Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Preschool Externalizing Behavior Predicts Gender-Specific Variation in Adolescent Neural Structure

  • Jessica Z. K. Caldwell ,

    Current address: Marquette General Hospital/Michigan State University, Marquette, MI, United States of America

    Affiliation Department of Psychology, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Jeffrey M. Armstrong,

    Affiliation Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Jamie L. Hanson,

    Affiliation Department of Psychology, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Matthew J. Sutterer,

    Affiliation Department of Psychology, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Diane E. Stodola,

    Affiliation Department of Psychology, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Michael Koenigs,

    Affiliation Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Ned H. Kalin,

    Affiliation Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Marilyn J. Essex ,

    Contributed equally to this work with: Marilyn J. Essex, Richard J. Davidson

    Affiliation Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin, United States of America

  • Richard J. Davidson

    Contributed equally to this work with: Marilyn J. Essex, Richard J. Davidson

    Affiliations Department of Psychology, University of Wisconsin–Madison, Madison, Wisconsin, United States of America, Department of Psychiatry, University of Wisconsin–Madison, Madison, Wisconsin, United States of America, Center for Investigating Healthy Minds, University of Wisconsin–Madison, Madison, Wisconsin, United States of America


Dysfunction in the prefrontal cortex, amygdala, and hippocampus is believed to underlie the development of much psychopathology. However, to date only limited longitudinal data relate early behavior with neural structure later in life. Our objective was to examine the relationship of early life externalizing behavior with adolescent brain structure. We report here the first longitudinal study linking externalizing behavior during preschool to brain structure during adolescence. We examined the relationship of preschool externalizing behavior with amygdala, hippocampus, and prefrontal cortex volumes at age 15 years in a community sample of 76 adolescents followed longitudinally since their mothers’ pregnancy. A significant gender by externalizing behavior interaction revealed that males—but not females—with greater early childhood externalizing behavior had smaller amygdala volumes at adolescence (t = 2.33, p = .023). No significant results were found for the hippocampus or the prefrontal cortex. Greater early externalizing behavior also related to smaller volume of a cluster including the angular gyrus and tempoparietal junction across genders. Results were not attributable to the impact of preschool anxiety, preschool maternal stress, school-age internalizing or externalizing behaviors, or adolescent substance use. These findings demonstrate a novel, gender-specific relationship between early-childhood externalizing behavior and adolescent amygdala volume, as well as a cross-gender result for the angular gyrus and tempoparietal junction.


The amygdala, hippocampus, and prefrontal cortex (PFC), especially ventromedial and orbitofrontal regions, are key neural substrates for processing and regulating emotion and social behavior. Altered structure and function within this neural circuit have been associated with diverse psychopathologies, including externalizing [14] and internalizing [5,6] disorders. Recent neurodevelopmental research has highlighted adolescence as a critical period for maturation of prefrontal structure and limbic/prefrontal connectivity [7,8]. Further, rates of psychiatric disorders increase in adolescence [9,10]. Thus, a major goal for affective and cognitive neuroscience is to link prefrontal and limbic structure and function with symptoms of psychopathology across development.

Longitudinal studies offer unique opportunities to examine markers and underlying mechanisms of vulnerability to psychopathology. However, few neuroimaging studies have utilized cohorts followed from early childhood, and most of these have emphasized internalizing behaviors [1116]. The single such study of childhood externalizing behavior showed a relationship between aggression across ages 7.5–11 and smaller adult male amygdala volume [16]. These findings complement cross-sectional studies of clinical antisocial behavior, which have most consistently implicated the amygdala and PFC [2,3,17,18], structures that are central in research on externalizing disorders. In particular, amygdala volume abnormalities have been suggested to underlie lower sensitivity to punishment and threat cues and reduced learning of socially reinforcing stimuli, and prefrontal differences to abnormalities in emotion regulation and prediction of consequences [19]. However, investigations have also implicated other structures in externalizing behavior, including the hippocampus [4,20,21], variations in which may contribute to reduced emotion-related learning [22]. Unexamined thus far is whether very early externalizing behaviors predict structural brain differences in adolescence, and whether these might be more significant predictors of later neural structure due to early brain plasticity and the documented role of early experience in shaping the brain [23]. It is also unknown whether these early behaviors explain later neural structure beyond contributions of early life internalizing behaviors, early life stress [2426], later mental health symptoms, and substance use [2730]. Moreover, despite well-established gender differences in externalizing behaviors [31,32], longitudinal neural correlates of female externalizing behaviors remain unexplored.

The present study examined the relationship of early-life externalizing behavior with adolescent amygdala, hippocampus, and PFC volumes in a community sample of 76 adolescents followed longitudinally since birth. We employed rigorous manual segmentation of the amygdala and hippocampus and whole-brain tensor-based morphometry, an analytic strategy that has yielded significant information on brain development [33,34]. Our primary hypothesis was that preschool externalizing behavior would be associated with smaller amygdala and PFC volumes at adolescence. The hippocampus was examined in an exploratory fashion, given previous mixed findings [4,21]. Next, we examined whether school-age externalizing behavior added to the explanatory power of early-life externalizing, with the expectation that early externalizing behavior would be more strongly associated with adolescent amygdala, hippocampus, and PFC volumes than later externalizing behavior. Finally, we considered other relevant variables including gender, maternal stress, preschool anxiety symptoms, school-age internalizing symptoms (i.e., anxiety and depression), school-age ADHD symptoms, and adolescent substance use. Overall, our aim was to add to the literature on neural correlates of early life externalizing behavior in males and females, and we achieved this aim.

Materials and Methods


A sample of 83 adolescents (45 female; mean age = 14.7 years) was recruited from the longitudinal Wisconsin Study of Families and Work [35]. While other aspects of this longitudinal study have been published, data contained here have not been published elsewhere. Original inclusion criteria were that mothers be over age 18, in the second trimester of pregnancy, living with the baby’s father, and employed or a full-time homemaker. Age 15 inclusion criteria were family’s ability to travel to the laboratory and standard MRI eligibility. The University of Wisconsin–Madison Institutional Review Board approved this research; participants’ parents gave written informed consent, and participants gave written informed assent. Research was conducted in accordance with principles expressed in the Declaration of Helsinki.

Post-scan exclusions yielded 76 subjects (42 female, mean age 14.7 years). Subjects were excluded for neural anomaly (1); history of infant febrile seizures (1); missing data, i.e., pubertal status (1), preschool externalizing (2), or more than 2 of the school-age assessments (1). One final data point was excluded due to a highly discrepant manually-traced amygdala volume thought to possibly relate to poor scan quality i.e., preliminary analyses, prior to the current manuscript, indicated this data point had studentized residual values > 4, which indicated a very significant outlier, and a Cook’s D value of nearly 0.5 (cutoff ∼0.05), indicating this point would significantly influence any further analyses (1).

Preschool Externalizing Symptoms

Preschool externalizing symptoms were assessed with the Preschool Behavior Questionnaire (PBQ) [37] using an 11-item scale that covers a broad range of externalizing behaviors (e.g., Is disobedient; Tells lies; Fights with other children). Mothers, fathers, and another familiar adult rated child externalizing behaviors at age 4.5 years on a 3-point scale (Does not apply, Applies sometimes, or Frequently applies); scores were averaged across reporters.

School-Age Externalizing Symptoms

School-age externalizing symptoms were assessed with multi-informant scores at child ages 7, 9, 11, 13, and 15 years. Adult- and child-report measures were developed in tandem to parallel one another. Mothers and teachers completed the MacArthur Health and Behavior Questionnaire (HBQ) [38,39]. Child report was obtained at age 7 via the Berkeley Puppet Interview [40] and subsequently with self-report HBQ [41]. Analyses employed the externalizing symptom scale from each measure, which covers oppositionality/defiance (6–9 items; e.g., Defiant, talks back to adults), conduct problems (9–15 items; e.g., Lies or cheats; Vandalizes), overt aggression (4–8 items; e.g., Gets in many fights), and relational aggression (6–7 items; e.g., Tries to get others to dislike a peer). At each age, principal components analysis (PCA) was used to create multi-informant scores [42]; to obtain an overall level of school-age symptoms as an assessment of trait-like externalizing behavior, PCA-derived scores were averaged across the 5 assessments and z-scored.

Control Variables

Preschool maternal stress. Maternal stress scores consisted of a five-domain, PCA-derived, maternal-report composite at child age 4.5, which encompassed maternal depressive symptoms, parenting stress, and role overload; family expressed anger; and financial stress [43].

Preschool anxiety symptoms. Preschool anxiety was assessed with a 9-item PBQ scale (e.g., Is worried, worries about many things). As with PBQ externalizing, responses were averaged across mother, father, and other-familiar-adult ratings at child age 4.5.

School-age mental health symptoms. HBQ generalized anxiety (7–13 items; e.g., Nervous, high strung, or tense), depression (6–16 items; e.g., Unhappy, sad, or depressed), and inattention/impulsivity (10–18 items; e.g., Cannot concentrate, cannot pay attention for long; Impulsive or acts without thinking) were multi-informant scores, averaged over 5 time-points as for HBQ externalizing behavior.

Adolescent pubertal status. At child age 15, a comprehensive puberty score was created using adolescent report of Tanner stages [44,45] and mother report of the Petersen Pubertal Development Scale [46] (5-point scale; scores range from 1 = no development to 5 = development seems complete) [47,48].

Adolescent substance use. At age 15, participant-reported substance use over the past 30 days was assessed. For alcohol and tobacco separately, total use was estimated, i.e., number of days used multiplied by typical number of drinks or cigarettes per day [49]; summary variables were z-scored.

MRI Parameters

High-resolution whole-brain anatomical images were collected using a GE 750 3T scanner (GE Medical Systems, Waukesha, WI, USA) (T1-weighted inversion recovery fast gradient echo, 124 axial slices, flip angle = 30°, Matrix = 256 x 192, FOV = 240, .9375x.9375x1.2 mm) and reconstructed using in-house software. Image inhomogeneities were smoothed with a multispectral segmentation/bias correction algorithm (FAST) [50], and images underwent skull and vessel correction, and contrast-adjustment. As in previous publications, images were reoriented to the “pathological plane” for more accurate comparison to atlases [51,52]. Specifically, images were first aligned to anterior commissure-posterior commissure (i.e., AC-PC) space and then rotated about the transverse axis such that in the mid-sagittal plane, the posterior junction of the tentorium of the cerebellum was aligned to the same horizontal plane as the inferior margin of the frontal lobe (i.e., as if the cerebrum were sitting on a microtome for sectioning).

Amygdala and Hippocampus Region of Interest (ROI) Definition

Given high variability and low validity of automated segmentation methods for regions like the amygdala [53,54], we employed rigorous manual segmentation of the amygdala and hippocampus. ROIs were manually traced by highly trained staff, who were blind to participant information, using in-house software.

Amygdala Region of Interest (ROI) Definition. Definition was completed according to previously established protocol [51]. Briefly, the optic tract, optic radiations, hippocampus, and inferior horn of the lateral ventricle were used to define the posterior border; temporal lobe white matter, cerebrospinal fluid (CSF), the anterior commissure, and entorhinal cortex were used to define the anterior boundaries. Following initial tracing in the axial plane, traces were refined in coronal and sagittal planes by comparison with ex-vivo atlas sections [55,56]. In particular, the sagittal view was used to confirm accurate separation of the amygdala from hippocampus, entorhinal cortex, optic radiations, and caudate/putamen, while the coronal view was used for refinement of the dorsolateral and inferomedial boundaries [51].

Interrater and spatial reliability were assessed according to established protocol [51]. Specifically, volumetric reliability was assessed via tracing of 6 randomly selected images (12 amygdalae) and examination of interrater intraclass correlation (ICC), which equaled 0.91. Spatial reliability assessment was conducted to ensure that high interrater reliability was not an artifact of numerically similar but spatially divergent amygdala masks; this statistic (intersection/union) averaged 0.84 over the 6 images.

Hippocampus Region of Interest (ROI) Definition. The most anterior aspect of the hippocampal head (HH) was defined as the coronal slice where the temporal horn of the lateral ventricle (TLV) appeared and the alveus was present (Fig. 1). Sagittal view was used for more accurate identification of the alveus (Fig. 1 A-B). The superior border to the HH was identified either by the ventricle arching above it or the white matter of the alveus, separating HH from amygdala. Where decreased image resolution or contrast prevented using these landmarks, the HH was defined superiorly by an arbitrary horizontal line from the superior margin of the TLV to the ambient cistern, in coronal view. The inferior border of the HH was defined as the white matter separating HH from parahippocampal gyrus. Laterally, the HH was defined by the TLV. In more anterior aspects of the HH, the medial border of the HH was defined by an arbitrary vertical line extending from the most medial aspect of the parahippocampal gyrus white matter to the ambient cistern. This line was used to separate the HH from the uncus, which consists mostly of entorhinal cortex and amygdala in this area. Beginning in the most anterior coronal slice where the uncal sulcus separates the hippocampus portion of the uncus from entorhinal cortex, the uncus was included in the HH; here the ambient cistern defined the HH medially.

Fig 1. Hippocampus tracing landmarks as viewed on coronal native space images.

Panels A-H, viewed from top to bottom and left to right in each row, show anterior to posterior progression through the hippocampus. Insets show areas of focus on full-brain sections (black box). An example region of interest is traced in black and labeled by hippocampal subregion: hippocampal head (HH), body (HB), and tail (HT). AMG = amygdala; FX = fornix; EC = entorhinal cortex; LV = lateral ventricle; PG = parahippocampal gyrus; PT = pulvinar of the thalamus; QC = quadrigeminal cistern; TLV = temporal horn of the lateral ventricle; UN = uncus.

The hippocampal body (HB) was defined anteriorly as the coronal slice in which the uncus was no longer present (Fig. 1 C-D). Where the uncus was difficult to distinguish from the hippocampus, three hierarchical criteria were used to identify the first slice of the HB: clearance of tissue from the ambient cistern and most medial section of the TLV, a more circular shape and lateral position of the hippocampus, and thinning of the inferomedial gray matter. Sagittal and axial views confirmed the HH/HT border. For all aspects, the HB followed the same general criteria as the HH, i.e., lateral, superior and medial borders were defined by the CSF of the TLV and ambient cistern; inferior border was defined by white matter. Careful attention was given to the superior HB border to avoid inclusion of the tail of the caudate, lateral geniculate and stria terminalis.

The anterior aspect of the hippocampal tail (HT) was defined generally following Malykhin’s guidelines [57]. In an effort to include as much tissue as possible in the HT, its anterior border was defined as the slice displaying at least 75% of the full profile of the crus of the fornix (Fig. 1 E-F). Similar to the HH and HB, the HT was defined as bordered laterally by the lateral ventricle, medially by either CSF or white matter, inferiorly by white matter, and superiorly by either lateral ventricle or white matter. Care was taken to avoid including the pulvinar of the thalamus and tail of caudate on the superior border. White matter of the fornix was excluded. Fasciolar gyrus and Andreas Retzius gyrus were included [58]. The posterior aspect of the HT was identified as the coronal slice where grey matter starts to appear inferiomedially to the lateral ventricle [59] (Fig. 1 G-H).

As for the amygdala, interrater and spatial reliability were assessed following established procedures [51,60,61]. Specifically, raters traced 5 randomly selected images (10 hippocampi); ICC was 0.97, and spatial reliability averaged 0.81.

Tensor-Based Morphometry (TBM)

Whole-brain TBM was employed to examine prefrontal associations, using Advanced Normalization Tools (ANTS) [62], one of the best available nonlinear warping algorithms [63], and FMRISTAT [64]. TBM is well-suited for a pediatric population, as it minimizes sources of variability such as brain tissue segmentation, uses a study-specific anatomical template, and yields high sensitivity at the voxel level.

Imaging processing and template creation. T1-weighted images were corrected for field inhomogeneity, masked to include brain tissue and ventricular CSF, and used to construct an optimal, study-specific template, via a diffeomorphic shape and intensity averaging technique [62,65]. The region-based cross-correlation similarity metric was employed. Processing yielded an unbiased average shape and appearance template and the set of diffeomorphisms and inverse diffeomorphisms that map from the template to each individual.

Symmetric diffeomorphic image normalization and tensor-based morphometry. After affine transformation, each individual brain was registered to the template using symmetric normalization available in the Advanced Normalization Tools (ANTS) package [62]. Symmetric normalization allows for large, yet physically reasonable, deformations and produces deformation tensor fields, defined in the template space, which chart voxelwise shape change from the template to each subject’s brain.

Jacobian determinants of the deformation field indicate fractional volume expansion and contraction at each voxel, quantifying magnitude of volume alterations required to match the template. Jacobian determinants were log transformed to increase proximity to the normal distribution [62] and smoothed using a Gaussian filter, yielding statistical maps with 8 mm at full-width, half-maximum spatial smoothness.

Data Analysis

Manually-segmented amygdala and hippocampus analyses. Kolmogorov-Smirnov tests confirmed normality for all variables. Amygdala and hippocampus ROI analyses employ residual volumes after regressing out cerebrum volume (i.e., total brain volume excluding brainstem and cerebellum) and pubertal status. Prior to analyses central to our hypotheses, we examined validity of evaluating the amygdala and hippocampus separately by brain hemisphere, and the hippocampus by subregion. A repeated measures general linear model (GLM) with hemisphere as a within-subjects factor and pubertal status and total brain volume as covariates showed no effect of hemisphere for either structure; therefore, findings are presented for the mean of left and right ROI volumes. Similarly, a repeated measures GLM with hippocampus subregion (i.e., head, body, and tail) as a within-subjects factor and gender and preschool externalizing as between-subjects factors showed no significant interaction of gender, externalizing, and subregion. Thus, total hippocampus volumes were employed.

Subsequently, linear regression analyses examined the relationships of preschool externalizing behavior with adolescent ROI volumes using centered variables. For each regression, gender and preschool externalizing, as well as the interaction of gender and preschool externalizing behavior, were entered as predictors of adolescent ROI volumes. A second hierarchical regression examined the additional contribution of school-age externalizing behavior by entering gender, preschool externalizing behavior, and the interaction of gender and preschool externalizing behavior in the first step and school-age externalizing behavior in a second step. Significant findings for preschool externalizing behavior were followed up with a hierarchical regression with gender, preschool externalizing, and the interaction of gender and preschool externalizing behavior predicting ROI volume in the first step, and potential confound variables (i.e., maternal stress at child age 4.5; preschool anxiety; school-age depression, anxiety, and ADHD symptoms; and adolescent alcohol use and tobacco use) entered via stepwise inclusion in a second hierarchical step. As these latter analyses revealed no significant contribution from any potential confounders, confound variables are not considered further.

Significant regression analyses were also subjected to regression diagnostics to assess for outliers of significant influence. First, the regression was run as above, saving studentized residual values and Cook’s Distance. Outliers were defined as data points with studentized residuals > 2. Outliers of influence were defined as data points with studentized residuals > 2 and Cook’s Distance > (4/N-k-1) cutoff; N = sample size; k = number of predictors) [36].

Tensor-based morphometry. Regression models were constructed examining behavioral variables of interest and variations in brain structure across the whole brain in FMRISTAT [64] for the full sample and for males and females separately. Whole-brain volume and pubertal status were entered into a linear regression model as nuisance variables. As in other work [25], an initial statistical threshold of p < .005 uncorrected was used in examining possible brain differences in relation to preschool externalizing behaviors. Signal above this threshold was corrected using Gaussian random-field theory [66] to limit type I error.


Descriptive Statistics

Participant pubertal status at age 15 was rated as nearly complete (mean = 4.47 on a 1–5 scale; SD = 0.48), with female development more advanced (Mann-Whitney z = -3.425; p = .001). Preschool externalizing scores had a mean of 5.18 (SD = 2.74) out of 22 possible points; gender differences were not significant (Mann-Whitney z = -1.642; p = .101). School-age externalizing z-scores had a mean of -0.0014 (SD = 0.872); gender differences were not significant (Mann-Whitney z = -1.740; p = .082). The level and variance of externalizing scores observed are consistent with other community samples in preschool [37,67] and elementary school [68,69], with scores covering a range that includes low, subclinical, and clinical levels. Correlations of ROI volumes with externalizing variables are presented in Table 1 and Fig. 2B. Descriptive statistics for raw whole brain volumes and amygdala and hippocampus ROI volumes are presented in Table 2.

Fig 2. Greater externalizing behavior in preschool correlates with smaller male amygdala volume.

A) Representative coronal view of the amygdala with key neuroanatomical landmarks denoted: AMG = amygdala; EC = entorhinal cortex; HH = hippocampal head; OT = optic tract; TLV = temporal horn of the lateral ventricle; TLWM = temporal lobe white matter. B) Correlation of preschool externalizing with amygdala volumes (upper panel: male r = -0.653, p = 0.000029; female r = -0.088, p = 0.578).

Table 1. Pearson correlations of neural volumes and externalizing behaviors.

Table 2. Descriptive statistics for raw computed and hand-traced neural volumes.

Whole brain volumes (mm3) were segmented automatically; all other volumes are uncorrected hand-traced volumes (mm3). See Statistical Analysis section in Materials and Methods for description of how raw amygdala and hippocampus volumes were corrected for total brain volume and pubertal status.

Preschool Externalizing Behavior and Amygdala Volume

We examined the hypothesis that preschool externalizing would be associated with smaller adolescent amygdala volume. A linear regression showed main effects for preschool externalizing behavior, such that higher preschool externalizing predicted smaller amygdala volumes (t = -3.161, p = .002). The effect of gender was not significant (t = -1.444, p = .153). Results revealed a significant gender by externalizing behavior interaction (t = 2.328, p = .023). Males, but not females, with greater preschool externalizing had smaller adolescent amygdala volumes (see Fig. 2A, B). Although regression diagnostics identified two outlying cases based on studentized residual values >2, and one case that also met influence criteria due to Cook’s D > (4/N-k-1) cutoff, the exclusion of that case did not reduce any findings to non-significance, thus this case was retained.

To determine whether school-age externalizing behaviors contribute to prediction of neural structure beyond what is explained by early-life externalizing behaviors, a second hierarchical regression analysis was completed with preschool externalizing predicting amygdala or hippocampus volume in the first step and school-age externalizing behaviors entered using stepwise inclusion in a second step. School-age externalizing did not contribute significant additional variance. As noted in Materials and Methods, no control variable—including preschool anxiety—explained significant variance.

Preschool Externalizing Behavior and Hippocampus Volume

Next, we examined the relationship of preschool externalizing with adolescent hippocampus volumes. Results revealed a main effect of gender, such that females had significantly smaller hippocampal volumes compared to males (t = -2.351, p = .021), but no main effect of externalizing (t = 0.796, p = 0.428). The interaction of gender and externalizing was not significant (t = 1.699, p = .094). Regression diagnostics showed three outlying cases based on studentized residual value >2; however, none of these three cases was of significant influence based on Cook’s D value, thus cases were retained.

Preschool Externalizing Behavior and Prefrontal Cortex

Voxelwise whole-brain regressions tested the hypothesis that early externalizing behavior would be associated with reduced PFC volume. No significant association with any PFC region was observed. The single significant association that emerged showed that greater preschool externalizing symptoms predicted smaller volumes within a cluster including right angular gyrus and temporoparietal junction (peak value: r = -4.18; p = 0.002; peak MNI coordinates: x = 55, y = -52, z = 45) (Fig. 3A, B). There were no significant results for the interaction of gender and externalizing behavior.

Fig 3. Greater preschool externalizing behavior relates to smaller right angular gyrus and temporoparietal junction, across genders.

A) Coronal, sagittal, and axial views; coordinates indicate cluster peak, within the angular gyrus. B) Correlation of individual subject cluster volume and preschool externalizing behavior (full sample: r = -0.536, p = 0.000000705; male: r = -0.584, p = 0.000355; female r = -0.607, p = 0.0000199).


The current investigation extends literature associating early-life behavior with adolescent and adult brain structure, focusing uniquely on preschool externalizing behavior in a community sample. Analyses employing manual segmentation of adolescent amygdala and hippocampus showed an association of greater preschool externalizing behavior with smaller male amygdala volumes, but no significant results for the hippocampus. Tensor-based morphometry analyses showed no relationship of preschool externalizing behavior with adolescent PFC volume, but revealed greater early externalizing behavior related to smaller volume of a region including the angular gyrus and temporoparietal junction, across genders. Other potentially relevant variables, including preschool maternal stress, preschool anxiety, school-age internalizing and externalizing behavior, and adolescent substance use did not account for these findings.

The observed relationship of greater preschool externalizing with smaller adolescent male amygdala volume is consistent with the single previous longitudinal investigation of childhood externalizing behavior and adult brain volume in a clinical sample [16] and also with work in a sample at high familial risk for externalizing behavior [70]. This result further supports cross-sectional investigations with samples of conduct-disordered children and adolescents [7173] and adults with a history of clinical externalizing [74,75]. Thus, the current study adds to previous work by showing measurable differences in a region important for emotion processing and social learning [7678], even in a community sample with generally modest levels of early-life externalizing behaviors, as early as age 4.5 years.

This very early life externalizing behavior predicted adolescent amygdala volume beyond the impact of later behavior. While consistent with amygdala sensitivity to early experience [23], the implications of this finding will require further investigation. Longitudinal neuroimaging studies are needed to determine when in the course of development the variations in amygdala volume first emerge and whether such variations represent negative consequences of, adaptive responses to, or risk factors for externalizing behaviors [4,79]. Future work should also examine the role of functional amygdala responses and behavioral differences influenced by the amygdala, such as sensitivity to punishment and threat, in the relationship between amygdala structure and externalizing behavior [80]. It is possible that early life behavior is particularly relevant to later brain structure in a community sample with primarily non-clinical externalizing behavior; it is an open question whether this pattern might be seen in samples of clinical externalizing disorders, or whether in those samples later trait behavior is equally or more predictive.

Of note, the present amygdala findings also complement longitudinal research relating early inhibited temperament—which has been linked to internalizing symptoms—to larger adult amygdala volume [14], and in doing so highlight overlap in neural circuits associated with childhood internalizing and externalizing disorders. Given the complex structure of the amygdala [81] and its role in emotion and behavior, it may be fruitful to examine timing, directionality, and degree of structural amygdala variation and the role this plays in comorbidity [82,83]. Gender may be particularly important to consider in this pursuit, given gender differences in prevalence and course of internalizing and externalizing disorders [84,85] as well as in neural development, structure, and function [32,8688].

Contrary to our a priori hypothesis, we did not find an association between externalizing behavior and smaller volume within the PFC using tensor-based morphometry. We did however uncover a relationship between greater early externalizing and smaller volumes near the right angular gyrus and temporoparietal junction, across genders. This finding is consistent with studies showing decreased temporal lobe [8991] and parietal volumes [71] in clinical samples and complements research related to clinical externalizing, including studies implicating the angular gyrus in making moral judgments [92] and the temporoparietal junction in decoding others’ emotional states [93] and acting altruistically [94]. In contrast, the present TBM results were not consistent with research showing reduced PFC (i.e., ventromedial PFC, orbitofrontal cortex) in clinical samples [2,3,95]. It is also noteworthy that TBM analyses did not uncover amygdala volume differences detected with manual segmentation. We believe this discrepancy is due to the greater precision of manual segmentation and possibly due to testing of the manually segmented amygdala as an a priori region of interest, thus not requiring correction for multiple comparisons.

Although several studies have linked variability in hippocampal volume with externalizing behavior [4,20,21], the present study did not. This null finding may suggest specificity for the relationship of early externalizing behavior with amygdala, but not hippocampal, volume at adolescence. Alternatively, effects may have reached significance with a larger sample or one with higher levels of severe externalizing behavior.

Several limitations of the current investigation deserve note. Particularly, the early behaviors studied here are not equivalent to or predictors of violent or clinically aggressive behaviors or psychopathic personality, and care must be taken not to equate these two phenotypes. The present study also cannot speak to causality—it remains an open question whether early-life behaviors are a symptom of early neural abnormalities, or a risk factor for neural dysregulation later in development. In addition, we cannot rule out that differences between our work and previous studies relate to differences in methodological approach to segmentation.

In summary, the present study provides unique evidence for a relationship between early childhood externalizing behavior and adolescent amygdala volume in males, as well as angular gyrus and temporoparietal junction volume across genders. This work augments existing literature on externalizing behavior by examining the relationship between early life externalizing and adolescent brain structure in a community sample. Strengths include longitudinal design, rigorous manual segmentation of key subcortical structures, TBM with particular strengths for pediatric populations, multi-informant measures of externalizing, and inclusion of both genders. Much future investigation will be required to clarify and expand our knowledge of externalizing syndromes and their neural correlates across genders.


The authors acknowledge the work of the investigators and staff of the Wisconsin Study of Families and Work and the Waisman Laboratory for Brain Imaging and Behavior as well as the statistical and methodological consultation of Brendon Nacewicz, Stacey Schaefer, and John Curtin of the University of Wisconsin–Madison and Birdie Shirtcliff of Iowa State University.

Author Contributions

Conceived and designed the experiments: JZKC MJE RJD. Performed the experiments: JZKC JMA DES MJS. Analyzed the data: JZKC JMA JLH DES MJS MJE RJD. Contributed reagents/materials/analysis tools: MJE RJD NHK. Wrote the paper: JZKC JMA JLH DES MJS MK NHK MJE RJD.


  1. 1. Yang Y, Raine A (2009) Prefrontal structural and functional brain imaging findings in antisocial, violent, and psychopathic individuals: a meta-analysis. Psychiatry Res 174: 81–88. pmid:19833485
  2. 2. Anderson NE, Kiehl KA (2012) The psychopath magnetized: insights from brain imaging. Trends Cogn Sci 16: 52–60. pmid:22177031
  3. 3. Blair RJR (2013) The neurobiology of psychopathic traits in youths. Nat Rev Neurosci 14: 786–799. pmid:24105343
  4. 4. Visser TAW, Ohan JL, Whittle S, Yücel M, Simmons JG, et al. (2013) Sex differences in structural brain asymmetry predict overt aggression in early adolescents. Soc Cogn Affect Neurosci. pmid:24369435
  5. 5. Etkin A, Wager TD (2007) Functional neuroimaging of anxiety: a meta-analysis of emotional processing in PTSD, social anxiety disorder, and specific phobia. Am J Psychiatry 164: 1476–1488. pmid:17898336
  6. 6. Hamilton JP, Etkin A, Furman DJ, Lemus MG, Johnson RF, et al. (2012) Functional neuroimaging of major depressive disorder: a meta-analysis and new integration of base line activation and neural response data. Am J Psychiatry 169: 693–703. pmid:22535198
  7. 7. Giedd JN, Stockman M, Weddle C, Liverpool M, Alexander-Bloch A, et al. (2010) Anatomic magnetic resonance imaging of the developing child and adolescent brain and effects of genetic variation. Neuropsychol Rev 20: 349–361. pmid:21069466
  8. 8. Somerville LH, Casey BJ (2010) Developmental neurobiology of cognitive control and motivational systems. Curr Opin Neurobiol 20: 236–241. pmid:20167473
  9. 9. Costello EJ, Copeland W, Angold A (2011) Trends in psychopathology across the adolescent years: what changes when children become adolescents, and when adolescents become adults? J Child Psychol Psychiatry 52: 1015–1025. pmid:21815892
  10. 10. Merikangas KR, He J-P, Burstein M, Swanson SA, Avenevoli S, et al. (2010) Lifetime prevalence of mental disorders in U.S. adolescents: results from the National Comorbidity Survey Replication—Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry 49: 980–989. pmid:20855043
  11. 11. Schwartz CE, Wright CI, Shin LM, Kagan J, Rauch SL (2003) Inhibited and uninhibited infants “grown up”: adult amygdalar response to novelty. Science 300: 1952–1953. pmid:12817151
  12. 12. Schwartz CE, Kunwar PS, Greve DN, Moran LR, Viner JC, et al. (2010) Structural differences in adult orbital and ventromedial prefrontal cortex predicted by infant temperament at 4 months of age. Arch Gen Psychiatry 67: 78–84. pmid:20048225
  13. 13. Schwartz CE, Kunwar PS, Greve DN, Kagan J, Snidman NC, et al. (2012) A phenotype of early infancy predicts reactivity of the amygdala in male adults. Mol Psychiatry 17: 1042–1050. pmid:21894151
  14. 14. Hill SY, Tessner K, Wang S, Carter H, McDermott M (2010) Temperament at 5 years of age predicts amygdala and orbitofrontal volume in the right hemisphere in adolescence. Psychiatry Res 182: 14–21. pmid:20236805
  15. 15. Burghy CA, Stodola DE, Ruttle PL, Molloy EK, Armstrong JM, et al. (2012) Developmental pathways to amygdala-prefrontal function and internalizing symptoms in adolescence. Nat Neurosci 15: 1736–1741. pmid:23143517
  16. 16. Pardini DA, Raine A, Erickson K, Loeber R (2014) Lower amygdala volume in men is associated with childhood aggression, early psychopathic traits, and future violence. Biol Psychiatry 75: 73–80. pmid:23647988
  17. 17. Wallace GL, White SF, Robustelli B, Sinclair S, Hwang S, et al. (2014) Cortical and subcortical abnormalities in youths with conduct disorder and elevated callous-unemotional traits. J Am Acad Child Adolesc Psychiatry 53: 456–465.e1. pmid:24655655
  18. 18. Raine A, Lencz T, Bihrle S, LaCasse L, Colletti P (2000) Reduced prefrontal gray matter volume and reduced autonomic activity in antisocial personality disorder. Arch Gen Psychiatry 57: 119–127; discussion 128–129. pmid:10665614
  19. 19. Crowe SL, Blair RJR (2008) The development of antisocial behavior: what can we learn from functional neuroimaging studies? Dev Psychopathol 20: 1145–1159. pmid:18838035
  20. 20. Laakso MP, Vaurio O, Koivisto E, Savolainen L, Eronen M, et al. (2001) Psychopathy and the posterior hippocampus. Behav Brain Res 118: 187–193. pmid:11164516
  21. 21. Raine A, Ishikawa SS, Arce E, Lencz T, Knuth KH, et al. (2004) Hippocampal structural asymmetry in unsuccessful psychopaths. Biol Psychiatry 55: 185–191. pmid:14732599
  22. 22. Maren S, Phan KL, Liberzon I (2013) The contextual brain: implications for fear conditioning, extinction and psychopathology. Nat Rev Neurosci 14: 417–428. pmid:23635870
  23. 23. Davidson RJ, McEwen BS (2012) Social influences on neuroplasticity: stress and interventions to promote well-being. Nat Neurosci 15: 689–695. pmid:22534579
  24. 24. Lupien SJ, McEwen BS, Gunnar MR, Heim C (2009) Effects of stress throughout the lifespan on the brain, behaviour and cognition. Nat Rev Neurosci 10: 434–445. pmid:19401723
  25. 25. Hanson JL, Chung MK, Avants BB, Shirtcliff EA, Gee JC, et al. (2010) Early stress is associated with alterations in the orbitofrontal cortex: A tensor-based morphometry investigation of brain structure and behavioral risk. J Neurosci 30: 7466–7472. pmid:20519521
  26. 26. Whittle S, Dennison M, Vijayakumar N, Simmons JG, Yücel M, et al. (2013) Childhood maltreatment and psychopathology affect brain development during adolescence. J Am Acad Child Adolesc Psychiatry 52: 940–952.e1. pmid:23972696
  27. 27. Martín-Santos R, Fagundo AB, Crippa JA, Atakan Z, Bhattacharyya S, et al. (2010) Neuroimaging in cannabis use: a systematic review of the literature. Psychol Med 40: 383–398. pmid:19627647
  28. 28. Bühler M, Mann K (2011) Alcohol and the human brain: a systematic review of different neuroimaging methods. Alcohol Clin Exp Res 35: 1771–1793. pmid:21777260
  29. 29. Schiffer B, Müller BW, Scherbaum N, Hodgins S, Forsting M, et al. (2011) Disentangling structural brain alterations associated with violent behavior from those associated with substance use disorders. Arch Gen Psychiatry 68: 1039–1049. pmid:21646569
  30. 30. Mackey S, Paulus M (2013) Are there volumetric brain differences associated with the use of cocaine and amphetamine-type stimulants? Neurosci Biobehav Rev 37: 300–316. pmid:23253945
  31. 31. Zahn-Waxler C, Shirtcliff EA, Marceau K (2008) Disorders of childhood and adolescence: gender and psychopathology. Annu Rev Clin Psychol 4: 275–303. pmid:18370618
  32. 32. Lenroot RK, Giedd JN (2010) Sex differences in the adolescent brain. Brain Cogn 72: 46. pmid:19913969
  33. 33. Yoon U, Perusse D, Lee J-M, Evans AC (2011) Genetic and environmental influences on structural variability of the brain in pediatric twin: Deformation based morphometry. Neurosci Lett 493: 8–13. pmid:21296128
  34. 34. Meintjes EM, Narr KL, van der Kouwe AJW, Molteno CD, Pirnia T, et al. (n.d.) A tensor-based morphometry analysis of regional differences in brain volume in Relation to prenatal alcohol exposure. NeuroImage Clin. Available: Accessed 24 May 2014.
  35. 35. Hyde JS, Klein MH, Essex MJ, Clark R (1995) Maternity leave and women’s mental health. Psychol Women Q 19: 257–285. .
  36. 36. Fox J, editor (1991) Regression diagnostics: an introduction. 1st ed. Newbury Park, CA: Sage Publications, Inc. 96 p. pmid:25144101
  37. 37. Behar L, Stringfield S (1974) A behavior rating scale for the preschool child. Dev Psychol 10: 601–610. .
  38. 38. Boyce WT, Essex MJ, Woodward HR, Measelle JR, Ablow JC, et al. (2002) The confluence of mental, physical, social and academic difficulties in middle childhood. I: Exploring the “headwaters” of early life moribdities. J Am Acad Child Adolesc Psychiatry 41: 580–587. pmid:12014791
  39. 39. Essex MJ, Boyce WT, Goldstein LH, Armstrong JM, Kraemer HC, et al. (2002) The confluence of mental, physical, social and academic difficulties in middle childhood. II: Developing the MacArthur Health and Behavior Questionnaire. J Am Acad Child Adolesc Psychiatry 41: 588–603. pmid:12014792
  40. 40. Ablow JC, Measelle JR, Kraemer HC, Harrington R, Luby J, et al. (1999) The MacArthur Three-City Outcome Study: evaluating multi-informant measures of young children’s symptomatology. J Am Acad Child Adolesc Psychiatry 38: 1580–1590. pmid:10596259
  41. 41. Burk LR, Armstrong JM, Park J-H, Zahn-Waxler C, Klein MH, et al. (2011) Stability of early identified aggressive victim status in elementary school and associations with later mental health problems and functional impairments. J Abnorm Child Psychol 39: 225–238. pmid:20811772
  42. 42. Kraemer HC, Measelle JR, Ablow JC, Essex MJ, Boyce WT, et al. (2003) A new approach to integrating data from multiple informants in psychiatric assessment and research: mixing and matching contexts and perspectives. Am J Psychiatry 160: 1566–1577. pmid:12944328
  43. 43. Essex MJ, Klein MH, Cho E, Kalin NH (2002) Maternal stress beginning in infancy may sensitize children to later stress exposure: effects on cortisol and behavior. Biol Psychiatry 52: 776–784. pmid:12372649
  44. 44. Tanner JM (1962) Growth at adolescence. (2nd ed.). Springfield, Ill.: Thomas. pmid:14041601
  45. 45. Morris NM, Udry JR (1980) Validation of a self-administered instrument to assess stage of adolescent development. J Youth Adolesc 9: 271–280. pmid:24318082
  46. 46. Petersen AC, Crockett L, Richards M, Boxer A (1988) A self-report measure of pubertal status: Reliability, validity, and initial norms. J Youth Adolesc 17: 117–133. pmid:24277579
  47. 47. Shirtcliff EA, Dahl RE, Pollak SD (2009) Pubertal development: correspondence between hormonal and physical development. Child Dev 80: 327–337. pmid:19466995
  48. 48. Ellis BJ, Shirtcliff EA, Boyce WT, Deardorff J, Essex MJ (2011) Quality of early family relationships and the timing and tempo of puberty: effects depend on biological sensitivity to context. Dev Psychopathol 23: 85–99. pmid:21262041
  49. 49. Burk LR, Armstrong JM, Goldsmith HH, Klein MH, Strauman TJ, et al. (2011) Sex, temperament, and family context: how the interaction of early factors differentially predict adolescent alcohol use and are mediated by proximal adolescent factors. Psychol Addict Behav 25: 1–15. pmid:21443307
  50. 50. Zhang Y, Brady M, Smith S (2001) Segmentation of brain MR images through a hidden Markov random field model and the expectation-maximization algorithm. IEEE Trans Med Imaging 20: 45–57. pmid:11293691
  51. 51. Nacewicz BM, Dalton KM, Johnstone T, Long MT, McAuliff EM, et al. (2006) Amygdala volume and nonverbal social impairment in adolescent and adult males with autism. Arch Gen Psychiatry 63: 1417–1428. pmid:17146016
  52. 52. Convit A, McHugh P, Wolf OT, de Leon MJ, Bobinski M, et al. (1999) MRI volume of the amygdala: a reliable method allowing separation from the hippocampal formation. Psychiatry Res 90: 113–123. pmid:10482383
  53. 53. Morey RA, Petty CM, Xu Y, Hayes JP, Wagner HR 2nd, et al. (2009) A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes. NeuroImage 45: 855–866. pmid:19162198
  54. 54. Hanson JL, Suh JW, Nacewicz BM, Sutterer MJ, Cayo AA, et al. (2012) Robust automated amygdala segmentation via multi-atlas diffeomorphic registration. Front Neurosci 6: 166. pmid:23226114
  55. 55. Mai JK, Assheuer J, Paxinos G (1997) Atlas of the human brain. San Diego, CA: Academic Press. 338 p. pmid:25165803
  56. 56. Duvernoy HM, Bourgouin P (1999) The human brain: surface, three-dimensional sectional anatomy with MRI, and blood supply. Springer. 452 p. pmid:25506965
  57. 57. Malykhin NV, Bouchard TP, Ogilvie CJ, Coupland NJ, Seres P, et al. (2007) Three-dimensional volumetric analysis and reconstruction of amygdala and hippocampal head, body and tail. Psychiatry Res 155: 155–165. pmid:17493789
  58. 58. Amaral DG, Scharfman HE, Lavenex P (2007) The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies). Prog Brain Res 163: 3–22. pmid:17765709
  59. 59. Pruessner JC, Li LM, Serles W, Pruessner M, Collins DL, et al. (2000) Volumetry of hippocampus and amygdala with high-resolution MRI and three-dimensional analysis software: minimizing the discrepancies between laboratories. Cereb Cortex N Y N 1991 10: 433–442. pmid:10769253
  60. 60. Rusch BD, Abercrombie HC, Oakes TR, Schaefer SM, Davidson RJ (2001) Hippocampal morphometry in depressed patients and control subjects: relations to anxiety symptoms. Biol Psychiatry 50: 960–964. pmid:11750892
  61. 61. Dalton KM, Nacewicz BM, Johnstone T, Schaefer HS, Gernsbacher MA, et al. (2005) Gaze fixation and the neural circuitry of face processing in autism. Nat Neurosci 8: 519–526. pmid:15750588
  62. 62. Avants B, Gee JC (2004) Geodesic estimation for large deformation anatomical shape averaging and interpolation. NeuroImage 23 Suppl 1: S139–S150. pmid:15501083
  63. 63. Klein A, Andersson J, Ardekani BA, Ashburner J, Avants B, et al. (2009) Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration. NeuroImage 46: 786–802. pmid:19195496
  64. 64. Worsley KJ, Liao CH, Aston J, Petre V, Duncan GH, et al. (2002) A general statistical analysis for fMRI data. NeuroImage 15: 1–15. pmid:11771969
  65. 65. Avants B, Duda JT, Kim J, Zhang H, Pluta J, et al. (2008) Multivariate analysis of structural and diffusion imaging in traumatic brain injury. Acad Radiol 15: 1360–1375. pmid:18995188
  66. 66. Worsley KJ, Taylor JE, Tomaiuolo F, Lerch J (2004) Unified univariate and multivariate random field theory. NeuroImage 23 Suppl 1: S189–S195. pmid:15501088
  67. 67. Bornstein MH, Hahn C-S, Suwalsky JTD (2013) Language and internalizing and externalizing behavioral adjustment: developmental pathways from childhood to adolescence. Dev Psychopathol 25: 857–878. pmid:23880396
  68. 68. Lemery-Chalfant K, Schreiber JE, Schmidt NL, Van Hulle CA, Essex MJ, et al. (2007) Assessing internalizing, externalizing, and attention problems in young children: validation of the MacArthur HBQ. J Am Acad Child Adolesc Psychiatry 46: 1315–1323. pmid:17885573
  69. 69. Kochanska G, Kim S (2012) Toward a new understanding of legacy of early attachments for future antisocial trajectories: evidence from two longitudinal studies. Dev Psychopathol 24: 783–806. pmid:22781855
  70. 70. Hill SY, De Bellis MD, Keshavan MS, Lowers L, Shen S, et al. (2001) Right amygdala volume in adolescent and young adult offspring from families at high risk for developing alcoholism. Biol Psychiatry 49: 894–905. pmid:11377407
  71. 71. Huebner T, Vloet TD, Marx I, Konrad K, Fink GR, et al. (2008) Morphometric brain abnormalities in boys with conduct disorder. J Am Acad Child Adolesc Psychiatry 47: 540–547. pmid:18356764
  72. 72. Fairchild G, Hagan CC, Walsh ND, Passamonti L, Calder AJ, et al. (2013) Brain structure abnormalities in adolescent girls with conduct disorder. J Child Psychol Psychiatry 54: 86–95. pmid:23082797
  73. 73. Sterzer P, Stadler C, Poustka F, Kleinschmidt A (2007) A structural neural deficit in adolescents with conduct disorder and its association with lack of empathy. NeuroImage 37: 335–342. pmid:17553706
  74. 74. Yang Y, Raine A, Narr KL, Colletti P, Toga AW (2009) Localization of deformations within the amygdala in individuals with psychopathy. Arch Gen Psychiatry 66: 986–994. pmid:19736355
  75. 75. Boccardi M, Frisoni GB, Hare RD, Cavedo E, Najt P, et al. (2011) Cortex and amygdala morphology in psychopathy. Psychiatry Res 193: 85–92. pmid:21676597
  76. 76. Blair RJR (2010) Neuroimaging of psychopathy and antisocial behavior: a targeted review. Curr Psychiatry Rep 12: 76–82. pmid:20425314
  77. 77. Debiec J, Díaz-Mataix L, Bush DEA, Doyère V, Ledoux JE (2010) The amygdala encodes specific sensory features of an aversive reinforcer. Nat Neurosci 13: 536–537. pmid:20348916
  78. 78. LeDoux J (2003) The emotional brain, fear, and the amygdala. Cell Mol Neurobiol 23: 727–738. pmid:14514027
  79. 79. Cheetham A, Allen NB, Whittle S, Simmons J, Yücel M, et al. (2014) Volumetric differences in the anterior cingulate cortex prospectively predict alcohol-related problems in adolescence. Psychopharmacology (Berl) 231: 1731–1742. pmid:24553579
  80. 80. Gao Y, Glenn AL, Schug RA, Yang Y, Raine A (2009) The neurobiology of psychopathy: a neurodevelopmental perspective. Can J Psychiatry Rev Can Psychiatr 54: 813–823. pmid:20047720
  81. 81. Sah P, Faber ESL, Lopez De Armentia M, Power J (2003) The amygdaloid complex: anatomy and physiology. Physiol Rev 83: 803–834. pmid:12843409
  82. 82. Angold A, Costello EJ, Erkanli A (1999) Comorbidity. J Child Psychol Psychiatry 40: 57–87. pmid:10102726
  83. 83. Bubier JL, Drabick DAG (2009) Co-occurring anxiety and disruptive behavior disorders: the roles of anxious symptoms, reactive aggression, and shared risk processes. Clin Psychol Rev 29: 658–669. pmid:19729235
  84. 84. Hicks BM, Blonigen DM, Kramer MD, Krueger RF, Patrick CJ, et al. (2007) Gender differences and developmental change in externalizing disorders from late adolescence to early adulthood: A longitudinal twin study. J Abnorm Psychol 116: 433–447. pmid:17696699
  85. 85. Kendler KS, Myers J (2014) The boundaries of the internalizing and externalizing genetic spectra in men and women. Psychol Med 44: 647–655. pmid:23574685
  86. 86. Paus T (2010) Sex differences in the human brain: a developmental perspective. Prog Brain Res 186: 13–28. pmid:21094883
  87. 87. Giedd JN, Raznahan A, Mills KL, Lenroot RK (2012) Review: magnetic resonance imaging of male/female differences in human adolescent brain anatomy. Biol Sex Differ 3: 19. pmid:22908911
  88. 88. Dennison M, Whittle S, Yücel M, Vijayakumar N, Kline A, et al. (2013) Mapping subcortical brain maturation during adolescence: evidence of hemisphere- and sex-specific longitudinal changes. Dev Sci 16: 772–791. pmid:24033581
  89. 89. Kruesi MJP, Casanova MF, Mannheim G, Johnson-Bilder A (2004) Reduced temporal lobe volume in early onset conduct disorder. Psychiatry Res 132: 1–11. pmid:15546698
  90. 90. Ermer E, Cope LM, Nyalakanti PK, Calhoun VD, Kiehl KA (2011) Aberrant paralimbic gray matter in criminal psychopathy. J Abnorm Psychol. Available: Accessed 4 June 2012.
  91. 91. Yang Y, Raine A, Colletti P, Toga AW, Narr KL (2011) Abnormal structural correlates of response perseveration in individuals with psychopathy. J Neuropsychiatry Clin Neurosci 23: 107–110. pmid:21304146
  92. 92. Miczek KA, de Almeida RMM, Kravitz EA, Rissman EF, de Boer SF, et al. (2007) Neurobiology of escalated aggression and violence. J Neurosci 27: 11803–11806. pmid:17978016
  93. 93. Zaki J, Ochsner KN, Ochsner K (2012) The neuroscience of empathy: progress, pitfalls and promise. Nat Neurosci 15: 675–680. pmid:22504346
  94. 94. Morishima Y, Schunk D, Bruhin A, Ruff CC, Fehr E (2012) Linking brain structure and activation in temporoparietal junction to explain the neurobiology of human altruism. Neuron 75: 73–79. pmid:22794262
  95. 95. Koenigs M (2012) The role of prefrontal cortex in psychopathy. Rev Neurosci 23: 253–262. pmid:22752782