An integrated multimodal model of alcohol use disorder generated by data-driven causal discovery analysis

摘要

Causal discovery analysis of the neurobehavioral determinants of AUD symptom severity in the HCP dataset was done using Greedy Fast Causal Inference (GFCI). GFCI returns a partial ancestral graph (PAG) depicting causal relationships between a set of variables, while assessing for unmeasured third variables in relationships (confounders). Standardized edge weights recovered via structural equation modeling (SEM) are displayed in text next to each edge in the graph. The overall SEM fit was good, RMSEA = 0.06, Tucker-Lewis Index = 0.91.