Contextual memory bias in emotional events: Neurobiological correlates and depression risk

Abstract

Background

Contextual memory loss of emotional events plays a critical role in depression psychopathology. Individuals with depression, clinical or subclinical, exhibit enhanced and impaired memory for emotionally negative stimuli and context in an event, respectively. This suggests that contextual encoding may fail because of attentional interference caused by concurrent negative stimuli, possibly leading to contextual memory loss as a depression risk. Amygdala–prefrontal connectivity and cortisol may underlie the mechanism; however, the relationships remain unknown.

Methods

One hundred twenty participants, including 34 with subclinical depression, underwent behavioral tasks, functional magnetic resonance imaging (fMRI) scans, and saliva collection. Encoding and 24 h later recollection performance of visuoperceptual/spatial/temporal context in a series of events, where fearful (vs. neutral) faces appeared, were measured via contextual memory tasks. Overgeneral autobiographical memory (OGM), a more remote form of contextual memory loss, was also assessed via the Autobiographical Memory Test. Amygdala connectivity was measured by fMRI during attentional interference by fearful (vs. neutral) faces to differentiate selective attention from encoding. Basal cortisol levels were assayed through saliva collected at encoding during the visit day and across 2 consecutive days in the following week (12 time points in total). We explored whether contextual memory encoding failure would explain depressive symptoms through OGM under possible moderation of amygdala connectivity and cortisol.

Results

In individuals with subclinical depression compared to those without, fearful faces disturbed memory encoding of the visuoperceptual context rather than 24 h later recollection, while neutral faces in their temporal proximity contrastingly augmented it. The more the contextual memory encoding bias (fearful vs. neutral) intensified, the more the amygdala’s functional connectivity with the ventromedial prefrontal cortex (vmPFC) weakened. Higher total cortisol output tended to be correlated with poorer 24-h later recollection of the temporal context. Moderated mediation effects of the amygdala-vmPFC connectivity and cortisol were not significant; however, contextual encoding bias explained depressive symptoms through negatively valenced OGM.

Conclusions

Negative stimuli appearing in an event might impair memory encoding of the visuoperceptual context under attentional interference, represented as weakened amygdala-vmPFC connectivity implicated in emotion-related attentional dysregulation. Conversely, negative stimuli might enhance temporally proximal visuoperceptual encoding after their disappearance. Contextual encoding bias could explain the overgeneralization (or lower coherence) of autobiographical memory and increase the risk of depression. The possible role of cortisol in recollecting the context of emotional events over time warrants further investigation.