Skip to main content
NCBI home page
European Journal of Psychotraumatology logoLink to European Journal of Psychotraumatology
. 2026 Oct 5;17(1):2730909. doi: 10.1080/20008066.2026.2730909

Childhood peer abuse moderates associations of inflammation and cortisol with subcortical brain volume

El abuso entre pares durante la infancia modera las asociaciones de la inflamación y el cortisol con el volumen cerebral subcortical

Laijun Bu a, Zhen Wang b,CONTACT
PMCID: PMC13644451  PMID: 42830709

ABSTRACT

Objective: Childhood peer abuse is a developmentally salient form of interpersonal adversity, yet its role in shaping links between peripheral biological stress markers and brain structure remains poorly understood. The present cross-sectional study examined whether retrospectively reported childhood peer verbal and physical abuse moderated associations of systemic inflammatory burden and hair cortisol with subcortical brain volume.

Methods: Participants from a community sample completed the Maltreatment and Abuse Chronology of Exposure scale and structural magnetic resonance imaging. Inflammatory burden was indexed by a composite of standardised IL-6, IL-8, TNF-alpha, and C-reactive protein values (n = 292), and hair cortisol was available in a subsample (n = 172). Bilateral amygdala, hippocampus, putamen, and caudate volumes were extracted from FreeSurfer segmentation. Linear regression models adjusted for gender, age, parental education, and estimated total intracranial volume; false discovery rate correction was applied across the four ROI tests within each marker-by-moderator family.

Results: Peer verbal abuse moderated associations of inflammatory burden with bilateral amygdala volume (β = −.095, pFDR = .032) and bilateral putamen volume (β = −.115, pFDR = .017). Peer physical abuse moderated the association between hair cortisol and bilateral amygdala volume (β = .227, pFDR = .035). Caudate models did not survive correction. Core findings remained significant after sensitivity adjustment for broader childhood adversity and current depression/anxiety symptoms.

Conclusions: Associations differed across brain regions, biological markers, and peer-abuse subtypes. These findings identify peer adversity as an important interpersonal context for understanding stress-related brain differences.

KEYWORDS: Peer abuse, peer victimisation, inflammation, cortisol, amygdala, putamen, caudate, subcortical volume

HIGHLIGHTS

  • Peer verbal abuse moderated inflammation-amygdala and putamen links.

  • Peer physical abuse moderated the hair cortisol-amygdala link.

  • Core effects persisted after adversity and symptom adjustment.

  • Findings support context-dependent stress marker-brain associations.

1. Introduction

Childhood peer abuse is a developmentally salient form of interpersonal adversity, yet it remains less fully integrated into neurobiological models than family-based maltreatment (Cubillo, 2022; Ke et al., 2022; Vaillancourt et al., 2013). Peer victimisation and peer abuse occur in the social world in which children and adolescents learn belonging, status, rejection, and safety outside the family. These experiences may be especially consequential because peer relationships become increasingly central across late childhood and adolescence, a developmental window in which stress-sensitive biological systems and cortico-subcortical circuits are still maturing. Accordingly, peer abuse may contribute to individual differences in stress biology and brain structure through mechanisms that overlap with, but are not reducible to, caregiver maltreatment.

Peer victimisation, including more severe forms of peer abuse, has been associated with persistent mental-health, social, and physical-health consequences across development and adulthood, including elevated risk for internalising symptoms, social withdrawal, and poorer health outcomes (Arseneault, 2018; Copeland et al., 2013; Takizawa et al., 2014; Wolke & Lereya, 2015). Neurobiological studies of peer victimisation remain comparatively fewer, but available work links peer adversity to structural and functional alterations in neural systems involved in threat processing, social evaluation, reward learning, and emotion regulation (Cubillo, 2022; Ke et al., 2022; Palamarchuk & Vaillancourt, 2022). A key unresolved question is whether peer abuse changes how peripheral biological stress markers relate to brain morphology, rather than simply exerting a uniform main effect on regional brain volume.

Two biological stress systems are particularly relevant in this context: systemic inflammation and hypothalamic-pituitary-adrenal (HPA) axis activity indexed by hair cortisol (Danese & Lewis, 2017; Murphy et al., 2022). Childhood adversity has been linked to elevated inflammatory activity and alterations in cortisol regulation, but these biomarkers are not simple causal readouts of adversity exposure. In cross-sectional young-adult samples, circulating inflammatory markers and hair cortisol index current or recent biological states shaped by many factors, including stress history, sleep, health, behaviour, and current psychological symptoms. Thus, their associations with brain structure may depend on developmental context, including whether an individual was exposed to peer abuse.

Theoretical accounts of stress biology emphasise that repeated interpersonal threat can recalibrate allostatic systems through immune, autonomic, and neuroendocrine pathways (McEwen, 1993, 1998; Miller et al., 2011; Slavich & Irwin, 2014). Neuroimmune models further propose that early adversity may become biologically embedded through coordinated changes across peripheral inflammation, neural development, and social-affective behaviour (Hostinar et al., 2018; Nusslock & Miller, 2016). Such models suggest that associations of inflammation and cortisol with brain structure may differ across adversity contexts. Prior studies suggest that early-life adversity can alter associations between peripheral biological stress markers and brain structure, although this evidence has largely focused on family-based maltreatment or composite adversity indices (Chiang et al., 2015; Frodl et al., 2012; Orellana et al., 2024). It remains unclear whether these biological marker-brain associations also vary with peer abuse exposure.

Peer abuse may differ from family-based maltreatment in both social meaning and developmental timing (Berman et al., 2022; Vaillancourt et al., 2013). Whereas family maltreatment occurs within caregiving relationships that organise safety, attachment, and dependence, peer abuse occurs within the social ecology of status, acceptance, exclusion, and group belonging. These social threats may be particularly relevant to neural systems involved in salience detection, memory for social threat, and reward-related learning. Distinguishing verbal and physical peer abuse allows a more precise test of these mechanisms. Verbal peer abuse, encompassing ridicule, humiliation, and social rejection, may be especially tied to social-evaluative threat, whereas physical peer abuse may more directly involve bodily threat and coercive dominance. Although these subtypes often co-occur, separating them can reveal whether biological marker-brain associations vary by the form of peer adversity experienced.

The present study focused on bilateral amygdala, hippocampus, putamen, and caudate volume as subcortical phenotypes relevant to stress-related variation in threat, contextual memory, and motivation. The amygdala is central to threat and salience processing and has shown heterogeneous associations with childhood adversity, with evidence for both larger and smaller volumes depending on exposure type, timing, and developmental stage (Teicher et al., 2016; Tottenham & Sheridan, 2009). The hippocampus supports contextual memory and HPA-axis regulation and is frequently implicated in adversity and stress-system research (Keding & Herringa, 2015; Riem et al., 2015). The putamen and caudate are components of striatal circuitry involved in reward learning, motivation, action selection, and social-affective behaviour; striatal structure and function have been linked to adversity and peer victimisation (Edmiston, 2011; Hanson et al., 2016; Quinlan et al., 2020). Because our hypotheses concerned subcortical stress, salience, memory, and reward-learning phenotypes, analyses were focused on this anatomically coherent subcortical ROI set rather than on a broad cortical regulatory network.

We examined whether retrospectively reported childhood peer verbal and physical abuse moderated associations of inflammatory burden and hair cortisol with bilateral amygdala, hippocampus, putamen, and caudate volumes. We expected peer abuse to moderate biological marker-brain associations, with the clearest directionally grounded predictions for smaller hippocampal and striatal volumes at higher levels of peer abuse. For the amygdala, we hypothesised altered rather than uniformly smaller volume associations because prior adversity studies have reported directionally heterogeneous amygdala findings across exposure types and developmental contexts. We further tested peer-abuse main effects, sensitivity models adjusting for broader childhood adversity and current symptoms, peer-versus-family specificity models, and sex/gender moderation analyses.

2. Methods

2.1. Participants

Participants were recruited from a community sample in Hangzhou, China. Before enrolment, participants were screened for psychiatric, personality, and neurological disorders and for MRI-related contraindications, including metallic implants and fixed orthodontic appliances. This was an independent community-based study. The primary inflammatory analyses included 292 participants with available structural MRI, peer abuse, inflammatory marker, and covariate data; their ages ranged from 13 to 28 years. Hair cortisol analyses were conducted in a subsample of 172 participants with available hair cortisol data; their ages ranged from 13 to 24 years. The analytic samples were defined by complete availability of the biological marker, peer-abuse scores, all four bilateral ROI volumes, and model covariates. Written informed consent was obtained from all participants. All study procedures were approved by the Ethics Committee of Hangzhou Normal University (approval number: 2021031001). Descriptive statistics for the inflammatory-marker and hair-cortisol samples are presented in Table 1.

Table 1.

Descriptive statistics for the inflammatory sample and cortisol subsample.

Variable Inflammatory sample Cortisol subsample
Gender, male/female, n (%) 109 (37.3) / 183 (62.7) 55 (32.0) / 117 (68.0)
Age, years 19.75 +/− 2.30 18.52 +/− 1.13
Parental education level 5.01 +/− 2.06 5.20 +/− 2.25
Childhood peer verbal abuse 3.50 +/− 3.08 3.17 +/− 3.10
Childhood peer physical abuse 1.09 +/− 1.98 0.83 +/− 1.70
Inflammatory burden composite 0.02 +/− 2.21 –
Hair cortisol, pg/mg – 8.25 +/− 6.03
Bilateral amygdala volume 3428.10 +/− 333.84 3422.19 +/− 333.00
Bilateral hippocampus volume 8321.95 +/− 718.65 8344.69 +/− 681.50
Bilateral putamen volume 10933.77 +/− 1075.69 10941.79 +/− 1065.38
Bilateral caudate volume 7459.04 +/− 822.07 7380.16 +/− 809.35
Open in a new tab

Note: Values are n (%) for categorical variables and mean +/− SD for continuous variables. Hair cortisol was available only in the hormone subsample. The inflammatory burden composite was based on standardised IL-6, IL-8, TNF-alpha, and CRP values. Parental education was coded from 2 to 12, with higher values indicating higher educational attainment. Brain volumes are bilateral summed volumes reported in mm³.

2.2. Assessment of childhood peer abuse

Retrospective exposure to childhood peer abuse was assessed using the Chinese version of the Maltreatment and Abuse Chronology of Exposure scale (MACE), which has demonstrated good psychometric properties in young adults (Chen et al., 2022; Teicher & Parigger, 2015). The present analyses focused on the MACE peer verbal abuse and peer physical abuse subscales. Subscale scores were expressed on the standard MACE 0–10 severity-scaled metric, with higher values reflecting greater retrospectively reported exposure severity. Because the sample was community-based and many participants reported low exposure, distributions of both peer abuse moderators are provided in Supplementary Figure S1.

2.3. MRI data acquisition, processing, and ROI definition

High-resolution structural MRI data were acquired on a 3.0-T Siemens Magnetom Trio A Tim System scanner equipped with a 12-channel phased-array head coil. T1-weighted anatomical images were obtained using a three-dimensional magnetization-prepared rapid gradient-echo sequence with the following parameters: repetition time = 2678 ms, echo time = 2.98 ms, flip angle = 7°, field of view = 256 × 256 mm, voxel size = 1 × 1 × 1 mm, and 192 sagittal slices. During scanning, participants were instructed to remain still, keep their eyes closed, stay relaxed, and avoid falling asleep. Structural MRI data were processed using the standard FreeSurfer pipeline, version 5.3 (Fischl, 2012). Automated preprocessing included skull stripping, intensity normalisation, spatial normalisation, subcortical segmentation, cortical surface reconstruction, and anatomical parcellation. Regional volumetric measures were extracted from the aseg subcortical segmentation output. Left and right hemisphere volumes were summed to represent total bilateral ROI volume; because all continuous variables were standardised before inference, using bilateral averages rather than sums would only rescale the raw bilateral measure and would not change standardised regression estimates. The ROI family included bilateral amygdala, hippocampus, putamen, and caudate volumes. Representative schematic illustrations are shown in Figure 1.

Figure 1.

Four schematic brain diagrams highlight the bilateral amygdala, hippocampus, putamen, and caudate in magenta.

Open in a new tab

Representative subcortical regions of interest (ROIs). The figure displays schematic illustrations of the bilateral amygdala, hippocampus, putamen, and caudate. All analysed ROI definitions were based on FreeSurfer aseg subcortical segmentation; left and right hemisphere volumes were summed to derive bilateral ROI volumes.

2.4. Biological stress markers

Systemic inflammatory burden was indexed using a composite derived from IL-6, IL-8, TNF-alpha, and C-reactive protein. This composite was designed to capture shared low-grade peripheral inflammatory activity across complementary markers rather than emphasising a single cytokine. Venous blood was drawn from the antecubital vein by clinical laboratory staff at an affiliated hospital of the study institution and collected into disposable vacuum tubes without anticoagulant. Serum IL-6, IL-8, and TNF-alpha were quantified using chemiluminescence immunoassays on an automated platform manufactured by Chongqing Keysmile Biological Technology Co., Ltd. (Chongqing, China). The TNF-alpha assay was a chemiluminescent sandwich immunoassay registered by Changsha Weimi Bio-Tech Co., Ltd. (Changsha, China). The manufacturer specified a linear range of 5–1000 pg/mL, a detection limit of ≤2 pg/mL, and acceptance criteria of ≤8% for within-batch CV and ≤10% for between-batch CV. These values represent manufacturer-specified acceptance criteria rather than study-specific precision estimates; study-specific intra- and inter-assay CVs were not available in the archived records. C-reactive protein was measured using a chemiluminescent enzyme immunoassay. Each inflammatory biomarker was z-standardised before aggregation, and standardised values were summed to form the inflammatory burden composite; higher values indicate greater inflammatory burden.

Hair cortisol was used as an index of cumulative cortisol exposure related to HPA-axis activity. A 3-cm proximal hair segment was cut as close as possible to the scalp, representing cortisol accumulation over approximately the preceding 3 months based on the commonly used average scalp-hair growth rate of approximately 1 cm per month (Stalder et al., 2017; Stalder & Kirschbaum, 2012). Hair segments were finely cut with scissors, and 20 mg of hair was processed and analysed using liquid chromatography-tandem mass spectrometry following previously published methods from the same laboratory (Chen et al., 2019; Chu et al., 2022; Mwanza et al., 2016). Published validation of this laboratory workflow reported good linearity (R2 > 0.99), a lower limit of quantitation of 0.3 pg/mg for cortisol, and intra-day and inter-day coefficients of variation below 15% (Chu et al., 2022). The archived study workbooks retained peak-area, internal-standard, calibration, and final concentration records for the analysed samples.

2.5. Statistical analysis

All analyses were conducted in R. Descriptive statistics were first calculated. All continuous variables were z-standardised before inferential analyses. Gender, age, parental education, and estimated total intracranial volume were included as covariates in all primary models. Parental education was coded from 2 to 12, with higher values indicating higher educational attainment.

Linear regression models were fitted to examine direct associations between each biological stress marker and regional brain volume. Biological stress markers included inflammatory burden and hair cortisol. Separate moderation models then tested whether peer verbal abuse or peer physical abuse moderated associations between each biological marker and each bilateral subcortical ROI. The primary interaction term was biological marker x peer abuse. False discovery rate correction was applied across the four ROI tests within each biological marker-by-peer abuse family (Benjamini & Hochberg, 1995).

When a significant interaction was observed, simple slope analyses were conducted to estimate biological marker-brain associations at low and high levels of the moderator, defined as −1 SD and +1 SD, respectively (Aiken & West, 1991). We also tested peer-abuse main effects on ROI volume. Sensitivity analyses examined whether key findings remained after additional adjustment for broader childhood adversity indexed by MACE_SUM_CTQ, defined as the total of MACE subscales corresponding to conventional CTQ-type maltreatment domains, and after adjustment for current DASS depression and anxiety symptoms or PTSD symptoms. Additional supplementary analyses tested sex/gender moderation, peer-versus-family adversity specificity, and within-model contrasts of interaction strength.

3. Results

3.1. Direct associations of biological markers with subcortical volume

Before testing interaction terms, we examined direct associations of biological stress markers with subcortical volume. Inflammatory burden was not significantly associated with bilateral amygdala volume, β = .057, p = .222, hippocampus volume, β = .008, p = .852, putamen volume, β = −.039, p = .417, or caudate volume, β = .023, p = .662. Hair cortisol was not significantly associated with amygdala volume, β = −.089, p = .128, putamen volume, β = −.041, p = .530, or caudate volume, β = −.014, p = .839; the negative association with hippocampus volume was nominal, β = −.136, p = .035.

3.2. Peer abuse moderated inflammation-subcortical volume associations

Table 2 summarises the moderation models testing whether childhood peer abuse moderated associations between inflammatory burden and subcortical volume. Peer verbal abuse significantly moderated associations of inflammatory burden with bilateral amygdala volume, beta_interaction = −.095, pFDR = .032, and bilateral putamen volume, beta_interaction = −.115, pFDR = .017. The corresponding hippocampus and caudate interactions were not significant after correction.

Table 2.

Moderation of inflammatory burden-subcortical volume associations by retrospectively reported childhood peer abuse.

ROI Peer verbal β t p pFDR Peer physical β t p pFDR
Bilateral amygdala −.095 −2.42 .016 .032 −.067 −1.16 .247 .330
Bilateral hippocampus −.018 −0.46 .648 .864 .004 0.07 .947 .947
Bilateral putamen −.115 −2.89 .004 .017 −.124 −2.08 .038 .152
Bilateral caudate .002 0.03 .972 .972 .075 1.16 .246 .330
Open in a new tab

Note: Standardised interaction coefficients are shown. Models adjusted for gender, age, parental education, and estimated total intracranial volume. pFDR values were corrected across the four ROIs within each peer abuse moderator.

Simple-slope analyses showed that, at lower levels of peer verbal abuse, greater inflammatory burden was associated with larger bilateral amygdala volume, β = .155, p = .013. At higher levels of peer verbal abuse, this association was not significant, β = −.035, p = .560 (Figure 2(A)). For putamen volume, inflammatory burden was not significantly associated with volume at lower peer verbal abuse, β = .079, p = .210, but was associated with smaller putamen volume at higher peer verbal abuse, β = −.151, p = .013 (Figure 2(B); Supplementary Table S2).

Figure 2.

Two interaction plots show associations of standardized inflammatory burden with standardized bilateral amygdala and putamen volumes at lower (negative 1 SD) and higher (+1 SD) peer verbal abuse, followed by a bar chart showing the observed peer verbal abuse score distribution. The upper-left plot shows that inflammatory burden is positively associated with bilateral amygdala volume at lower peer verbal abuse (beta equals 0.15, p equals 0.013) but not at higher peer verbal abuse (beta equals negative 0.03, p equals 0.560). The upper-right plot shows that inflammatory burden is not significantly associated with bilateral putamen volume at lower peer verbal abuse (beta equals 0.08, p equals 0.210) but is negatively associated at higher peer verbal abuse (beta equals negative 0.15, p equals 0.013). Shaded bands indicate 95% confidence intervals. The lower bar chart shows that lower peer verbal abuse scores are more frequent than higher scores.

Open in a new tab

Simple slopes for significant interactions between inflammatory burden and retrospectively reported childhood peer verbal abuse. Panel A shows bilateral amygdala volume, and Panel B shows bilateral putamen volume. Lines represent estimated associations at low (−1 SD) and high (+1 SD) peer verbal abuse. Shaded areas represent 95% confidence intervals, and gray bars show the observed peer verbal abuse score distribution.

Peer physical abuse showed a nominal interaction with inflammatory burden for bilateral putamen volume, β_interaction = −.124, p = .038, but this effect did not survive correction across the four ROI tests, pFDR = .152. This secondary pattern is interpreted cautiously; its simple slopes are provided in Supplementary Figure S2 and Supplementary Table S2.

3.3. Peer physical abuse moderated the cortisol-amygdala association

Table 3 summarises moderation models testing whether childhood peer abuse moderated associations between hair cortisol and subcortical volume. Peer physical abuse significantly moderated the association between hair cortisol and bilateral amygdala volume, β_interaction = .227, pFDR = .035. No other hair cortisol interaction survived FDR correction across the four ROI tests.

Table 3.

Moderation of hair cortisol-subcortical volume associations by retrospectively reported childhood peer abuse.

ROI Peer verbal β t p pFDR Peer physical β t p pFDR
Bilateral amygdala .047 0.66 .513 .761 .227 2.65 .009 .035
Bilateral hippocampus −.088 −1.11 .267 .761 .115 1.21 .230 .306
Bilateral putamen .025 0.30 .761 .761 −.088 −0.90 .368 .368
Bilateral caudate .026 0.31 .757 .761 .176 1.76 .080 .160
Open in a new tab

Note: Standardised interaction coefficients are shown. Models adjusted for gender, age, parental education, and estimated total intracranial volume. pFDR values were corrected across the four ROIs within each peer abuse moderator. Cortisol analyses were conducted in the hormone subsample.

Simple-slope analyses showed that, at lower levels of peer physical abuse, higher hair cortisol was associated with smaller bilateral amygdala volume, β = −.256, p = .003. At higher levels of peer physical abuse, the hair cortisol-amygdala association was positive but not statistically significant, β = .197, p = .108 (Figure 3; Supplementary Table S2). Peer verbal abuse did not significantly moderate hair cortisol associations after correction.

Figure 3.

An interaction plot shows the association of standardized hair cortisol with standardized bilateral amygdala volume at lower (negative 1 SD) and higher (+1 SD) peer physical abuse, followed by a bar chart showing the observed peer physical abuse score distribution. Hair cortisol is negatively associated with bilateral amygdala volume at lower peer physical abuse (beta equals negative 0.26, p equals 0.003) but is not significantly associated with bilateral amygdala volume at higher peer physical abuse (beta equals 0.20, p equals 0.108). Shaded bands indicate 95 percent confidence intervals. The lower bar chart shows that lower peer physical abuse scores are more frequent than higher scores.

Open in a new tab

Simple slopes for the significant interaction between hair cortisol and retrospectively reported childhood peer physical abuse predicting bilateral amygdala volume. Lines represent estimated associations at low (−1 SD) and high (+1 SD) peer physical abuse. Shaded areas represent 95% confidence intervals, and gray bars show the observed peer physical abuse score distribution.

3.4. Supplementary main-effect, sensitivity, specificity, and sex/gender analyses

Peer verbal and physical abuse were not significantly associated with bilateral amygdala, hippocampus, putamen, or caudate volume after FDR correction in either the inflammatory sample or the cortisol subsample (Supplementary Table S1). The three primary interactions that survived four-ROI FDR correction in the main analyses, inflammatory burden × peer verbal abuse for amygdala and putamen volume and hair cortisol × peer physical abuse for amygdala volume, remained statistically significant in the prespecified sensitivity models adjusting for MACE_SUM_CTQ, DASS depression and anxiety symptoms, or PTSD symptoms (Supplementary Table S3). The p values in these targeted sensitivity models were not subjected to a new FDR correction. The nominal inflammatory burden × peer physical abuse interaction for putamen volume was attenuated after current symptom adjustment and is therefore treated as secondary.

Exploratory sex/gender moderation analyses of the four focal interaction patterns did not identify significant three-way interactions after FDR correction (Supplementary Table S4). Exploratory peer-versus-family adversity models provided partial evidence for peer-adversity specificity, with the clearest specificity observed for the hair cortisol × peer physical abuse association with amygdala volume (Supplementary Table S5). Exploratory within-model contrasts did not indicate clear differences between peer verbal and peer physical abuse interactions in the inflammatory models, whereas in the cortisol subsample the peer physical abuse interaction was stronger than the peer verbal abuse interaction for amygdala and hippocampus volume (Supplementary Table S6). The p values for the formal contrasts in Supplementary Tables S5 and S6 were not adjusted for multiple comparisons.

4. Discussion

The present study suggests that childhood peer abuse is associated with variation in how peripheral biological stress markers relate to subcortical brain volume. Direct biomarker-volume associations were largely null, apart from a nominal negative association between hair cortisol and hippocampal volume. Inflammatory burden interacted with peer verbal abuse in relation to amygdala and putamen volume, whereas hair cortisol interacted with peer physical abuse in relation to amygdala volume. Adding caudate to the same corrected ROI family did not change these findings, and the caudate results did not survive correction.

The putamen finding is consistent with models linking interpersonal adversity, inflammation, and motivational circuitry. At higher levels of peer verbal abuse, greater inflammatory burden was associated with smaller putamen volume; a similar direction was observed for peer physical abuse but did not survive correction. The putamen participates in reward learning, action selection, and motivated behaviour, processes that may be especially relevant when peer relationships repeatedly communicate social threat, humiliation, or exclusion. Hanson et al. (2016) reported that cumulative childhood stress was associated with blunted reward-related brain activity in adulthood, supporting the broader idea that adversity may alter reward-system development. Nusslock and Miller's (2016) neuroimmune network hypothesis further proposes that early adversity can coordinate immune signalling with neural systems relevant to reward, threat, and social behaviour. In this light, the inflammation-putamen association observed under higher peer verbal abuse may reflect a coupling between peripheral inflammatory state and subcortical motivational circuitry rather than an isolated inflammatory main effect.

The amygdala findings require a developmental and exposure-specific interpretation. Peer verbal abuse moderated the inflammation-amygdala association, but the pattern was not a simple higher-inflammation-smaller-volume effect. Instead, inflammatory burden was positively associated with amygdala volume at lower levels of peer verbal abuse and unrelated at higher levels. This pattern is consistent with evidence that adversity-related amygdala findings vary across exposure types, timing, and developmental stages (Teicher et al., 2016; Tottenham & Sheridan, 2009). These results suggest that the association between inflammation and amygdala volume may differ according to peer verbal abuse exposure.

Hair cortisol showed a different pattern in relation to peer physical abuse. At lower levels of peer physical abuse, higher hair cortisol was associated with smaller amygdala volume, whereas this association was no longer negative at higher levels of peer physical abuse. Hair cortisol reflects integrated HPA-axis activity over a recent window and is affected by multiple biological and psychosocial determinants (Stalder et al., 2017; Stalder & Kirschbaum, 2012). Murphy et al. (2022) highlighted heterogeneity in cortisol findings across populations and trauma characteristics. The positive simple slope at high peer physical abuse was not statistically significant and should not be interpreted as evidence for amygdala enlargement.

The sensitivity and specificity analyses qualified these findings. The primary interactions remained statistically significant after adjustment for broader CTQ-type maltreatment burden. Peer-versus-family adversity models provided the clearest specificity for the hair cortisol x peer physical abuse association with amygdala volume, whereas the inflammation-putamen findings showed only partial specificity relative to family-adversity comparators. The exploratory contrasts were not adjusted for multiple comparisons and should not be interpreted as definitive evidence of specificity. The absence of corrected caudate interactions suggests that the findings were regionally constrained.

Several limitations warrant emphasis. First, peer abuse was assessed retrospectively and may be affected by recall bias and current affective state (Baldwin et al., 2019; Reuben et al., 2016). Second, the cross-sectional design precludes causal or developmental ordering among peer abuse, biological stress markers, and brain volume. Third, the community sample had relatively low peer physical abuse exposure, and the hair cortisol subsample was smaller than the inflammatory sample. Fourth, inflammatory markers and hair cortisol were assessed once and may be affected by health behaviour, infection, medication, sleep, and other unmeasured factors. Study-specific intra- and inter-assay CVs for the inflammatory assays were not available in the archived records. Fifth, regional volume is a macrostructural measure and cannot identify neuronal, glial, or vascular mechanisms. Sixth, the formal contrasts in Supplementary Tables S5 and S6 were exploratory and were not corrected for multiple comparisons. Finally, sex/gender analyses were exploratory, were limited to four focal interaction patterns, and used the binary variable available in the dataset.

In conclusion, childhood peer abuse was associated with differences in how inflammatory burden and hair cortisol related to subcortical volume. The FDR-corrected findings involved interactions of inflammatory burden with peer verbal abuse for amygdala and putamen volume, and of hair cortisol with peer physical abuse for amygdala volume. Longitudinal studies are needed to clarify the timing, mechanisms, and clinical significance of these associations.

Supplementary Material

Supplementary_Materials_revised_FINAL.docx
ZEPT_A_2730909_SM1842.docx (267.6KB, docx)

Funding Statement

This study was supported by Guangzhou Xinhua University 2025 Key Project in Humanities and Social Sciences (Project No.: 2025KYZDSK03), titled: Construction of a Risk Warning Model and Exploration of Cognitive Behavioral Intervention Strategies for Adolescent Self-Injury Behavior in the Context of ‘Healthy China'.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The raw and processed research data supporting the findings of this study can be obtained from the corresponding author following a reasonable request and the author’s formal consent.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/20008066.2026.2730909.

References

  1. Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Sage. [Google Scholar]
  2. Arseneault, L. (2018). Annual research review: The persistent and pervasive impact of being bullied in childhood and adolescence: Implications for policy and practice. Journal of Child Psychology and Psychiatry, 59(4), 405–421. 10.1111/jcpp.12841 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baldwin, J. R., Reuben, A., Newbury, J. B., & Danese, A. (2019). Agreement between prospective and retrospective measures of childhood maltreatment: A systematic review and meta-analysis. JAMA Psychiatry, 76(6), 584. 10.1001/jamapsychiatry.2019.0097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society Series B: Statistical Methodology, 57(1), 289–300. 10.1111/j.2517-6161.1995.tb02031.x [DOI] [Google Scholar]
  5. Berman, I. S., McLaughlin, K. A., Tottenham, N., Godfrey, K., Seeman, T., Loucks, E., Suomi, S., Danese, A., & Sheridan, M. A. (2022). Measuring early life adversity: A dimensional approach. Development and Psychopathology, 34(2), 499–511. 10.1017/S0954579421001826 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen, Y., Wang, Z., Zheng, X., Wu, Z., & Zhu, J. (2022). The Chinese version of the Maltreatment and Abuse Chronology of Exposure (MACE) scale: Psychometric properties in a sample of young adults. PLoS One, 17(6), e0270709. 10.1371/journal.pone.0270709 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chen, Z., Zhang, Q., Chen, S., Wang, W., Liu, G., & Deng, H. (2019). Determination, intercorrelation and intraindividual stability of five steroids in hair, saliva and urine among Chinese college students. Steroids, 149, 108418. 10.1016/j.steroids.2019.05.010 [DOI] [PubMed] [Google Scholar]
  8. Chiang, J. J., Taylor, S. E., & Bower, J. E. (2015). Early adversity, neural development, and inflammation. Developmental Psychobiology, 57(8), 887–907. 10.1002/dev.21329 [DOI] [PubMed] [Google Scholar]
  9. Chu, L., Liu, S., Wu, Y., Yang, J., Qiao, S., Zhou, Y., Deng, H., Li, X., & Shen, Z. (2022). Hair levels of steroid, endocannabinoid, and the ratio biomarkers predict viral suppression among people living with HIV/AIDS in China. Clinica Chimica Acta, 535, 143–152. 10.1016/j.cca.2022.08.023 [DOI] [PubMed] [Google Scholar]
  10. Copeland, W. E., Wolke, D., Angold, A., & Costello, E. J. (2013). Adult psychiatric outcomes of bullying and being bullied by peers in childhood and adolescence. JAMA Psychiatry, 70(4), 419. 10.1001/jamapsychiatry.2013.504 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cubillo, A. (2022). Neurobiological correlates of the social and emotional impact of peer victimization: A review. Frontiers in Psychiatry, 13, 866926. 10.3389/fpsyt.2022.866926 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Danese, A., & Lewis, S. J. (2017). Psychoneuroimmunology of early-life stress: The hidden wounds of childhood trauma? Neuropsychopharmacology, 42(1), 99–114. 10.1038/npp.2016.198 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Edmiston, E. E. (2011). Corticostriatal-limbic gray matter morphology in adolescents with self-reported exposure to childhood maltreatment. Archives of Pediatrics & Adolescent Medicine, 165(12), 1069. 10.1001/archpediatrics.2011.565 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fischl, B. (2012). Freesurfer. NeuroImage, 62(2), 774–781. 10.1016/j.neuroimage.2012.01.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Frodl, T., Carballedo, A., Hughes, M. M., Saleh, K., Fagan, A., Skokauskas, N., McLoughlin, D. M., Meaney, J., O’Keane, V., & Connor, T. J. (2012). Reduced expression of glucocorticoid-inducible genes GILZ and SGK-1: High IL-6 levels are associated with reduced hippocampal volumes in major depressive disorder. Translational Psychiatry, 2(3), e88–e88. 10.1038/tp.2012.14 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hanson, J. L., Albert, D., Iselin, A.-M. R., Carré, J. M., Dodge, K. A., & Hariri, A. R. (2016). Cumulative stress in childhood is associated with blunted reward-related brain activity in adulthood. Social Cognitive and Affective Neuroscience, 11(3), 405–412. 10.1093/scan/nsv124 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hostinar, C. E., Nusslock, R., & Miller, G. E. (2018). Future directions in the study of early-life stress and physical and emotional health: Implications of the neuroimmune network hypothesis. Journal of Clinical Child & Adolescent Psychology, 47(1), 142–156. 10.1080/15374416.2016.1266647 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Ke, T., De Simoni, S., Barker, E., & Smith, P. (2022). The association between peer-victimisation and structural and functional brain outcomes: A systematic review. JCPP Advances, 2(2), e12081. 10.1002/jcv2.12081 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Keding, T. J., & Herringa, R. J. (2015). Abnormal structure of fear circuitry in pediatric post-traumatic stress disorder. Neuropsychopharmacology, 40(3), 537–545. 10.1038/npp.2014.239 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. McEwen, B. S. (1993). Stress and the individual: Mechanisms leading to disease. Archives of Internal Medicine, 153(18), 2093. 10.1001/archinte.1993.00410180039004 [DOI] [PubMed] [Google Scholar]
  21. McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338(3), 171–179. 10.1056/NEJM199801153380307 [DOI] [PubMed] [Google Scholar]
  22. Miller, G. E., Chen, E., & Parker, K. J. (2011). Psychological stress in childhood and susceptibility to the chronic diseases of aging: Moving toward a model of behavioral and biological mechanisms. Psychological Bulletin, 137(6), 959–997. 10.1037/a0024768 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Murphy, F., Nasa, A., Cullinane, D., Raajakesary, K., Gazzaz, A., Sooknarine, V., Haines, M., Roman, E., Kelly, L., O’Neill, A., Cannon, M., & Roddy, D. W. (2022). Childhood trauma, the HPA axis and psychiatric illnesses: A targeted literature synthesis. Frontiers in Psychiatry, 13, 748372. 10.3389/fpsyt.2022.748372 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mwanza, C., Chen, Z., Zhang, Q., Chen, S., Wang, W., & Deng, H. (2016). Simultaneous HPLC-APCI-MS/MS quantification of endogenous cannabinoids and glucocorticoids in hair. Journal of Chromatography B, 1028, 1–10. 10.1016/j.jchromb.2016.06.002 [DOI] [PubMed] [Google Scholar]
  25. Nusslock, R., & Miller, G. E. (2016). Early-life adversity and physical and emotional health across the lifespan: A neuroimmune network hypothesis. Biological Psychiatry, 80(1), 23–32. 10.1016/j.biopsych.2015.05.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Orellana, S. C., Bethlehem, R. A. I., Simpson-Kent, I. L., Van Harmelen, A.-L., Vértes, P. E., & Bullmore, E. T. (2024). Childhood maltreatment influences adult brain structure through its effects on immune, metabolic, and psychosocial factors. Proceedings of the National Academy of Sciences, 121(16), e2304704121. 10.1073/pnas.2304704121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Palamarchuk, I. S., & Vaillancourt, T. (2022). Integrative brain dynamics in childhood bullying victimization: Cognitive and emotional convergence associated With stress psychopathology. Frontiers in Integrative Neuroscience, 16, 782154. 10.3389/fnint.2022.782154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Quinlan, E. B., IMAGEN Consortium, Barker, E. D., Luo, Q., Banaschewski, T., Bokde, A. L. W., Bromberg, U., Büchel, C., Desrivières, S., Flor, H., Frouin, V., Garavan, H., Chaarani, B., Gowland, P., Heinz, A., Brühl, R., Martinot, J.-L., Martinot, M.-L. P., Nees, F., … Schumann, G. (2020). Peer victimization and its impact on adolescent brain development and psychopathology. Molecular Psychiatry, 25(11), 3066–3076. 10.1038/s41380-018-0297-9 [DOI] [PubMed] [Google Scholar]
  29. Reuben, A., Moffitt, T. E., Caspi, A., Belsky, D. W., Harrington, H., Schroeder, F., Hogan, S., Ramrakha, S., Poulton, R., & Danese, A. (2016). Lest we forget: Comparing retrospective and prospective assessments of adverse childhood experiences in the prediction of adult health. Journal of Child Psychology and Psychiatry, 57(10), 1103–1112. 10.1111/jcpp.12621 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Riem, M. M. E., Alink, L. R. A., Out, D., Van Ijzendoorn, M. H., & Bakermans-Kranenburg, M. J. (2015). Beating the brain about abuse: Empirical and meta-analytic studies of the association between maltreatment and hippocampal volume across childhood and adolescence. Development and Psychopathology, 27(2), 507–520. 10.1017/S0954579415000127 [DOI] [PubMed] [Google Scholar]
  31. Slavich, G. M., & Irwin, M. R. (2014). From stress to inflammation and major depressive disorder: A social signal transduction theory of depression. Psychological Bulletin, 140(3), 774–815. 10.1037/a0035302 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Stalder, T., & Kirschbaum, C. (2012). Analysis of cortisol in hair – State of the art and future directions. Brain, Behavior, and Immunity, 26(7), 1019–1029. 10.1016/j.bbi.2012.02.002 [DOI] [PubMed] [Google Scholar]
  33. Stalder, T., Steudte-Schmiedgen, S., Alexander, N., Klucken, T., Vater, A., Wichmann, S., Kirschbaum, C., & Miller, R. (2017). Stress-related and basic determinants of hair cortisol in humans: A meta-analysis. Psychoneuroendocrinology, 77, 261–274. 10.1016/j.psyneuen.2016.12.017 [DOI] [PubMed] [Google Scholar]
  34. Takizawa, R., Maughan, B., & Arseneault, L. (2014). Adult health outcomes of childhood bullying victimization: Evidence from a five-decade longitudinal British birth cohort. American Journal of Psychiatry, 171(7), 777–784. 10.1176/appi.ajp.2014.13101401 [DOI] [PubMed] [Google Scholar]
  35. Teicher, M. H., & Parigger, A. (2015). The ‘Maltreatment and Abuse Chronology of Exposure’ (MACE) scale for the retrospective assessment of abuse and neglect during development. PLoS One, 10(2), e0117423. 10.1371/journal.pone.0117423 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Teicher, M. H., Samson, J. A., Anderson, C. M., & Ohashi, K. (2016). The effects of childhood maltreatment on brain structure, function and connectivity. Nature Reviews Neuroscience, 17(10), 652–666. 10.1038/nrn.2016.111 [DOI] [PubMed] [Google Scholar]
  37. Tottenham, N., & Sheridan, M. A. (2009). A review of adversity, the amygdala and the hippocampus: A consideration of developmental timing. Frontiers in Human Neuroscience, 3, Article 68. 10.3389/neuro.09.068.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Vaillancourt, T., Hymel, S., & McDougall, P. (2013). The biological underpinnings of peer victimization: Understanding why and how the effects of bullying can last a lifetime. Theory Into Practice, 52(4), 241–248. 10.1080/00405841.2013.829726 [DOI] [Google Scholar]
  39. Wolke, D., & Lereya, S. T. (2015). Long-term effects of bullying. Archives of Disease in Childhood, 100(9), 879–885. 10.1136/archdischild-2014-306667 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary_Materials_revised_FINAL.docx
ZEPT_A_2730909_SM1842.docx (267.6KB, docx)

Data Availability Statement

The raw and processed research data supporting the findings of this study can be obtained from the corresponding author following a reasonable request and the author’s formal consent.

Articles from European Journal of Psychotraumatology are provided here courtesy of Taylor & Francis

RESOURCES