Perceived social support is a crucial protective factor, improving physical and psychological health and reducing stress responses. It has been linked to lower allostatic load and mitigated stress responses to life events, potentially increasing resilience against mental health issues, especially following childhood maltreatment (CM). Conversely, social conflict negatively impacts well-being, particularly in individuals lacking social support. Neuroimaging studies demonstrate a link between social support and decreased amygdala and insula activity, resulting in fewer negative emotions during social exclusion. In contrast, CM significantly increases the risk of various mental disorders like major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). Studies have shown reduced hippocampal volume and increased limbic activity towards negative stimuli in individuals with CM, mirroring patterns found in MDD and PTSD patients. These neural changes, detectable even in childhood, emphasize the importance of early limbic development. Although social support can mitigate the negative consequences of CM, this interplay remains under-researched neuroimaging-wise. Previous studies have shown varying results regarding social support's buffering effects, particularly regarding developmental stage; early childhood seems to be a critical period for both the adverse effects of CM and the positive effects of social support. The current study aimed to investigate the relationship between perceived social support, CM, and limbic activity towards negative facial stimuli in healthy adults. The focus was on the hippocampus and amygdala due to their roles in emotion processing and their established connection to CM and social support. The hypothesis was a positive association between CM and limbic activity, a negative association between perceived social support and limbic activity, and a moderating effect of CM on the association between social support and limbic activity.

The literature review extensively examined the established links between perceived social support and positive mental and physical health outcomes, contrasting these with the known negative effects of childhood maltreatment (CM) on mental health and brain structure and function. Numerous studies highlighted the association between social support and reduced stress responses and allostatic load. In contrast, CM was consistently linked to increased limbic activity, reduced hippocampal volume, and an increased risk of various mental health disorders. The review also discussed the potentially mediating or moderating role of neural mechanisms, specifically focusing on the amygdala and hippocampus, in the interplay between social support and the long-term effects of CM. Discrepancies in previous studies' findings regarding the buffering effects of social support, especially concerning the age of participants, were noted and discussed as justification for the study's focus on healthy adults. The review provided the basis for the study's hypotheses, which predicted the independent effects of social support and CM on limbic activity and the moderating role of CM on the association between social support and limbic activity.

This study used data from the Münster Neuroimaging Cohort (MNC), initially including 212 healthy individuals (18–65 years) recruited via public notices and newspaper ads. Exclusion criteria included lifetime mental disorders, neurological abnormalities, head injuries, chronic diseases, organic mental disorders, psychotropic medication use, and MRI contraindications. Participants underwent fMRI scans while performing a negative emotional face processing task and a sensorimotor control task. Questionnaires, including the Childhood Trauma Questionnaire (CTQ) and the Social Support Questionnaire (FSOZU-K-22), assessed CM and perceived social support. The sample was initially divided into groups with and without CM based on CTQ cut-offs established by Walker et al. [56]. To account for the impact of depressive symptoms, a 1:1 matching procedure based on Beck Depression Inventory (BDI-I) scores, using Matchlt software, was implemented, resulting in a final sample of 65 participants in each group. fMRI data were preprocessed using SPM8, and region-of-interest (ROI) analysis focused on the bilateral amygdala-hippocampus-complex (AHC). Statistical analyses included Pearson correlations to assess the relationship between CM and social support, regression analyses to investigate their associations with AHC activity, and an ANCOVA to examine the interaction between social support and group (CM vs. no CM) on limbic activity. Whole-brain analysis using a minimum cluster size of *k* ≥ 100 voxels was also conducted for exploratory purposes. Significance thresholds were obtained at cluster-level by threshold-free cluster enhancement (TFCE). Robustness checks included analyses using continuous CM scores and analyses of unmatched samples.

A significant negative correlation was observed between perceived social support and CM (*r* = −0.293, *p* < 0.001). Higher CTQ scores (indicating greater CM) were significantly associated with increased bilateral AHC activity during negative emotion processing (Left: *p*_FWE = 0.006, *r* = 0.239; Right: *p*_FWE = 0.009, *r* = 0.216). Higher perceived social support showed a trend towards lower left AHC activity (*p*_FWE = 0.050, *r* = −0.136). The ANCOVA revealed a significant social support × group interaction on bilateral AHC activity (*p*_FWE ≤ 0.024), driven by a significant negative association between perceived social support and bilateral AHC activity in the group without CM (Left: *p*_FWE = 0.021, *r* = −0.429; Right: *p*_FWE = 0.010, *r* = −0.411). No significant association was found in the CM group. Whole-brain analyses confirmed these interactions in clusters including the hippocampus, parahippocampal gyrus, temporal gyri, and amygdala, again driven by the negative association in the non-maltreated group. Robustness checks using continuous CM scores and unmatched samples confirmed the key findings. Additional analyses suggested that the moderating effect of CM was primarily driven by abuse, not neglect, and was robust to the effects of perceived stress.

The study's findings demonstrate that CM significantly moderates the relationship between perceived social support and limbic activity during negative emotion processing in healthy adults. The protective effect of perceived social support on limbic activity is evident only in individuals without or with low to moderate CM, particularly those who experienced abuse. This lack of a buffering effect in individuals with high levels of CM is likely due to the long-lasting neurobiological changes caused by CM during critical developmental periods, potentially altering social perception and information processing. The observed positive trend between social support and limbic activity in individuals with high CM, when analyzing CM continuously, could tentatively suggest a detrimental effect of perceived social support in this context. These results extend previous findings by demonstrating that the buffering effects of social support on limbic activity are most pronounced in the absence of significant CM. They also highlight the potential for maladaptive responses to social support in individuals with CM.

This study uniquely demonstrates that CM moderates the link between perceived social support and limbic responses to negative stimuli in healthy adults. The protective effect of perceived social support is apparent only in individuals without significant CM. This finding underscores the need for interventions focused on improving social support perception, coping mechanisms, and self-esteem in individuals affected by CM to potentially mitigate the long-term consequences of CM on brain function and mental well-being. Future research should explore these relationships in clinical samples and across different age groups.

The study's cross-sectional design limits causal inferences. The reliance on retrospective self-report questionnaires for CM and social support introduces potential biases. The use of a binary categorization of CM and potentially gender-biased cut-off scores may have introduced group heterogeneity and information loss. The focus on a sample of healthy adults, potentially a more resilient subgroup, limits the generalizability of findings to other populations, including individuals with mental health disorders. Finally, using a shortened version of the FSOZU-K-22 prevented a detailed analysis of specific social support subscales.