Multimoding MRI Study On The Personality Dimension Of Harm Avoidance

Objective:Harm avoidance (HA) is a personality dimension involving the tendency to respond intensely to signals of aversive stimuli. While its neural substrates remain elusive. We combined voxel-based morphometry (VBM) analysis, independent component analysis (ICA) and resting-state functional connectivity (rsFC) based on ROI analysis of fMRI data to investigate neural correlates of HA scores in a large sample of healthy young adults.Subjects and Methods:A total of291right-handed healthy young adults were recruited in this study. After completing the tests of the temperament of the Tridimensional Personality Questionnaire (TPQ) and the Beck Depression Inventory (BDI), the subjects’ resting-state fMRI and sagittal structural images were collected using a GE3.0T Signa HDX scanner. During the resting-state scanning, all subjects were explicitly instructed to keep their eyes closed, relax, as motionless as possible and think of nothing.1. All preprocessing steps were carried out using the statistical parametric mapping (SPM8) based on Matlab.2. Voxel-based partial correlation analyses were carried out to test correlations between the HA scores and the GMV while controlling age, gender and BDI scores.3. After ICA by the GIFT software based on Matlab, we chooseed eleven RSNs from the results of ICA. The partial correlation analyses were carried out to test correlations between the HA scores and rsFCs within the eleven RSNs while age, gender and BDI scores were entered as covariates.4. The amygdala subregions were extracted using the maximum probabilistic map (MPM). The relationships between HA scores and amygdala subregion rsFCs were analyzed using a multiple regression model within the significant mask of each amygdala subregion while controlling for age and BDI scores.5. The multiple comparisons were statistically corrected by Monte Carlo simulation. Finally the corrected statistical parameter mapping was overlapped onto MNI standard template and the coordinate, size and peak t values of each significant cluster were recorded.Results:1. The HA scores were negatively correlated with the GMVs in the right inferior temporal gyrus, right insular cortex, the bilateral precuneus and right superior parietal lobule.2. The relationships between HA scores and RSNs were found in dorsal attention network, the ventral attention network, the right frontoparietal network, the sensorimotor network.3. The rsFC patterns of the amygdalar subregions showed different. HA scores were correlated with rsFCs between the LB and temporal and occipital cortices related to emotional information input, between the CM and the frontal cortices associated with emotional output control, and between the SF and the frontal and temporal areas involved in both functions.Conclusion:1. HA scores were related to gray matter volumes of several brain areas.2. HA scores were also related to resting-state functional networks.3. These findings not only confirmed associations between HA score and the rsFCs of amygdalar subregions, but also found sex differences in these correlations.