OUP user menu

Size matters: Increased grey matter in boys with conduct problems and callous–unemotional traits

Stéphane A. De Brito, Andrea Mechelli, Marko Wilke, Kristin R. Laurens, Alice P. Jones, Gareth J. Barker, Sheilagh Hodgins, Essi Viding
DOI: http://dx.doi.org/10.1093/brain/awp011 843-852 First published online: 17 March 2009

Summary

Brain imaging studies of adults with psychopathy have identified structural and functional abnormalities in limbic and prefrontal regions that are involved in emotion recognition, decision-making, morality and empathy. Among children with conduct problems, a small subgroup presents callous–unemotional traits thought to be antecedents of psychopathy. No structural brain imaging study has examined this subgroup of children. The present study used voxel-based morphometry to compare whole brain grey matter volumes and concentrations of boys with elevated levels of callous–unemotional conduct problems and typically developing boys and explored four a priori regions of interest. sMRI scans were collected from 23 boys with elevated levels of callous–unemotional conduct problems (mean age = 11 years 8 months) and 25 typically developing boys (mean age = 11 years 6 months) selected from a community sample of children. Data were analysed using optimized voxel-based morphometry. Study-specific probability maps were created and four a priori regions of interest identified (orbitofrontal, anterior cingulate and anterior insular cortices and amygdala). Both grey matter volume and concentration were examined controlling for cognitive ability and hyperactivity–inattention symptoms. Boys with callous–unemotional conduct problems, as compared with typically developing boys, presented increased grey matter concentration in the medial orbitofrontal and anterior cingulate cortices, as well as increased grey matter volume and concentration in the temporal lobes bilaterally. These findings may indicate a delay in cortical maturation in several brain areas implicated in decision making, morality and empathy in boys with callous–unemotional conduct problems.

  • conduct problems
  • callous–unemotional traits
  • orbitofrontal cortex
  • anterior cingulate cortex
  • voxel-based morphometry

Introduction

In the last decade, researchers have sought to study the precursors of adult psychopathy in children and have succeeded in identifying a group of children with conduct problems who display callous–unemotional traits (Frick and Marsee, 2006). The increasing evidence base regarding callous–unemotional conduct problems has resulted in its consideration as a possible sub-typing index for the forthcoming fifth edition of the Diagnostic and Statistical Manual of Mental Disorder (Moffitt et al., 2008). Children with callous–unemotional conduct problems present behavioural, temperamental and neuro-cognitive features consistent with the syndrome of adult psychopathy and thus are thought to be at risk of developing the disorder (Blair et al., 2006). They commit violent crimes at a younger age and display more severe and more varied conduct problems than other children with conduct problems (Frick and Marsee, 2006). Several lines of evidence suggest that boys with conduct problems and callous–unemotional traits also exhibit a distinct temperament that is consistent with characteristics of adults with psychopathy. These children have a preference for novel and dangerous activities, exhibit a lack of emotional responsiveness to negative emotional stimuli and a lack of sensitivity to punishment suggestive of low fearfulness (Blair et al., 2006; Frick and Marsee, 2006). Callous–unemotional traits show moderate to strong heritability among children and youth with conduct problems (Taylor et al., 2003; Viding et al., 2005; Larsson et al., 2006) suggesting that there may be genetic vulnerability to callous–unemotional temperament. In addition, antisocial behaviour is more highly heritable among children with callous–unemotional traits than among children with only conduct problems (Viding et al., 2005).

Adults with psychopathy present impairments on neuropsychological tasks tapping the functionality of the ventrolateral/orbitofrontal cortex (e.g. reversal learning) and the amygdala (e.g. recognition of emotional facial expression of fear) (Blair et al., 2005). Recent findings suggest that boys with callous–unemotional conduct problems are characterized by neurocognitive deficits especially in their performance on tests tapping ventrolateral/orbitofrontal cortex and amygdala functioning (Blair et al., 2006).

Consistent with the literature on neuropsychological test performance, studies using functional magnetic resonance imaging (fMRI) techniques have identified limbic-prefrontal circuit abnormalities in children, adolescents and adults with high levels of psychopathy (Kiehl et al., 2001; Birbaumer et al., 2005; Lotze et al., 2007; Finger et al., 2008; Marsh et al., 2008; Jones et al., 2009). Regions implicated in these studies include ventrolateral/orbitofrontal cortex, amygdala, rostral anterior cingulate cortex and hippocampus. These regions are thought to play critical roles in emotion regulation, empathic processing and fear conditioning (Davidson et al., 2000; Sanders et al., 2003; Singer et al., 2004).

Structural magnetic resonance imaging (sMRI) has shown that adults with high levels of psychopathy present increased white matter volumes in the corpus callosum (Raine et al., 2003), asymmetric and decreased grey matter volume within the hippocampus (Laakso et al., 2001; Raine et al., 2004) and reductions in prefrontal grey matter volume (see Yang et al., 2005; although see Laakso et al., 2002). Importantly, these sMRI studies used manual tracing or semi-automated region of interest guided measurement of brain structures, thereby possibly introducing an observer bias. Recent sMRI studies have relied on voxel-based morphometry (VBM), a whole brain, fully automated and unbiased technique for characterizing both regional brain volume and tissue concentration on a voxel-wise basis (Good et al., 2001). In VBM, concentration analysis of grey matter compares the proportion of grey matter to all tissue types within a specific region, whereas volume analysis of grey matter quantifies the absolute volume of grey matter (Mechelli et al., 2005). Although these neurobiological markers tend to be correlated, with greater grey matter concentration associated with greater grey matter volume, they refer to different aspects of the data and may provide complementary information. Two sMRI studies using VBM have compared individuals with high psychopathy scores to healthy individuals. de Oliveira-Souza et al. (2008) reported decreased grey matter concentration in the frontopolar, orbital and anterior temporal cortices, superior temporal sulcus region and insula among psychiatric out-patients with high psychopathy scores as compared with healthy men. Tiihonen et al. (2008) observed focal, symmetrical, bilateral grey matter volume reductions in the postcentral gyri, frontopolar cortex and the orbitofrontal cortex and reduced volume in right anterior insula in violent offenders with high psychopathy scores as compared with healthy men. The finding of reduced insula volume is particularly interesting as this region has been implicated in empathic processing (Singer et al., 2004). Using lobar analyses, the violent offenders with high psychopathy scores, as compared with the healthy men, presented larger white matter volumes, bilaterally, in the occipital and parietal lobes and in the left cerebellum, and larger grey matter volume in the right cerebellum. Analyses in this study did not control for cognitive abilities. These sMRI studies have identified abnormalities in brain structures that have been shown to be functioning differently in fMRI studies of adults with psychopathy.

There is currently no sMRI data focusing on children with callous–unemotional conduct problems. Four sMRI studies (two using VBM methods) have been published on children and adolescents with conduct problems, but none has assessed callous–unemotional traits. A recent VBM study reported a 6% reduction in total grey matter volume in children with conduct problems (Huebner et al., 2008), while two earlier studies assessing whole-brain volume in much smaller samples failed to find a difference between individuals with conduct problems compared with typically developing children and adolescents (Bussing et al., 2002; Kruesi et al., 2004). Some studies have observed decreased grey matter volume in the temporal lobes (Kruesi et al., 2004; Huebner et al., 2008) and in the left amygdala (Sterzer et al., 2007; Huebner et al., 2008) in children and adolescents with conduct problems, as compared with typically developing children and adolescents. However, evidence for structural differences within prefrontal regions, such as the orbitofrontal cortex, has been less consistent. A recent VBM study reported decreased orbitofrontal cortex and increased cerebellar grey matter volume in children and adolescents with conduct problems (Huebner et al., 2008), while other studies have not found significant group differences in these regions. Sterzer et al. (2007) also reported that adolescents with conduct problems have smaller grey matter volume in the anterior insula bilaterally.

The existing sMRI studies on children suffer from some important limitations. With the exception of Huebner et al. (2008), the studies examined small samples. Some did not control for important factors such as cognitive ability (Bussing et al., 2002; Kruesi et al., 2004) and co-morbid attention-deficit/hyperactivity disorder (Kruesi et al., 2004). Finally, most of the studies included participants with a wide age range [e.g. 9.75–20.5 years for Kruesi et al. (2004); 12–17 years for Huebner et al. (2008); 9–15 years for Sterzer et al. (2007)]. Although all studies matched groups for age, the distributions of the ages of the children within each group were not reported. This constitutes a potential problem. Grey matter volume decreases from ages 4 to 21 in a linear, but region specific manner, in several cortical areas (Gogtay et al., 2004) and may thus confound group differences.

The present study examined a community sample of boys, and compared grey matter volume and grey matter concentration of boys with callous–unemotional conduct problems and typically developing boys. All analyses controlled for cognitive ability and hyperactivity–inattention symptoms, as both low cognitive ability and hyperactivity–inattention symptoms are known to be elevated among children with conduct problems (Raine et al., 2005) and are associated with brain morphology (Wilke et al., 2003b; Shaw et al., 2007). Four regions of interest were identified based on the previous empirical and theoretical literature: the ventrolateral/orbitofrontal cortex, amygdala, anterior cingulate cortex and anterior insula. The ventrolateral/orbitofrontal cortex and the amygdala were selected, as these regions have been identified as dysfunctional in neuropsychological and fMRI studies of children, adolescents and adults with psychopathy (Blair et al., 2006; Finger et al., 2008; Marsh et al., 2008; Jones et al., 2009). Also, these regions are central to the most clearly articulated neurocognitive model of developmental psychopathy (Blair et al., 2005). The anterior cingulate cortex and anterior insula were included because structural differences within these regions have been identified in fMRI and sMRI studies of adults with psychopathy and children with conduct problems. The anterior cingulate cortex has been implicated in the regulation of emotional responses (Stein et al., 2007). Both the anterior cingulate cortex and anterior insula are thought to be involved in empathic processing (Singer et al., 2004), which has been shown to be impaired in adults with psychopathy and children with conduct problems and callous–unemotional traits (Blair et al., 2006). We predicted that boys with callous–unemotional conduct problems would show grey matter differences in these key regions of interest, as compared with typically developing boys. As this was the first sMRI study of callous–unemotional conduct problems in childhood, we did not make specific predictions regarding the direction of the differences or any additional brain areas.

Methods and materials

Participants

Participants were recruited from the Twins Early Development Study, a community sample of twins (Trouton et al., 2002). Parents completed screening questionnaires about MRI contra-indicators and provided consent to be contacted about the study. The study and recruitment procedure were approved by the Institute of Psychiatry and South London and Maudsley NHS Foundation Trust Research Ethics Committee. After complete description of the study to the children and their parents, written informed consent was obtained from both.

One hundred and thirty-eight boys attended the scanning facility as part of a large twin imaging study. 130 completed a scan. All boys came from a same sex twin pair (one twin per pair), were aged 10–13 years (mean age 11 years 7 months), had no psychiatric, neurological or medical problems, and a full scale IQ of at least 80. Based on scores from the Conduct Problem sub-scale of the Strengths and Difficulties Questionnaire (Goodman, 1997) and the Callous–Unemotional trait scale of the Antisocial Process Screening Device (Frick and Hare, 2001), boys with scores in appropriate range were assigned to either the callous–unemotional conduct problems group (n = 23) or typically developing boys group (n = 25) resulting in a total sample of 48 boys for the analysis (See Supplementary material for details).

Measures

The Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) is a widely used measure of childhood psychopathology with good reliability and validity (Goodman, 2001).

The Antisocial Process Screening Device (APSD) (Frick and Hare, 2001) assesses the antecedents of psychopathy among children. One factor measures callous–unemotional traits.

The Wechsler Abbreviated Scales of Intelligence (Wechsler, 1999) includes the Vocabulary and Matrix Reasoning subscales to provide an estimate of Full Scale Intelligence Quotient (FSIQ).

Image acquisition

Structural brain images were acquired using a General Electric Signa 3.0 Telsa Excite II MRI scanner (GE Medical systems, Milwaukee, WI, USA) at the Centre for Neuroimaging Science, Institute of Psychiatry, London, UK. A high-resolution, 3D T1-weighted dataset was acquired using an inversion recovery prepared spoiled gradient echo sequence. Imaging parameters were TR = 8 ms; TE = 2.9 ms; TI = 450 ms; excitation aflip angle = 20°. The in-plane matrix size was 256 × 192 over a 280 × 210 mm field of view, reconstructed to 256 × 256 over 280 × 280 mm. In plane pixel size was thus 1.09375 × 1.09375 mm. Two hundred through plane partitions (each 1.1-mm thick) were collected, with two partitions being discarded at each end of the imaging volume to minimize wrap-round artefacts. Partial k-space coverage (0.75 NEX) was used. The scanning time was 6 min.

Image processing

Data were processed and analysed on a Sun Ultra 60 workstation (Sun Microsystems, Mountain View, CA, USA) using MATLAB 7.1 (MathWorks, Natick, MA, USA) and statistical parametric mapping software (SPM5, Wellcome Department of Imaging Neuroscience, London, UK; http://www.fil.ion.ucl.ac.uk/spm). SPM5 provides an integrated spatial normalization and segmentation routine that combines bias correction, tissue classification and non-linear warping (Ashburner and Friston, 2005). While this procedure is less-influenced by the prior tissue probability maps, we decided to construct our own reference data due to the substantial differences between paediatric and adult brains and the issues that arise when processing paediatric imaging data based on inappropriate reference data (Wilke et al., 2002, 2003a). Consequently, in order to ensure appropriate processing of the input images, we first proceeded to construct custom reference data for segmentation and spatial normalization (Good et al., 2001). The creation of the customized priors generally follows the approach taken by Good et al. (2001) and is described in full detail in the Supplementary material. The normalization step was implemented with and without modulation to assess grey matter volume and grey matter concentration, respectively. The modulation step consists of multiplying (or modulating) the spatially normalized grey matter by its relative volume before and after spatial normalization; this compensates for the effect of non-linear spatial normalization, which might expand or shrink the volume of certain brain regions, and preserves the volume of grey matter within a voxel (Mechelli et al., 2005). In this study, we analysed both modulated and unmodulated data in order to examine the differences in volume and concentration, respectively. Differences in grey matter concentration were also reported in this study so that its results could be compared with the results of a previous VBM study of adults with psychopathy that reported grey matter concentration (see de Oliveira-Souza et al., 2008). All images were written out to 1 × 1 × 1 mm isotropic voxel in standard anatomical space (Montreal Neurological Institute, http://www.bic.mni.mcgill.ca/brainweb). Both modulated and unmodulated images were convolved with an 8 mm Full-Width at Half-Maximum Gaussian kernel. While other authors have employed larger kernels e.g. 12 mm (Tiihonen et al., 2008), this kernel was selected because it is consistent with one other VBM analysis of children (Sterzer et al., 2007), and because the a priori regions of interests included both cortical structures and a smaller subcortical structure (i.e. the amygdala), which is relevant due to the matched filter theorem (Jones et al., 2005).

Statistical analyses

Employing the framework of the General Linear Model, we performed statistical group analysis on a voxel-by-voxel basis. Regionally specific between-group differences in grey matter concentration and grey matter volume were then assessed using analysis of covariance with global grey matter, Full Scale Intelligence Quotient (FSIQ), and hyperactivity–inattention symptoms as covariates of no interest. For each contrast, statistical parametric maps were computed on a voxel-by-voxel basis to test for morphological differences between groups. Consistent with previous VBM analyses, results for the whole-brain analysis were considered significant at the threshold of P < 0.001, uncorrected for multiple comparisons, using an extent threshold of 100 contiguous voxels (equivalent to 0.8 cm3) to protect against type I error (Ananth et al., 2002; Farrow et al., 2005; Schmitz et al., 2006; de Oliveira-Souza et al., 2008). In addition to the whole-brain analysis, using the SPM5 tool we applied a small volume correction in the a priori regions of interest using a threshold of P < 0.05 after False Discovery Rate [FDR; (Genovese et al., 2002)] correction for multiple comparisons. Following de Oliveira-Souza et al. (2008), the radius of the spheres for the small volume correction was 6 mm in the ventrolateral/orbitofrontal cortex and anterior insula and 7 mm in the amygdala and the anterior cingulate cortex.

Results

Demographic and behavioural characteristics

As presented in Table 1, boys with callous–unemotional conduct problems presented higher levels of conduct problems, a lower FSIQ than the typically developing boys. However, they did not differ from the typically developing boys on the Emotional Symptoms subscale of the SDQ, which provides a screen for anxiety and mood disorders. In addition, although both groups had lower than average birth weights in line with their twin status, they did not differ from each other on birth weight.

View this table:
Table 1

Demographic and behavioural characteristics of the participant groups

Typically developing (n = 25)Conduct problems and Callous–unemotional traits (n = 23)P-value
Mean (SD)Mean (SD)
Age (in months)141.0 (11.8)139.3 (8.1)0.55a
FSIQ106.9 (10.59)95.4 (10.59)<0.001
APSD—Callous-unemotional traits3.0 (0.96)7.9 (0.96)<0.001
SDQ—Conduct Problem symptoms0.8 (0.85)4.4 (1.48)<0.001
SDQ—Hyperactivity symptoms3.4 (2.68)7.0 (2.46)<0.001
SDQ—Emotional symptoms1.8 (1.91)2.1 (1.77)0.54
Birth weight (in grams)2552.75 (589.80)2567.88 (578.84)0.93
  • a Student t-test.

  • APSD = Antisocial Process Screening Device; FSIQ = Full Scale Intelligence Quotient; SDQ = Strengths and Difficulties Questionnaire.

Regions of interest

Boys with callous–unemotional traits exhibited increased grey matter concentration in the posterior medial orbitofrontal cortex, dorsal and rostral anterior cingulate cortices (Fig. 1 and Table 2) as compared with the typically developing boys. In the other two ROIs—the anterior insula and the amygdala—no significant group differences were observed. Boys with callous–unemotional conduct problems did not exhibit decreased or increased grey matter volume relative to the typically developing boys. To explore whether the discrepancies between the results of grey matter concentration and grey matter volume analyses were due to a statistical threshold effect or whether they represented real differences, the statistical threshold was lowered (P < 0.005, uncorrected). This revealed that, consistent with the grey matter concentration results, boys with callous–unemotional conduct problems exhibited increased grey matter volume in the posterior medial orbitofrontal cortex (x, y, z = 2, 18, −20) and dorsal anterior cingulate cortex (x, y, z = 14, 27, 20); however no trend was detected in the rostral anterior cingulate cortex.

Figure 1

Foci of significant greater grey matter concentration among the boys with conduct problems and callous–unemotional traits (n = 23) relative to the typically developing boys (n = 25) in a priori regions of interest are highlighted in yellow (P < 0.001, uncorrected for multiple comparisons for the whole brain). The coloured bar in the centre indicates the z-scores. Significant clusters of concentration abnormalities are shown on the (A) medial orbitofrontal cortex; (B) right dorsal anterior cingulate cortex; (C) left rostral anterior cingulate cortex; and (D) left dorsal anterior cingulate cortex. The coordinates of voxels of maximal statistical significance in each of these clusters, as well as their size and peak z-scores, are provided in Table 2.

View this table:
Table 2

Results of the a priori region of interest analyses showing higher grey matter concentration for boys with conduct problems and callous–unemotional traits compared with the typically developing boys

Anatomical region and Brodmann areasMNI coordinates (x, y, z)Cluster sizeP (corr)aP (unc)Z
Posterior medial OFC (BA 11)320−195320.001<0.0014.29
Right dorsal ACC (BA 24)1029222740.005<0.0013.77
Left rostral ACC (BA 32)−342121680.008<0.0013.53
Left dorsal ACC (BA 24)−72404410.009<0.0013.45
  • a FDR correction for multiple comparison within regions of interest (see Methods and materials section).

  • OFC = orbitofrontal cortex; ACC = anterior cingulate cortex; MNI = Montreal Neurological Institute.

Whole-brain analyses

Other brain regions outside the regions of interest such as the superior parietal lobule, superior temporal gyrus bilaterally, postcentral gyrus, superior frontal gyrus, left posterior hippocampus and cuneus were larger among the boys with callous–unemotional conduct problems than the typically developing boys, both in grey matter volume and grey matter concentration (Table 3). The boys with callous–unemotional conduct problems, as compared with the typically developing boys, displayed increased grey matter concentration in the cerebellum, parahippocampal gyrus, insula and posterior cingulate cortex and increased grey matter volume in the middle frontal gyrus. Again, to explore whether the discrepancies between the results of grey matter concentration and grey matter volume analyses were due to the statistical threshold or whether they represented real differences, we lowered the statistical threshold (P < 0.005, uncorrected). This revealed that boys with callous–unemotional conduct problems had increased grey matter volume in the insula (x, y, z = 43, −10, 11) and the posterior hippocampus (x, y, z = −32, −39, −2) and increased grey matter concentration in the middle frontal gyrus (x, y, z = 40, 8, 60 and x, y, z = 44, 35, 42): no trend was detected in the cerebellum, parahippocampal gyrus and posterior cingulate cortex. No regions of reduced grey matter concentration or grey matter volume were detected among the boys with callous–unemotional conduct problems as compared with the typically developing boys.

View this table:
Table 3

Results of whole-brain analyses showing higher grey matter volume and grey matter concentration (uncorrected P < 0.001) for boys with conduct problems and callous–unemotional traits compared with the typically developing boys

Anatomical region and Brodmann AreasMNI coordinates (x, y, z)Cluster sizeP (unc)Z
Volume
    Left superior parietal lobule (BA 7)−37−6357569<0.0014.44
    Left superior temporal gyrus (BA 38)−161−41473<0.0014.02
    Right superior frontal gyrus (BA 8)171546234<0.0013.99
    Right postcentral gyrus (BA 3)66−1632756<0.0013.97
    Right precentral gyrus (BA 4)48−923107<0.0013.76
    Right superior temporal gyrus (BA 22)67−4010158<0.0013.73
    Right superior temporal gyrus (BA 38)2119−341090<0.0013.18
    Right uncus (BA 28)10−1−29<0.0013.66
    Right middle frontal gyrus (BA 6)40960135<0.0013.57
    Right middle frontal gyrus (BA 9)433542162<0.0013.51
    Left posterior hippocampus−32−39−2111<0.0013.41
    Right cuneus (BA 19)4−9324219<0.0013.37
Concentration
    Left cerebellum−2−69−26696<0.0014.47
    Left superior temporal gyrus/uncus (BA 38/36)−221−391992<0.0014.35
    Right calcarine sulcus1−76−11395<0.0014.29
    Left parahippocampal gyrus (BA 30)−16−34−9318<0.0014.06
        (BA 35)−21−38−22<0.0013.92
    Right uncus (BA 28)17−3−294111<0.0014.04
    Right superior temporal gyrus (BA 38)2715−34<0.0013.86
    Right superior frontal gyrus (BA 8)171645242<0.0013.91
    Right cuneus (BA 19/18)3−8425376<0.0013.91
    Left inferior temporal gyrus (BA 20)−47−10−37395<0.0013.83
    Right insula (BA 13)44−812342<0.0013.77
    Left precuneus (BA 7)−19−6645213<0.0013.74
    Left posterior hippocampus−28−38−4206<0.0013.70
    Left superior parietal lobule (BA 7)−38−6459397<0.0013.67
    Right intraparietal sulcus27−6343128<0.0013.67
    Left posterior cingulate cortex (BA 30)−8−598266<0.0013.61
    Left superior parietal lobule (BA 7)−29−5359127<0.0013.60
    Left inferior parietal lobule (BA 40)−40−3846113<0.0013.60
    Right postcentral gyrus49−823135<0.0013.53
    Left posterior cingulate cortex (BA 24)−1−2539165<0.0013.52
    Right superior temporal gyrus (BA 22)58−24−1203<0.0013.43
  • Text in bold indicates regions characterized by both higher grey matter volume and higher grey matter concentration.

Post hoc analyses

As our groups were closely matched in their age ranges, we investigated the possibility that the grey matter concentration differences in the regions of interest, where callous–unemotional conduct problems boys showed larger grey matter concentration than typically developing boys, might reflect a maturational lag. Lobar analyses document loss of grey matter in the frontal cortex starting between the ages 10–11 years in boys (Lenroot et al., 2007). The age range in both of our study groups comprised 10–13.3 years, thus covering a period of time where grey matter loss in the frontal cortex should begin. Analyses were conducted with the SPSS package 15 for Windows (SPSS) to examine the effect of age, group and the interaction of age by group for the two regions of interest in which grey matter concentration were significantly different for the callous–unemotional conduct problems and typically developing boys. We conducted regression analyses in which group (typically developing; callous–unemotional conduct problems), age (standardized) and the interaction term (group × age) were entered simultaneously as predictor variables of grey matter concentration in the orbitofrontal cortex, the left dorsal anterior cingulate cortex, the right dorsal anterior cingulate cortex and the rostral anterior cingulate cortex. Two potentially interesting results emerged. For orbitofrontal cortex grey matter concentration, there was a significant main effect of group (β = 0.39, P = 0.002), age (β = −0.52, P = 0.001), and a significant group by age interaction (β = 0.32, P = 0.035). As can be seen in Fig. 2A, typically developing boys showed the expected pattern of age-related grey matter loss (r = −0.57, P = 0.003), whereas among boys with callous–unemotional conduct problems grey matter concentration changed little with age (r = 0.06, P = 0.78). For the grey matter concentration in the left dorsal anterior cingulate cortex there was a significant main effect of group (β = 0.28, P = 0.047), age (β = −0.33, P = 0.046), and a significant group by age interaction (β = 0.39, P = 0.02). As can be seen in Fig. 2B, typically developing boys showed the expected pattern of age related grey matter loss (r = −0.43, P = 0.03), whereas among boys with callous–unemotional conduct problems grey matter concentration increased with age (r = 0.29, P = 0.19). As an additional check, total grey matter volume was entered as a regressor into the analyses. The group by age interaction in the orbitofrontal cortex became only marginally significant following this step (β = 0.27, P = 0.073), but the group by age interaction in the left dorsal anterior cingulate cortex remained statistically significant (β = 0.39, P = 0.026). The regression analysis for the right dorsal anterior cingulate cortex showed no significant effects of age or of age by group. In the rostral anterior cingulate cortex, there was a significant effect of age (β = −0.51, P = 0.002), but no significant age by group interaction.

Figure 2

Scatter plots showing for the typically developing (TD) boys and the callous–unemotional conduct problems (CP/CU+) boys the correlation between age and grey matter concentration in: (A) the medial orbitofrontal cortex and (B) the left dorsal anterior cingulate cortex. The local maxima of grey matter were extracted from the unmodulated data and correlation analyses were performed in the SPSS package 15 for Windows (SPSS).

Discussion

This article reported the first sMRI study of boys characterized by callous–unemotional conduct problems. Increased grey matter concentration was observed in boys with callous–unemotional conduct problems as compared with typically developing boys in two of the four regions of interest: the medial orbitofrontal cortex and the anterior cingulate cortex (both rostrally and dorsally). Whole-brain analyses also confirmed grey matter concentration and grey matter volume increases in several other brain areas. These group differences emerged after statistically controlling for FSIQ and hyperactivity–inattention symptoms. No regions of decreased grey matter concentration or grey matter volume were detected in boys with callous–unemotional conduct problems.

In line with two previous VBM studies of adults with high psychopathy scores (de Oliveira-Souza et al., 2008; Tiihonen et al., 2008), in the present study boys with callous–unemotional conduct problems exhibited structural differences within the orbitofrontal cortex, as compared with typically developing boys. However, unlike adult psychopaths who exhibited decreases in the orbitofrontal cortex grey matter concentration and grey matter volume, boys with callous–unemotional conduct problems showed increased grey matter concentration in the orbitofrontal cortex. Abnormal orbitofrontal cortex activation has been observed in offenders with psychopathy during classical aversive conditioning (Veit et al., 2002; Birbaumer et al., 2005), in individuals from the community with high psychopathy scores participating in social interaction and affective paradigms (Gordon et al., 2004; Rilling et al., 2006; Lotze et al., 2007), and in children with callous–unemotional conduct problems during punished reversal errors in a probabilistic response reversal task (Finger et al., 2008). In addition, children with callous–unemotional conduct problems, like adults with psychopathy (Mitchell et al., 2002; Budhani et al., 2006), are impaired, albeit to a lesser extent, in reversal learning and behavioural extinction paradigms (Fisher and Blair, 1998; Blair et al., 2001; Budhani and Blair, 2005) that measure two types of instrumental learning mediated by the orbitofrontal cortex (Kringelbach and Rolls, 2004; Murray et al., 2007). Finally, several models derived from human lesion studies and fMRI work in healthy individuals have identified the medial orbitofrontal cortex, among other brain regions, as one of the core regions for human moral cognition (Moll et al., 2005; Blair, 2007). Adults with psychopathy and children with callous–unemotional conduct problems are impaired on tasks assessing moral reasoning (Blair, 1995, 1997). Thus, our findings, in conjunction with those of two other VBM studies, in adults with high psychopathy scores (de Oliveira-Souza et al., 2008; Tiihonen et al., 2008), suggest that structural differences in the medial orbitofrontal cortex may be associated with the impaired moral reasoning seen in adults with psychopathy and children with callous–unemotional conduct problems. As shown by the post hoc analysis, however, the group difference observed in the present study appears to result from a lack of age-related decreases in the orbitofrontal cortex grey matter concentration among the callous–unemotional conduct problems boys.

The boys with callous–unemotional conduct problems also presented with increased grey matter concentration in the dorsal and rostral anterior cingulate cortices, two regions involved in the regulation of cognitive and emotional behaviour, respectively (Bush et al., 2000; Singer et al., 2004). Previous fMRI studies of adults with psychopathy have also implicated these areas (Kiehl et al., 2001; Veit et al., 2002; Muller et al., 2003; Birbaumer et al., 2005). Abnormal activation within the rostral anterior cingulate cortex has been observed among offenders with a diagnosis of psychopathy during classical aversive conditioning (Veit et al., 2002; Birbaumer et al., 2005), an affective memory task (Kiehl et al., 2001), and when viewing negative affect-laden pictures (Muller et al., 2003). In the latter study, abnormal activation was also observed in the dorsal anterior cingulate cortex (Muller et al., 2003). In healthy adults, the rostral anterior cingulate cortex is closely connected with the amygdala and is directly involved in affective processing of aversive stimuli and the experience of empathy (Bush et al., 2000; Singer et al., 2004; Schunck et al., 2008). Thus, structural abnormalities within the rostral anterior cingulate cortex, and connected brain regions (e.g. dorsal anterior cingulate cortex), might partially explain the deficits in the processing of aversive stimuli and empathy that characterize adults with psychopathy and children with callous–unemotional conduct problems (Blair et al., 2006; Patrick, 2006). The results of the post hoc analysis of the left dorsal anterior cingulate cortex showed that the group difference observed in the left dorsal grey matter concentration is due to an increase in grey matter concentration in the callous–unemotional conduct problems boys and the expected grey matter concentration decrease among the typically developing boys. While this analysis was based on a relatively small number of boys at each age, it suggests a distinct pattern of brain development in this region among the boys with callous–unemotional conduct problems.

Contrary to our hypothesis, we did not find structural differences in two of our regions of interest: the amygdala and the anterior insula. Although, two previous VBM studies comparing boys with conduct problems and typically developing boys have observed grey matter volume reduction in the left amygdala in those with conduct problems (Sterzer et al., 2007; Huebner et al., 2008), it must be noted that, to date, no sMRI study in adults with psychopathy has found evidence of structural abnormalities in the amygdala. The absence of structural differences in the amygdala does not preclude functional differences, which have been reported in neuropsychological and fMRI studies of adults with psychopathy and children with callous–unemotional conduct problems (e.g. Blair et al., 2006; Jones et al., 2009). As for the insula, we did find that boys with callous–unemotional conduct problems, as compared with the typically developing boys, had increased grey matter volume in the right insula, but this difference lied in the posterior section and not in the anterior insula, the region of interest.

Since this was the first sMRI study of children with callous–unemotional conduct problems, differences in brain regions not included in our a priori regions of interest were examined. No regions of decreased grey matter volume or grey matter concentration were identified, but several regions of increased grey matter volume and grey matter concentration were seen across the brain. For example, boys with callous–unemotional conduct problems presented increased grey matter concentration in the left posterior hippocampus and posterior cingulate cortex—two brain regions involved in long-term memory and moral reasoning, respectively. Although not part of the critical neural circuit specified in the neurocognitive model of developmental psychopathy proposed by Blair et al. (2005), both hippocampus and posterior cingulate cortex differences have been reported in previous sMRI and fMRI studies of adult psychopathy (Kiehl et al., 2001; Laakso et al., 2001; Muller et al., 2003; Raine et al., 2004; de Oliveira-Souza et al., 2008). Boys with callous–unemotional conduct problems also presented increased grey matter concentration in the cerebellum. This finding is in line with the reported increases of grey matter volume in adult violent offenders with psychopathy (Tiihonen et al., 2008). Generally, the cerebellum is best known for its role in the coordination of motor behaviour (Bastian et al., 1999). However, there is growing evidence that it plays a crucial role in emotion processing and fear conditioning via its connection with limbic structures (e.g. amygdala and hippocampus, among others) and the hypothalamic-pituitary-adrenal axis (Schutter and van Honk, 2005). A recent fMRI study found that, in comparison to healthy adults, adults with high psychopathy scores showed less cerebellar activity to fearful faces (Deeley et al., 2006). However, this finding should be interpreted with caution due to the small sample size and poor matching of the groups.

Two previous VBM studies of children with conduct problems, in which callous–unemotional traits were not measured, found decreased grey matter volume instead of increased grey matter volume in several brain areas implicated in our study (Sterzer et al., 2007; Huebner et al., 2008). For example, Huebner et al. (2008) found decreased orbitofrontal cortex grey matter volume in children with conduct problems in comparison to typically developing boys, while Sterzer et al. (2007) did not find any differences between boys with conduct problems and typically developing boys. This different pattern of results highlights the importance of sub-typing children with conduct problems when studying the neurobiological correlates of conduct problems (Frick and Marsee, 2006; Moffitt et al., 2008).

Some methodological limitations of this study must be considered when interpreting the results. First, while our VBM analyses were based on a larger sample than previous studies, 48 participants constitutes a relatively small cohort for a VBM study. Second, our results may not generalize to girls. Third, the post hoc analyses examining the effect of group, age and age by group were based on cross-sectional and as such should be considered with caution. Our hypothesis that the increased grey matter in the boys with callous–unemotional conduct problems reflects a maturational lag will have to be tested in a longitudinal study. Fourth, our group were not defined by an expert rated questionnaire or a structured clinical psychiatric interview that would have provided more detailed information about psychopathological features in our participants. The study is characterized, however, by several strengths. First, in contrast to some previous VBM studies, analyses were controlled for the effects of cognitive abilities and hyperactivity–inattention symptoms. Second, customized probability maps for the segmentation of structural images were created. Third, results were not confounded by factors such as treatment with medication or institutionalization. Although substance abuse was not formally assessed within our sample, a community sampling of this age group is unlikely to include children with prolonged substance abuse problems (Fuller, 2007). A fourth strength of the present study was the recruitment of participants from a non-clinical sample thus allowing wider generalization of results than is possible from studies of clinical samples that often suffer from selection bias and co-morbidity issues.

In conclusion, this first sMRI study to compare boys with callous–unemotional conduct problems to typically developing boys provides preliminary evidence of increased grey matter concentration in the two critical regions of interest, medial orbitofrontal cortex and anterior cingulate cortex (rostrally and dorsally). Whole-brain analysis implicated grey matter concentration and grey matter volume increases in additional brain areas including temporal lobes and the cerebellum. Given recent evidence of grey matter volume loss during brain development, the findings may be interpreted to suggest a delay in cortical maturation. These grey matter increases in boys with callous–unemotional conduct problems were observed in areas implicated in decision-making, morality, and empathy. Deficits in decision making, morality and empathy have been shown to characterize children, adolescents and adults with psychopathic traits. Our data points towards the candidate neurobiological underpinnings of these observable psychopathological features.

Supplementary material

Supplementary material is available at Brain online.

Funding

Medical Research Council (UK) (G0401170 to E.V.); Department of Health Forensic Mental Health Programme (MRD 12-73 to E.V.); Medical Research Council PhD studentship (to S.A.D.B); National Institute for Health Research Postdoctoral Fellowship Award (to K.R.L.); Royal Society Wolfson Merit Award (to S.H.); National Institute for Health Research Biomedical Research Centre South London (to K.R.L. and S.H.); Maudsley NHS Foundation Trust/Institute of Psychiatry (King's College London) (to K.R.L. and S.H.).

Acknowledgements

We thank the TEDS children, parents and teachers participating in this research. We would also like to thank Prof. Robert Plomin, Mrs Patricia Busfield, Mr Andrew McMillan and staff at the Centre for Neuroimaging Sciences, Institute of Psychiatry for their generous help with this research.

Footnotes

  • Abbreviations:
    Abbreviations:
    fMRI
    functional magnetic resonance imaging
    FSIQ
    Full Scale Intelligence Quotient
    SDQ
    Strengths and Difficulties Questionnaire
    sMRI
    structural magnetic resonance imaging
    VBM
    voxel-based morphometry

References

View Abstract