Major depressive disorder (MDD) is a leading cause of disability worldwide 1. Impairments in emotion regulation and biases in memory processes are core facets underlying MDD 2–7. While there is extensive research documenting these impairments in depressed individuals, the neural basis for these processes, as well as how they mechanistically contribute to depressive symptoms, is not well understood.
Autobiographical memories represent a critical area of dysfunction in MDD 2, 7. How individuals interpret, process, and recall personal events impacts their thoughts about themselves and their future 2, 8. For example, it has been shown that individuals with MDD have a bias to both encode and recall negative memories 2, 9. Depressed individuals recall negative memories faster and more frequently compared with positive memories 2, 9. Compared with never-depressed individuals, those with a history of depression also have impoverished recall for positive events: positive events are less vivid, less specific, and provide less benefit emotionally 2, 9–13. Additionally, depressed individuals exhibit dysfunctional rumination and avoidance patterns with negative memories 2, 14. The ability to effectively regulate one’s emotional response to negative events and memories is critical for maintained well-being 2, 4. Therefore, it is clear that dysfunction in memory processes, coupled with impairments in emotion regulation, contribute to the cognitive and mood symptoms of MDD 2, 7, 9.
Impaired emotion regulation is also well documented in depression 4–6. Depressed individuals employ more ineffective emotion regulation strategies (e.g., suppression, avoidance) and underuse effective strategies, like cognitive reappraisal and acceptance 4–6. Additionally, lab-based studies have shown that neural processes underlying emotion regulation, which involves both higher level cognitive structures and subcortical limbic structures, differ in MDD compared with healthy volunteers for negative stimuli 15. This work has largely been conducted using standardized stimulus sets such as sad faces, negatively-valanced words, and upsetting scenes 15. However, differences in the neural mechanisms of negative affect and emotion regulation during more personalized or clinically relevant stimuli, like personal negative autobiographical memories, are less well understood.
Neural processes for regulating negative affect in response to upsetting memories in individuals with MDD appear to differ from that of healthy volunteers (HVs) 16. Prior work found increased amygdala-hippocampal connectivity mediated the relationship between higher self-report negative affect ratings in depressed subjects compared with HVs 16, suggesting that processes supported by the hippocampus may contribute to the negative affect experienced while recalling negative autobiographical memories. Additionally, depressed subjects showed a larger reduction in posterior hippocampal activation compared to HVs during reappraisal of autobiographical memories 16, indicating that reductions in hippocampal activity may be related to an emotion regulation strategy employed specifically by depressed individuals 16. When done correctly, cognitive reappraisal involves adjusting one’s emotional reaction to a negative memory without having to suppress the memory 2, 5, 17, 18. However, because of the relationship between memory engagement and negative affect 16, individuals with MDD may have difficulty reappraising, and may suppress the memory itself rather than reappraising their emotional response to the memory. While these findings provide initial support for neural differences in emotion regulation and negative affect during negative memories, a major limitation of prior work is that it has relied on static measures for assessing these complex psychological processes. The use of static measures (i.e., a single self-report rating, a single beta weight representing 10–20 second of task) does not allow examination of the dynamics of the relationship between memory, negative affect, and emotion regulation. In order to understand how memory engagement impacts these processes, we need to examine the temporal relationship between memory, negative affect, and emotion regulation.
Advances in analysis techniques, as well as scanner acquisition parameter options, e.g., shorter pulse sequence repetition time (TRs), have allowed for the examination of neural proxies of psychological processes, like negative affect, memory, and emotion regulation, at second-to-second intervals. Machine learning has allowed for identification of sensitive and specific neural signatures for complex psychological processes like negative affect and cognitive reappraisal 19–21. Unlike traditional univariate approaches that examine overall magnitude of BOLD activity within a set of voxels or clusters, machine learning-based approaches identify a spatially distributed multivoxel pattern as a proxy for a given mental process 21. Neural signatures have also been shown to be a more robust predictor of psychological processes compared with region of interest (ROI) brain regions or resting-state networks (e.g., salience network, default mode network) 19. The Picture-Induced Negative Affect Signature (PINES) and the Emotion Regulation Signature (ERS) are validated neural signatures that serve as a proxy for negative emotion and emotional reappraisal 19, 20. The PINES and ERS consist of weighted maps distributed across cortical and subcortical regions that can be applied to functional magnetic resonance imaging (fMRI) data to produce a value that represents the degree to which someone is engaging in negative affect or emotion regulation at each TR. Using these values, we can examine fluctuations in these psychological processes at the moment-to-moment level.
The current study examined whether hippocampal activity during a negative autobiographical memory fMRI task predicted subsequent negative affect measured by PINES for unmedicated depressed individuals and healthy volunteers. We also examined whether depressed individuals differed from HVs in their neural approaches to downregulating negative affect during this task. During the fMRI task, participants were asked to recall the negative memory and then instructed to immerse themselves in it or distance themselves from the negative memory (Fig. 1). We hypothesized that hippocampal activity would predict subsequent change in negative affect signature expression for MDDs and HVs during the memory cue and immerse conditions, with greater hippocampal activity predicting increased negative affect signature expression. During the distance condition, we hypothesized that there would be an interaction of group by hippocampal activity, such that greater hippocampal activity would predict subsequent increased negative affect signature expression for depressed individuals but not for healthy volunteers. We also explored differences in reappraisal strategies. During the distance condition, we hypothesized that there would be an interaction of group by hippocampal activity, such that greater hippocampal activity would predict subsequent increases in emotion regulation as measured by ERS for HVs, but not for individuals with MDD. Additionally, we examined whether reduction in hippocampal activity was used as a regulatory strategy by depressed individuals. We predicted that MDD participants would show a greater reduction in hippocampal activity when instructed to distance from the negative memory compared with HVs.