association of childhood violence exposure with adolescent

Original Investigation | Pediatrics

Association of Childhood Violence Exposure With Adolescent NeuralNetwork DensityLeigh G. Goetschius, MS; Tyler C. Hein, MD; Sara S. McLanahan, PhD; Jeanne Brooks-Gunn, PhD; Vonnie C. McLoyd, PhD; Hailey L. Dotterer, MS; Nestor Lopez-Duran, PhD;Colter Mitchell, PhD; Luke W. Hyde, PhD; Christopher S. Monk, PhD; Adriene M. Beltz, PhD

Abstract

IMPORTANCE Adverse childhood experiences are a public health issue with negative sequelae thatpersist throughout life. Current theories suggest that adverse childhood experiences reflectunderlying dimensions (eg, violence exposure and social deprivation) with distinct neuralmechanisms; however, research findings have been inconsistent, likely owing to variability in how theenvironment interacts with the brain.

OBJECTIVE To examine whether dimensional exposure to childhood adversity is associated withperson-specific patterns in adolescent resting-state functional connectivity (rsFC), defined assynchronized activity across brain regions when not engaged in a task.

DESIGN, SETTING, AND PARTICIPANTS A sparse network approach in a large sample withsubstantial representation of understudied, underserved African American youth was used toconduct an observational, population-based longitudinal cohort study. A total of 183 adolescentsaged 15 to 17 years from Detroit, Michigan; Toledo, Ohio; and Chicago, Illinois, who participated in theFragile Families and Child Wellbeing Study were eligible for inclusion. Environmental data from birthto adolescence were collected via telephone and in-person interviews, and neuroimaging datacollected at a university lab. The study was conducted from February 1, 1998, to April 26, 2017, anddata analysis was performed from January 3, 2019, to May 22, 2020.

EXPOSURES Composite variables representing violence exposure and social deprivation createdfrom primary caregiver reports on children at ages 3, 5, and 9 years.

MAIN OUTCOMES AND MEASURES Resting-state functional connectivity person-specific networkmetrics (data-driven subgroup membership, density, and node degree) focused on connectivityamong a priori regions of interest in 2 resting-state networks (salience network and default mode)assessed with functional magnetic resonance imaging.

RESULTS Of the 183 eligible adolescents, 175 individuals (98 girls [56%]) were included in theanalysis; mean (SD) age was 15.88 (0.53) years and 127 participants (73%) were African American.Adolescents with high violence exposure were 3.06 times more likely (95% CI, 1.17-8.92) to be in asubgroup characterized by high heterogeneity (few shared connections) and low network density(sparsity). Childhood violence exposure, but not social deprivation, was associated with reduced rsFCdensity (β = −0.25; 95% CI, −0.41 to −0.05; P = .005), with fewer salience network connections(β = −0.26; 95% CI, −0.43 to −0.08; P = .005) and salience network-default mode connections(β = −0.20; 95% CI, −0.38 to −0.03; P = .02). Violence exposure was associated with node degree ofright anterior insula (β = −0.29; 95% CI, −0.47 to −0.12; P = .001) and left inferior parietal lobule(β = −0.26; 95% CI, −0.44 to −0.09; P = .003).

(continued)

Key PointsQuestion Are violence exposure and

social deprivation associated with

person-specific patterns

(heterogeneity) of adolescent resting-

state functional connectivity?

Findings In this cohort study of 175

adolescents, childhood violence

exposure, but not social deprivation,

was associated with reduced adolescent

resting-state density of the salience and

default mode networks. A data-driven

algorithm, blinded to childhood

adversity, identified youth with

heightened violence exposure based on

resting-state connectivity patterns.

Meaning Childhood violence exposure

appears to be associated with

adolescent functional connectivity

heterogeneity, which may reflect

person-specific neural plasticity and

should be considered in neuroscience-

based interventions.

+ Video

+ Supplemental content

Author affiliations and article information arelisted at the end of this article.

Open Access. This is an open access article distributed under the terms of the CC-BY License.

JAMA Network Open. 2020;3(9):e2017850. doi:10.1001/jamanetworkopen.2020.17850 (Reprinted) September 23, 2020 1/13

Downloaded From: https://jamanetwork.com/ on 05/22/2022

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamanetworkopen.2020.17850&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2020.17850


Abstract (continued)

CONCLUSIONS AND RELEVANCE The findings of this study suggest that childhood violenceexposure is associated with adolescent neural network sparsity. A community-detection algorithm,blinded to child adversity, grouped youth exposed to heightened violence based only on patterns ofrsFC. The findings may have implications for understanding how dimensions of adverse childhoodexperiences impact individualized neural development.

JAMA Network Open. 2020;3(9):e2017850.

Corrected on December 7, 2020. doi:10.1001/jamanetworkopen.2020.17850

Introduction

Adversity during childhood is a common, detrimental public health issue. Adverse childhoodexperiences negatively impact physical and mental health, and effects likely persist into adulthood.1-3

Early adverse environments have underlying dimensions, such as violence exposure (eg,neighborhood violence) and social deprivation (eg, neglect),4,5 which have distinct neural correlatesrelated to emotion, fear, and reward processing.4,6 For instance, violence exposure and socialdeprivation are associated with blunted amygdala and ventral striatum reactivity, respectively.5

However, it is unclear how these dimensions affect neural circuitry. Adverse childhood experiencesnot measured dimensionally are associated with alterations in resting-state functional connectivity(rsFC) of the salience network (SN), which is a task-positive network including the anterior insula thatis involved in identifying and integrating salient input,7 and the default mode network (DMN), whichis a task-negative network including the inferior parietal lobule that is linked to internal thought,memory, and social-cognitive processes.8-11 However, inferences are limited by relatively small,homogeneous samples focused on few brain regions using retrospective reports of adversity.12 Thus,there are significant knowledge gaps concerning the ways in which early violence exposure and socialdeprivation impact later functioning of neural circuits and how that impact varies across individuals.

Neural circuits are typically studied using a network framework, with key features includingdensity (ie, number of connections13) and node degree (ie, number of connections involving aspecific brain region).14,15 Across development, network density increases between distal nodes andnode degree increases for hub regions, such as the anterior insula.16 Understanding how earlyadversity relates to network density has a potential for revealing how the environment affects braindevelopment.14

These influences are likely to be person specific because there is considerable variability inneural responses to environmental stress,17 and thus, mean-based analyses may not accuratelyreflect an individual’s circuitry.18 Data from neuroimaging projects, such as the Midnight Scan Club,have illustrated that the organization of an individual’s rsFC is unique and qualitatively different fromthe group average.19 Moreover, consistent with behavioral studies of early adversity,20 averageestimates of adversity’s effects on the brain often have high variances.21-23 This variance raises thequestion of whether there is information about individual differences and their causes—important foreventual prevention and intervention—that is not being conveyed by mean-based conclusions.

In the present study, we examined the association between dimensional indexes of childhoodexposure and individualized adolescent rsFC networks. We used a large, longitudinal sample ofadolescents with a substantial representation of African American and low-income participants, whoare often underrepresented in neuroimaging research,12 and a person-specific rsFC approach thatdetects meaningful connections among brain regions while identifying subgroups of participantsthat share network features (group iterative multiple model estimation [GIMME]).24,25 Wehypothesized that childhood violence exposure and social deprivation would be associated withperson-specific indexes of SN and DMN density, respectively. This study was preregistered with theCenter for Open Science.

JAMA Network Open | Pediatrics Association of Childhood Violence Exposure With Adolescent Neural Network Density




Methods

Participants were from the Fragile Families and Child Wellbeing Study (FFCWS), a population-basedcohort study of children born in large US cities, with an oversample of nonmarital births as well as alarge proportion of families of racial/ethnic minorities with low resources.26 This study followed theStrengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelinefor cohort studies. This study was approved by the University of Michigan Institutional Review Board.Caregivers and participants provided informed consent or assent. Participants received financialcompensation.

In the FFCWS, data were collected from the primary caregiver (94% biological mothers) andfocal child at their birth and at ages 1, 3, 5, 9, and 15 years through in-home visits and phone calls.27

Data for the violence exposure and social deprivation composites, which have been previouslyreported,5,6,28 came from surveys collected at ages 3, 5, and 9 years. During wave 6 (when the focalchild was approximately age 15 years), 237 adolescents from Detroit, Michigan; Toledo, Ohio; andChicago, Illinois, participated in a supplementary visit during which rsFC data were collected. Thestudy was conducted from February 1, 1998, to April 26, 2017, and data analysis was performed fromJanuary 3, 2019, to May 22, 2020. To our knowledge, rsFC data and their association with earlyadversity have not been previously published. Owing to the representative sampling of the FFCWS,youth were not preselected based on their willingness to participate in a magnetic resonance imaging(MRI) study, which is a common procedure in neuroimaging. This sampling approach led to missingor incomplete MRI data (n = 54) but was not statistically significantly different from the recruitedsample (eMethods and eTable 1 in the Supplement).

Participants and ProcedureA total of 183 adolescents aged 15 to 17 years were eligible for inclusion. The participants and theirprimary caregiver came to a university lab and completed questionnaires regarding the focal child’scurrent life stress, pubertal development, and other demographic variables. During an MRI scan, 8minutes of data were collected while the adolescents were instructed to remain still and focus on awhite fixation cross on a black screen.

Violence exposure and social deprivation were each operationalized by composite z scorescalculated from FFCWS data at ages 3, 5, and 9 years; thus, 0 is the approximate mean. Bothconstructs included primary caregiver’s report of experiences that directly (eg, physical abuse) andindirectly (eg, community support) affect the child (eMethods in the Supplement). Violenceexposure was operationalized as physical or emotional abuse directed at the child, exposure tointimate partner violence, and witnessing or being victimized by community violence. Socialdeprivation was operationalized as emotional or physical neglect, lack of romantic partner supportfor the mother, and lack of neighborhood cohesion.5 To reflect a comprehensive assessment ofcumulative, dimensional childhood exposure to violence and social deprivation, both constructsincluded experiences with several levels of proximity to the child (eg, home and neighborhood) atmultiple ages.5,29

To address potential confounding factors, sensitivity analyses adjusted for sex, race, pubertaldevelopment,30 adolescent life stress,31-33 maternal educational level at the child’s birth, andmaternal marital status at the child’s birth26 (eMethods in the Supplement).

Neuroimaging MeasuresMR AcquisitionMagnetic resonance data were acquired using a 3T scanner with an 8-channel head coil (GEDiscovery MR750, GE Healthcare). Head movement was minimized through instructions to theparticipant and padding was placed around the head. Functional T2*-weighted blood oxygenationlevel–dependent images were acquired using a reverse spiral sequence34 of 40 contiguous axial3-mm slices (eMethods in the Supplement).




https://www.equator-network.org/reporting-guidelines/strobe/





Imaging Data AnalysisPreprocessing was primarily conducted in FSL, version 5.0.7.35,36 Structural images were skull-stripped and segmented. Functional images were skull-stripped, spatially smoothed, registered tosubject-specific structural and Montreal Neurological Institute space, and corrected for motion usingMCFLIRT37 and ICA-AROMA.38 Nuisance signal from white matter and cerebrospinal fluid wasremoved. Data were high-pass filtered. As an additional precaution against motion-related artifacts,participants with an average relative framewise displacement greater than 0.5 mm before motionpreprocessing were excluded (n = 4) (eMethods in the Supplement).

Participant-specific time series (235 functional volumes) from 7 regions of interest (ROIs) perhemisphere (14 total) were extracted. The ROIs and their locations were selected using Neurosynth39

and preregistered. Four bilateral ROIs defined the SN: amygdala, insula, dorsal anterior cingulatecortex, and dorsal prefrontal cortex. Three bilateral ROIs defined the DMN: inferior parietal lobule(IPL), posterior cingulate cortex, and medial temporal gyrus. Regions of interest were 6.5-mmspheres around central coordinates (eTable 2 in the Supplement) linearly adjusted for participantbrain volume.

GIMMESubgrouping GIMME (S-GIMME, version 0.5.1)24,25,40 in R, version 3.5.1 (The R Project for StatisticalComputing) was used for rsFC analyses (eFigure 1 in the Supplement). Beginning with an emptynetwork, S-GIMME fits person-specific unified structural equation models25 in a data-driven mannerby using Lagrange multiplier tests41 to add directed connections among ROIs that arecontemporaneous (occurring at the same functional volume) or lagged (occurring at the previousvolume, including autoregressives) and that apply to the group level (everyone in the sample), thesubgroup level (everyone in a data-derived subsample), or individual level (unique to an individual).S-GIMME is a sparse mapping approach in which only connections that account for a significantamount of variance are added to each participant’s network until the model fits the data wellaccording to standard fit indexes (root-mean-square error of approximation �0.05; standard rootmean residual �0.05; confirmatory fit index �0.95; and nonnormed fit index �0.95).42,43 Duringmodel generation, connections that have become nonsignificant with the addition of newconnections are pruned. S-GIMME uses a community detection algorithm (Walktrap) to detectsubgroups of participants and determine their shared connectivity patterns. S-GIMME has beendescribed and validated in large-scale simulations and applied to empirical data.40,42,44,45 Networkdensity and node degree of only the contemporaneous connections were extracted from person-specific S-GIMME networks, because lagged connections control for sequential dependencies (eg,hemodynamic response function).25,46

Statistical AnalysisInferential analyses were completed in R, version 3.5.1 and examined whether childhood adversitystatistically significantly predicted adolescent neural network features. First, binary logisticregression was used to statistically predict S-GIMME–detected subgroup membership fromchildhood violence exposure and social deprivation. Second, multiple regression was used tostatistically predict density (ie, number of connections) within the SN, within the DMN, and betweenthe SN and DMN from childhood violence exposure and social deprivation. Third, multiple regressionwas used to statistically predict node degree (ie, number of connections involving a specified node)for each of the 14 ROIs from violence exposure and social deprivation using a Bonferroni-correctedsignificance threshold (P < .004). In follow-up sensitivity analyses, covariates were added to allregressions to assess the robustness of observed effects.







Results

Of the 183 participants, 175 adolescents aged 15 to 17 years (mean [SD] age, 15.88 [0.53] years) wereincluded in the analysis. Of the 183 SAND adolescents with resting-state MRI data, 3 were excludedowing to artifacts in the functional or structural MRI data, 4 had excessive motion (as defined byaverage relative framewise displacement >0.5 mm), and 1 had signal dropout in the areas of the brainincluded in the present analysis. Of these 175 adolescents, 98 participants (56%) were girls, 77participants (44%) were boys, and 127 participants (73%) were African American. All person-specificresting state networks fit the data well, according to average indexes: root-mean-square error ofapproximation = 0.06, standard root mean residual = 0.05, confirmatory fit index = 0.93, andnonnormed fit index = 0.96 (individual indexes in eTable 7 in the Supplement). Group-levelconnections were detected within and between the SN and DMN (Figure 1A), 2 subgroups ofparticipants with subgroup-level connections were identified (Figure 1B, C), and all person-specificmaps contained individual-level connections (mean [SD], 11.61 [5.32]) (eFigure 2 in the Supplementand the Video show individual maps). Final maps revealed that the first subgroup (n = 42) (Figure 1B,D) was qualitatively homogeneous with 27 subgroup-level connections and few individual-levelconnections (mean [SD], 5.6 [03.19]), and that the second subgroup (n = 133) Figure 1C, E) wasqualitatively heterogeneous, with 8 subgroup-level connections and many individual-levelconnections (mean [SD], 13.50 [4.36]). Possible extreme outliers or overfit models did not impactresults (eMethods in the Supplement).

Violence exposure was associated with subgroup membership (b = 1.12, P = .03). With a unitincrease in violence exposure, participants were 3.06 (95% CI, 1.17-8.92) times more likely to beclassified in the larger, heterogeneous subgroup (Table 1). In sensitivity analyses, the odds ratio was2.54 (eTable 6 in the Supplement). Members of the heterogeneous subgroup experienced higherlevels of childhood violence exposure (mean [SD], 0.09 [0.53]) than those in the homogeneoussubgroup (mean [SD], −0.15 [0.43]). Social deprivation was not associated with subgroupmembership (Table 1).

Childhood violence exposure was related to reduced density (ie, sparsity) in the person-specificmaps (β = −0.25; 95% CI, −0.41 to −0.05; P = .005; adjusted R2 = 0.059) (Figure 2). Specifically,violence exposure was associated with sparsity within the SN (β = −0.26; 95% CI, −0.43 to −0.08;P = .005) and between the SN and DMN (β = −0.20; 95% CI, −0.38 to −0.03; P = .02) (Table 2),including in sensitivity analyses (eTable 3 in the Supplement). Violence exposure was not associatedwith DMN density, and social deprivation was not related to network metrics (Table 2).

Childhood violence exposure was related to reduced node degree for the right insula(β = −0.29; 95% CI, −0.47 to −0.12; P = .001) and left IPL (β = −0.26; 95% CI, −0.44 to −0.09;P = .003) using a Bonferroni-corrected significance threshold (Table 2; eTable 4 in the Supplement),including in sensitivity analyses (eTable 5 in the Supplement). There were no significant associationswith social deprivation.

Discussion

Results from a predominantly understudied and underserved sample with high rates of povertysuggest that childhood violence exposure, but not social deprivation, is associated with adolescentneural circuitry. Data-driven analyses identified a subset of adolescents with heterogeneous patternsof connectivity (ie, few shared and many individual connections) in 2 key neural networks associatedwith salience detection, attention, and social-cognitive processes (ie, the SN and DMN).7,8 Thissubgroup of adolescents was exposed to more violence in childhood than the other subgroup, whosepatterns of neural connectivity were relatively more homogeneous (ie, had many connections incommon), suggesting that violence exposure may lead to more person-specific alterations in neuralcircuitry. Beyond subgroups, network density within the SN and between the SN and DMN wassparse for adolescents with high violence exposure, likely due to few connections involving the right











Figure 1. S-GIMME Connectivity Results

Group pathsA

Homogeneous subgroup pathsB Heterogeneous subgroup pathsC

Individual-level connection, subgroup 1D Individual-level connection, subgroup 2E

A, Connections fit at the group-level (ie, statistically meaningful for �75% of sample)(n = 175). B, Subgroup-level connections for first algorithm-detected subgroup (n = 42).C, Subgroup-level connections for second algorithm-detected subgroup (n = 133). D,Individual-level connections for illustrative participant in first subgroup. E, Individual-

level connections for illustrative participant in second subgroup. All connections aredirected and contemporaneous. Red nodes are part of the salience network. Blue nodesare part of the default mode network.




insula and the left IPL. These factors could not be accounted for by social deprivation, in-scannermotion, race, sex, pubertal development, current life stress, or maternal marital status or educationallevel at the time of the participant’s birth.

Findings regarding the neural network subgroups are noteworthy because the communitydetection algorithm within GIMME detected rsFC patterns in the brain from exposures that occurredat least 6 years earlier. Moreover, high childhood violence exposure in the subgroup characterizedby neural heterogeneity likely reflects the person-specific outcomes of early adversity on the brainand suggests that research on the developmental sequelae of adverse childhood experiences shouldconsider individual differences in neural compensatory responses to stress.17 Although it is importantto replicate these findings in other samples, S-GIMME has reliably classified subgroups in empiricaldata,40,45 and there is evidence from simulations that modeling connections at the subgroup level, inaddition to the group level, improves the validity and reliability of results.40

Considering the sample as a whole, results also suggest that violence exposure is associatedwith blunted connectivity within the SN and between the SN and DMN. As expected, the observedreduced SN density in adolescents with heightened childhood violence exposure differs from typicaldevelopmental patterns that show stronger rsFC within SN nodes and increased density ofconnections with hub regions, such as the anterior insula, as the brain matures.8,16 It is difficult,however, to align the present findings with previous work that reported increased SN rsFC in trauma-exposed youth9,10 because those samples were small, used different metrics of connectivity, and haddifferent sample compositions. Moreover, the present sample was likely experiencing chronicadversity, and research from animal models of chronic stress proposes that, over time, the body’sstress response (eg, hypothalamic pituitary adrenal axis reactivity) becomes blunted or habituatedto typical stressors.47 Previous research on hypothalamic pituitary adrenal axis reactivity in thissample revealed a blunted cortisol response in adolescents with heightened childhood violenceexposure,28 and work in other high-risk samples showed blunted activation of the amygdala, an SNnode, to threatening stimuli.48,49 The present study expands this notion to the function of threatdetection neural circuits, and future research should examine whether this is compensatory or evenadaptive.

Table 1. Logistic Regression Results for Association Between Violence Exposure and Social Deprivationand Subgroup Membership While Controlling for Motiona

Variable b (SE) Odds ratio (95% CI)

Intercept 0.55 (0.32) 1.73 (0.90-3.22)

Violence exposureb 1.12 (0.52) 3.06 (1.17- 8.92)

Social deprivation −0.49 (0.46) 0.61 (0.25-1.54)

Motionb,c 7.96 (3.59) 2860.05 (6.33- 8 236 598)

a Heterogeneous or homogeneous subgroupmembership.

b Significant factor associated with subgroupmembership.

c Motion indicated by mean relative framewisedisplacement.

Figure 2. Association Between Childhood Violence Exposure and Reduced Network Density

20–1 1 2 3

70

60

Tota

l con

nect

ions

, No.

Violence exposure, density

50

40

30

0 Number of connections modeled for each child.Adjusted R2 = 0.059.




Beyond density, childhood violence exposure was associated with reduced node degree of theright anterior insula and left IPL. These results are consistent with the extant literature because theright anterior insula in the SN facilitates shifting between the DMN and central executive network,50

which contributes to higher-level executive function.8 Moreover, early life stress has been linked toinsular connectivity within the SN,9 DMN (specifically, the left IPL, which plays a role in workingmemory51), and other neural ROIs.52 These results also show differences in the way that the anteriorinsula is integrated within and between neural networks in youth exposed to violence in their homesand neighborhoods using longitudinal data from a population-based sample.

This study represents a person-specific approach to the neuroscientific investigation of thesequelae of early adversity. Past research on early adversity and rsFC assumed that the sameconnectivity patterns characterize all, or a majority of participants, but if this assumption is violated(as is likely the case in studies of diverse populations and biopsychosocial phenomena), then resultsmay not accurately describe any individual.18,53 The presence of group- and subgroup-levelconnections in the present study suggests that there is some consistency in the connections withinand between the SN and DMN, aligning with an assumption of homogeneity that is prevalent in rsFCresearch, but the large number of individual-level connections, especially in adolescents with highlevels of early violence exposure, show that there was also notable heterogeneity that requiredperson-specific analyses to accurately reflect rsFC, encouraging future research using person-specificmodeling approaches.

Table 2. Regression Results for Association Between Dimensional Exposure to Adversity and Network Densityand Node Degreea,b

Variable b (95% CI)c β (95% CI)d re FitTotal density

Intercept 41.88 (40.57 to 43.19) R2 = 0.075; 95% CI, 0.01-0.15

Violence exposuref −3.08 (−5.20 to −0.96) −0.25 (−0.41 to −0.05) −0.12

Social deprivation 1.81 (−0.28 to 3.91) 0.15 (−0.05 to 0.30) .06

Motionf,g 12.98 (3.55 to 22.40) 0.21 (0.06 to 0.37) 0.17f

Salience network density

Intercept 15.92 (15.47 to 16.37) R2 = 0.088; 95% CI, 0.02-0.16


Social deprivation 0.65 (−0.08 to 1.37) 0.15 (−0.02 to 0.32) 0.07

Motionf,g 5.17 (1.91 to 8.43) 0.24 (0.09 to 0.39) 0.20

Density between salience and default mode networks

Intercept 16.03 (15.32 to 16.74) R2 = 0.079; 95% CI, 0.01-0.15



Motionf,g 8.58 (3.48 to 13.67) 0.25 (0.10 to 0.40) 0.23

Default mode network density

Intercept 9.93 (9.50 to 10.36) R2 = 0.024; 95% CI, 0.00-0.07

Violence exposure −0.66 (−1.35 to 0.03) −0.17 (−0.35 to 0.01) −0.13

Social deprivation 0.40 (−0.28 to 1.09) 0.10 (−0.07 to 0.28) 0

Motiona −0.77 (−3.85 to 2.30) −0.04 (−0.19 to 0.12) −0.06

Left inferior parietal lobule degree

Intercept 7.33 (6.98 to 7.68) R2 = 0.050; 95% CI, 0.00-0.11

Violence exposureh −0.85 (−1.41 to −0.28) −0.26 (−0.44 to −0.09) −0.17


Motiong 1.16 (−1.35 to 3.68) 0.07 (−0.08 to 0.22) 0.03

Right insula degree

Intercept 7.72 (7.48 to 7.96) R2 = 0.067; 95% CI, 0.01-0.14

Violence exposureh −0.65 (−1.04 to −0.26) −0.29 (−0.47 to −0.12) −0.16

Social deprivation 0.42 (0.03 to 0.80) 0.19 (0.01 to 0.36) 0.06

Motiong 1.38 (−0.36 to 3.12) 0.12 (−0.03 to 0.27) 0.08

a Significant findings on nodes are reported here;results for all nodes are in eTable 4 in theSupplement.

b Network density is the number of connectionsmodeled by network; node degree, the sum of themodeled connections involving each node.

c Unstandardized regression weights.d Standardized regression weights.e Zero-order correlation.f Significantly associated with network density at

P < .05.g Motion indicated by mean relative framewise

displacement prior to motion correction.h Indicates significant association with node degree at

a Bonferroni-corrected P value <.05 divided by 14(the number of statistical comparisons) to get thenew Bonferroni-corrected P threshold nodes(P < .004).





All significant findings concerned violence exposure, and there were no detected associationsbetween social deprivation and rsFC. This set of results could indicate that social deprivation has aless salient influence on patterns of spontaneous neural fluctuations. Some studies have identifiedlinks between social deprivation and functional connectivity, but they concerned extreme,nonnormative deprivation (eg, previous institutionalization).21,54 This deprivation may bequalitatively different from deprivation operationalized in the present study, and may operatethrough different mechanisms. In addition, because a hypothesis-driven approach to node selectionwas taken in this study, it is possible that deprivation is associated with rsFC of SN or DMN nodes notmeasured here, with other networks (eg, central executive), or in different populations (eg, withextreme or heightened variability of deprivation). It is also tenable that there are other dimensions ofadversity that would have differential associations with rsFC (eg, those linked to emotionality), whichfuture research should explore. Nonetheless, these findings present evidence for dimensionalframeworks of adversity5,55 because there were distinct neural correlates for violence exposure.

LimitationsThis study had limitations. Based on the demographic characteristics of the sample (eg, 73% AfricanAmerican, born in Midwestern cities), it is not clear whether findings will generalize beyondlow-income, urban, African American youth; nonetheless, the present work is important becausethese populations are often underrepresented in neuroimaging research and underserved by themedical community.12 Resting-state functional MRI was collected on only a single occasion inadolescence; thus, it is unclear whether connectivity patterns reflect stable or changing neuralfeatures. In addition, it is not possible to know the direction of association (eg, whether neuraldifferences predate exposure to adversity). Violence exposure and social deprivation compositeswere derived from parent reports. Exposures between the FFCWS collection waves at ages 9 and 15years could not be accounted for in this study. Owing to changes in the FFCWS questionnaire at year15, current adversity could not be controlled using the composite scores created for earlier ages.5 Tocompensate, a life stress scale was used as a covariate; however, that confounding variable did notimpact associations. The ecologic pattern of poverty-related adversity is complex; thus, there areunmeasured variables that may explain these associations or contribute to cascades of risk (eg,parental psychopathologic factors).

Conclusions

In this prospective, longitudinal study, childhood violence exposure, but not social deprivation, wasassociated with person-specific differences in how the adolescent brain functions in regions involvedin salience detection and higher-level cognitive processes. These differences were potent enoughthat a data-driven algorithm, blinded to child adversity, grouped youth with heightened violenceexposure based on the heterogeneity of their neural networks, suggesting that the impact ofviolence exposure may have divergent and personalized associations with functional neuralarchitecture. These findings have implications for understanding how dimensions of adversity affectbrain development, which may inform future neuroscience-based policy interventions.

ARTICLE INFORMATIONAccepted for Publication: July 13, 2020.

Published: September 23, 2020. doi:10.1001/jamanetworkopen.2020.17850

Correction: This article was corrected on December 7, 2020, to fix an error in the subject area.

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 GoetschiusLG et al. JAMA Network Open.





https://jamanetwork.com/journals/jamanetworkopen/pages/instructions-for-authors#SecOpenAccess/?utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2020.17850

Corresponding Author: Christopher S. Monk, PhD, Department of Psychology, University of Michigan, 530Church St, 2000 E Hall, Ann Arbor, MI 48109 ([email protected]).

Author Affiliations: Department of Psychology, University of Michigan, Ann Arbor (Goetschius, Hein, McLoyd,Dotterer, Lopez-Duran, Hyde, Monk, Beltz); Serious Mental Illness Treatment Resource and Evaluation Center,Office of Mental Health and Suicide Prevention, Department of Veterans Affairs, Ann Arbor, Michigan (Hein);Department of Sociology, Princeton University, Princeton, New Jersey (McLanahan); Teachers College & Collegeof Physicians and Surgeons, Columbia University, New York, New York (Brooks-Gunn); Survey Research Center ofthe Institute for Social Research, University of Michigan, Ann Arbor (Mitchell, Hyde, Monk); Population StudiesCenter of the Institute for Social Research, University of Michigan, Ann Arbor (Mitchell); Neuroscience GraduateProgram, University of Michigan, Ann Arbor (Monk); Department of Psychiatry, University of Michigan, AnnArbor (Monk).

Author Contributions: Ms Goetschius had full access to all of the data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analysis.

Concept and design: Goetschius, Hein, Brooks-Gunn, McLoyd, Mitchell, Hyde, Monk, Beltz.

Acquisition, analysis, or interpretation of data: Goetschius, Hein, McLanahan, Brooks-Gunn, Dotterer, Lopez-Duran, Mitchell, Hyde, Monk, Beltz.

Drafting of the manuscript: Goetschius, Dotterer, Monk, Beltz.

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

Statistical analysis: Goetschius, Hein, Brooks-Gunn, Dotterer, Beltz.

Obtained funding: Lopez-Duran, Mitchell, Hyde, Monk.

Administrative, technical, or material support: McLanahan, Dotterer, Lopez-Duran, Mitchell, Hyde, Monk, Beltz.

Supervision: Brooks-Gunn, Lopez-Duran, Mitchell, Hyde, Monk, Beltz.

Conflict of Interest Disclosures: Dr McLanahan reported receiving grants from the National Institutes of Health(NIH) during the conduct of the study. Dr Mitchell reported receiving grants from the NIH and the JacobsFoundation during the conduct of the study. Dr Hyde reported receiving grants from the NIH during the conductof the study. Dr Monk reported receiving grants from the NIH during the conduct of the study. Dr Beltz reportedreceiving grants from the Jacobs Foundation during the conduct of the study. No other disclosures were reported.

Funding/Support: This study was supported by NIH grants R01MH103761 (Monk), T32HD007109 (McLoyd andMonk), and S10OD012240 (Noll), as well as a Doris Duke Fellowship for the Promotion of Child Well-being (Hein).Investigators have also been funded by the Jacobs Foundation (Mitchell and Beltz). Funding for the Fragile Familiesand Child Wellbeing Study was provided by National Institute of Child Health and Human Development grantsR01-HD36916, R01-HD39135, and R01-HD40421 and a consortium of private foundations.

Role of the Funder/Sponsor: The NIH, Doris Duke Foundation, and Jacobs Foundation had no role in design andconduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, orapproval of the manuscript; or decision to submit the manuscript for publication.

Additional Information: The Center for Open Science preregistration can be viewed at https://osf.io/mrwhn/, andthe data will be openly available at https://nda.nih.gov/edit_collection.html?id=2106.

REFERENCES1. Finkelhor D, Turner HA, Shattuck A, Hamby SL. Prevalence of childhood exposure to violence, crime, and abuse:results from the National Survey of Children’s Exposure to Violence. JAMA Pediatr. 2015;169(8):746-754. doi:10.1001/jamapediatrics.2015.0676

2. Green JG, McLaughlin KA, Berglund PA, et al. Childhood adversities and adult psychiatric disorders in thenational comorbidity survey replication I: associations with first onset of DSM-IV disorders. Arch Gen Psychiatry.2010;67(2):113-123. doi:10.1001/archgenpsychiatry.2009.186

3. Nusslock R, Miller GE. Early-life adversity and physical and emotional health across the lifespan: a neuroimmunenetwork hypothesis. Biol Psychiatry. 2016;80(1):23-32. doi:10.1016/j.biopsych.2015.05.017

4. McLaughlin KA, Sheridan MA, Lambert HK. Childhood adversity and neural development: deprivation andthreat as distinct dimensions of early experience. Neurosci Biobehav Rev. 2014;47:578-591. doi:10.1016/j.neubiorev.2014.10.012

5. Hein TC. Dimensions of early adversity as distinct predictors of adolescent brain development. Published 2019.Accessed March 19, 2019. https://deepblue.lib.umich.edu/bitstream/handle/2027.42/149945/heint_1.pdf?sequence=1&isAllowed=y




mailto:[email protected]

https://osf.io/mrwhn/

https://nda.nih.gov/edit_collection.html?id=2106

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamapediatrics.2015.0676&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2020.17850

https://jama.jamanetwork.com/article.aspx?doi=10.1001/jamapediatrics.2015.0676&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2020.17850

https://jama.jamanetwork.com/article.aspx?doi=10.1001/archgenpsychiatry.2009.186&utm_campaign=articlePDF%26utm_medium=articlePDFlink%26utm_source=articlePDF%26utm_content=jamanetworkopen.2020.17850

https://dx.doi.org/10.1016/j.biopsych.2015.05.017

https://dx.doi.org/10.1016/j.neubiorev.2014.10.012

https://dx.doi.org/10.1016/j.neubiorev.2014.10.012

https://deepblue.lib.umich.edu/bitstream/handle/2027.42/149945/heint_1.pdf?sequence=1&isAllowed=y

https://deepblue.lib.umich.edu/bitstream/handle/2027.42/149945/heint_1.pdf?sequence=1&isAllowed=y

6. Goetschius LG, Hein TC, Mitchell C, et al. Childhood violence exposure and social deprivation predict adolescentamygdala-orbitofrontal cortex white matter connectivity. Dev Cogn Neurosci. Published online July 8, 2020. doi:10.1016/j.dcn.2020.100822

7. Uddin LQ. Functions of the salience network. In: Salience Network of the Human Brain. Academic Press;2017:11-16. doi:10.1016/B978-0-12-804593-0.00003-5

8. Menon V. Large-scale brain networks and psychopathology: a unifying triple network model. Trends Cogn Sci.2011;15(10):483-506. doi:10.1016/j.tics.2011.08.003

9. Marusak HA, Etkin A, Thomason ME. Disrupted insula-based neural circuit organization and conflictinterference in trauma-exposed youth. Neuroimage Clin. 2015;8:516-525. doi:10.1016/j.nicl.2015.04.007

10. Marusak HA, Martin KR, Etkin A, Thomason ME. Childhood trauma exposure disrupts the automatic regulationof emotional processing. Neuropsychopharmacology. 2015;40(5):1250-1258. doi:10.1038/npp.2014.311

11. van der Werff SJA, Pannekoek JN, Veer IM, et al. Resting-state functional connectivity in adults with childhoodemotional maltreatment. Psychol Med. 2013;43(9):1825-1836. doi:10.1017/S0033291712002942

12. Falk EB, Hyde LW, Mitchell C, et al. What is a representative brain? neuroscience meets population science.Proc Natl Acad Sci U S A. 2013;110(44):17615-17622. doi:10.1073/pnas.1310134110

13. Beltz AM, Gates KM. Network mapping with GIMME. Multivariate Behav Res. 2017;52(6):789-804. doi:10.1080/00273171.2017.1373014

14. Sporns O. Networks of the Brain. MIT Press; 2010.

15. Power JD, Fair DA, Schlaggar BL, Petersen SE. The development of human functional brain networks. Neuron.2010;67(5):735-748. doi:10.1016/j.neuron.2010.08.017

16. Fair DA, Cohen AL, Power JD, et al. Functional brain networks develop from a “local to distributed”organization. PLoS Comput Biol. 2009;5(5):e1000381. doi:10.1371/journal.pcbi.1000381

17. Marder E, Goaillard J-M. Variability, compensation and homeostasis in neuron and network function. Nat RevNeurosci. 2006;7(7):563-574. doi:10.1038/nrn1949

18. Molenaar PCM. A manifesto on psychology as idiographic science: bringing the person back into scientificpsychology, this time forever. Meas Interdiscip Res Perspect. 2004;2(4):201-218. doi:10.1207/s15366359mea0204_1

19. Gordon EM, Laumann TO, Gilmore AW, et al. Precision functional mapping of individual human brains. Neuron.2017;95(4):791-807.e7. doi:10.1016/j.neuron.2017.07.011

20. Hughes K, Bellis MA, Hardcastle KA, et al. The effect of multiple adverse childhood experiences on health:a systematic review and meta-analysis. Lancet Public Health. 2017;2(8):e356-e366. doi:10.1016/S2468-2667(17)30118-4

21. Gee DG, Gabard-Durnam LJ, Flannery J, et al. Early developmental emergence of human amygdala-prefrontalconnectivity after maternal deprivation. Proc Natl Acad Sci U S A. 2013;110(39):15638-15643. doi:10.1073/pnas.1307893110

22. Hanson JL, Nacewicz BM, Sutterer MJ, et al. Behavioral problems after early life stress: contributions of thehippocampus and amygdala. Biol Psychiatry. 2015;77(4):314-323. doi:10.1016/j.biopsych.2014.04.020

23. Hanson JL, Hariri AR, Williamson DE. Blunted ventral striatum development in adolescence reflects emotionalneglect and predicts depressive symptoms. Biol Psychiatry. 2015;78(9):598-605. doi:10.1016/j.biopsych.2015.05.010

24. Lane ST, Gates KM, Molenaar PCM. GIMME: group iterative multiple model estimation. 2019. AccessedFebruary 28, 2019. https://CRAN.R-project.org/package=gimme

25. Gates KM, Molenaar PCM, Hillary FG, Ram N, Rovine MJ. Automatic search for fMRI connectivity mapping: analternative to Granger causality testing using formal equivalences among SEM path modeling, VAR, and unifiedSEM. Neuroimage. 2010;50(3):1118-1125. doi:10.1016/j.neuroimage.2009.12.117

26. Reichman NE, Teitler JO, Garfinkel I, McLanahan SS. Fragile families: sample and design. Child Youth Serv Rev.2001;23(4-5):303-326. doi:10.1016/S0190-7409(01)00141-4

27. Princeton University. Fragile families and child wellbeing study. Published 2019. Accessed October 20, 2019.https://fragilefamilies.princeton.edu/data-and-documentation/public-data-documentation

28. Peckins MK, Roberts AG, Hein TC, et al. Violence exposure and social deprivation is associated with cortisolreactivity in urban adolescents. Psychoneuroendocrinology. 2020;111:104426. doi:10.1016/j.psyneuen.2019.104426

29. McLaughlin KA, Sheridan MA. Beyond cumulative risk: a dimensional approach to childhood adversity. Curr DirPsychol Sci. 2016;25(4):239-245. doi:10.1177/0963721416655883




https://dx.doi.org/10.1016/j.dcn.2020.100822

https://dx.doi.org/10.1016/B978-0-12-804593-0.00003-5

https://dx.doi.org/10.1016/j.tics.2011.08.003

https://dx.doi.org/10.1016/j.nicl.2015.04.007

https://dx.doi.org/10.1038/npp.2014.311

https://dx.doi.org/10.1017/S0033291712002942

https://dx.doi.org/10.1073/pnas.1310134110

https://dx.doi.org/10.1080/00273171.2017.1373014

https://dx.doi.org/10.1080/00273171.2017.1373014

https://dx.doi.org/10.1016/j.neuron.2010.08.017

https://dx.doi.org/10.1371/journal.pcbi.1000381

https://dx.doi.org/10.1038/nrn1949

https://dx.doi.org/10.1207/s15366359mea0204_1

https://dx.doi.org/10.1207/s15366359mea0204_1

https://dx.doi.org/10.1016/j.neuron.2017.07.011

https://dx.doi.org/10.1016/S2468-2667(17)30118-4

https://dx.doi.org/10.1016/S2468-2667(17)30118-4






https://CRAN.R-project.org/package=gimme

https://dx.doi.org/10.1016/j.neuroimage.2009.12.117

https://dx.doi.org/10.1016/S0190-7409(01)00141-4

https://fragilefamilies.princeton.edu/data-and-documentation/public-data-documentation

https://dx.doi.org/10.1016/j.psyneuen.2019.104426

https://dx.doi.org/10.1016/j.psyneuen.2019.104426

https://dx.doi.org/10.1177/0963721416655883

30. Carskadon MA, Acebo C. A self-administered rating scale for pubertal development. J Adolesc Health. 1993;14(3):190-195. doi:10.1016/1054-139X(93)90004-9

31. Shaw DS, Gilliom M, Ingoldsby EM, Nagin DS. Trajectories leading to school-age conduct problems. DevPsychol. 2003;39(2):189-200. doi:10.1037/0012-1649.39.2.189

32. Masten AS, Neemann J, Andenas S. Life events and adjustment in adolescents: the significance of eventindependence, desirability, and chronicity. J Res Adolesc. 1994;4(1):71-97. doi:10.1207/s15327795jra0401_5

33. Farrell AD, Ampy LA, Meyer AL. Identification and assessment of problematic interpersonal situations forurban adolescents. J Clin Child Psychol. 1998;27(3):293-305. doi:10.1207/s15374424jccp2703_6

34. Glover GH, Law CS. Spiral-in/out BOLD fMRI for increased SNR and reduced susceptibility artifacts. MagnReson Med. 2001;46(3):515-522. doi:10.1002/mrm.1222

35. Jenkinson M, Beckmann CF, Behrens TEJ, Woolrich MW, Smith SM. FSL. Neuroimage. 2012;62(2):782-790.doi:10.1016/j.neuroimage.2011.09.015

36. Beltz AM, Dotterer HL, Goetschius LG. GIMME preprocessing: initial release. Published May 9, 2019.

37. Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust and accurate linearregistration and motion correction of brain images. Neuroimage. 2002;17(2):825-841. doi:10.1006/nimg.2002.1132

38. Pruim RHR, Mennes M, van Rooij D, Llera A, Buitelaar JK, Beckmann CF. ICA-AROMA: a robust ICA-basedstrategy for removing motion artifacts from fMRI data. Neuroimage. 2015;112(suppl C):267-277. doi:10.1016/j.neuroimage.2015.02.064

39. Yarkoni T, Poldrack RA, Nichols TE, Van Essen DC, Wager TD. Large-scale automated synthesis of humanfunctional neuroimaging data. Nat Methods. 2011;8(8):665-670. doi:10.1038/nmeth.1635

40. Gates KM, Lane ST, Varangis E, Giovanello K, Guskiewicz K. Unsupervised classification during time-seriesmodel building. Multivariate Behav Res. 2017;52(2):129-148. doi:10.1080/00273171.2016.1256187

41. Sörbom D. Model modification. Psychometrika. 1989;54(3):371-384. doi:10.1007/BF02294623

42. Gates KM, Molenaar PCM. Group search algorithm recovers effective connectivity maps for individuals inhomogeneous and heterogeneous samples. Neuroimage. 2012;63(1):310-319. doi:10.1016/j.neuroimage.2012.06.026

43. Brown TA. Confirmatory Factor Analysis for Applied Research. Guilford Press; 2006.

44. Gates KM, Molenaar PCM, Iyer SP, Nigg JT, Fair DA. Organizing heterogeneous samples using communitydetection of GIMME-derived resting state functional networks. PLoS One. 2014;9(3):e91322. doi:10.1371/journal.pone.0091322

45. Price RB, Lane S, Gates K, et al. Parsing heterogeneity in the brain connectivity of depressed and healthyadults during positive mood. Biol Psychiatry. 2017;81(4):347-357. doi:10.1016/j.biopsych.2016.06.023

46. Smith SM. The future of FMRI connectivity. Neuroimage. 2012;62(2):1257-1266. doi:10.1016/j.neuroimage.2012.01.022

47. McCarty R. Learning about stress: neural, endocrine and behavioral adaptations. Stress. 2016;19(5):449-475.doi:10.1080/10253890.2016.1192120

48. Gard AM, Waller R, Shaw DS, Forbes EE, Hariri AR, Hyde LW. The long reach of early adversity: parenting,stress, and neural pathways to antisocial behavior in adulthood. Biol Psychiatry Cogn Neurosci Neuroimaging.2017;2(7):582-590. doi:10.1016/j.bpsc.2017.06.005

49. Holz NE, Boecker-Schlier R, Buchmann AF, et al. Ventral striatum and amygdala activity as convergence sitesfor early adversity and conduct disorder. Soc Cogn Affect Neurosci. 2017;12(2):261-272. doi:10.1093/scan/nsw120

50. Sridharan D, Levitin DJ, Menon V. A critical role for the right fronto-insular cortex in switching between central-executive and default-mode networks. Proc Natl Acad Sci U S A. 2008;105(34):12569-12574. doi:10.1073/pnas.0800005105

51. Philip NS, Sweet LH, Tyrka AR, et al. Exposure to childhood trauma is associated with altered n-back activationand performance in healthy adults: implications for a commonly used working memory task. Brain ImagingBehav. 2016;10(1):124-135. doi:10.1007/s11682-015-9373-9

52. Teicher MH, Anderson CM, Ohashi K, Polcari A. Childhood maltreatment: altered network centrality ofcingulate, precuneus, temporal pole and insula. Biol Psychiatry. 2014;76(4):297-305. doi:10.1016/j.biopsych.2013.09.016

53. Eavani H, Satterthwaite TD, Filipovych R, Gur RE, Gur RC, Davatzikos C. Identifying sparse connectivitypatterns in the brain using resting-state fMRI. Neuroimage. 2015;105:286-299. doi:10.1016/j.neuroimage.2014.09.058




https://dx.doi.org/10.1016/1054-139X(93)90004-9

https://dx.doi.org/10.1037/0012-1649.39.2.189

https://dx.doi.org/10.1207/s15327795jra0401_5

https://dx.doi.org/10.1207/s15374424jccp2703_6

https://dx.doi.org/10.1002/mrm.1222


https://dx.doi.org/10.1006/nimg.2002.1132




https://dx.doi.org/10.1038/nmeth.1635

https://dx.doi.org/10.1080/00273171.2016.1256187

https://dx.doi.org/10.1007/BF02294623



https://dx.doi.org/10.1371/journal.pone.0091322

https://dx.doi.org/10.1371/journal.pone.0091322




https://dx.doi.org/10.1080/10253890.2016.1192120

https://dx.doi.org/10.1016/j.bpsc.2017.06.005

https://dx.doi.org/10.1093/scan/nsw120



https://dx.doi.org/10.1007/s11682-015-9373-9





54. Chugani HT, Behen ME, Muzik O, Juhász C, Nagy F, Chugani DC. Local brain functional activity following earlydeprivation: a study of postinstitutionalized Romanian orphans. Neuroimage. 2001;14(6):1290-1301. doi:10.1006/nimg.2001.0917

55. Sheridan MA, McLaughlin KA. Dimensions of early experience and neural development: deprivation andthreat. Trends Cogn Sci. 2014;18(11):580-585. doi:10.1016/j.tics.2014.09.001

SUPPLEMENT.eMethods. Detailed MethodseReferenceseTable 1. Participant DemographicseTable 2. MNI Coordinates for ROIseTable 3. Regression Results for Network DensityeTable 4. Node Degree for ROIs That Were Not Significantly Associated With Violence Exposure or SocialDeprivation Using the Bonferroni-Corrected Significance ThresholdeTable 5. Node Density Adjusted for CovariateseTable 6. Logistic Regression Results Adjusted for CovariateseFigure 1. Flowchart of the S-GIMME Analytical StepseFigure 2. Individual Connectivity MapseTable 7. Model Fit for Each Participant






https://dx.doi.org/10.1016/j.tics.2014.09.001

association of childhood violence exposure with adolescent

Documents