Stressful life events (SLEs) are a significant predictor of major depressive disorder (MDD) onset and recurrence. The diathesis-stress model posits that individuals with pre-existing vulnerabilities (genetic predispositions, childhood maltreatment (CM)) are more susceptible to developing MDD when exposed to SLEs. While cross-sectional studies have linked SLEs and CM to gray matter volume (GMV) alterations in various brain regions, longitudinal studies are needed to understand the temporal dynamics of these relationships. This study aimed to investigate the longitudinal associations between SLEs and GMV changes in MDD patients and HCs over a 2-year period, exploring the moderating role of CM. The hypotheses were that HCs would show greater GMV reductions in response to SLEs compared to MDD patients, and that CM would moderate this relationship only in MDD patients. The study also explored the relationship in MDD patients with more severe depression and those experiencing at least one depressive episode during the follow-up period, considering their increased vulnerability to SLEs. The potential moderating role of C-reactive protein (CRP), a marker of inflammation, was also investigated.

Prior cross-sectional MRI studies have demonstrated associations between recent SLEs and GMV alterations in healthy individuals, primarily in the insula, anterior cingulate, and medial prefrontal cortices. MDD patients showed a different pattern, with fewer GMV alterations in the medial orbitofrontal cortex than HCs. Another line of research has shown that both HCs and MDD patients with self-reported CM exhibit smaller GMV in several brain regions compared to those without CM. CM has also been found to moderate the relationship between SLEs and GMV in MDD patients but not in HCs, suggesting a brain structural basis for the diathesis-stress model in MDD. However, the absence of longitudinal studies limits our understanding of whether observed brain structural correlates are a consequence or predictor of stress. Two existing longitudinal studies have shown GMV reductions in response to SLEs in HCs, but the adaptive or maladaptive nature of these changes remains unclear, particularly in the context of MDD.

This study utilized data from the Marburg-Münster Affective Disorder Cohort Study (MACS), encompassing 754 participants (392 HCs, 362 MDD patients) aged 18–65 years. Participants underwent T1-weighted MRI scans at baseline (T1) and follow-up (T2) approximately 2 years later. Clinical assessments included the Life Events Questionnaire (LEQ) to measure SLEs during the 2-year interval, the Childhood Trauma Questionnaire (CTQ) for CM, and various other clinical and psychosocial measures (Hamilton Depression Rating Scale (HAM-D), State-Trait Anxiety Inventory (STAI-S), Global Assessment of Functioning (GAF), Perceived Stress Scale (PSS), etc.). MRI data were pre-processed using CAT12 toolbox in SPM12, including steps such as realignment, bias correction, tissue classification, spatial normalization, and smoothing. Voxel-based morphometry (VBM) was employed to assess longitudinal GMV changes. Statistical analyses included 2x2 repeated measures ANCOVAs to examine the association between SLEs and GMV changes in HCs and MDD patients, with age, sex, and interscan interval as covariates. Linear regression analyses were used to explore the moderation effects of CM, CRP, and other clinical variables. Further analyses investigated the relationship in MDD subgroups (with vs. without depressive episodes during the follow-up period), operationalizing severity through hospitalization, duration of episodes, etc. Additional control analyses were performed to explore potential confounding influences of clinical, psychosocial, and other factors. Finally, cross-sectional analyses were conducted to investigate predictive or retrospective associations between baseline or follow-up GMV and SLEs.

The repeated measures ANCOVA revealed two significant clusters showing differential longitudinal GMV changes in response to SLEs between HCs and MDD patients. HCs exhibited significant GMV reductions in the left middle frontal gyrus and left precentral/postcentral gyri with increasing SLEs, whereas MDD patients did not show such reductions. Moderation analyses showed no significant three-way interaction between SLEs, CM, and group on GMV change in either cluster, suggesting that CM did not differentially moderate the impact of SLEs on GMV between groups. However, exploratory analyses revealed a significant three-way interaction between SLEs, CM, and recurrence group (MDD patients with at least one episode vs. those without vs. HCs) on GMV change in the middle frontal, precentral, and postcentral gyri. MDD patients with an episode showed significant GMV increases with increasing SLEs and CM, unlike those without an episode or HCs. This three-way interaction remained significant even after adjusting for the number of depressive episodes. Predictive cross-sectional analysis indicated that baseline GMV in the precentral and postcentral gyri could predict future SLEs, with HCs showing larger GMV at baseline in association with higher SLEs during the follow-up period. MDD patients showed an opposite pattern, with smaller baseline GMV linked to higher SLEs, suggesting differential stress vulnerabilities between groups. No significant correlations were found between cluster values and other clinical and psychosocial factors at follow-up.

This study provides longitudinal evidence for distinct neural responses to stress in MDD patients and HCs. The GMV reductions observed in HCs might represent adaptive responses to stress, possibly involving reallocation of neural resources. The lack of such reductions in MDD patients might reflect impaired adaptive mechanisms. The GMV increases observed in MDD patients with an episode, high SLEs and CM may represent maladaptive changes, potentially involving glial cell proliferation and neuroinflammation. This aligns with the diathesis-stress model, where pre-existing vulnerabilities (CM, elevated CRP) interact with adult SLEs to influence brain structure and increase the risk of MDD recurrence. The predictive cross-sectional analysis suggests that baseline GMV may reflect individual differences in stress vulnerability, impacting the experience of SLEs and subsequent brain structural changes. The different patterns in HCs and MDD patients highlight the complex interaction between pre-existing vulnerabilities, environmental stressors, and the course of MDD.

This longitudinal study provides novel insights into the neurobiological underpinnings of the diathesis-stress model in MDD. The findings highlight distinct GMV changes in response to SLEs in HCs and MDD patients, suggesting adaptive vs. maladaptive responses to stress. The three-way interaction between SLEs, CM, and recurrence in MDD patients with episodes points to the crucial role of pre-existing vulnerabilities and recent stressors in shaping brain structure and MDD course. Future research should investigate the mechanisms underlying these findings, such as neuroinflammation and synaptic plasticity, utilizing longer follow-up periods, multiple assessment points, and extending the investigation to other psychiatric disorders.

The study's limitations include the reliance on self-reported measures of SLEs and CM, which could be prone to recall bias and measurement error. While several potential confounders were accounted for, other factors (physical health, lifestyle) may have influenced the findings. The cross-sectional analyses, although suggestive, do not definitively establish causality between GMV and SLEs. Furthermore, the study population might not be fully representative of the broader MDD population.