In NeuroImage ; h5-index 117.0
Mediation analysis is used to investigate the role of intermediate variables (mediators) that lie in the path between an exposure and outcome variable. While significant research has focused on developing methods for assessing the influence of mediators on the exposure-outcome relationship, current approaches do not easily extend to settings where the mediator is high-dimensional (e.g., neuroimaging, genomics, and metabolomics). Here we introduce a novel machine learning based method for identifying high-dimensional mediators. The proposed algorithm is agnostic to the machine learning model used, providing significant flexibility in the types of situations it can be applied. We illustrate the proposed methodology using data from two functional Magnetic Resonance Imaging studies. In both, our multivariate mediation model links exposure variables, high dimensional brain measures and behavioral outcomes into a single unified model. Using the proposed approach, we identify brain-based measures that simultaneously encode the exposure variable and correlate with the behavioral outcome.
Nath Tanmay, Caffo Brian, Wager Tor, Lindquist Martin A
2022-Dec-28
