one expected by chance. The proposed approach produces low-dimensional brain partition representation suggesting the existence of common functionally-based brain organization. Further investigation at the subject-by-region-specific level shows biologically meaningful differences in brain regions' co-activation during the gustatory stimulation. The methods developed provide a novel approach to study the co-activation dynamics arising from both task-based and resting state fMRI data