No abstract

Behavioral research has found that trait anxiety is associated with a lower propensity for risk‐taking. However, the neural mechanisms underlying this relation are still unknown. To address this question, we employed voxel‐based morphometry (VBM) analysis and resting‐state functional connectivity (rs‐FC) to examine the influence of trait anxiety on risk‐taking. We theorized that trait anxiety may affect risk‐taking via negative prospection during episodic future thinking, which is known to be mediated by episodic memory systems including the hippocampus. We measured risk‐taking using the Balloon Analogue Risk Task (BART) and found that risk‐taking in this task was negatively correlated with trait anxiety. The VBM results suggested a positive correlation between trait anxiety and grey matter volumes in the hippocampus, consistent with previous results. Functional connectivity results indicated that functional connectivity between a right hippocampus cortex (RHPC) seed region and left insula (LInsula) was positively correlated with trait anxiety scores but negatively correlated with risk‐taking. Critically, mediation analysis showed that trait anxiety played a completely mediating role in the relation between the functional connectivity of RHPC‐LInsula and risk‐taking. These results suggested that trait anxiety can affect risk‐taking via episodic future thinking mechanisms subserved by the hippocampal cortex acting in concert with emotional and motivational control mechanisms subserved by the insular cortex.

Compulsion is one of core symptoms of obsessive–compulsive disorder (OCD). Although many studies have investigated the neural mechanism of compulsion, no study has used brain-based measures to predict compulsion. Here, we used connectome-based predictive modeling (CPM) to identify networks that could predict the levels of compulsion based on whole-brain functional connectivity in 57 OCD patients. We then applied a computational lesion version of CPM to examine the importance of specific brain areas. We also compared the predictive network strength in OCD with unaffected first-degree relatives (UFDR) of patients and healthy controls. CPM successfully predicted individual level of compulsion and identified networks positively (primarily subcortical areas of the striatum and limbic regions of the hippocampus) and negatively (primarily frontoparietal regions) correlated with compulsion. The prediction power of the negative model significantly decreased when simulating lesions to the prefrontal cortex and cerebellum, supporting the importance of these regions for compulsion prediction. We found a similar pattern of network strength in the negative predictive network for OCD patients and their UFDR, demonstrating the potential of CPM to identify vulnerability markers for psychopathology.

No abstract