Functional Mediation Analysis of Functional Magnetic Resonance Imaging Experiments

• Activity Number: 243 - Brain Structural and Functional Connectivity Analysis
• Type: Contributed
• Date/Time: Monday, July 30, 2018 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Yi Zhao* and Xi Luo and Martin A Lindquist and Brian Caffo
• Companies: Johns Hopkins Bloomberg School of Public Health and Brown University and Johns Hopkins University and Johns Hopkins University
• Keywords: causal mediation; functional data; dynamic connectivity

Abstract

In this study, we introduce a functional mediation analysis framework in which the three key variables, the treatment, the mediator and the outcome, are continuous functions. Causal mediation analysis is widely utilized to delineate the causal effect of treatment into the direct effect on the outcome and the indirect effect passing through an intermediate variable (mediator). With functional measures, causal assumptions and interpretations are not immediately well-defined. Motivated by a functional magnetic resonance imaging (fMRI) study, we propose two functional mediation models based on the influence of the mediator: (1) concurrent and (2) historical. We further discuss causal assumptions, and elucidate causal interpretations. Our proposed models enable the estimate of individual causal effect curves, where both the direct and indirect effects vary across time. Applied to a task-based fMRI study, our functional mediation framework provides a new perspective of studying dynamic brain connectivity.