Hidden nodes of personality: functional brain networks and their trait correlates

Abstract This study investigates the associations between personality traits and the topology of resting-state functional brain networks, aiming to identify trait-specific global, nodal, and predictive network markers. Graph-theoretical metrics of functional connectivity were related to psychometric measures of personality traits using correlational analysis, unsupervised clustering, principal component analysis (PCA), and predictive modeling. Nonlinear relevance vector machine (RVM) and linear regression models were applied to assess the predictability of traits from network measures. Psychopathic traits were moderately and negatively associated with global integration measures, including mean node strength ( r = − 0.264) and mean clustering coefficient ( r = − 0.247), indicating reduced local segregation and integration. Masochistic traits showed a negative correlation with mean betweenness centrality ( r = − 0.230), suggesting a more distributed network architecture. At the nodal level, eigenvector centrality was positively associated with masochistic traits in frontal mid regions ( r = 0.28–0.32), left inferior frontal gyrus ( r = 0.27–0.31), left postcentral gyrus ( r = 0.26), and supramarginal gyrus ( r = 0.27–0.31). Psychopathic traits were positively related to eigenvector centrality of Heschl’s gyrus ( r = 0.25–0.26). A replicable association was observed between oral traits, node strength, and the left nucleus accumbens ( r = 0.26), indicating increased centrality of reward-related circuits. PCA revealed three latent components explaining 75% of total variance, separating psychopathic–masochistic, oral–rigid, and masochistic trait dimensions. In predictive analyses, RVM models achieved the lowest errors (mean RMSE < 0.4), with the highest accuracy for psychopathic traits using functional connectivity (RMSE = 0.40) and for masochistic traits using node strength (RMSE = 0.346). Linear regression yielded exceptionally low error for psychopathic traits predicted by betweenness centrality (RMSE = 0.122). Our work offers a modest contribution to the development of a nomological network of traits within the bioenergetic analytical paradigm in the domain of personality neuroscience. Personality traits are associated with distinct and partially nonlinear patterns of functional brain network organization. Both global topology and trait-specific nodal centrality contribute to the neural expression and prediction of personality dimensions, supporting a multilevel network-based model of personality.