Figure 2.
Macroscale network connectivities of FA in predicting drinking PC1. (A) Positive (red) and negative (blue) networks correlated with PC1, where larger and smaller nodes represent more and less connections, respectively. In the positive network, higher edge weights were associated with more severe drinking, whereas in the negative network, lower edge weights were associated with more severe drinking. (B) Correlation between actual (X-axis) and predicted (Y-axis) drinking PC1 values generated using CPM. (C) Mediation analysis illustrating the relationship between negative network characteristics, PC1, and rule-breaking behavior scores. CPM, connectome-based predictive modeling; FA, fractional anisotropy; L, left; R, right; Par, parietal; Cer, cerebellum; Mot, motor; Sub, subcortical; PFC, prefrontal; Ins, insula; Tem, temporal; Occ, occipital; Lim, limbic; PC1, first principal component (obtained of PCAs of all drinking measures); PCA, principal component analysis.