[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Estimated Trajectories of Hippocampus Volume by Preschool Adverse Childhood Experiences (ACEs) and School-age Maternal Support
Estimated Trajectories of Hippocampus Volume by Preschool Adverse Childhood Experiences (ACEs) and School-age Maternal Support
Figure 2.
Estimated Trajectories of Amygdala Volume by Preschool Adverse Childhood Experiences (ACEs) and School-age Maternal Support
Estimated Trajectories of Amygdala Volume by Preschool Adverse Childhood Experiences (ACEs) and School-age Maternal Support
Figure 3.
Estimated Trajectories of Caudate Volume by Preschool Adverse Childhood Experiences (ACEs) and Preschool Maternal Support
Estimated Trajectories of Caudate Volume by Preschool Adverse Childhood Experiences (ACEs) and Preschool Maternal Support
Table 1.  
Characteristics of Participants Included in MLM Analyses
Characteristics of Participants Included in MLM Analyses
Table 2.  
Multilevel Models of Brain Volumes Across Scans 1 to 4 by ACEs and Maternal Support Covarying for Sex and Whole Brain Volume
Multilevel Models of Brain Volumes Across Scans 1 to 4 by ACEs and Maternal Support Covarying for Sex and Whole Brain Volume
1.
Bick  J, Nelson  CA.  Early adverse experiences and the developing brain.  Neuropsychopharmacology. 2016;41(1):177-196. doi:10.1038/npp.2015.252PubMedGoogle ScholarCrossref
2.
Hair  NL, Hanson  JL, Wolfe  BL, Pollak  SD.  Association of child poverty, brain development, and academic achievement.  JAMA Pediatr. 2015;169(9):822-829. doi:10.1001/jamapediatrics.2015.1475PubMedGoogle ScholarCrossref
3.
Luby  J, Belden  A, Botteron  K,  et al.  The effects of poverty on childhood brain development: the mediating effect of caregiving and stressful life events.  JAMA Pediatr. 2013;167(12):1135-1142. doi:10.1001/jamapediatrics.2013.3139PubMedGoogle ScholarCrossref
4.
Johnson  SB, Riis  JL, Noble  KG.  State of the art review: poverty and the developing brain.  Pediatrics. 2016;137(4):e20153075. doi:10.1542/peds.2015-3075PubMedGoogle Scholar
5.
Hostinar  CE, Johnson  AE, Gunnar  MR.  Parent support is less effective in buffering cortisol stress reactivity for adolescents compared to children.  Dev Sci. 2015;18(2):281-297. doi:10.1111/desc.12195PubMedGoogle ScholarCrossref
6.
Teicher  MH, Samson  JA, Anderson  CM, Ohashi  K.  The effects of childhood maltreatment on brain structure, function and connectivity.  Nat Rev Neurosci. 2016;17(10):652-666. doi:10.1038/nrn.2016.111PubMedGoogle ScholarCrossref
7.
Teicher  MH, Anderson  CM, Ohashi  K,  et al.  Differential effects of childhood neglect and abuse during sensitive exposure periods on male and female hippocampus.  Neuroimage. 2018;169:443-452. doi:10.1016/j.neuroimage.2017.12.055PubMedGoogle ScholarCrossref
8.
De Bellis  MD.  The psychobiology of neglect.  Child Maltreat. 2005;10(2):150-172. doi:10.1177/1077559505275116PubMedGoogle ScholarCrossref
9.
Whittle  S, Vijayakumar  N, Simmons  JG,  et al.  Role of positive parenting in the association between neighborhood social disadvantage and brain development across adolescence.  JAMA Psychiatry. 2017;74(8):824-832. doi:10.1001/jamapsychiatry.2017.1558PubMedGoogle ScholarCrossref
10.
Hanson  JL, Nacewicz  BM, Sutterer  MJ,  et al.  Behavioral problems after early life stress: contributions of the hippocampus and amygdala.  Biol Psychiatry. 2015;77(4):314-323. doi:10.1016/j.biopsych.2014.04.020PubMedGoogle ScholarCrossref
11.
Rao  U, Chen  LA, Bidesi  AS, Shad  MU, Thomas  MA, Hammen  CL.  Hippocampal changes associated with early-life adversity and vulnerability to depression.  Biol Psychiatry. 2010;67(4):357-364. doi:10.1016/j.biopsych.2009.10.017PubMedGoogle ScholarCrossref
12.
Edmiston  EE, Wang  F, Mazure  CM,  et al.  Corticostriatal-limbic gray matter morphology in adolescents with self-reported exposure to childhood maltreatment.  Arch Pediatr Adolesc Med. 2011;165(12):1069-1077. doi:10.1001/archpediatrics.2011.565PubMedGoogle ScholarCrossref
13.
Carrión  VG, Haas  BW, Garrett  A, Song  S, Reiss  AL.  Reduced hippocampal activity in youth with posttraumatic stress symptoms: an fMRI study.  J Pediatr Psychol. 2010;35(5):559-569. doi:10.1093/jpepsy/jsp112PubMedGoogle ScholarCrossref
14.
Shaked  D, Millman  ZB, Moody  DLB,  et al.  Sociodemographic disparities in corticolimbic structures.  PLoS One. 2019;14(5):e0216338. doi:10.1371/journal.pone.0216338PubMedGoogle Scholar
15.
McDermott  CL, Seidlitz  J, Nadig  A,  et al.  Longitudinally mapping childhood socioeconomic status associations with cortical and subcortical morphology.  J Neurosci. 2019;39(8):1365-1373. doi:10.1523/JNEUROSCI.1808-18.2018PubMedGoogle ScholarCrossref
16.
Hsieh  J-Y, Wang  H-W, Chang  S-J,  et al.  Mesenchymal stem cells from human umbilical cord express preferentially secreted factors related to neuroprotection, neurogenesis, and angiogenesis.  PLoS One. 2013;8(8):e72604. doi:10.1371/journal.pone.0072604PubMedGoogle Scholar
17.
Hart  H, Rubia  K.  Neuroimaging of child abuse: a critical review.  Front Hum Neurosci. 2012;6:52. doi:10.3389/fnhum.2012.00052PubMedGoogle ScholarCrossref
18.
Yang  J, Liu  H, Wei  D,  et al.  Regional gray matter volume mediates the relationship between family socioeconomic status and depression-related trait in a young healthy sample.  Cogn Affect Behav Neurosci. 2016;16(1):51-62. doi:10.3758/s13415-015-0371-6PubMedGoogle ScholarCrossref
19.
Kelly  PA, Viding  E, Wallace  GL,  et al.  Cortical thickness, surface area, and gyrification abnormalities in children exposed to maltreatment: neural markers of vulnerability?  Biol Psychiatry. 2013;74(11):845-852. doi:10.1016/j.biopsych.2013.06.020PubMedGoogle ScholarCrossref
20.
Noble  KG, Houston  SM, Kan  E, Sowell  ER.  Neural correlates of socioeconomic status in the developing human brain.  Dev Sci. 2012;15(4):516-527. doi:10.1111/j.1467-7687.2012.01147.xPubMedGoogle ScholarCrossref
21.
Luby  JL, Belden  A, Harms  MP, Tillman  R, Barch  DM.  Preschool is a sensitive period for the influence of maternal support on the trajectory of hippocampal development.  Proc Natl Acad Sci U S A. 2016;113(20):5742-5747. doi:10.1073/pnas.1601443113PubMedGoogle ScholarCrossref
22.
Luby  JL, Barch  DM, Belden  A,  et al.  Maternal support in early childhood predicts larger hippocampal volumes at school age.  Proc Natl Acad Sci U S A. 2012;109(8):2854-2859. doi:10.1073/pnas.1118003109PubMedGoogle ScholarCrossref
23.
Luby  JL.  Treatment of anxiety and depression in the preschool period.  J Am Acad Child Adolesc Psychiatry. 2013;52(4):346-358. doi:10.1016/j.jaac.2013.01.011PubMedGoogle ScholarCrossref
24.
Brody  GH, Yu  T, Nusslock  R,  et al.  The protective effects of supportive parenting on the relationship between adolescent poverty and resting-state functional brain connectivity during adulthood.  Psychol Sci. 2019;30(7):1040-1049. doi:10.1177/0956797619847989PubMedGoogle ScholarCrossref
25.
Bangalore  L.  Brain Development. New York, NY: Infobase Publishing; 2009.
26.
Egger  H, Ascher  B, Angold  A.  Preschool Age Psychiatric Assessment (PAPA): Version 1.1. Durham, NC: Center for Developmental Epidemiology, Department of Psychiatry and Behavioral Sciences, Duke University Medical Center; 1999.
27.
Angold  A, Prendergast  M, Cox  A, Harrington  R, Simonoff  E, Rutter  M.  The Child and Adolescent Psychiatric Assessment (CAPA).  Psychol Med. 1995;25(4):739-753. doi:10.1017/S003329170003498XPubMedGoogle ScholarCrossref
28.
Maxwell  ME.  Manual for the Family Interview for Genetic Studies (FIGS). Bethesda, MD: Clinical Neurogenetics Branch, Intramural Research Program, National Insititute of Mental Health; 1992.
29.
Luby  JL, Belden  AC, Jackson  JJ,  et al.  Early childhood depression and alterations in the trajectory of gray matter maturation in middle childhood and early adolescence.  JAMA Psychiatry. 2016;73(1):31-38. doi:10.1001/jamapsychiatry.2015.2356PubMedGoogle ScholarCrossref
30.
Peake  RWA, Law  T, Esposito  CL, Kellogg  MD.  Towards a random-access LC-MS/MS model for busulfan analysis.  Clin Chem Lab Med. 2017;55(7):e154-e157. doi:10.1515/cclm-2016-0727PubMedGoogle ScholarCrossref
31.
Belden  AC, Luby  JL.  The Emotion Reactivity Questionnaire. St Louis, MO: Washington University School of Medicine, Early Emotional Development Program; 2005.
32.
Belden  AC, Sullivan  JP, Luby  JL.  Depressed and healthy preschoolers’ internal representations of their mothers’ caregiving: associations with observed caregiving behaviors one year later.  Attach Hum Dev. 2007;9(3):239-254. doi:10.1080/14616730701455395PubMedGoogle ScholarCrossref
33.
Glasser  MF, Smith  SM, Marcus  DS,  et al.  The Human Connectome Project’s neuroimaging approach.  Nat Neurosci. 2016;19(9):1175-1187. doi:10.1038/nn.4361PubMedGoogle ScholarCrossref
34.
Reuter  M, Rosas  HD, Fischl  B.  Highly accurate inverse consistent registration: a robust approach.  Neuroimage. 2010;53(4):1181-1196. doi:10.1016/j.neuroimage.2010.07.020PubMedGoogle ScholarCrossref
35.
Destrieux  C, Fischl  B, Dale  A, Halgren  E.  Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature.  Neuroimage. 2010;53(1):1-15. doi:10.1016/j.neuroimage.2010.06.010PubMedGoogle ScholarCrossref
36.
Turkheimer  E, Haley  A, Waldron  M, D’Onofrio  B, Gottesman  II.  Socioeconomic status modifies heritability of IQ in young children.  Psychol Sci. 2003;14(6):623-628. doi:10.1046/j.0956-7976.2003.psci_1475.xPubMedGoogle ScholarCrossref
37.
Tottenham  N, Hare  TA, Quinn  BT,  et al.  Prolonged institutional rearing is associated with atypically large amygdala volume and difficulties in emotion regulation.  Dev Sci. 2010;13(1):46-61. doi:10.1111/j.1467-7687.2009.00852.xPubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 3,052
    Original Investigation
    Pediatrics
    September 18, 2019

    Association of Timing of Adverse Childhood Experiences and Caregiver Support With Regionally Specific Brain Development in Adolescents

    Author Affiliations
    • 1Department of Psychiatry (Child), School of Medicine, Washington University in St Louis, St Louis, Missouri
    • 2Department of Psychiatry, School of Medicine, Washington University in St Louis, St Louis, Missouri
    • 3Department of Psychological and Brain Sciences, Washington University in St Louis, St Louis, Missouri
    JAMA Netw Open. 2019;2(9):e1911426. doi:10.1001/jamanetworkopen.2019.11426
    Key Points español 中文 (chinese)

    Question  Is there developmental timing and regional specificity to the associations among adverse childhood experiences, caregiver support, and structural brain development in childhood?

    Findings  This cohort study of 211 children and their caregivers during 4 waves of neuroimaging and behavioral assessments from preschool to adolescence found an association between the interaction of preschool adverse childhood experiences and support during school age and the structural development of the amygdala and hippocampus, with volumes of these regions being the largest in children with low adverse childhood experiences and high caregiver support. In contrast, both preschool adverse childhood experiences and support were independently associated with the development of the caudate.

    Meaning  Findings may inform the timing and targets for prevention of poor outcomes related to exposure to adversity.

    Abstract

    Importance  Few data are available to inform the associations and timing of the associations between adversity, caregiver support, and brain outcomes. Consideration of timing has important public health implications to inform more precise prevention strategies.

    Objective  To evaluate the timing and regional specificity of the association between adverse childhood experiences (ACEs) and caregiver support to structural development of limbic and striatal brain regions in middle childhood and adolescence.

    Design, Setting, and Participants  This 15-year developmental, neuroimaging cohort study included 211 children and their caregivers screened from day care centers and preschools in the St Louis, Missouri, metropolitan area during the preschool period, with an additional 4 waves of neuroimaging at school age through adolescence from November 14, 2007, to August 29, 2017. The cohort was oversampled for preschoolers with elevated symptoms of depression using a brief screener. Data analysis was performed from March 19, 2019, to July 26, 2019.

    Main Outcomes and Measures  Volumes in adolescence and developmental trajectories of volumes of the amygdala, hippocampus, caudate, subgenual cingulate, and insula during 4 waves of scanning; ACEs and observed caregiver support at preschool and school age; and volumes of amygdala, hippocampus, insula, and subgenual cingulate during 4 waves of scanning.

    Results  A total of 211 children (107 [50.7%] male) completed at least 1 scan. At preschool (mean [SD] age, 5.5 [0.8] years), ACE data were available for 164 children (84 [51.2%] male) and maternal support data for 155 children; at school age (mean [SD], 8.3 [1.2] years), ACE data were available for 172 children and maternal support data for 146 children. Unique patterns of the association between ACEs and support were found, with an association between the interaction of preschool ACEs and school-age support and the development of the hippocampus (t = −2.27; P = .02) and amygdala (t = −2.12; P = .04). A buffering hypothesis was not confirmed because high caregiver support was more strongly associated with the development of these regions only in the context of low ACEs. In contrast, preschool ACEs (t = −2.30; P = .02) and support (t = 2.59; P = .01) had independent associations with the development of the caudate.

    Conclusions and Relevance  The findings suggest that there are unique regional associations of support and adversity with key brain structures important for emotional regulation. Results may inform the timing and potential targets of preventive action for the range of poor developmental outcomes.

    Introduction

    Increasing evidence of the associations between early-life experience and human brain development has emerged during the past 2 decades.1 Building on animal studies and retrospective data in humans, several prospective neuroimaging studies2-6 have further elucidated the negative associations of varying forms of adversity and the positive associations of nurturing caregiving in early childhood with the key brain regions critical for adaptive functioning. Related to these findings, sensitive periods during which environmental factors have a more powerful association with brain development have been found and are known to vary by region.7 These risk and protective factors and their associations with the function and structure of limbic and frontal regions have been directly associated with mental health outcomes.8-10 These studies underscore the association of potentially modifiable environmental factors with childhood brain development and represent a paradigm change in the understanding of the development of brain regions subserving key emotional and cognitive capacities previously believed to be largely genetically driven.

    Prospective studies10-14 have provided evidence that early-childhood adversity, often manifest as exposure to poverty and/or experiences of abuse and neglect, are associated with the structural development of limbic regions, including the hippocampus and amygdala. Although less well replicated, data also suggest similar associations in the basal ganglia,12,15,16 with retrospective and some prospective data also suggesting associations with the dorsal prefrontal cortex, anterior cingulate, and subgenual cingulate.9,14,15,17-20

    Although these negative associations between adversity and neural development are increasingly clear, there is also evidence that supportive caregiving has important positive associations with brain development and may in some situations serve as a protective or buffering factor that mitigates this risk. Data from the longitudinal developmental neuroimaging study21,22 examined here have demonstrated positive associations between early caregiver support and later childhood hippocampal volume, with preschool identified as a sensitive period. In the same study sample, negative associations between poverty and limbic outcomes were mediated by caregiving support and stressful life events.23 Whittle and colleagues9 reported positive associations between supportive caregiving and dorsal frontal lobe and amygdala development in adolescents exposed to poverty. Positive associations between caregiver support and resting state connectivity have also been reported in a longitudinal study24 of children raised in poverty. However, more information is needed on the interactive effects of adversity and caregiver support.

    The extant literature offers few data to inform the associations and timing of the associations between adversity and caregiver support and brain outcomes. Consideration of timing has important public health implications to inform more precise prevention strategies. Myelination and experience-dependent processes, such as pruning, vary by brain region, with different areas reaching peak synaptogenesis at different times, suggesting that there may be developmental specificity to associations of adversity and nurturance that varies by region.25 Prospective data that inform the developmental timing of when adversity and support have their most powerful associations with the development of key brain regions and if and how they might interact are needed. We tested the following hypotheses: (1) adversity and caregiver support will each have independent associations with brain development; (2) the timing (preschool, school age) of adversity and caregiver support will influence associations with brain development, with stronger associations in the preschool period; and (3) caregiver support will moderate (interact with) the association between early adversity and brain development, such that among children with higher caregiver support there will be less of a negative association between early adversity and brain development in cortical and subcortical regions known to be related to early adversity.

    Methods
    Participants

    Data from the Preschool Depression Study, a 15-year (June 1, 2003, onward), neuroimaging cohort study that began when participants were preschoolers, were used for this analysis (eFigure in the Supplement gives a study overview). This study included 211 children and their caregivers screened from day care centers and preschools in the St Louis, Missouri, metropolitan area during the preschool period, with an additional 4 waves of neuroimaging at school age through adolescence from November 14, 2007, to August 29, 2017. During 1 preschool and 2 school-age waves, observational measures of parenting during mildly stressful laboratory tasks were obtained. During each study wave (ages during waves are given in the eFigure in the Supplement), measures of experiences of adversity since the last wave were collected. Neuroimaging was conducted with child participants during 4 waves that spanned middle childhood to young adulthood completed to date. Data analysis was performed from March 19, 2019, to July 26, 2019. Written parental consent and child assent or consent (depending on age) to a protocol approved by the institutional review board at Washington University in St Louis were obtained before study participation. All data were deidentified. This report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    Measures
    Adverse Childhood Experiences

    We created adverse childhood experience (ACE) scores using variables from the life events section of the Preschool Age Psychiatric Assessment (PAPA)26 or Child and Adolescent Psychiatric Assessment (CAPA),27 parental psychopathologic variables from the Family Interview for Genetic Studies,28 and exposure to poverty based on family income to needs ratio at each assessment wave. The ACE scores were similar to those previously reported.29 The PAPA was conducted at each assessment that occurred between the ages of 3 and 7 years based on parent report, and the CAPA was used at 8 years of age based on parent report; both parent and child reports were obtained for 9 years or more. The ACE variable was adapted from the original Felitti definition but also included exposure to poverty because of its established role in brain development. Poverty was defined as an income to needs ratio less than 1, in accordance with federal guidelines. The variables included in the ACE sum scores and methods used to create the sum scores are given in eTable 1 in the Supplement.

    Preschool and School-age Maternal Support

    Maternal support in the present study was defined as the degree to which caregivers (92% mothers) approached and interacted with their children with positive regard and nurturance as well as their efforts to be emotionally and developmentally supportive. Preschool maternal support was measured during the second annual study wave when children were aged 3 to 6 years; their caregivers engaged in a mildly stressful laboratory task. This task, known as the waiting task, which was designed to evoke mild dyadic stress, required the child to wait for 8 minutes before opening a brightly wrapped gift within reach while the parent completed questionnaires. Staff trained on reliability coded the interaction for supportive caregiving strategies used to help regulate the child’s frustration in the face of thwarted ability to open the gift. Prior findings21,22,30-32 suggest that caregiving support coded during this task has good psychometric properties.

    A different developmentally appropriate mildly stressful task, the puzzle task, was completed by children and their caregivers during the third annual study wave (when children were 6-7 years of age) and again during their fifth assessment wave 4 to 5 years later (8-12 years of age) as a measure of school-age maternal support. The child was asked to attempt to put together a puzzle without being able to see the pieces while relying on the parent (who could see the puzzle board and placement of the child’s hands) for instructions. Parents gave the child instructions on where to place each piece, and the dyad earned a prize if they finished the puzzle in 5 minutes. Both of these measures of maternal support have been previously published.21

    Magnetic Resonance Imaging Acquisition

    Structural images were collected as part of longer scan sessions that also included acquisition of task-based and functional connectivity data. For the first 3 waves, imaging data were collected using a 3-T Siemens TIM TRIO. T1-weighted structural images were acquired in the sagittal plane using 2 magnetization-prepared rapid gradient-echo (MPRAGE) 3-dimensional sequences (repetition time, 2400 milliseconds; echo time, 3.16 milliseconds; flip angle, 8°; 176 sections; field of view, 256 mm; 1.0-mm3 voxels). For scan 4, we used a 3-T Siemens PRISMA with a 32-channel head coil using Human Connectome Project–style acquisitions,33 using a sagittally acquired MPRAGE 3-dimensional sequence (repetition time, 2300 milliseconds; echo time, 3.16 milliseconds; flip angle, 8°; 160 sections; field of view, 256 mm; 1.0-mm3 voxels). Correlations of whole brain volumes across scans were high (0.954 to 0.981 for total gray-matter volume, 0.984 to 0.991 for supratentorial volume, and 0.917 to 0.999 for intracranial volume), suggesting that scanner changes should not influence our ability to examine associations of ACEs and caregiver support to brain development.

    Brain volumes were generated using the FreeSurfer software, version 5.3 processing stream described by Luby et al29 (further details are given in eTable 1 in the Supplement). Data on the volume of the left and right hippocampus, amygdala, caudate, nucleus accumbens, and putamen in the participant’s native space were obtained using the FreeSurfer aseg.stats report.33,34 Volumes of the anterior insula, subgenual cingulate, dorsolateral prefrontal cortex, dorsal anterior cingulate, and rostral cingulate were taken from Destrieux Atlas (eTable 1 in the Supplement).35 We did not have a priori hypotheses about left or right hemispheres, and thus, we averaged the 2. These regions of interest were chosen based on the evidence in the extant literature reviewed in the Introduction section. Whole brain volume was used as a covariate to assess specificity.

    Statistical Analysis

    We used a hierarchical analytical approach to identify brain regions with associations with ACEs and caregiver support. We started with 4 sets of general linear models (GLMs) evaluating the volumes of the regions of interest at scan 4, when participants were adolescents or young adults. Each set included covariates of scan 4 (age, sex, and whole brain volume) and 1 of the 4 following variables of interest: ACEs and maternal support at preschool and school age. The results were false discovery rate (FDR) corrected within each set of GLMs, with an FDR with a 2-sided P < .05 indicating significance. In cases in which an independent variable was associated with brain volume at one age but not another, an additional GLM was conducted that included both age periods to assess the specificity of the association with the one age period.

    Any variables significant after FDR correction in the GLMs described above were then included in multilevel models (MLMs) to examine whether ACEs and/or maternal support at preschool and/or school age were significantly associated with the trajectory of brain volumes across scans 1 to 4. The MLMs used random intercept and slope components and unstructured covariance. Time was defined as age at scan, which was centered at 12 years of age, and the covariates were sex and whole brain volume. When the interaction between an independent variable and age was not significant (P ≥ .05), the variable was removed from the model to test for main effects. To test hypotheses about interactions between ACEs and maternal support, 3-way interactions among ACEs, maternal support, and age were tested and included in the final models when significant.

    Results

    A total of 211 children completed at least 1 scan. At preschool (mean [SD] age, 5.5 [0.8] years), data on ACEs were available for 164 children (84 [51.2%] male) and maternal support data for 155 children; at school age (mean [SD], 8.3 [1.2] years), data on ACEs were available for 172 children and maternal support data for 146 children. Correlations among these 4 variables are given in eTable 2 in the Supplement. Table 1 provides demographic and diagnostic characteristics of the samples included in each of the MLMs. Participants were not included in analyses if they had unusable scan data at all scan waves or were missing data on preschool ACEs because this was an independent variable in every MLM. Missing data were not imputed. Comparisons between participants included (n = 164) and not included (n = 47) in analyses are presented in eTable 3 in the Supplement.

    Preschool or School-age ACEs and Adolescent and Young Adult Brain Volume

    As shown in eTable 4 in the Supplement, after FDR, higher preschool ACE scores were significantly associated with decreased insula, hippocampus, amygdala, subgenual cingulate, and caudate volume at scan 4. Greater school-age ACEs were significantly associated only with decreased insula volume. As detailed in eTable 5 in the Supplement, when both preschool and school-age ACEs were included in these models, preschool ACEs continued to have the only significant association with the brain volume for amygdala (mean [SE] estimate, −0.067 [0.024]; t = −2.77; P = .007), subgenual cingulate (mean [SE] estimate, −0.121 [0.060]; t = −2.03; P = .04), and caudate (mean [SE] estimate, −0.165 [0.065]; t = −2.55; P = .01).

    Preschool or School-age Support and Adolescent and Young Adult Brain Volume

    After FDR correction, greater preschool maternal support was significantly associated with increased hippocampus (mean [SE] estimate, 0.017 [0.005]; t = 3.38; FDR P = .005) and caudate (mean [SE] estimate, 0.019 [0.005]; t = 3.57; FDR P = .005) volume at scan 4, and greater school-age maternal support was significantly associated with increased insula (mean [SE] estimate, 0.011 [0.004]; t = 2.72; FDR P = .04), hippocampus (mean [SE] estimate, 0.015 [0.005]; t = 3.03; FDR P = .03), and amygdala (mean [SE] estimate, 0.005 [0.002]; t = 2.58; FDR P = .04) volume at scan 4 (eTable 6 in the Supplement). As detailed in eTable 7 in the Supplement, when both preschool and school-age maternal support was included, maternal support during preschool continued to have the only significant association with scan 4 caudate volume (mean [SE] estimate, 0.019 [0.006]; t = 2.94; P = .004), whereas only school-age maternal support was significantly associated with scan 4 insula (mean [SE] estimate, 0.011 [0.005]; t = 2.36; P = .02) and amygdala (mean [SE] estimate 0.005 [0.002]; t = 1.97; P = .05) volume.

    ACEs and Maternal Support and Trajectories of Brain Volume
    Insula Volume Trajectories

    Greater preschool (mean [SE] estimate, −0.123 [0.033]; t = −3.78; FDR P = .002) and school-age (mean [SE] estimate, −0.106 [0.032]; t = −3.35; FDR P = .01) ACE scores and less school-age maternal support (mean [SE] estimate, 0.011 [0.004]; t = 2.72; FDR P = .04) were each significantly associated with reduced insula volume at scan 4. As indicated in Table 2, when all 3 of these independent variables were included in an MLM of insula volume across scans 1 to 4, only the main association of preschool ACEs was significant (mean [SE] estimate, −0.0626 [0.0296]; t = −2.11; P = .04), with higher preschool ACEs associated with smaller insula volume (smaller intercept). The interactions between variables and age (slope) were not significant and not included in the final model.

    Hippocampus Volume Trajectories

    Greater preschool ACEs (mean [SE] estimate, −0.102 [0.040]; t = −2.53; FDR P = .03) and less maternal support at preschool (mean [SE] estimate, 0.017 [0.005]; t = 3.38; FDR P = .005) and school age (mean [SE] estimate, 0.015 [0.005]; t = 3.03; FDR P = .03) were each significantly associated with reduced scan 4 hippocampal volume. As indicated in Table 2, when all these variables were in the model, the 3-way interaction among preschool ACEs, school-age maternal support, and age was significant (mean [SE] estimate, −0.0011 [0.0005]; t = −2.27; P = .02), indicating that preschool ACEs and school-age maternal support in interaction were associated with hippocampal trajectories over time.

    To parse the source of this interaction and test our hypothesis about buffering, we centered the maternal support variable at −1 SD, mean, and +1 SD and examined the association between preschool ACEs and age. As shown in Figure 1, the interaction between preschool ACEs and age (eg, slope) was not significantly associated with hippocampus volume at low levels of maternal support (mean [SE] estimate, −0.0038 [0.0052]; t = −0.74; P = .46), but the association was significant at moderate (mean [SE] estimate, −0.0133 [0.0045]; t = −2.92; P = .004) and high (mean [SE] estimate, −0.0227 [0.0070]; t = −3.24; P = .002) levels of support. The largest estimated hippocampal volumes were at the combination of low values of preschool ACEs and high values of school-age maternal support.

    Amygdala Volume Trajectories

    Preschool ACEs (mean [SE] estimate, −0.043 [0.017]; t = −2.58; FDR P = .003) and school-age maternal support (mean [SE] estimate, 0.005 [0.002]; t = 2.58; FDR P = .04) were each significantly associated with scan 4 amygdala volume. As indicated in Table 2, the 3-way interaction among preschool ACEs, school-age maternal support, and age was significant in the MLM (mean [SE] estimate, −0.0006 [0.0003]; t = −2.12; P = .04). To parse the source of this interaction and test our hypothesis about buffering, we again centered the maternal support variable at −1 SD, mean, and +1 SD and examined the association between preschool ACEs and age. As shown in Figure 2, the association between preschool ACEs and age (eg, slope) was not significantly associated with amygdala volume at low levels of maternal support (mean [SE] estimate, −0.0020 [0.0030]; t = −0.67; P = .50), but the association was significant at moderate (mean [SE] estimate, −0.0072 [0.0026]; t = −2.78; P = .006) and high (mean [SE] estimate, −0.0125 [0.0040]; t = −3.08; P = .002) levels of support. The largest estimated amygdala volumes were at the combination of low values of preschool ACEs and high values of school-age maternal support.

    Subgenual Cingulate Volume Trajectories

    Only preschool ACEs had a significant association with scan 4 subgenual cingulate volume (mean [SE] estimate, −0.106 [0.040]; t = −2.66; FDR P = .03). However, in the MLM across scans 1 to 4, preschool ACEs were not associated with the intercept or slope of subgenual cingulate over time.

    Caudate Volume Trajectories

    Both preschool ACEs (mean [SE] estimate, −0.101 [0.043]; t = −2.32; FDR P = .04) and maternal support (mean [SE] estimate, 0.019 [0.005]; t = 3.57; FDR P = .005) were significantly associated with scan 4 caudate volume. As indicated in Table 2, when both these independent variables were included in an MLM of caudate volume at scans 1 to 4, lower preschool ACEs (mean [SE] estimate, −0.0776 [0.0337]; t = −2.30; P = .02) and greater preschool maternal support (mean [SE] estimate, 0.0105 [0.0040]; t = 2.59; P = .01) were independently significantly associated with greater caudate volume (Table 2 and Figure 3). The interaction between these independent variables and age (slope) was not significant.

    Discussion

    Findings from this longitudinal, developmental, neuroimaging cohort study replicate and extend findings in the extant literature regarding the associations of ACEs and maternal support with limbic and striatal brain regions previously reported to be associated with early psychosocial variables. Greater preschool ACEs were associated with lower volumes of hippocampus, amygdala, insula, subgenual cingulate, and caudate in late adolescence and early adulthood (scan 4), whereas school-age ACEs were only associated with the insula volume. With regard to maternal support, greater preschool support was associated with increased hippocampus and caudate volumes, whereas school-age support had positive associations with insula, hippocampus, and amygdala volumes. These findings provide new evidence, to our knowledge, of developmental specificity in the associations of psychosocial experience to brain development regionally when examined cross-sectionally.

    Regionally varying patterns also emerged when trajectories of brain volumes during 4 scan waves were examined. The buffering hypothesis was not supported for any brain region. Instead, for hippocampus and amygdala, high support was associated with patterns of growth typically associated with better outcomes (eg, greater volumes) only in the context of low ACEs. This finding contrasts with previously reported buffering effects in the hippocampus3 but might be consistent with prior data suggesting that children realize their genetic potential at greater rates in the context of environmental advantage.36 A similar pattern emerged for amygdala volumes. These findings underscore the long-term deleterious associations between exposure to ACEs during the preschool period and the developmental trajectory of the hippocampus and amygdala but suggest that the positive association of maternal support with brain development is most apparent in the context of low preschool ACEs.

    These findings differ somewhat from prior results21 in that both the hippocampus and amygdala volumes were negatively associated with ACEs experienced during the preschool period but positively associated with support at school age when all variables were in the model simultaneously, rather than preschool support, as in our prior hippocampal findings that did not consider ACEs. However, when considered independently, both preschool maternal support and school-age maternal support were associated with hippocampal volume at the last scan wave. This pattern suggests that preschool ACEs are associated with the variance in hippocampal volumes associated with preschool maternal support, whereas at school age, maternal support is associated with additional variance whether considered independently or in combination with other factors. Our findings with regard to amygdala volume also contrast with a previous report37 of institutionalized young children, which found associations with larger amygdala volumes. The type of neglect and deprivation experienced by institutionalized children may be qualitatively different from ACEs or the type of maternal support that we measured and therefore may have different associations with brain development. Taken together, these patterns suggest that adversity early in life may be associated with hippocampus and amygdala development but that caregiver support during childhood may be most positively associated with brain development in these regions in the context of low adversity. Findings therefore underscore the critical importance of determining whether preventing exposure to ACEs, particularly during the preschool period, may benefit childhood brain development.

    In contrast, for the caudate, a key region known in reward processing and related risk for mood disorders, both lower preschool ACEs and greater preschool maternal support were independently associated with greater caudate volume, with no interaction between the 2. These findings suggest that for the caudate growth trajectory, preschool maternal support has positive associations with development even in the context of exposure to ACEs and that greater ACEs have negative associations with caudate development regardless of the level of maternal support.

    For the insula, both preschool and school-age ACEs were associated with scan 4 volumes, but the only significant association with insula volume across all scans was that of preschool ACEs. Furthermore, despite the association of preschool ACEs with subgenual cingulate volume cross-sectionally at scan 4, there was no association with the trajectory of subgenual cingulate volume during the 4 scan waves. Thus, the associations of ACEs and support at different ages varied considerably across the range of brain regions previously found to be associated with these psychosocial factors.

    Our findings suggest regional specificity to the timing of when ACEs and support have significant associations with limbic and striatal brain regions. Both low ACEs and high support were associated with healthy development of the hippocampus and amygdala, whereas each was independently associated with positive development of the caudate. The timing of the positive associations with support also varied by region, with school-age support having more of an association with hippocampus and amygdala volumes in the context of low ACEs, whereas preschool support and low ACEs were independently associated with caudate developmental trajectories. Furthermore, a different pattern was evident for the insula and subgenual cingulate, with preschool ACEs associated with the former and neither environmental variable associated with the trajectory of the latter.

    Limitations

    This study has limitations. The study is limited by the fact that children were ascertained at preschool age and underwent scanning at school age. We therefore were not able to consider brain developmental outcomes earlier in life, when even stronger associations may be present. In addition, several other aspects of psychosocial experience were not measured, such as the home environment, community factors, and the presence of other caregivers, which could meaningfully be associated with brain development. In addition, our support variable focused primarily on maternal support, and more work is needed to understand the associations with paternal support.

    Conclusions

    Our findings suggest that there is a nuanced and regionally specific pattern of the psychosocial factors associated with healthy development of limbic and striatal brain regions key to adaptive emotion processing. Future studies that account for other key environmental factors, including environmental toxins and other forms of support, appear to be needed to further elucidate important associations with these trajectories. Such data could inform the development of more precise prevention approaches for a range of poor developmental outcomes and for the enhancement of brain development in children.

    Back to top
    Article Information

    Accepted for Publication: July 27, 2019.

    Published: September 18, 2019. doi:10.1001/jamanetworkopen.2019.11426

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Luby JL et al. JAMA Network Open.

    Corresponding Author: Joan L. Luby, MD, Department of Psychiatry (Child), School of Medicine, Washington University in St Louis, 660 S Euclid, PO Box 8511, St Louis, MO 63110 (lubyj@wustl.edu).

    Author Contributions: Drs Luby and Barch had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Luby, Barch.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: All authors.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Tillman, Barch.

    Obtained funding: Luby, Barch.

    Administrative, technical, or material support: Luby.

    Supervision: Luby, Barch.

    Conflict of Interest Disclosures: Dr Barch reported receiving grants from the National Institute of Mental Health during the conduct of the study. No other disclosures were reported.

    Funding/Support: This study was supported by grant R01 MH090786 from the National Institute of Mental Health (Drs Luby and Barch).

    Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.

    Additional Contributions: We appreciate the participation of the study population, the research assistants who collected the data, and the senior statisticians who contributed to this effort. The study participants and research assistants were compensated.

    References
    1.
    Bick  J, Nelson  CA.  Early adverse experiences and the developing brain.  Neuropsychopharmacology. 2016;41(1):177-196. doi:10.1038/npp.2015.252PubMedGoogle ScholarCrossref
    2.
    Hair  NL, Hanson  JL, Wolfe  BL, Pollak  SD.  Association of child poverty, brain development, and academic achievement.  JAMA Pediatr. 2015;169(9):822-829. doi:10.1001/jamapediatrics.2015.1475PubMedGoogle ScholarCrossref
    3.
    Luby  J, Belden  A, Botteron  K,  et al.  The effects of poverty on childhood brain development: the mediating effect of caregiving and stressful life events.  JAMA Pediatr. 2013;167(12):1135-1142. doi:10.1001/jamapediatrics.2013.3139PubMedGoogle ScholarCrossref
    4.
    Johnson  SB, Riis  JL, Noble  KG.  State of the art review: poverty and the developing brain.  Pediatrics. 2016;137(4):e20153075. doi:10.1542/peds.2015-3075PubMedGoogle Scholar
    5.
    Hostinar  CE, Johnson  AE, Gunnar  MR.  Parent support is less effective in buffering cortisol stress reactivity for adolescents compared to children.  Dev Sci. 2015;18(2):281-297. doi:10.1111/desc.12195PubMedGoogle ScholarCrossref
    6.
    Teicher  MH, Samson  JA, Anderson  CM, Ohashi  K.  The effects of childhood maltreatment on brain structure, function and connectivity.  Nat Rev Neurosci. 2016;17(10):652-666. doi:10.1038/nrn.2016.111PubMedGoogle ScholarCrossref
    7.
    Teicher  MH, Anderson  CM, Ohashi  K,  et al.  Differential effects of childhood neglect and abuse during sensitive exposure periods on male and female hippocampus.  Neuroimage. 2018;169:443-452. doi:10.1016/j.neuroimage.2017.12.055PubMedGoogle ScholarCrossref
    8.
    De Bellis  MD.  The psychobiology of neglect.  Child Maltreat. 2005;10(2):150-172. doi:10.1177/1077559505275116PubMedGoogle ScholarCrossref
    9.
    Whittle  S, Vijayakumar  N, Simmons  JG,  et al.  Role of positive parenting in the association between neighborhood social disadvantage and brain development across adolescence.  JAMA Psychiatry. 2017;74(8):824-832. doi:10.1001/jamapsychiatry.2017.1558PubMedGoogle ScholarCrossref
    10.
    Hanson  JL, Nacewicz  BM, Sutterer  MJ,  et al.  Behavioral problems after early life stress: contributions of the hippocampus and amygdala.  Biol Psychiatry. 2015;77(4):314-323. doi:10.1016/j.biopsych.2014.04.020PubMedGoogle ScholarCrossref
    11.
    Rao  U, Chen  LA, Bidesi  AS, Shad  MU, Thomas  MA, Hammen  CL.  Hippocampal changes associated with early-life adversity and vulnerability to depression.  Biol Psychiatry. 2010;67(4):357-364. doi:10.1016/j.biopsych.2009.10.017PubMedGoogle ScholarCrossref
    12.
    Edmiston  EE, Wang  F, Mazure  CM,  et al.  Corticostriatal-limbic gray matter morphology in adolescents with self-reported exposure to childhood maltreatment.  Arch Pediatr Adolesc Med. 2011;165(12):1069-1077. doi:10.1001/archpediatrics.2011.565PubMedGoogle ScholarCrossref
    13.
    Carrión  VG, Haas  BW, Garrett  A, Song  S, Reiss  AL.  Reduced hippocampal activity in youth with posttraumatic stress symptoms: an fMRI study.  J Pediatr Psychol. 2010;35(5):559-569. doi:10.1093/jpepsy/jsp112PubMedGoogle ScholarCrossref
    14.
    Shaked  D, Millman  ZB, Moody  DLB,  et al.  Sociodemographic disparities in corticolimbic structures.  PLoS One. 2019;14(5):e0216338. doi:10.1371/journal.pone.0216338PubMedGoogle Scholar
    15.
    McDermott  CL, Seidlitz  J, Nadig  A,  et al.  Longitudinally mapping childhood socioeconomic status associations with cortical and subcortical morphology.  J Neurosci. 2019;39(8):1365-1373. doi:10.1523/JNEUROSCI.1808-18.2018PubMedGoogle ScholarCrossref
    16.
    Hsieh  J-Y, Wang  H-W, Chang  S-J,  et al.  Mesenchymal stem cells from human umbilical cord express preferentially secreted factors related to neuroprotection, neurogenesis, and angiogenesis.  PLoS One. 2013;8(8):e72604. doi:10.1371/journal.pone.0072604PubMedGoogle Scholar
    17.
    Hart  H, Rubia  K.  Neuroimaging of child abuse: a critical review.  Front Hum Neurosci. 2012;6:52. doi:10.3389/fnhum.2012.00052PubMedGoogle ScholarCrossref
    18.
    Yang  J, Liu  H, Wei  D,  et al.  Regional gray matter volume mediates the relationship between family socioeconomic status and depression-related trait in a young healthy sample.  Cogn Affect Behav Neurosci. 2016;16(1):51-62. doi:10.3758/s13415-015-0371-6PubMedGoogle ScholarCrossref
    19.
    Kelly  PA, Viding  E, Wallace  GL,  et al.  Cortical thickness, surface area, and gyrification abnormalities in children exposed to maltreatment: neural markers of vulnerability?  Biol Psychiatry. 2013;74(11):845-852. doi:10.1016/j.biopsych.2013.06.020PubMedGoogle ScholarCrossref
    20.
    Noble  KG, Houston  SM, Kan  E, Sowell  ER.  Neural correlates of socioeconomic status in the developing human brain.  Dev Sci. 2012;15(4):516-527. doi:10.1111/j.1467-7687.2012.01147.xPubMedGoogle ScholarCrossref
    21.
    Luby  JL, Belden  A, Harms  MP, Tillman  R, Barch  DM.  Preschool is a sensitive period for the influence of maternal support on the trajectory of hippocampal development.  Proc Natl Acad Sci U S A. 2016;113(20):5742-5747. doi:10.1073/pnas.1601443113PubMedGoogle ScholarCrossref
    22.
    Luby  JL, Barch  DM, Belden  A,  et al.  Maternal support in early childhood predicts larger hippocampal volumes at school age.  Proc Natl Acad Sci U S A. 2012;109(8):2854-2859. doi:10.1073/pnas.1118003109PubMedGoogle ScholarCrossref
    23.
    Luby  JL.  Treatment of anxiety and depression in the preschool period.  J Am Acad Child Adolesc Psychiatry. 2013;52(4):346-358. doi:10.1016/j.jaac.2013.01.011PubMedGoogle ScholarCrossref
    24.
    Brody  GH, Yu  T, Nusslock  R,  et al.  The protective effects of supportive parenting on the relationship between adolescent poverty and resting-state functional brain connectivity during adulthood.  Psychol Sci. 2019;30(7):1040-1049. doi:10.1177/0956797619847989PubMedGoogle ScholarCrossref
    25.
    Bangalore  L.  Brain Development. New York, NY: Infobase Publishing; 2009.
    26.
    Egger  H, Ascher  B, Angold  A.  Preschool Age Psychiatric Assessment (PAPA): Version 1.1. Durham, NC: Center for Developmental Epidemiology, Department of Psychiatry and Behavioral Sciences, Duke University Medical Center; 1999.
    27.
    Angold  A, Prendergast  M, Cox  A, Harrington  R, Simonoff  E, Rutter  M.  The Child and Adolescent Psychiatric Assessment (CAPA).  Psychol Med. 1995;25(4):739-753. doi:10.1017/S003329170003498XPubMedGoogle ScholarCrossref
    28.
    Maxwell  ME.  Manual for the Family Interview for Genetic Studies (FIGS). Bethesda, MD: Clinical Neurogenetics Branch, Intramural Research Program, National Insititute of Mental Health; 1992.
    29.
    Luby  JL, Belden  AC, Jackson  JJ,  et al.  Early childhood depression and alterations in the trajectory of gray matter maturation in middle childhood and early adolescence.  JAMA Psychiatry. 2016;73(1):31-38. doi:10.1001/jamapsychiatry.2015.2356PubMedGoogle ScholarCrossref
    30.
    Peake  RWA, Law  T, Esposito  CL, Kellogg  MD.  Towards a random-access LC-MS/MS model for busulfan analysis.  Clin Chem Lab Med. 2017;55(7):e154-e157. doi:10.1515/cclm-2016-0727PubMedGoogle ScholarCrossref
    31.
    Belden  AC, Luby  JL.  The Emotion Reactivity Questionnaire. St Louis, MO: Washington University School of Medicine, Early Emotional Development Program; 2005.
    32.
    Belden  AC, Sullivan  JP, Luby  JL.  Depressed and healthy preschoolers’ internal representations of their mothers’ caregiving: associations with observed caregiving behaviors one year later.  Attach Hum Dev. 2007;9(3):239-254. doi:10.1080/14616730701455395PubMedGoogle ScholarCrossref
    33.
    Glasser  MF, Smith  SM, Marcus  DS,  et al.  The Human Connectome Project’s neuroimaging approach.  Nat Neurosci. 2016;19(9):1175-1187. doi:10.1038/nn.4361PubMedGoogle ScholarCrossref
    34.
    Reuter  M, Rosas  HD, Fischl  B.  Highly accurate inverse consistent registration: a robust approach.  Neuroimage. 2010;53(4):1181-1196. doi:10.1016/j.neuroimage.2010.07.020PubMedGoogle ScholarCrossref
    35.
    Destrieux  C, Fischl  B, Dale  A, Halgren  E.  Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature.  Neuroimage. 2010;53(1):1-15. doi:10.1016/j.neuroimage.2010.06.010PubMedGoogle ScholarCrossref
    36.
    Turkheimer  E, Haley  A, Waldron  M, D’Onofrio  B, Gottesman  II.  Socioeconomic status modifies heritability of IQ in young children.  Psychol Sci. 2003;14(6):623-628. doi:10.1046/j.0956-7976.2003.psci_1475.xPubMedGoogle ScholarCrossref
    37.
    Tottenham  N, Hare  TA, Quinn  BT,  et al.  Prolonged institutional rearing is associated with atypically large amygdala volume and difficulties in emotion regulation.  Dev Sci. 2010;13(1):46-61. doi:10.1111/j.1467-7687.2009.00852.xPubMedGoogle ScholarCrossref
    ×