Identifying shared and distinct alterations in brain structure across internalising and externalising disorders through the ENIGMA consortium

: (Alternative Format Thesis)

Abstract

Internalising and externalising disorders often first manifest during childhood and adolescence, which increases a person’s risk of being diagnosed with further psychiatric conditions in adulthood. To better understand psychopathology in youth, and in turn identify potential treatment targets, it is crucial to examine the neurobiological patterns associated with these psychiatric problems. While most neuroimaging studies adopt case-control designs which conceptualise mental health conditions as distinct categories (e.g., based on DSM or ICD diagnoses), this approach is challenged by the substantial heterogeneity found within diagnostic groups, and the overlapping symptoms and patterns of comorbidity seen between disorders. Accordingly, psychiatric research is seeing a shift towards ‘transdiagnostic’ approaches which cut across diagnostic boundaries. With increasing challenges to the categorical diagnostic paradigm, these approaches can offer new insight into neurobiology which might be shared or distinct across youth psychopathology, with implications for future mental health research and our diagnostic systems.

However, the identification of shared and distinct brain alterations across psychiatric disorders is currently hindered by several limitations. These include a reliance on small sample sizes, insufficient consideration of participant characteristics (e.g., sex, childhood maltreatment history), differences in data acquisition and analysis methods, and limited research with child and adolescent samples. Furthermore, although dimensional approaches have the potential to provide complementary evidence on psychopathology-brain relationships, the existing evidence comes from US-based samples with low levels of psychopathology. To address these limitations, the aims of this thesis were 1) identify shared and distinct structural brain alterations among internalising and externalising disorders in youth, 2) identify associations between brain structure and dimensional psychopathology in an international sample including clinical and high-risk youth, and 3) increase understanding of the impact of participant characteristics on observed brain alterations, such as age and sex, and previously unmeasured sources of heterogeneity.

One challenge of identifying robust disorder-specific associations with brain structure is the substantial heterogeneity which exists within each diagnostic category. In our first study, we investigated the impact of childhood maltreatment on white matter microstructural alterations in Conduct Disorder (CD) to provide a test of the ecophenotype hypothesis that maltreated and non-maltreated individuals with the same disorder may differ at a neurobiological level. Using diffusion tensor imaging data from the European multi-site FemNAT-CD study, we found that youth with CD (n=100) showed higher fractional anisotropy in the corpus callosum than controls (n=169), but when divided based on their maltreatment histories, only the non- maltreated CD subgroup (n=61) displayed these effects. The maltreated CD subgroup (n=39), on the other hand, showed no alterations relative to controls, but exhibited higher right superior longitudinal fasciculus axial diffusivity compared to their non-maltreated counterparts. This study suggests that, despite sharing the same diagnosis, maltreated and non-maltreated forms of CD are associated with distinct white matter microstructural alterations. Maltreatment experiences may therefore be an important source of heterogeneity which should be considered in future neuroimaging studies.

In Study 2, we pooled data from 68 international samples within the ENIGMA consortium to conduct the first mega-analytic comparison of cortical and subcortical structure between youth with Attention-Deficit/Hyperactivity Disorder (ADHD; n=1,317), CD (n=1,172), anxiety disorders (n=1,044), or depression (n=504), along with healthy controls (n=4,743). All four disorders were characterised by lower surface area in the insula, entorhinal cortex, and middle temporal gyrus, and lower amygdala volume. Externalising-specific (i.e., ADHD and CD-related) reductions in fronto-parietal surface area were observed, but internalising- specific associations were not found. Some evidence of ‘disorder-specific’ alterations were identified for ADHD, CD, and anxiety disorders, but there were no cases where a diagnostic group significantly differed both from controls and the other three disorders. We also identified age-by-diagnosis interactions in a few regions for ADHD, anxiety, and depression, but no sex-by-diagnosis interactions were found. These results demonstrate that youth with different internalising and externalising disorders show common alterations in cortical surface area and subcortical volume relative to healthy controls. As well as extending our understanding of the neurobiological basis of youth psychopathology, these findings provide further evidence for transdiagnostic associations across conventional diagnostic boundaries.

Finally, in Study 3 we collated data from 11,207 children and adolescents from 16 international case-control, high-risk, and community-/population-based cohorts across the ENIGMA consortium. We used a bifactor model to estimate specific dimensions (internalising and externalising), which capture variance shared between related symptoms, and a general psychopathology dimension, which represents variance which is shared across different forms of psychopathology, and then investigated associations between each of these dimensions and brain structure. General psychopathology was negatively associated with surface area in 15 (out of 34) regions, hippocampal volume, and total surface area and intracranial volume. The externalising dimension was associated with lower surface area in 15 regions across the temporal, frontal, parietal, and occipital lobes, as well as lower total surface area and intracranial volume. Conversely, the internalising dimension was negatively associated with surface area in just one region (the paracentral gyrus), and this was not robust to sensitivity analyses. No interactions with either age or sex were observed. Overall, these findings suggest that general psychopathology is associated with widespread brain alterations in youth, and provide evidence for distinct associations with externalising psychopathology, but not internalising psychopathology.

Taken together, the current thesis substantially advances our understanding of the neurobiological basis of youth psychopathology. Our results suggest that common brain structural alterations are associated with both internalising and externalising disorders in youth, with little evidence for disorder-specific effects. We identified further issues with categorical psychiatric diagnoses by showing that maltreatment-related CD may represent a distinct variant of the disorder. Finally, we identified widespread associations between general psychopathology and brain structure, suggesting that common variance across disorders may have confounded previous case-control designs. As well as identifying key brain regions and networks which may be implicated across internalising and externalising psychopathology in youth, these findings challenge the prevailing use of case-control designs in psychiatric neuroimaging research by suggesting that brain structural alterations do not respect the diagnostic boundaries of our current psychiatric nosology. This may have further implications for our conceptualisation of mental health problems, as these findings present additional challenges to the current categorical psychiatric nosology.

Date of Award	19 Feb 2025
Original language	English
Awarding Institution	University of Bath
Supervisor	Graeme Fairchild (Supervisor) & Esther Walton (Supervisor)