Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 May 8.
Published in final edited form as: Psychol Med. 2023 Sep 26;54(5):1045–1056. doi: 10.1017/S0033291723002866

Reciprocal relationships between stress and depressive symptoms: the essential role of the nucleus accumbens

Yizhou Ma 1, Peter Kochunov 1, Mark D Kvarta 1, Tara LeGates 2, Bhim M Adhikari 1, Joshua Chiappelli 1, Andrew van der Vaart 1, Eric L Goldwaser 3, Heather Bruce 1, Kathryn S Hatch 4, Si Gao 1, Shuo Chen 1, Ann Summerfelt 1, Thomas E Nichols 5, L Elliot Hong 1
PMCID: PMC11078439  NIHMSID: NIHMS1988932  PMID: 37750294

Abstract

Background.

Stress and depression have a reciprocal relationship, but the neural underpinnings of this reciprocity are unclear. We investigated neuroimaging phenotypes that facilitate the reciprocity between stress and depressive symptoms.

Methods.

In total, 22 195 participants (52.0% females) from the population-based UK Biobank study completed two visits (initial visit: 2006–2010, age = 55.0 ± 7.5 [40–70] years; second visit: 2014–2019; age = 62.7 ± 7.5 [44–80] years). Structural equation modeling was used to examine the longitudinal relationship between self-report stressful life events (SLEs) and depressive symptoms. Cross-sectional data were used to examine the overlap between neuroimaging correlates of SLEs and depressive symptoms on the second visit among 138 multimodal imaging phenotypes.

Results.

Longitudinal data were consistent with significant bidirectional causal relationship between SLEs and depressive symptoms. In cross-sectional analyses, SLEs were significantly associated with lower bilateral nucleus accumbal volume and lower fractional anisotropy of the forceps major. Depressive symptoms were significantly associated with extensive white matter hyperintensities, thinner cortex, lower subcortical volume, and white matter microstructural deficits, mainly in corticostriatal-limbic structures. Lower bilateral nucleus accumbal volume were the only imaging phenotypes with overlapping effects of depressive symptoms and SLEs (B = −0.032 to −0.023, p = 0.006–0.034). Depressive symptoms and SLEs significantly partially mediated the effects of each other on left and right nucleus accumbens volume (proportion of effects mediated = 12.7–14.3%, p < 0.001–p = 0.008). For the left nucleus accumbens, post-hoc seed-based analysis showed lower resting-state functional connectivity with the left orbitofrontal cortex (cluster size = 83 voxels, p = 5.4 × 10−5) in participants with high v. no SLEs.

Conclusions.

The nucleus accumbens may play a key role in the reciprocity between stress and depressive symptoms.

Keywords: depressive symptoms, multimodal neuroimaging, nucleus accumbens, reward processing, stress, UK Biobank, ventral striatum

Introduction

Depression is a leading cause of disability worldwide with over 10% of the population experiencing one or more major depressive episodes in their lifetime (Bromet et al., 2011; World Health Organization, 2020). A well-known risk factor for depression is stress: individuals with a history of childhood adversity, stressful life events (SLEs), or trauma are more likely to develop depression, relapse, and become treatment resistant (Amital, Fostick, Silberman, Beckman, & Spivak, 2008; Chapman et al., 2004; Kendler, Karkowski, & Prescott, 1999; Widom, DuMont, & Czaja, 2007). The tie between stress and depression is also reciprocal: individuals with depression are at a higher risk of encountering stress in part due to their experience with depression and its sequelae including functional impairment (Conway, Hammen, & Brennan, 2012; Hammen, 1991, 2006; van Os & Jones, 1999). Despite this bidirectional relationship, the underlying neural substrates remain unclear. Identifying the neural basis of this reciprocal and potentially reinforcing loop may help develop more effective interventions that build resilience against both stress and depression.

Both depression and stress involve neurobiological pathways in the limbic, paralimbic, and prefrontal structures and their orchestrated regulation of stress-related hormones and neurotransmitters (Godoy, Rossignoli, Delfino-Pereira, Garcia-Cairasco, & Umeoka, 2018; Gold, Machado-Vieira, & Pavlatou, 2015; Maletic et al., 2007). Neuroimaging meta-analyses show that both major depression and stress-related conditions (e.g. childhood adversity, post-traumatic stress disorder) are associated with lower hippocampal volume, lower paralimbic cortical thickness or volume, reduced white matter integrity in the corpus callosum, and altered connectivity in frontostriatal, orbitofrontal, and limbic networks (Bao et al., 2021; Calem, Bromis, McGuire, Morgan, & Kempton, 2017; Dennis et al., 2021; Drysdale et al., 2017; Kraynak, Marsland, Hanson, & Gianaros, 2019; Lim, Howells, Radua, & Rubia, 2020; Lim, Radua, & Rubia, 2014; Logue et al., 2018; Schmaal et al., 2017, 2016; van Velzen et al., 2020; Wang et al., 2021). Despite these overlaps, few human studies have directly compared the neural correlates of stress and depression in the same individuals.

In rodents, stress-induced depressive-like behavior has been linked to neurobiological changes in regions such as the hippocampus, amygdala, and ventral striatum (Brenes, Rodríguez, & Fornaguera, 2008; Hollis, Wang, Dietz, Gunjan, & Kabbaj, 2010; Zan et al., 2021). In humans, history of stress is found to independently contribute to hippocampal, prefrontal, temporal, and white matter alterations in depression (Chaney et al., 2014; Frodl, Reinhold, Koutsouleris, Reiser, & Meisenzahl, 2010; Kronmüller et al., 2008; Meinert et al., 2019; Tozzi et al., 2020; van Harmelen et al., 2010; Vythilingam et al., 2002). However, a prevailing focus of existing neurobiological research is the unidirectional effect of stress on depression, while the reverse is over-looked where depression increases the vulnerability to stress. To our knowledge, no studies have attempted to identify overlapping brain regions in stress and depression and their role in the reciprocal stress–depression relationship. In this study, we explored this question in a large sample of community adults. We chose to study depressive symptoms, rather than clinical depression, as it can be measured continuously and fluctuate in individuals regardless of a diagnosis of depression. We hypothesized that SLEs experienced by individuals would predict their depressive symptoms over time and vice versa. We conducted brain-wide association analyses across multimodal neuroimaging phenotypes to capture brain structures shared between SLEs and depressive symptoms. Based on the literature, we predicted that SLEs and depressive symptoms would share neuroimaging correlates in structures including the hippocampus, amygdala, hypothalamus, ventral striatum, and prefrontal cortex (Brenes et al., 2008; Chaney et al., 2014; Frodl et al., 2010; Hollis et al., 2010; Kronmüller et al., 2008; Meinert et al., 2019; Tozzi et al., 2020; van Harmelen et al., 2010; Vythilingam et al., 2002; Zan et al., 2021), which may be the key region(s) to understand their reciprocity.

Materials and methods

Participants

In total, 22 195 community adults (52.0% female) in the UK Biobank (UKBB)’s neuroimaging data release version 1.6 were included. We used data from the initial assessment visit (v0, 2006–2010) and the second visit with brain imaging (v2, 2014–2019). The first repeat assessment visit (v1), completed by only 24.5% participants, was not used. Participants’ age was 55.0 ± 7.5 (range 40–70) years on v0 and 62.7 ± 7.5 (44–80) years on v2. Interval between visits was 7.7 ± 1.5 (4–12) years. Behavioral and neuroimaging protocols were previously described (Miller et al., 2016; Sudlow et al., 2015). The UKBB was approved by the North West Multi-center Research Ethics Committee. All participants provided written informed consent. We received approval from the UKBB to access and analyze the data. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

Stressful life events (SLEs)

The UKBB does not use a validated inventory for SLEs but includes a screening question on stressful events such as illness, injury, and bereavement. On both visits, participants reported if they experienced the following in the last two years: ‘serious illness, injury, or assault to yourself’, ‘serious illness, injury, or assault of a close relative’, ‘death of a close relative’, ‘death of a spouse or partner’, ‘marital separation/divorce’, and ‘financial difficulties’. Research in the UKBB showed that SLEs measured by this question was significantly positively associated with psycho-pathology and chronic pain (Davis et al., 2020; Macfarlane, Beasley, & Macfarlane, 2014). Following a previous study (Macfarlane et al., 2014), we defined SLEs as the number of events reported, and recoded values larger than 3 as 3 to reduce right skewness. Alternatively, we used log transformation to reduce skewness and reported findings in online Supplementary Results. In total, 97.1% of the participants had data for both visits.

As illustrated in online Supplementary Fig. S1, on both visits, most participants reported no SLEs. On average, around 30% of the participants reported one SLE; 8% reported two SLEs; and 1.5% reported three or more SLEs (3 SLEs: 1.3%, 4 SLEs: 0.2%, 5 SLEs: 0.01%). The two most reported SLEs were ‘death of a close relative’ (around 20%) and ‘serious illness, injury or assault of a close relative’ (around 13%). The least reported SLE was ‘death of a spouse or partner’ (less than 1.5%). On both v0 and v2, more SLEs were associated with younger age (r = −0.11 and −0.14, p < 2 × 10−16), female sex (B = 0.08 and 0.10, p < 2 × 10−16), and non-White British ancestry (B = 0.06 and 0.04, p = 0.0002 and 0.006).

Depressive symptoms

The UKBB does not use a validated tool to measure depressive symptoms on each visit. Following a previous study (Arnau-Soler et al., 2019), we measured participants’ depressive symptoms over the preceding 2 weeks with four touchscreen questions. Two questions were from the Patient Health Questionnaire (PHQ)-2 (Kroenke, Spitzer, & Williams, 2003) and measured core depressive symptoms of anhedonia (‘had little interest or pleasure in doing things’) and depressed mood (‘felt down, depressed, or hopeless’). One question was from the PHQ-9 (Spitzer, Kroenke, Williams, & Group, 1999) and measured fatigue (‘felt tired or had little energy’). The last question measured psychomotor agitation (‘felt tense, fidgety, or restless’). On both visits, participants rated their symptoms over the past 2 weeks from ‘not at all’ to ‘nearly every day’ (0–3). Cronbach’s α was 0.8 on v0 and 0.78 on v2. Following the previous study (Arnau-Soler et al., 2019), we transformed the sum of depressive symptoms to a four-point scale to correct for right skewness (0 = 0, 1 = 1–2; 2 = 3–5, 3 = 6 or more). Findings when log transformation was used to correct for skewness were included in online Supplementary Results. In total, 88.4% of the participants had data for both visits.

To further examine the validity of the depressive symptoms measure, we compared depressive symptoms across participants with different lifetime history of depression (Smith et al., 2013). On both visits, participants with a history of probable recurrent depression had significantly more depressive symptoms than participants with a history of probable single episode depression, followed by subclinical depression and mood disorder controls, supporting the validity of the depressive symptoms measure. Details of this analysis, including criteria used to determine lifetime history of depression, can be found in the online Supplementary Methods.

Online Supplementary Fig. S3 illustrates the distribution of depressive symptoms across the two visits. On average, around 42% of the participants reported no depressive symptoms; around 34% scored 1; 14% scored 2; and 3% scored 3. On both visits, more depressive symptoms were associated with younger age (r = −0.18 and −0.16, p < 2 × 10−16), female sex (B = 0.14 and 0.16, p < 2 × 10−16), and non-White British ancestry (B = 0.04 and 0.06, p = 0.04 and 0.007).

Adverse childhood experiences (ACEs)

As SLEs may be influenced by ACEs (Gheorghe, Li, Gallacher, & Bauermeister, 2021), we considered ACEs reported by a subsample of participants (71.3%). We measured ACEs as the responses to the Childhood Trauma Screener (Glaesmer et al., 2013; Witt et al., 2022), which surveys participants’ experiences of emotional neglect, physical neglect, emotional abuse, physical abuse, and sexual abuse in childhood. We dichotomized each item and summed them to result in total ACEs ranging from 0 to 5 (Gheorghe et al., 2021).

Multimodal imaging phenotypes

All participants completed multimodal neuroimaging on v2. Given that no neuroimaging was performed at v0, all analyses involving neuroimaging were cross-sectional using v2 data. We examined 138 imaging phenotypes generated by the UKBB (online Supplementary Table S4), including (1) cortical thickness (CT) of 68 regions; (2) intracranial volume; (3) subcortical volume (SV) of 14 structures; (4) total volume of white matter hyperintensities; and (5) weighted-mean fractional anisotropy (FA) and mean diffusivity (MD) of 27 white matter tracts. Details of image acquisition, quality control, and processing were previously described (Alfaro-Almagro et al., 2018; Miller et al., 2016). Briefly, participants completed T1-MPRAGE (voxel size = 1 mm3), T2-FLAIR (voxel size = 1.05 × 1 × 1 mm3), and diffusion (50 × b = 1000 and 2000 s/mm2, voxel size = 2mm3) magnetic resonance imaging (MRI). CT, SV, and intracranial volume were derived from UKBB’s T1 pipeline (Alfaro-Almagro et al., 2018) and generated with FreeSurfer (Desikan et al., 2006; Fischl et al., 2004) and FSL’s FIRST tool (Patenaude, Smith, Kennedy, & Jenkinson, 2011). White matter hyperintensities were derived from UKBB’s T2 pipeline (Alfaro-Almagro et al., 2018). FA and MD were derived from UKBB’s diffusion pipeline (Alfaro-Almagro et al., 2018) and generated with FSL’s TBSS tool (Smith et al., 2006). For each phenotype, observations outside of six mean absolute deviations from the median were excluded.

Resting-state functional MRI (rsfMRI)

Participants completed 6 min of rsfMRI on v2 (TR = 735 ms, TE = 39 ms, flip angle = 52°, voxel size = 2.4 mm3). Images acquisition, quality control, and processing were previously described (Alfaro-Almagro et al., 2018; Miller et al., 2016), which included artifact removal with FMRIB’s ICA-based X-noiseifier (Beckmann & Smith, 2004; Griffanti et al., 2014; Salimi-Khorshidi et al., 2014). We additionally completed the following in FSL: spatial registration to the 2 mm MNI152 space, high-pass filtering (FWHM = 2355 s), and spatial smoothing (FWHM = 4 mm, iso-tropic Gaussian kernel).

Statistical analysis

Figure 1a illustrates the data structure used to explore the longitudinal reciprocity between stress and depressive symptoms. We hypothesized that depressive symptoms on v0 would predict more subsequent SLEs when controlling for previous SLEs (here proxied by SLEs within 2 years before v2 and v0, respectively). We also hypothesized that SLEs within 2 years before v2 would predict more subsequent depressive symptoms when controlling for previous depressive symptoms (here proxied by depressive symptoms within 2 weeks before v2 and v0, respectively). We implemented longitudinal structural equation modeling using the lavaan package (Rosseel, 2012) in R (R Core Team, 2019), with maximum likelihood estimation with robust standard errors and Satorra-Bentler scaled test statistics.

Figure 1.

Figure 1.

Open in a new tab

Relationship between SLEs and depressive symptoms based on structural equation modeling. (a) Timeline of repeated assessments across two visits and four time windows. (b) Structural equation model showing bidirectional relationship between depressive symptoms and SLEs. SLEs: stressful life events. β, standardized coefficient; S.E., standard error; v0, initial assessment visit; v2, second visit that included brain imaging.

We identified imaging phenotypes associated with SLEs on v2 by regressing each standardized imaging phenotype on SLEs in R, correcting for 138 phenotypes with false discovery rate (FDR) q < 0.05 (Benjamini & Hochberg, 2000). The same analysis was repeated for depressive symptoms. We then identified which structure(s) were significantly associated with both SLEs and depressive symptoms.

To examine whether imaging phenotypes significantly associated with both SLEs and depressive symptoms were implicated in their reciprocity, we tested two mediation models using v2 data. One model tested the effect of depressive symptoms mediating the relationship between SLEs and imaging phenotypes. The other tested the effect of SLEs mediating the relationship between depressive symptoms and imaging phenotypes. We used the mediation package (Tingley, Yamamoto, Hirose, Keele, & Imai, 2014) in R with 1000 nonparametric bootstraps.

We further performed post-hoc resting-state functional connectivity (rsFC) analysis on v2 using structures significantly associated with both SLEs and depressive symptoms as seeds. RsFC was the Fisher’s z transformed Pearson’s correlation between a voxel’s timeseries and a seed’s mean timeseries. To reduce computational and storage burden, we selected a subsample a priori (N = 1932) to form a balanced 2 × 2 factorial design between SLEs and depressive symptoms. Specifically, we selected participants with high SLEs (HS, when SLEs ≥ 2) v. no SLEs (NS) and high depressive symptoms (HD, when depressive symptoms ≥ 2) v. no depressive symptoms (ND). Among participants who completed rsfMRI, 483 were HS and ND, 509 HS and HD, 1498 NS and HD, and 5795 NS and ND. The final subsample thus included 483 participants from each group, with NS and HD and NS and ND one-on-one matched to HS and ND for age (±3 years) and sex. Because the HS and HD pool was small, we could not fully match this group to HS and ND on age (HS and ND = 60.9 ± 6.7 years, HS and HD = 58.0 ± 6.5 years, p = 8.1 × 10−12) and sex (HS and ND = 59.0% female, HS and HD = 67.7%, p = 0.006), and we included age and sex as covariates. We set threshold contrasts for SLEs and depressive symptoms with the cluster command in FSL (voxelwise threshold: p = 0.001, clusterwise threshold: p = 0.05). A post-hoc power analysis suggested that with the current sample size, the voxelwise power to detect the main effect of SLEs or depressive symptoms on rsFC when effect size was small (i.e. Cohen’s f = 0.1, equivalent to d = 0.2) was 0.86.

Table 1 summarizes the participants used in each analysis. All statistical analyses controlled for age, age2, sex, and White British ancestry. Models involving imaging also controlled for handedness (left/right/mixed), scanning site, and x, y, and z head positions in the scanner. Additionally, CT analysis controlled for whether T2_FLAIR was used with T1 in FreeSurfer preprocessing; SV analysis controlled for intracranial volume; rsFC analysis controlled for head motion. See details of covariates in online Supplementary Table S4.

Table 1.

Demographics of the participants included in each analysis

Analyses Bidirectional relationship, structural phenotypes, mediation Functional connectivity
N 22 189 1932
Age (years, v2) 62.7 ± 7.5 58.8 ± 6.6
% Female 52.0% 65.6%
White British ancestry (%) 86.1% 84.2%
College education 47.7% 48.4%
Household income 2.8 ± 1.1 2.8 ± 1.1
Open in a new tab

Note: Household income: (1) less than £18 000; (2) £18 000–£30 999; (3) £31 000–£51 999; (4) £52 000–£100 000; (5) greater than £ 100 000.

Results

Relationship between SLEs and depressive symptoms

Structural equation modeling revealed a significant bidirectional relationship between SLEs and depressive symptoms (Fig. 1b). Controlling for SLEs reported on v0, depressive symptoms on v0 were positively associated with SLEs reported on v2 (β = 0.079, p < 0.001), suggesting that more severe depression predicted more future SLEs. Conversely, controlling for depressive symptoms on v0, SLEs reported on v2 were positively associated with depressive symptoms on v2 (β = 0.112, p < 0.001), suggesting that past SLEs predicted depressive symptoms. An alternative model additionally supported a direct effect of SLEs reported on v0 on depressive symptoms on v2, consistent with lasting effects of SLEs on depressive symptoms (online Supplementary Fig. S4).

SLEs, depressive symptoms, and imaging phenotypes

More SLEs were associated with three imaging phenotypes (Fig. 2, online Supplementary Table S5), including lower forceps major FA (B = −0.039, pFDR = 0.023) and lower bilateral nucleus accumbens volume (left: B = −0.032, pFDR = 0.023; right: B = −0.030, pFDR = 0.034; B: differences in standardized imaging phenotypes when SLEs increased by 1). As expected, SLEs and ACEs were significantly positively correlated (r = 0.074, p < 2 × 10−16). Among the three imaging phenotypes associated with SLEs, more ACEs were only associated with lower forceps major FA (B = −0.024, p = 0.0007), not bilateral nucleus accumbens volume (left: B = 0.0002, p = 0.97, right: B = −0.006, p = 0.35). After controlling for ACEs, the above effects of SLEs remained nominally significant but did not survive FDR correction (forceps major FA: B = −0.037, p = 0.003; nucleus accumbens volume left: B = −0.028, p = 0.008, right: B = −0.025, p = 0.019). Controlling for ACEs led to a reduction of the absolute B values by 5.7, 12.0, and 16.8%, respectively, which was expected as more ACEs are known to be linked to more SLEs in adulthood (Halonen et al., 2017) that is also supported by their significant correlation in this cohort.

Figure 2.

Figure 2.

Open in a new tab

Imaging phenotypes associates of depressive symptoms and SLEs. All p values were false discovery rate (FDR) corrected. Dashed line: −log(p) corresponding to corrected p = 0.05. CT, cortical thickness; ICV, intracranial volume; SV, subcortical volume; WMH, white matter hyperintensities; FA, fractional anisotropy; MD, mean diffusivity; SLEs, stressful life events. For names of the imaging phenotypes, see online Supplementary Table S4.

Higher depressive symptoms were associated with 44 imaging phenotypes including (1) thinner cortex in 23 regions with the strongest effects in the left rostral anterior cingulate and medial orbitofrontal cortices; (2) smaller volume in six subcortical structures with the strongest effects in bilateral nucleus accumbens; (3) higher MD in 11 tracts with the strongest effects in the bilateral anterior thalamic radiation; (4) lower FA in three tracts with the strongest effects in the bilateral posterior thalamic radiation; and (5) higher total white matter hyperintensities (Fig. 2, online Supplementary Table S5).

Lower bilateral nucleus accumbens volumes were the only imaging phenotypes significantly associated with both depressive symptoms (left: B = −0.026, pFDR = 0.006; right: B = −0.023, pFDR = 0.016) and SLEs (see above). This remained true when using log transformation instead of the recoding method to correct for right skewness in SLEs and depressive symptoms (online Supplementary Results, Table S6).

Mediation effects

Depressive symptoms partially and significantly mediated the relationship between SLEs and bilateral nucleus accumbens volume (left: indirect effect = −0.004, p < 0.001, 95% confidence interval [CI] −0.007 to −0.001, 12.8% of the total effect was mediated; right: indirect effect = −0.004, p = 0.012, 95% CI −0.006 to −0.001, 12.7% mediated) (Fig. 3a). Conversely, SLEs also partially and significantly mediated the relationship between depressive symptoms and bilateral nucleus accumbens volume (left: indirect effect = −0.004, p = 0.002, 95% CI −0.006 to −0.001, 14.3% mediated; right: indirect effect = −0.003, p = 0.008, 95% CI −0.006 to −0.001, 14.3% mediated) (Fig. 3b). The mediation effects in both directions accounted for similar proportion of the total effects.

Figure 3.

Figure 3.

Open in a new tab

Mediation effects. (a) Depressive symptoms significantly mediated the relationship between SLEs and bilateral nucleus accumbens volume. (b) SLEs significantly mediated the relationship between depressive symptoms and bilateral nucleus accumbens volume. Depressive symptoms and SLEs were measured at v2. All paths controlled for age, age2, sex, and White British ancestry. Paths involving the nucleus accumbens volume additionally controlled for handedness, scanning site, x, y, and z head positions in the scanner, and intracranial volume. SLEs: stressful life events. L, left; R, right; Vol, volume. B values may be slightly different from those reported in online Supplementary Table S5 due to slightly different numbers of cases with complete data.

Ideally, the model in Fig. 3b would use depressive symptoms reported immediately prior to SLEs, while we used depressive symptoms on v2 as a proxy (Chiappelli, Nugent, Thangavelu, Searcy, & Hong, 2014). However, when using depressive symptoms on v0 instead, the mediation effects remained largely the same (left: indirect effect = −0.003, p < 0.001, 95% CI −0.005 to −0.001, 14.5% mediated; right: indirect effect = −0.003, p = 0.004, 95% CI −0.004 to −0.001, 11.5% mediated), supporting that longitudinally experienced depressive symptoms may have influenced the nucleus accumbens via increased occurrences of SLEs. To further test this hypothesis, we tentatively categorized SLEs based on their likelihood to be a consequence of depressive symptoms (Brown & Harris, 1978; Paykel, 1987). Independent SLEs (iSLEs), or events that are more likely independent from depressive symptoms, were ‘serious illness, injury, or assault of a close relative’, ‘death of a close relative’, and ‘death of a spouse or partner’. Dependent SLEs (dSLEs), or events that are more likely partly dependent on depressive symptoms, were ‘marital separation/divorce’ and ‘financial difficulties’. ‘Serious illness, injury, or assault to yourself’ was excluded as it could be either an iSLE or dSLE. Consistent with our hypothesis, dSLEs significantly mediated the relationship between depressive symptoms and nucleus accumbens volume (left: indirect effect = −0.003, p = 0.05, 95% CI −0.005 to −0.001, 10.2% mediated; right: indirect effect = −0.003, p = 0.018, 95% CI −0.006 to −0.001, 13.3% mediated), but not iSLEs (left: p = 0.130; right: p = 0.360). However, caution must be taken to interpret this finding given the putative classification of SLEs.

Nucleus accumbens functional connectivity

Post-hoc rsFC analysis revealed significantly lower rsFC in HS than NS between the left nucleus accumbens seed and a cluster in the left orbitofrontal cortex (OFC) (Fig. 4; cluster size = 83, clusterwise p = 5.4 × 10−5, peak F value = 21.34, MNI coordinate = [−6, 62, −12]). No significant differences were found between HD and ND or for the right nucleus accumbens seed.

Figure 4.

Figure 4.

Open in a new tab

SLEs were associated with lower resting-state functional connectivity (rsFC) between the left nucleus accumbens region-of-interest (green) and the left orbitofrontal cortex (red). Voxel-wise threshold, p = 0.001, cluster-wise threshold, p = 5.4 × 10−5. Imbedded box is the plot comparing rsFC in the high SLEs (HS) and no SLEs (NS) groups. RsFC values are residuals after regressing out age, age2, sex, and White British ancestry, handedness, scanning site, x, y, and z head positions in the scanner, and head motion. L, left; R, right; SLE, stressful life event.

Replication after excluding neurological cases

Finally, we repeated our analyses after excluding participants with neurological conditions (N = 448, 2.0% of the total sample; online Supplementary Table S7). Bilateral nucleus accumbens volume remained the only imaging phenotypes significantly and negatively associated with both SLEs and depressive symptoms (online Supplementary Table S8). Conclusions from the rsFC and mediation analyses were not affected.

Discussion

We examined the brain structures implicated in the reciprocal relationship between stress and depressive symptoms by studying SLEs and depressive symptoms in a large community sample. We replicated the bidirectional relationship between SLEs and depressive symptoms reported by prior studies (Amital et al., 2008; Chapman et al., 2004; Conway et al., 2012; Hammen, 1991; Kendler et al., 1999; van Os & Jones, 1999; Widom et al., 2007). Neuroimaging correlates of SLEs included three structures and functional connectivity between the left nucleus accumbens and the left OFC. Neuroimaging correlates of depressive symptoms included widespread morphological and white matter microstructural changes. Bilateral nucleus accumbens were the only regions with overlapping effects of SLEs and depressive symptoms, and SLEs and depressive symptoms partially mediated the effect of each other on nucleus accumbens volume. The nucleus accumbens may be a neurobiological nexus for the bidirectional relationship between stress and depressive symptoms.

Stress exposure has long been recognized as a risk factor for depression (Amital et al., 2008; Chapman et al., 2004; Kendler et al., 1999; Widom et al., 2007). Reciprocally, the stress generation hypothesis posits that depressive symptoms put an individual at a higher risk for stressful experiences due to depression-related symptoms, behaviors, characteristics, and social context (Hammen, 1991, 2006). Structural equation modeling of longitudinal data was consistent with this reciprocity between SLEs and depressive symptoms. Bidirectional causal relationship is difficult to test because the proposed causal event, by definition, can be affected by the proposed consequence. Using a longitudinal design allowed this reciprocity to be examined.

We found SLEs to be associated with lower white matter integrity in the forceps major, a tract that involves the splenium part of the corpus callosum. Disruptions in the macro- and microstructure of the corpus callosum are well documented in stress and trauma (De Bellis & Keshavan, 2003; De Bellis et al., 2002; Dennis et al., 2021; Jensen et al., 2018). One mechanism can be stress-induced glucocorticoid increase that affects myelination (Huang, Harper, Evans, Newnham, & Dunlop, 2001). We also found SLEs to be associated with weaker rsFC between the left nucleus accumbens and the left OFC, an area central to emotion regulation and decision making (Bechara, Damasio, & Damasio, 2000). OFC alteration has been previously reported in stress (Hanson et al., 2010; Muhammad, Carroll, & Kolb, 2012) and reduced coordination between the OFC and the nucleus accumbens may reflect impaired self-regulation after stress exposure (Meyer & Bucci, 2016).

We replicated widespread alterations associated with depressive symptoms reported by previous studies (Milak et al., 2005; Schmaal et al., 2017, 2016; Shen et al., 2020; van Velzen et al., 2020). Some discrepancies, such as significantly lower hippocampal volume in recurrent depression (Schmaal et al., 2016) but not in the current sample, may be explained by our focus on depression symptomatology as opposed to diagnosis in an older, mostly subclinical population. Most of the imaging associates of depressive symptoms were not significantly associated with SLEs, which is expected as depressive symptoms can result from various factors other than SLEs (Saveanu & Nemeroff, 2012). One caveat though is that the brief UKBB SLEs assessment could not have captured all sources of stress, and additional brain correlates of stress may not have been discovered.

Animal models of post-traumatic stress disorder and major depression overlap extensively in hippocampus, ventral striatum, PFC, and hypothalamic-pituitary-adrenal axis pathology (Ploski & Vaidya, 2021). Here, we found that the nucleus accumbens was the only brain region associated with both SLEs and depressive symptoms. Notably, both the left and right nucleus accumbens showed overlapping effects of SLEs and depressive symptoms after multiple comparison correction, suggesting that this finding is unlikely fortuitous. Moreover, SLEs partially mediated the effect of depressive symptoms on bilateral nucleus accumbens volume and vice versa. While neuroimaging studies have shown that the brain abnormalities in depression can be attributed to previous history of stress (Meinert et al., 2019; Vythilingam et al., 2002), to our knowledge this is the first study to show the reverse, i.e. the neural associates of stress may be accounted for by preexisting depressive symptoms. Furthermore, we show that the mediation effect of SLEs was stable using depressive symptoms measured across two visits several years apart and was only significant for SLEs that are more likely influenced by depressive symptoms (i.e. dSLEs). Thus, our results are consistent with the stress generation theory of depression.

There are strong neurobiological and psychopathological bases for why the nucleus accumbens stood out as a brain structure involved in stress, depressive symptoms, and their reciprocity. The nucleus accumbens has been associated with stress and depression in humans (Edmiston et al., 2011; Gheorghe et al., 2021; Liu et al., 2021; Wacker, Dillon, & Pizzagalli, 2009; Walsh et al., 2014; Whittle et al., 2014). It plays a central role in the brain’s mesolimbic neurotransmission pathways (Baik, 2020; Fox & Lobo, 2019) by receiving dopaminergic, GABAergic, and glutamatergic projections from the ventral tegmentum, amygdala, and other areas for salience, reward, and punishment processing (Bongioanni et al., 2021; Kohls et al., 2013; Lowes et al., 2021). It is connected to essentially all limbic areas for emotional/motivational responses (Floresco, 2015; Saddoris, Cacciapaglia, Wightman, & Carelli, 2015). The nucleus accumbens is therefore a hub for converging neural processes that regulate stress response and depression formation and the impact of stress (depressive symptoms) on this region increases the vulnerability to the other. Smaller nucleus accumbens volume may be the macroscopic manifestation of these underlying processes.

This study has several limitations. First, the measures of SLEs and depressive symptoms have not been previously validated and are susceptible to recall bias (Monroe, 1982; Monroe, Slavich, Torres, & Gotlib, 2007) (although both have been used in previous studies, see Arnau-Soler et al. [2019]; Davis et al. [2020]; Harshfield et al. [2020]; Lehto et al. [2020]; Macfarlane et al. [2014]). Moreover, given restraints imposed by the UKBB design, we could only probe the longitudinal effects between SLEs and depressive symptoms with proxy measures. These issues, although practical for a very large sample, have likely introduced noises and compromised the size of relationships that could be observed, which cannot be fully compensated for by sample size. Second, while we used longitudinal analysis to probe the relationship between SLEs and depressive symptoms, it cannot be taken as direct evidence of causality. Similarly, our neuroimaging findings used cross-sectional data and can only be interpreted as correlational. Third, the 138 imaging phenotypes surveyed in our analyses were not exhaustive and additional shared neural correlates between SLEs and depressive symptoms may exist, such as rsFC phenotypes. Machine-learning approaches with unsupervised learning may be better suited to identify distributed networks that link SLEs and depression (Genon, Eickhoff, & Kharabian, 2022; Marek et al., 2022). Fourth, the effect sizes were small, and the SLEs findings did not survive FDR correction after controlling for ACEs, suggesting that findings in this study may also be partly attributed to early stressful experiences. Fifth, our analyses did not account for factors such as medication use, psychotherapy, and physical comorbidity. Last, the UKBB sample was predominantly healthy white volunteers (Fry et al., 2017), and generalizability to groups with more pathology and lower socioeconomic status is unclear.

To summarize, a reciprocal relationship between SLEs and depressive symptoms was supported using a longitudinal epidemiological dataset, and this bidirectional susceptibility was found to be associated with the nucleus accumbens. To our knowledge, this is the first study to identify the neural underpinnings of the bidirectional stress–depressive symptoms relationship in humans. Further research may be important for developing new therapeutics to disrupt this relationship and increase resilience for both stress-induced depressive symptoms and depressive symptoms-related stressful experiences.

Supplementary Material

supplement

Funding statement.

This work was supported by the National Institutes of Health grants R01MH112180, R01MH116948, R01EB015611, and R01MH117601 awarded to L. E. H. These funding sources provided financial support to enable design and conduct of the study or collection, management, or analysis of the data. None of the funding agencies had a role in the interpretation of the data. None had a role in the preparation, review, or approval of the manuscript. None had a role in the decision to submit the manuscript for publication.

Competing interests.

L. E. H. has received or plans to receive research funding or consulting fees on research projects from Mitsubishi, Your Energy Systems LLC, Neuralstem, Taisho, Heptares, Pfizer, Luye Pharma, Sound Pharma, Takeda, and Regeneron. None was involved in the design, analysis, or outcomes of the study. All other authors declare no financial relationships with commercial interests.

Footnotes

Supplementary material. The supplementary material for this article can be found at https://doi.org/10.1017/S0033291723002866.

Code availability. Associated codes are available upon reasonable request from the corresponding author.

Data availability statement.

Data supporting the findings of this study are available upon reasonable request from the corresponding author. Raw data are available through the UK Biobank.

References

  1. Alfaro-Almagro F, Jenkinson M, Bangerter NK, Andersson JLR, Griffanti L, Douaud G, … Smith SM (2018). Image processing and Quality Control for the first 10,000 brain imaging datasets from UK Biobank. NeuroImage, 166(April 2017), 400–424. 10.1016/j.neuroimage.2017.10.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Amital D, Fostick L, Silberman A, Beckman M, & Spivak B (2008). Serious life events among resistant and non-resistant MDD patients. Journal of Affective Disorders, 110(3), 260–264. [DOI] [PubMed] [Google Scholar]
  3. Arnau-Soler A, Macdonald-Dunlop E, Adams MJ, Clarke TK, MacIntyre DJ, Milburn K, … Thomson PA (2019). Genome-wide by environment interaction studies of depressive symptoms and psychosocial stress in UK Biobank and Generation Scotland. Translational Psychiatry, 9(1), 14. 10.1038/s41398-018-0360-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Baik JH (2020). Stress and the dopaminergic reward system. Experimental and Molecular Medicine, 52(12), 1879–1890. 10.1038/s12276-020-00532-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bao W, Gao Y, Cao L, Li H, Liu J, Liang K, … Huang X (2021). Alterations in large-scale functional networks in adult posttraumatic stress disorder: A systematic review and meta-analysis of resting-state functional connectivity studies. Neuroscience and Biobehavioral Reviews, 131, 1027–1036. 10.1016/j.neubiorev.2021.10.017 [DOI] [PubMed] [Google Scholar]
  6. Bechara A, Damasio H, & Damasio AR (2000). Emotion, decision making and the orbitofrontal cortex. Cerebral Cortex, 10(3), 295–307. 10.1093/cercor/10.3.295 [DOI] [PubMed] [Google Scholar]
  7. Beckmann CF, & Smith SM (2004). Probabilistic independent component analysis for functional magnetic resonance imaging. IEEE Transactions on Medical Imaging, 23(2), 137–152. 10.1109/TMI.2003.822821 [DOI] [PubMed] [Google Scholar]
  8. Benjamini Y, & Hochberg Y (2000). On the adaptive control of the false discovery rate in multiple testing with independent statistics. Journal of Educational and Behavioral Statistics, 25(1), 60–83. [Google Scholar]
  9. Bongioanni A, Folloni D, Verhagen L, Sallet J, Klein-Flügge MC, & Rushworth MFS (2021). Activation and disruption of a neural mechanism for novel choice in monkeys. Nature, 591(7849), 270–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Brenes JC, Rodríguez O, & Fornaguera J (2008). Differential effect of environment enrichment and social isolation on depressive-like behavior, spontaneous activity and serotonin and norepinephrine concentration in prefrontal cortex and ventral striatum. Pharmacology Biochemistry and Behavior, 89(1), 85–93. 10.1016/j.pbb.2007.11.004 [DOI] [PubMed] [Google Scholar]
  11. Bromet E, Andrade LH, Hwang I, Sampson NA, Alonso J, De Girolamo G, … Iwata N (2011). Cross-national epidemiology of DSM-IV major depressive episode. BMC Medicine, 9(1), 1–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Brown GW, & Harris T (1978). Social origins of depression: A reply. Psychological Medicine, 8(4), 577–588. [DOI] [PubMed] [Google Scholar]
  13. Calem M, Bromis K, McGuire P, Morgan C, & Kempton MJ (2017). Meta-analysis of associations between childhood adversity and hippocampus and amygdala volume in non-clinical and general population samples. NeuroImage: Clinical, 14, 471–479. 10.1016/j.nicl.2017.02.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chaney A, Carballedo A, Amico F, Fagan A, Skokauskas N, Meaney J, & Frodl T (2014). Effect of childhood maltreatment on brain structure in adult patients with major depressive disorder and healthy participants. Journal of Psychiatry and Neuroscience, 39(1), 50–59. 10.1503/jpn.120208 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Chapman DP, Whitfield CL, Felitti VJ, Dube SR, Edwards VJ, & Anda RF (2004). Adverse childhood experiences and the risk of depressive disorders in adulthood. Journal of Affective Disorders, 82(2), 217–225. [DOI] [PubMed] [Google Scholar]
  16. Chiappelli J, Nugent KL, Thangavelu K, Searcy K, & Hong LE (2014). Assessment of trait and state aspects of depression in schizophrenia. Schizophrenia Bulletin, 40(1), 132–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Conway CC, Hammen C, & Brennan PA (2012). Expanding stress generation theory: Test of a transdiagnostic model. Journal of Abnormal Psychology, 121(3), 754–766. 10.1037/a0027457 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Davis KAS, Coleman JRI, Adams M, Allen N, Breen G, Cullen B, … Hotopf M (2020). Mental health in UK Biobank – development, implementation and results from an online questionnaire completed by 157 366 participants: A reanalysis. BJPsych Open, 6(2), 1–8. 10.1192/bjo.2019.100 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. De Bellis MD, & Keshavan MS (2003). Sex differences in brain maturation in maltreatment-related pediatric posttraumatic stress disorder. Neuroscience and Biobehavioral Reviews, 27(1–2), 103–117. 10.1016/S0149-7634(03)00013-7 [DOI] [PubMed] [Google Scholar]
  20. De Bellis MD, Keshavan MS, Shifflett H, Iyengar S, Beers SR, Hall J, & Moritz G (2002). Brain structures in pediatric maltreatment-related posttraumatic stress disorder: A sociodemographically matched study. Biological Psychiatry, 52(11), 1066–1078. 10.1016/S0006-3223(02)01459-2 [DOI] [PubMed] [Google Scholar]
  21. Dennis EL, Disner SG, Fani N, Salminen LE, Logue M, Clarke EK, … Morey RA (2021). Altered white matter microstructural organization in posttraumatic stress disorder across 3047 adults: Results from the PGC-ENIGMA PTSD consortium. Molecular Psychiatry, 26(8), 4315–4330. 10.1038/s41380-019-0631-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Desikan RS, Ségonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D, … Killiany RJ (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 31(3), 968–980. 10.1016/j.neuroimage.2006.01.021 [DOI] [PubMed] [Google Scholar]
  23. Drysdale AT, Grosenick L, Downar J, Dunlop K, Mansouri F, Meng Y, … Liston C (2017). Resting-state connectivity biomarkers define neuro-physiological subtypes of depression. Nature Medicine, 23(1), 28–38. 10.1038/nm.4246 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Edmiston EE, Wang F, Mazure CM, Guiney J, Sinha R, Mayes LC, & Blumberg HP (2011). Corticostriatal-limbic gray matter morphology in adolescents with self-reported exposure to childhood maltreatment. Archives of Pediatrics and Adolescent Medicine, 165(12), 1069–1077. 10.1001/archpediatrics.2011.565 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Fischl B, Van Der Kouwe A, Destrieux C, Halgren E, Ségonne F, Salat DH, … Dale AM (2004). Automatically parcellating the human cerebral cortex. Cerebral Cortex, 14(1), 11–22. 10.1093/cercor/bhg087 [DOI] [PubMed] [Google Scholar]
  26. Floresco SB (2015). The nucleus accumbens: An interface between cognition, emotion, and action. Annual Review of Psychology, 66, 25–52. [DOI] [PubMed] [Google Scholar]
  27. Fox ME, & Lobo MK (2019). The molecular and cellular mechanisms of depression: A focus on reward circuitry. Molecular Psychiatry, 24(12), 1798–1815. 10.1038/s41380-019-0415-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Frodl T, Reinhold E, Koutsouleris N, Reiser M, & Meisenzahl EM (2010). Interaction of childhood stress with hippocampus and prefrontal cortex volume reduction in major depression. Journal of Psychiatric Research, 44(13), 799–807. 10.1016/j.jpsychires.2010.01.006 [DOI] [PubMed] [Google Scholar]
  29. Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T, … Allen NE (2017). Comparison of sociodemographic and health-related characteristics of UK Biobank participants with those of the general population. American Journal of Epidemiology, 186(9), 1026–1034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Genon S, Eickhoff SB, & Kharabian S (2022). Linking interindividual variability in brain structure to behaviour. Nature Reviews Neuroscience, 23(5), 307–318. 10.1038/s41583-022-00584-7 [DOI] [PubMed] [Google Scholar]
  31. Gheorghe DA, Li C, Gallacher J, & Bauermeister S (2021). Associations of perceived adverse lifetime experiences with brain structure in UK Biobank participants. Journal of Child Psychology and Psychiatry and Allied Disciplines, 62(7), 822–830. 10.1111/jcpp.13298 [DOI] [PubMed] [Google Scholar]
  32. Glaesmer H, Schulz A, Häuser W, Freyberger HJ, Brähler E, & Grabe H-J (2013). The childhood trauma screener (CTS)-development and validation of cut-off-scores for classificatory diagnostics. Psychiatrische Praxis, 40 (4), 220–226. [DOI] [PubMed] [Google Scholar]
  33. Godoy LD, Rossignoli MT, Delfino-Pereira P, Garcia-Cairasco N, & Umeoka E. H. de L. (2018). A comprehensive overview on stress neurobiology: Basic concepts and clinical implications. Frontiers in Behavioral Neuroscience, 12(July), 1–23. 10.3389/fnbeh.2018.00127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Gold PW, Machado-Vieira R, & Pavlatou MG (2015). Clinical and biochemical manifestations of depression: Relation to the neurobiology of stress. Neural Plasticity, 2015, 7–9. 10.1155/2015/581976 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Griffanti L, Salimi-Khorshidi G, Beckmann CF, Auerbach EJ, Douaud G, Sexton CE, … Smith SM (2014). ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging. NeuroImage, 95, 232–247. 10.1016/j.neuroimage.2014.03.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Halonen JI, Kivimäki M, Vahtera J, Pentti J, Virtanen M, Ervasti J, … Lallukka T (2017). Childhood adversity, adult socioeconomic status and risk of work disability: A prospective cohort study. Occupational and Environmental Medicine, 74(9), 659–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Hammen C (1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100(4), 555. [DOI] [PubMed] [Google Scholar]
  38. Hammen C (2006). Stress generation in depression: Reflections on origins, research, and future directions. Journal of Clinical Psychology, 62(9), 1065–1082. [DOI] [PubMed] [Google Scholar]
  39. Hanson JL, Chung MK, Avants BB, Shirtcliff EA, Gee JC, Davidson RJ, & Pollak SD (2010). Early stress is associated with alterations in the orbitofrontal cortex: A tensor-based morphometry investigation of brain structure and behavioral risk. Journal of Neuroscience, 30(22), 7466–7472. 10.1523/JNEUROSCI.0859-10.2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Harshfield EL, Pennells L, Schwartz JE, Willeit P, Kaptoge S, Bell S, … Wassertheil-Smoller S (2020). Association between depressive symptoms and incident cardiovascular diseases. JAMA, 324(23), 2396–2405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Hollis F, Wang H, Dietz D, Gunjan A, & Kabbaj M (2010). The effects of repeated social defeat on long-term depressive-like behavior and short-term histone modifications in the hippocampus in male Sprague-Dawley rats. Psychopharmacology, 211(1), 69–77. 10.1007/s00213-010-1869-9 [DOI] [PubMed] [Google Scholar]
  42. Huang WL, Harper CG, Evans SF, Newnham JP, & Dunlop SA (2001). Repeated prenatal corticosteroid administration delays myelination of the corpus callosum in fetal sheep. International Journal of Developmental Neuroscience, 19(4), 415–425. [DOI] [PubMed] [Google Scholar]
  43. Jensen SKG, Pangelinan M, Björnholm L, Klasnja A, Leemans A, Drakesmith M, … Paus T (2018). Associations between prenatal, childhood, and adolescent stress and variations in white-matter properties in young men. NeuroImage, 182(June 2017), 389–397. 10.1016/j.neuroimage.2017.10.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Kendler KS, Karkowski LM, & Prescott CA (1999). Causal relationship between stressful life events and the onset of major depression. American Journal of Psychiatry, 156(6), 837–841. [DOI] [PubMed] [Google Scholar]
  45. Kohls G, Perino MT, Taylor JM, Madva EN, Cayless SJ, Troiani V, … Schultz RT (2013). The nucleus accumbens is involved in both the pursuit of social reward and the avoidance of social punishment. Neuropsychologia, 51(11), 2062–2069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Kraynak TE, Marsland AL, Hanson JL, & Gianaros PJ (2019). Retrospectively reported childhood physical abuse, systemic inflammation, and resting corticolimbic connectivity in midlife adults. Brain, Behavior, and Immunity, 82, 203–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Kroenke K, Spitzer RL, & Williams JBW (2003). The patient health questionnaire-2: Validity of a two-item depression screener. Medical Care, 41(11), 1284–1292. [DOI] [PubMed] [Google Scholar]
  48. Kronmüller KT, Pantel J, Götz B, Köhler S, Victor D, Mundt C, … Schröder J (2008). Life events and hippocampal volume in first-episode major depression. Journal of Affective Disorders, 110(3), 241–247. 10.1016/j.jad.2008.01.022 [DOI] [PubMed] [Google Scholar]
  49. Lehto K, Hägg S, Lu D, Karlsson R, Pedersen NL, & Mosing MA (2020). Childhood adoption and mental health in adulthood: The role of gene-environment correlations and interactions in the UK Biobank. Biological Psychiatry, 87(8), 708–716. 10.1016/j.biopsych.2019.10.016 [DOI] [PubMed] [Google Scholar]
  50. Lim L, Howells H, Radua J, & Rubia K (2020). Aberrant structural connectivity in childhood maltreatment: A meta-analysis. Neuroscience and Biobehavioral Reviews, 116(January), 406–414. 10.1016/j.neubiorev.2020.07.004 [DOI] [PubMed] [Google Scholar]
  51. Lim L, Radua J, & Rubia K (2014). Gray matter abnormalities in childhood maltreatment: A voxelwise meta-analysis. American Journal of Psychiatry, 171(8), 854–863. 10.1176/appi.ajp.2014.13101427 [DOI] [PubMed] [Google Scholar]
  52. Liu R, Wang Y, Chen X, Zhang Z, Xiao L, & Zhou Y (2021). Anhedonia correlates with functional connectivity of the nucleus accumbens subregions in patients with major depressive disorder. NeuroImage: Clinical, 30, 102599. 10.1016/j.nicl.2021.102599 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Logue MW, van Rooij SJH, Dennis EL, Davis SL, Hayes JP, Stevens JS, … Morey RA (2018). Smaller hippocampal volume in post-traumatic stress disorder: A multisite ENIGMA-PGC study: Subcortical volumetry results from posttraumatic stress disorder consortia. Biological Psychiatry, 83(3), 244–253. 10.1016/j.biopsych.2017.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Lowes DC, Chamberlin LA, Kretsge LN, Holt ES, Abbas AI, Park AJ, … Enikolopov AG (2021). Ventral tegmental area GABA neurons mediate stress-induced blunted reward-seeking in mice. Nature Communications, 12(1), 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Macfarlane TV, Beasley M, & Macfarlane GJ (2014). Self-reported facial pain in UK Biobank study: Prevalence and associated factors. Journal of Oral and Maxillofacial Research, 5(3), 1–13. 10.5037/jomr.2014.5302 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Maletic V, Robinson M, Oakes T, Iyengar S, Ball SG, & Russell J (2007). Neurobiology of depression: An integrated view of key findings. International Journal of Clinical Practice, 61(12), 2030–2040. 10.1111/j.1742-1241.2007.01602.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Marek S, Tervo-Clemmens B, Calabro FJ, Montez DF, Kay BP, Hatoum AS, … Hendrickson TJ (2022). Reproducible brain-wide association studies require thousands of individuals. Nature, 603(7902), 654–660. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Meinert S, Repple J, Nenadic I, Krug A, Jansen A, Grotegerd D, … Dannlowski U (2019). Reduced fractional anisotropy in depressed patients due to childhood maltreatment rather than diagnosis. Neuropsychopharmacology, 44(12), 2065–2072. 10.1038/s41386-019-0472-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Meyer HC, & Bucci DJ (2016). Imbalanced activity in the orbitofrontal cortex and nucleus accumbens impairs behavioral inhibition. Current Biology, 26(20), 2834–2839. 10.1016/j.cub.2016.08.034 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Milak MS, Parsey RV, Keilp J, Oquendo MA, Malone KM, & Mann JJ (2005). Neuroanatomic correlates of psychopathologic components of major depressive disorder. Archives of General Psychiatry, 62(4), 397–408. [DOI] [PubMed] [Google Scholar]
  61. Miller KL, Alfaro-Almagro F, Bangerter NK, Thomas DL, Yacoub E, Xu J, … Andersson JLR (2016). Multimodal population brain imaging in the UK Biobank prospective epidemiological study. Nature Neuroscience, 19(11), 1523–1536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Monroe SM (1982). Assessment of life events: Retrospective vs concurrent strategies. Archives of General Psychiatry, 39(5), 606–610. 10.1001/archpsyc.1982.04290050074014 [DOI] [PubMed] [Google Scholar]
  63. Monroe SM, Slavich GM, Torres LD, & Gotlib IH (2007). Major life events and major chronic difficulties are differentially associated with history of major depressive episodes. Journal of Abnormal Psychology, 116 (1), 116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Muhammad A, Carroll C, & Kolb B (2012). Stress during development alters dendritic morphology in the nucleus accumbens and prefrontal cortex. Neuroscience, 216, 103–109. 10.1016/j.neuroscience.2012.04.041 [DOI] [PubMed] [Google Scholar]
  65. Patenaude B, Smith SM, Kennedy DN, & Jenkinson M (2011). A Bayesian model of shape and appearance for subcortical brain segmentation. NeuroImage, 56(3), 907–922. 10.1016/j.neuroimage.2011.02.046 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Paykel ES (1987). Methodology of life events research. Research Paradigms in Psychosomatic Medicine, 17, 13–29. [DOI] [PubMed] [Google Scholar]
  67. Ploski JE, & Vaidya VA (2021). The neurocircuitry of posttraumatic stress disorder and major depression: Insights into overlapping and distinct circuit dysfunction – a tribute to Ron Duman. Biological Psychiatry, 90(2), 109–117. 10.1016/j.biopsych.2021.04.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. R Core Team. (2019). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. [Google Scholar]
  69. Rosseel Y (2012). Lavaan: An R package for structural equation modeling and more. Version 0.5–12 (BETA). Journal of Statistical Software, 48(2), 1–36. [Google Scholar]
  70. Saddoris MP, Cacciapaglia F, Wightman RM, & Carelli RM (2015). Differential dopamine release dynamics in the nucleus accumbens core and shell reveal complementary signals for error prediction and incentive motivation. Journal of Neuroscience, 35(33), 11572–11582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Salimi-Khorshidi G, Douaud G, Beckmann CF, Glasser MF, Griffanti L, & Smith SM (2014). Automatic denoising of functional MRI data: Combining independent component analysis and hierarchical fusion of classifiers. NeuroImage, 90, 449–468. 10.1016/j.neuroimage.2013.11.046 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Saveanu RV, & Nemeroff CB (2012). Etiology of depression: Genetic and environmental factors. Psychiatric Clinics of North America, 35(1), 51–71. 10.1016/j.psc.2011.12.001 [DOI] [PubMed] [Google Scholar]
  73. Schmaal L, Hibar DP, Sämann PG, Hall GB, Baune BT, Jahanshad N, … Veltman DJ (2017). Cortical abnormalities in adults and adolescents with major depression based on brain scans from 20 cohorts worldwide in the ENIGMA Major Depressive Disorder Working Group. Molecular Psychiatry, 22(6), 900–909. 10.1038/mp.2016.60 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Schmaal L, Veltman DJ, Van Erp TGM, Smann PG, Frodl T, Jahanshad N, … Hibar DP (2016). Subcortical brain alterations in major depressive disorder: Findings from the ENIGMA Major Depressive Disorder working group. Molecular Psychiatry, 21(6), 806–812. 10.1038/mp.2015.69 [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Shen X, Howard DM, Adams MJ, Hill WD, Clarke TK, McIntosh AM, … Whalley HC (2020). A phenome-wide association and Mendelian Randomisation study of polygenic risk for depression in UK Biobank. Nature Communications, 11(1), 1–16. 10.1038/s41467-020-16022-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Smith DJ, Nicholl BI, Cullen B, Martin D, Ul-Haq Z, Evans J, … Pell JP (2013). Prevalence and characteristics of probable major depression and bipolar disorder within UK Biobank: Cross-sectional study of 172,751 participants. PLoS ONE, 8(11), 1–8. 10.1371/journal.pone.0075362 [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Smith SM, Jenkinson M, Johansen-Berg H, Rueckert D, Nichols TE, Mackay CE, … Behrens TEJ (2006). Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data. NeuroImage, 31(4), 1487–1505. 10.1016/j.neuroimage.2006.02.024 [DOI] [PubMed] [Google Scholar]
  78. Spitzer RL, Kroenke K, Williams JBW, & the Patient Health Questionnaire Primary Care Study Group (1999). Validation and utility of a self-report version of PRIME-MD: The PHQ primary care study. JAMA, 282(18), 1737–1744. [DOI] [PubMed] [Google Scholar]
  79. Sudlow C, Gallacher J, Allen N, Beral V, Burton P, Danesh J, … Landray M (2015). UK Biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Medicine, 12(3), e1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Tingley D, Yamamoto T, Hirose K, Keele L, & Imai K (2014). Mediation: R package for causal mediation analysis. Journal of Statistical Software, 59(5), 1–38.26917999 [Google Scholar]
  81. Tozzi L, Garczarek L, Janowitz D, Stein DJ, Wittfeld K, Dobrowolny H, … Frodl T (2020). Interactive impact of childhood maltreatment, depression, and age on cortical brain structure: Mega-analytic findings from a large multi-site cohort. Psychological Medicine, 50(6), 1020–1031. 10.1017/S003329171900093X [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. van Harmelen AL, van Tol MJ, van der Wee NJA, Veltman DJ, Aleman A, Spinhoven P, … Elzinga BM (2010). Reduced medial prefrontal cortex volume in adults reporting childhood emotional maltreatment. Biological Psychiatry, 68(9), 832–838. 10.1016/j.biopsych.2010.06.011 [DOI] [PubMed] [Google Scholar]
  83. van Os J, & Jones PB (1999). Early risk factors and adult person–environment relationships in affective disorder. Psychological Medicine, 29(5), 1055–1067. [DOI] [PubMed] [Google Scholar]
  84. van Velzen LS, Kelly S, Isaev D, Aleman A, Aftanas LI, Bauer J, … Schmaal L (2020). White matter disturbances in major depressive disorder: A coordinated analysis across 20 international cohorts in the ENIGMA MDD working group. Molecular Psychiatry, 25(7), 1511–1525. 10.1038/s41380-019-0477-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  85. Vythilingam M, Heim C, Newport J, Miller AH, Anderson E, Bronen R, … Douglas Bremner J (2002). Childhood trauma associated with smaller hippocampal volume in women with major depression. American Journal of Psychiatry, 159(12), 2072–2080. 10.1176/appi.ajp.159.12.2072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  86. Wacker J, Dillon DG, & Pizzagalli DA (2009). The role of the nucleus accumbens and rostral anterior cingulate cortex in anhedonia: Integration of resting EEG, fMRI, and volumetric techniques. NeuroImage, 46(1), 327–337. 10.1016/j.neuroimage.2009.01.058 [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. Walsh ND, Dalgleish T, Lombardo MV, Dunn VJ, Van Harmelen AL, Ban M, & Goodyer IM (2014). General and specific effects of early-life psychosocial adversities on adolescent grey matter volume. NeuroImage: Clinical, 4, 308–318. 10.1016/j.nicl.2014.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  88. Wang X, Xie H, Chen T, Cotton AS, Salminen LE, Logue MW, … Liberzon I (2021). Cortical volume abnormalities in posttraumatic stress disorder: An ENIGMA-psychiatric genomics consortium PTSD workgroup mega-analysis. Molecular Psychiatry, 26(8), 4331–4343. 10.1038/s41380-020-00967-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  89. Whittle S, Lichter R, Dennison M, Vijayakumar N, Schwartz O, Byrne ML, … Alle NB (2014). Structural brain development and depression onset during adolescence: A prospective longitudinal study. American Journal of Psychiatry, 171(5), 564–571. 10.1176/appi.ajp.2013.13070920 [DOI] [PubMed] [Google Scholar]
  90. Widom CS, DuMont K, & Czaja SJ (2007). A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up. Archives of General Psychiatry, 64(1), 49–56. [DOI] [PubMed] [Google Scholar]
  91. Witt A, Öz Y, Sachser C, Brähler E, Glaesmer H, & Fegert JM (2022). Validation and standardization of the childhood trauma screener (CTS) in the general population. Child and Adolescent Psychiatry and Mental Health, 16(1), 1–12. 10.1186/s13034-022-00506-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  92. World Health Organization. (2020). Global health estimates 2020: Disease burden by cause, age, sex, by country, and by region, 2000–2019. In World Health Organization (WHO). https://www.who.int/data/gho/data/themes/mortality-and-global-health-estimates/global-health-estimates-leading-causes-of-dalys [Google Scholar]
  93. Zan G, Sun X, Wang Y, Liu R, Wang C, Du W, … Liu J (2021). Amygdala dynorphin/κ opioid receptor system modulates depressive-like behavior in mice following chronic social defeat stress. Acta Pharmacologica Sinica, 43(3), 577–587. 10.1038/s41401-021-00677-6. [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

supplement
NIHMS1988932-supplement-supplement.zip (439.9KB, zip)

Data Availability Statement

Data supporting the findings of this study are available upon reasonable request from the corresponding author. Raw data are available through the UK Biobank.

RESOURCES