Gopikrishna Deshpande1, Simon Lacey2,
Henrik Hagtvedt3, Venessa Patrick4, Amy Anderson2,
Randall Stilla2, Joo Ricardo Sato5, Srinivas Reddy6,
K. Sathian2, Xiaoping Hu7
1AU MRI Research Center,
Department of Electrical & Computer Engineering, Auburn University,
Auburn, AL, United States; 2Department of Neurology, Emory
University, Atlanta, GA, United States; 3Carroll School of
Management, Boston College, Chestnut Hill, MA, United States; 4C.
T. Bauer College of Business, University of Houston, Houston, TX, United
States; 5Center of Mathematics, Computation & Cognition,
Universidade Federal do ABC, Santo Andr, Brazil; 6Centre for
Marketing Excellence, Singapore Management University, Singapore; 7Coulter
Department of Biomedical Engineering, Georgia Institute of Technology &
Emory University, Atlanta, GA, United States
Hemodynamic variability can affect the validity of inferences obtained from Granger causality (GC) analysis of fMRI. Also, it is difficult to obtain context-dependent and/or dynamic connectivity from traditional GC analysis of short fMRI time series data. In order to alleviate these problems, we developed a stimulus-entrained, dynamic GC approach which not only models the time-varying connectivity but also determines whether the dynamics are entrained to external stimuli. Using simulations, we show that this approach is not affected by hemodynamic variability. Also, we demonstrate the experimental utility of this approach using a visual art paradigm.