The neural correlates of guilt highlight preclinical manifestations between bipolar and major depressive disorders.

Background: Over 25 % of bipolar disorder (BD) patients are misdiagnosed with major depressive disorder (MDD). An urgent need exists for a biomarker to differentiate BD from MDD. Various manifestations and intensities of maladaptive guilt processing might uniquely contribute to the pathogenesis of BD compared to MDD.

Method: This study adopted a first-person perspective guilt-provoking functional magnetic resonance imaging (fMRI) task, respectively induced by painful and ambiguous scenarios in 111 individuals: 35 with remitted MDD, 38 with remitted bipolar I disorder (BD-I), and 38 matched controls.

Results: A significant interaction between group and sense of agency in predicting guilt ratings for ambiguous, rather than painful, scenarios. The association between sense of agency and guilt was significant in MDD but not in BD-I patients or controls. Activation in the dorsomedial prefrontal cortex (dmPFC), pregenual anterior cingulate cortex (pgACC), and right inferior parietal lobule (IPL) was higher in BD-I than MDD subjects in response to ambiguous scenarios, whereas these were comparable to painful ones. The correlation between guilt ratings and activation in the dorsal anterior cingulate cortex (dACC) to ambiguous scenarios was significant in MDD, but not in BD-I or controls. The results of the multivariate pattern classification analysis showed that in the ambiguous scenarios, the accuracy of using brain activation patterns in the dmPFC, pgACC, and IPL to distinguish between participants with MDD and BD-I was 70.0 %, 71.5 %, and 68.7 %, respectively. An additional test of the classification model, employing a combined mask of the three ROIs to distinguish between the two mood disorders in ambiguous scenarios, achieved an accuracy of 78.6 % and an AUC value of 0.84.

Conclusions: Subjective reports and neural correlates of guilt in ambiguous social situations, as well as a sense of agency, could provide to be a potential biomarker to help distinguish between BD-I and MDD even in the remitted stage.