The structure, function and connectivity of the human brain are topographically organized. This topographic organization provides profound insight into cortical information processing, representation of mental states, and accounts for individual variation in behavioral traits and cognition. Whereas classical models of brain topography focus on distinct cortical patches defined by discrete boundaries, contemporary evidence from neuroimaging suggests that topographic variation may be better conceptualized in terms of a set of continuous gradients of gradual change that overlap in space. My work aims to reconcile these two conceptualizations of brain topography, particularly with respect to the insular cortex, a topographically complex and functionally heterogeneous cortical lobe whose organization has remained disputed for centuries. Using modern functional neuroimaging techniques, I showed that the insula’s topography is best conceptualized as a continuum of gradual change oriented along an anterior-posterior axis. I found that individual variation in the insula’s functional topography associates with human cognitive and emotional traits as well as somatosensory functions. Having characterized the functional architecture of the insular cortex in healthy adults, my next aim was to investigate whether neuropsychiatric illness is associated with alterations in the insula’s functional organization. To this end, I compared the insula’s functional connectivity gradients between individuals with schizophrenia and healthy controls. I found evidence suggesting subtle reorganization of the insula’s functional topography in schizophrenia. In particular, the connectivity profile along the anterior-posterior topographic axis of the insular cortex was altered and less differentiated in individuals with schizophrenia. I showed that the extent of reorganization of the insula’s functional topography significantly associates with the severity of clinical symptoms, particularly negative symptoms of psychosis and intellectual impairment. Finally, I applied the new methodology that I developed to map the insula’s topography to study other brain regions, including the entire human subcortex. This unveiled four hierarchical scales of subcortical organization, recapitulating well-known anatomical nuclei at the coarsest scale and delineating 27 new bilateral regions at the finest. Based on this work, I developed a new MRI subcortical atlas to enable holistic connectome mapping and characterization of cortico-subcortical circuits. The new subcortex atlas was personalized to account for connectivity differences across individuals and utilized to uncover a reproducible association between subcortical functional connectivity and tobacco use. Overall, this thesis provides fundamental insight into the functional organization of the human insular cortex and subcortex in health and neuropsychiatric illness, particularly focusing on the distinction between classical models of topographic variation based on discrete regions and contemporary representations involving continuous gradients. The new methodology that I developed is not limited to the insular cortex and the subcortex and can be applied to other cortical and subcortical regions in humans as well as other species.

Introduction and Aims Generalised anxiety disorder (GAD) comprises a debilitating cluster of psychological and physiological symptoms that markedly impairs quality of life. GAD is characterised by hallmark cognitions of persistent worry and anticipatory anxiety. Evidence exists for dysregulation in excitatory/inhibitory neurobiological pathways in prefrontal and limbic brain regions, with the dorsal anterior cingulate cortex an area of particular interest. However, limited research exists assessing regional activations and the role of metabolites such as gamma-aminobutyric acid in these regions, nor modulations as a function of treatment. The aim of the thesis was to investigate the functional and metabolic features of this region, and to assess the role of neuroimaging biomarkers of anxiolytic treatment response. Methods Two investigations were conducted utilising structural features of the region of interest: task-based functional magnetic resonance blood oxygen level-dependant signal activation and GABA levels via magnetic resonance spectroscopy together with relevant psychometric and psychiatric measures. The first study was a cross-sectional investigation undertaken to compare neuroimaging biomarkers in 41 participants with GAD with 35 healthy control participants. The second study was an 8-week RCT sub-study involving 41 participants randomised to either daily 240mg of kavalactones Piper methysticum (Kava) extract or a matching placebo. This proof-of-concept study assessed the aforementioned outcomes and whether these markers signal the plant’s anxiolytic activity. Results The results of the first investigation did not reveal group differences in GABA level (p = .302). The relationship between GABA and anxiety severity was different for each group; a significant positive correlation in GAD (e.g., HAM-A, p = .018) and a negative correlation in healthy controls (e.g., trait anxiety, p = .019). The functional task was successful in eliciting regional BOLD signal differences between valent congruency conditions. Two regions exhibited significant group differences (at p < .05), showing hyperactivation in GAD and reduced activation in healthy controls. In both groups BOLD signal significantly predicted severity of state anxiety (GAD p = .027; HC p = .041). Gender, age, and comorbidity in the GAD group also influenced the biomarker-anxiety relationships. The results of the second study showed that Kava treatment was associated with a reduction to GABA levels at eight weeks (p = .049). The treatment was not associated with anxiety symptom, nor fMRI signal change, measured at eight weeks. Discussion This research investigated regional brain properties in GAD for biomarker utility, before testing them in a ‘proof of concept’ study using the purported anxiolytic agent, Kava. Metabolic and functional data were successful in producing differences in the dorsal ACC that could be (if replicated in a larger study) be utilised as biomarkers to aid in the management of GAD symptoms. Limitations of the studies were small sample sizes, GABA signal quality and equivocal toolbox results. The neurobiological effects of Kava have not been directly studied using MRI imaging in humans. The findings of a reduction to GABA levels after treatment may potentially reflect a normalising of the GABA system similar to healthy control data observed in the first study. GAD is a prevalent psychiatric disorder that is under-diagnosed and under-treated. While a great deal of work is inherent in establishing biomarkers for clinical benefit, this research contributes MRI evidence of biological differences, and insight into the mechanisms of Kava, together with a translational rationale for the study of novel anxiolytics as potential GAD treatments. The outcomes and findings of this research fit well with the current affective disorder literature and exceed contemporary work in the field of GAD biomarker and treatment research.

Communication between neural elements underpins all aspects of brain functioning. Large-scale neural signalling unfolds atop the human connectome, the complex network that describes how gray matter regions are interconnected by white matter projections. The mechanisms governing the propagation and communication of signals across the connectome remain unknown. The main focus of this thesis is the investigation of network communication models aimed at elucidating how the anatomical substrate of nervous systems facilitates and constrains functional interactions between gray matter regions. To date, the vast majority of network neuroscience studies have assumed neural signalling occurs via topological shortest paths. This is reflected by the widespread use of graph measures such as global efficiency, betweenness centrality and the small world coefficient. In recent years, researchers have begun to question this assumption on the basis that communication via shortest paths is contingent on centralized knowledge of connectome topology, and thus may not be a biologically realistic signalling model. This has led to the exploration of decentralized strategies of network propagation. Most efforts in this direction are focused on diffusive communication, which typically models neural signalling from the perspective of random walk processes. While these approaches do not mandate knowledge assumptions about network organization, they fail to promote efficient and energetically frugal neural information transfer. Therefore, the literature on brain network communication models is currently concentrated on the opposing strategies of shortest path routing and diffusive communication. This thesis aims to reconsider this dichotomous state by investigating alternative brain network communication models. In Chapter 3, we explore the concept of navigation in mammalian connectomes. Navigation is a greedy routing strategy in which information is propagated based on the spatial positioning of brain regions. Using human, macaque and mouse brain networks, we provide evidence that connectome organization is conducive to decentralized efficient communication under navigation. Specifically, the combination of empirical connectome topology and geometry was necessary for successful network navigation, with disruptions to either attribute resulting in marked decreases of navigation efficiency. These findings suggest that brain network architecture may have evolved to facilitate efficient decentralized information transfer, and indicate a three-way relationship between topology, geometry and communication in nervous systems. Decentralized network communication models can be asymmetric. This means that the efficiency of signalling paths may vary depending on the direction of information flow. Importantly, this behaviour occurs even in undirected networks. This unexplored facet of network communication provides the opportunity to study directional patterns of signalling in the human structural connectome, in which all connections are considered bidirectional due to the inability of diffusion imaging to resolve axonal directionality. In Chapter 4, we develop the statistical framework of send-receive asymmetry and demonstrate that it contributes novel insight into large-scale neural signalling directionality. Crucially, this chapter provides cross-modal evidence for the utility of decentralized communication models by demonstrating a statistical association between send-receive asymmetry and the directionality of effective connectivity. Lastly, in Chapter 5 we perform a systematic evaluation of the main neural signalling models proposed in the network neuroscience literature. We evaluate models in terms of their (i) predictive utility of interindividual variation in human behaviour and (ii) structure-function coupling strength. We hypothesize that communication models performing better in these criteria may provide more parsimonious characterizations of information transfer mechanisms in the human brain. Importantly, we benchmark communication models against structural connectivity, and provide evidence that accounting for polysynaptic communication improves the behavioural and functional predictions derived from direct anatomical connections alone. Combining behavioral and functional results into a single ranking of communication models positioned navigation as the top model, suggesting that it may more faithfully recapitulate biological neural signalling patterns. The results in this chapter contribute to elucidating the relationship between human behaviour, functional connectivity and connectome communication. Collectively, the findings reported in this thesis further our knowledge of large-scale neural signalling, promoting a unified understanding of brain structure, function and communication.

Adolescent mental health research is a developing area. Existing treatment guidelines for adolescent mental health care in most western countries emphasise the role of inpatient care when needed. Inpatient units are the most widely used acute element of adolescent mental health services internationally. Yet little is known about inpatient units, their therapeutic operations, models of care and perceived helpfulness. Less is known about general adolescent inpatient units from the perspectives of adolescents, caregivers and clinicians. In order to address this gap in understanding an adolescent inpatient model of care in operation, a prospective mixed-methods approach was adopted. Health of the Nation Outcome Scales for Children and Adolescents (HoNOSCA) were collected, measuring global functioning at T1 (admission) and T2 (discharge). Semi-structured interviews were conducted with 16 adolescents and 12 caregivers at T1, T2 and T3 (six months post discharge). Qualitative data were first analysed thematically followed by a trajectory analysis. One-off semi-structured interviews were conducted with clinicians (n=10) and analysed using thematic analysis. Data collection began in May 2017 and ceased in October 2018. The majority of adolescents (n=72) were 16 years of age (26%), female (82%) and with a primary diagnosis of a mood disorder (57%). HoNOSCA data were completed by clinicians (n=57) and adolescents (n=56). Most adolescents improved at the time of discharge. Self-injury and emotional symptoms had greater reductions according to clinician and adolescent-self-ratings (p<0.01). Mean change (improvement) in HoNOSCA total score was 7.3 (SD 7.5) based on clinician ratings and 7.2 (SD 9.5) for adolescent-self-ratings. The mean length of stay was 28 days (SD 15.8). Interviews with clinicians resulted in the identification of three thematic features of the model of care relating to containment, engagement and therapy. These included; (a) an environment conducive to containment, (b) adolescent engagement through shared experiences and (c) dialectical behaviour therapy embedded culture. Adolescent and caregiver experiences described followed a recovery narrative consisting of three key phases which included, ‘waiting for help’ (T1), ‘help arrived’ (T2) and having ‘returned to regular life’ (T3). The overarching trajectory theme was a ‘winding road to recovery’. Findings revealed the admission was helpful for many young people who were on the winding road to recovery. These findings provide insights into the lived experiences from adolescents who have had an inpatient stay and their caregivers of an adolescent specific inpatient model of care. These findings should be used to improve clinical services and inform research aiming to articulate exemplary adolescent inpatient models of care. Furthermore, the findings provide guidance and practical information to commissioners, clinicians and policy makers implementing models of care. Finally, the detailed findings provide a foundation for planning inpatient care that is valued by clinicians, young people and their families.

The idea that receiving social support in the aftermath of a disaster is beneficial to survivors’ psychological outcomes is widely accepted by both researchers and practitioners. However previous studies assessing the association between received social support and psychological outcomes have produced mixed results. One limitation of these studies is that they only assessed the quantity of support received, but not its quality. However, current evidences suggested that receiving social support is not always beneficial, this can lead to both positive and negative consequences. This research was intended to examine the effects of postdisaster received social support on psychological outcomes, and to explore the determinants of recipients’ perception of social support interactions. In the context of a longitudinal mixed method design, three studies were conducted among survivors of the Lushan earthquake in China: a longitudinal quantitative study examining the impact of post-disaster received social support on posttraumatic stress symptoms and psychological distress; a longitudinal quantitative analysis examining the impact of post-disaster social support on posttraumatic growth; and a qualitative study describing survivors’ experiences of receiving postdisaster social support, and exploring what matters to recipients in their evaluation of these experiences. A convenience sample of Lushan earthquake survivors was invited to complete a questionnaire 7 months after the earthquake (n =199) and was followed up 31 months later (n =161). Face-to-face semi-structured interviews were conducted by the author among a sub-group of participants (n =11) in the follow-up survey. This thesis found that greater quality of social support received 7 months after disaster predicted lower levels of posttraumatic stress symptoms and psychological distress two years later, however quantity of received social support was not significant in predicting these two outcomes. The association between quantity of received social support and PTG was moderated by the quality of received social support. Specifically, for survivors who received high quality of post-disaster social support, greater amounts of support received facilitated their posttraumatic growth. However, for those survivors who received poor quality of post-disaster support, greater quantity of support impeded PTG. Social support recipients cared about features related to outcomes, interaction, provider and management. First, recipients cared about features of the outcomes of social support interactions; that is, whether the support that was received fulfilled their needs, and also whether the distribution of support was equal and fair. Second, they cared about whether the interaction goes smoothly and easily; whether the interaction was initiated by providers; whether the interaction was public and therefore witnessed by others; whether they were treated with dignity in the interaction; and whether they had opportunities to reciprocate. Third, characteristics of help providers, including whether they sincerely wanted to help them, and what they devoted in order to provide assistance, also matter. Lastly, the efficacy and appropriateness regarding the management of disaster relief resources was concerned by recipients.

Schizophrenia is a severe neuropsychiatric disorder, arising in adolescence and early adulthood and characterised by hallucinations, delusions, blunted affect and disorganised thought patterns. One of the most enduring features of schizophrenia and psychosis are structural brain deficits, whose pathophysiological mechanism is unknown. Accumulating evidence indicates that inflammation both peripherally and centrally in the form of increased activation of the brain’s immune cells, microglia, may be a potential cause of structural deficits in psychosis. The evidence is multi-faceted ranging from mouse models that demonstrate increased numbers of microglia, to clinical studies of patients with schizophrenia showing increased pro- inflammatory molecules within peripheral blood. However, there are still many questions that remain unanswered, including whether inflammation varies across stages of psychosis, whether it is related to structural brain deficits and symptomatology and how inflammation identified in schizophrenia relates to other candidate pathways implicated in psychosis. In this thesis a multi-disciplinary approach was adopted, considered appropriate to tackling the complexity of these questions. Firstly, to determine whether inflammation was associated with other candidate pathways implicated in psychosis, we conducted an animal study, utilising an mGluR5 KO mouse model of psychosis. Glutamate has been shown to influence neuroinflammation, with cellular studies demonstrating that mGluR5 can regulate microglial numbers and activation. At the time of conducting the study, there were no satisfactory mGluR5 PET ligands that enabled in vivo monitoring within the clinical population. The mGluR5 KO mouse had been shown to display neuropsychiatric endophenotypes related to schizophrenia and thereby offered an alternative approach to gaining further insight into the role of mGluR5 in neuroinflammation and how this may impact symptoms associated with psychosis. Our aim was to determine whether neuroinflammation, in the form of increased microglial numbers and activation was present in the mGluR5 KO mouse model thereby giving further insight as to the potential interaction of the glutamatergic system and in particular, mGluR5, with microglial homeostasis. Secondly, we aimed to determine whether peripheral inflammation was related to brain structure and clinical symptomatology. This was executed by conducting two clinical studies, that examined peripheral pro- and anti-inflammatory cytokines and complement proteins in relation to brain regional thickness and volume measurements. We used a multiplex enzyme linked immune-absorbent assays (ELISA) in serum to quantify peripheral cytokines and complement proteins across various stages of psychosis ranging from those at ultra-high risk of psychosis (UHR), to individuals experiencing their first episode (FEP) and subjects with chronic schizophrenia. We sought to determine whether circulating cytokine and complement protein levels were associated with clinical symptomatology and measurements of thickness and brain volume detected using structural magnetic resonance imaging (MRI). This thesis aimed to investigate: 1) whether mGluR5 KO mice, which demonstrate phenotypic features of schizophrenia displayed neuroinflammation in the form of increased microglial numbers when compared to their wildtype littermates 2) whether there was a relationship between cytokine or complement proteins and structural brain measurements across UHR, FEP and chronic schizophrenia 3) whether peripheral inflammatory markers (cytokine or complement proteins) were increased or decreased across stages of psychosis and examine their relationship with clinical symptoms. We found that mGluR5 KO mice have increased microglial numbers compared to WT. This agreed with our hypothesis that animals lacking mGluR5, would show higher rates of inflammation in the brain, in accordance with an anti-inflammatory effect of increased mGluR5 signalling and the psychotic endophenotype of these mice. Our findings indicate that mGluR5 may affect microglial homeostasis in the context of neurodevelopment and may impact on psychosis related behaviours exhibited by mGluR5 KO mice. Secondly, our clinical studies showed that cytokines and complement proteins were related to several brain structures implicated in psychosis, including the frontal cortex and ventricles. We revealed a positive correlation between several anti-inflammatory cytokines such as IL4 and IL13 and increases in frontal cortical thickness, which was absent in patients with psychosis. Conversely, increases in pro-inflammatory cytokine IL5 were associated with decreases in whole brain volume in FEP individuals. Thirdly, we found that while peripheral cytokines did not differ significantly between patients and controls, complement proteins were elevated in UHR and chronic schizophrenia patients. While there were no associations between cytokine proteins and clinical symptoms, we identified a molecular pattern of increased C4 and decreased C3 protein, which was associated with increases in positive and negative symptoms. Taken together, the work of this thesis suggests that inflammation is present in psychosis both in the brain and peripherally but that this depends on the proteins and stage of illness examined. Moreover, we revealed that complement proteins C3 and C4 were associated with alterations in brain structure across the combined cohort; in the case of cytokines however, the positive association between elevated anti-inflammatory cytokines and increased frontal thickness was not preserved or reversed in patient groups, indicating a potential imbalance of pro- and anti- inflammatory cytokines may influence brain structure in psychosis. Finally, we have shown that peripheral inflammation in the form of cytokine and complement proteins, may influence both brain structure and clinical symptomatology, which provides fertile ground for future longitudinal exploration of neuroinflammation in schizophrenia and psychosis.

Schizophrenia is a chronic disabling disorder with complex multifactorial aetiology. It is associated with childhood adversity and glutamatergic genes, both of which contribute to brain development and cognition. However, the relationships between these factors are not fully understood and must still be elucidated. This thesis addresses gaps in understanding of this complex link. These findings will be informative for early identification and treatment of those with schizophrenia. Chapter one provides a conceptual framework for the models used in this thesis. A literature review on schizophrenia, childhood adversity, glutamatergic genes, brain development, and cognition is included. The links between these factors are described and the aims of the thesis are justified. Chapter two aimed to identify the association between metabotropic glutamate receptor 3 genetic variation and schizophrenia and explored potential population stratification. This meta-analysis study consisted of 14 single nucleotide polymorphisms of metabotropic glutamate receptor 3 from a total of 11318 schizophrenia cases, 13820 controls, and 486 parent proband trios. We found significant associations for three single nucleotide polymorphisms. We also found evidence for population stratification in that the risk allele was dependent on the population under study. These findings support the genome wide-implicated link between metabotropic glutamate receptor 3 genetic variation and schizophrenia risk, and further support the notion that alleles conferring this risk may be population specific. Chapter three aimed to examine the extent to which the association between childhood adversity and cognition is mediated by structural brain volumes and moderated by glutamatergic polygenic risk score in the context of brain volumes as a mediator. A total of 176 schizophrenia patients and 118 healthy controls participants were assessed for a history of childhood adversity and underwent cognitive testing and structural neuroimaging. Six glutamatergic genes were genotyped, and a weighted glutamatergic polygenic risk score was calculated. Mediation and moderated-mediation models were tested. We found that that there were significant mediation effects of intracranial and total brain volumes on the association between childhood adversity and delayed memory in the overall sample, as well as in the schizophrenia patients. There was also a significant mediation effect of subcortical volume on the association between childhood adversity and working memory in the schizophrenia patients, but not healthy controls. However, there was no significant moderation effect of glutamatergic polygenic risk score on the association between childhood adversity and cognition in the context of brain volume as a mediator. This study demonstrated that childhood adversity exerts a negative impact on intracranial, total brain, and subcortical volumes in schizophrenia. Adversity encountered during childhood may pre-program the brain for subsequent memory performance in adulthood. The effect of glutamatergic polygenic on the association between childhood adversity, brain volume, and cognition in schizophrenia could be related to illness stage or severity. Chapter four aimed to examine interrelationships between childhood adversity, glutamatergic polygenic risk score, frontal lobe volume, and spatial working memory in 51 treatment-resistant schizophrenia patients and 40 healthy controls from the Cooperative Research Centre for Mental Health psychosis study cohort. We found that treatment-resistant schizophrenia patients displayed impairment in spatial working memory between search errors, spatial working memory strategy, and spatial span relative to healthy controls. A significant moderation effect of glutamatergic polygenic risk score was found on the association between childhood adversity and the spatial working memory factor which comprising spatial working memory between search errors, spatial working memory strategy, and spatial span in the treatment-resistant schizophrenia group, but not in the healthy controls. The conditional effects on the association between childhood adversity and spatial working memory indicated that, in the presence of higher childhood adversity, treatment-resistant schizophrenia patients with higher glutamatergic polygenic risk score demonstrated poorer spatial working memory, while those with lower glutamatergic polygenic risk score showed better spatial working memory. Synergistic effects between childhood adversity and glutamatergic polygenic risk score on spatial working memory performance in treatment-resistant schizophrenia patients suggests that lower glutamatergic polygenic risk score may, in part, protect patients from the detrimental effects of childhood adversity on spatial working memory performance, while higher glutamatergic polygenic risk score increases the risk. Chapter five summarises the main findings of each study and highlights the clinical implications and future directions of this critical research area so as to improve mental health for children subjected to adversity.

Aims: This study aimed to investigate the prevalence of the common mental health disorders associated with increased exposure to potentially traumatic events in career and volunteer firefighters, and to identify which individual, acute stressor, operational and organisational factors predict mental health outcomes. Method: Four Australian services participated in a prospective study, with 335 firefighters completing an online survey twice (12-months apart). The survey comprised demographic and fire service information, self-report measures for PTSD, depression, anxiety, alcohol use, exposure to personal trauma and life stressors, number and types of firefighter-related potentially traumatic events experienced in the previous 12-months, job satisfaction related to operational and organisational characteristics of their role, and a measure of the priority their fire service placed on their psychological wellbeing. Structured clinical interviews were conducted with a sample of survey respondents to assess for PTSD, depression, generalised anxiety disorder and alcohol use disorder. Results: From 297 clinical interviews (91 career firefighters and 206 volunteer firefighters), 24% (n = 22) of the career firefighters met diagnosis for any psychiatric disorder. Of these, 3% met criteria for PTSD, 5% for depression, 4% for generalised anxiety disorder, and almost 12% for alcohol use disorder. The primary regression analyses indicated the following findings for the career firefighters. The main variable associated with each of the four disorders was the respective level of symptoms at baseline. In addition, exposure to more potentially traumatic events in the previous 12 months contributed to, and high job satisfaction associated with the operational aspects of their role protected them against the development of PTSD and depression. Finally, the findings indicated that experiencing more recent life events in the previous 12 months and rank (being a firefighter) contributed to more symptoms of Alcohol Use Disorder. From the interviews 17% (n = 35) of the volunteer firefighters met diagnosis for any psychiatric disorder. Of these, 2% met criteria for PTSD, 4% for depression, 5% for generalised anxiety disorder, and 6% for alcohol use disorder. The primary regression analyses indicated that the only predictor of PTSD, generalised anxiety disorder and alcohol use disorder was the respective baseline level for each disorder. The main predictors of depression were baseline level of symptoms of depression and experiencing more recent life events in the previous 12 months. Conclusions: The relatively low prevalence rates for the career and volunteer firefighters in this study indicate their reasonably good mental health across the four disorders. The rates compare favourably with other firefighter studies and are comparable with the rates in the general population for most disorders. An exception to this was volunteer firefighters’ high rate of alcohol dependence and low rate of PTSD. The findings indicate the importance of fire services developing cultures that support and encourage the early identification and management of symptoms and have systems in place to monitor the types of and frequency of firefighters’ exposure to potentially traumatic events. Finally, particularly for the career firefighters, the findings highlight the protective nature of high job satisfaction associated with operational characteristics of their role and reduced mental health symptoms, and the importance for firefighters and managers to recognise and address reductions in job satisfaction.

Background: Greater levels of internalising symptoms during childhood and adolescence increase the risk of developing depression later in life. As such, investigating risk factors for depression in young people is important for understanding the aetiology of the disorder and for informing prevention strategies. Whilst it is known that extreme forms of childhood adversities, such as childhood trauma and abuse, represent a risk factor for depression, less has been done in investigating the effect of parenting behaviour. Emerging research suggests that negative parenting practices (e.g., parental rejection), as well as the lack of positive parental behaviours, increase the risk of internalising symptoms during childhood and adolescence. One of the proposed mechanisms linking poor parenting practices and depression is the effect of parenting behaviour on brain function and brain structure, particularly in regions involved in processing and regulating responses to emotional stimuli, such as the amygdala, prefrontal cortex and hippocampus. Animal studies suggest that parenting may affect the brain via effects on the hypothalamic-pituitary-adrenal (HPA) axis, which has a central role in the response to stress. Genetic factors are also likely to play a role, whereby genetic variants in HPA axis genes may contribute to the regulation/dysregulation of the system. To date, however, there is a lack of research focusing on genes related to HPA axis function, and their interaction with environmental factors (including parenting) in predicting internalising symptoms. Further, no research has investigated the neural mechanisms by which parenting behaviours, HPA genes, and their interaction, exert their influence on internalising symptoms. The aim of this thesis was to investigate a) whether parenting behaviour interacts with HPA genes to influence internalising symptoms in children/adolescents, and b) whether brain structure/function of regions involved in emotion processing (particularly the amygdala and hippocampus) mediates this link. Method: Data from two longitudinal studies were used in this thesis: the Adolescent Development Study (ADS) and the Families and Childhood Transitions Study (FACTS), to explore aims across three experimental chapters. The first study included 98 adolescents from the ADS, for whom hippocampal structural development was measured from magnetic resonance imaging (MRI) scans performed across three waves (W): W1 (mean age =12.6 years), W2 (mean age=16.5 years), W3 (mean age =18.8 years). Maternal negative behaviour was measured at W1 from an observed interaction task. Adolescents’ depressive symptom severity was measured at W1 and W3 with self-reported questionnaires, and genetic risk was calculated using a composite HPA genetic risk score. The second and third studies included 86 (mean age=10.1 years) and 80 (mean age=10.0 years) children, respectively, from FACTS. Observational measures of maternal parenting behaviour were collected during mother-child interactions. Children underwent functional MRI (fMRI) while performing an implicit emotion-processing task. Self-reported and parent-reported measures of child internalising symptoms were also collected. HPA genetic risk was calculated in a similar fashion to the first study. Results: Across both studies, neither HPA genetic risk score nor the interaction between HPA genetic risk score parenting behaviour predicted internalising symptoms in children/adolescents. For study 1, we did not find support for a mediating role of hippocampal structure in the relationship between parenting behaviour and internalising symptoms. Rather, we found a moderation effect such that negative maternal parenting behaviour predicted depressive symptoms longitudinally in adolescents with flattened hippocampal growth and HPA genetic risk. While maternal behaviour was not associated with hippocampal development, in study 2 it was associated with activity/connectivity in brain regions involved in emotion processing. In particular, maternal negative behaviour during a problem-solving interaction was associated with increased amygdala reactivity and connectivity with parietal cortex in children. Maternal negative behaviour during an event-planning interaction was associated with decreased activity in the lingual gyrus in girls. Maternal communicative behaviour was associated with increased medial orbitofrontal cortex activity. These parenting-related neural findings were not associated with child internalising symptoms. In study 3, HPA genetic risk predicted increased amygdala-precuneus connectivity, which in turn was associated with greater internalising symptoms in children. Moreover, HPA genetic risk interacted with negative maternal parenting behaviour to predict increased connectivity between amygdala and superior frontal gyrus, anterior cingulate cortex and parietal cortex. Conclusion: These results suggest that there is a complex interplay between neurobiological, environmental and genetic factors in predicting internalising symptoms in children and adolescents. Neurobiological factors (e.g., hippocampal development) may act as an independent risk factor for depressive symptom severity, but they may also be influenced by environmental and genetic factors. Parenting behaviour, particularly maternal negative and communicative behaviour, may influence brain activity and connectivity in regions involved in emotion processing. The lack of association with internalising symptoms hampers our ability to draw conclusions on the significance of the parenting-related neural findings for children’s mental health. However, given that depression is characterised by impaired emotion processing, including increased sensitivity towards negative stimuli, we speculate that differences associated with negative parenting behaviour, such as amygdala heightened reactivity to negative stimuli, may represent a risk factor for the disorder. HPA genetic variants may moderate the effect of parenting behaviour on these circuits (particularly the frontoamygdala circuitry), such that not all children are sensitive to parenting effects on neurobiology, and may confer risk for depression via altering corticolimbic connectivity. This study highlights the importance of parenting behaviour for children emotional neurocircuitry development and demonstrates the moderating role of genetic factors.

Background Adolescence is widely reported to be a time of increased risk for depression and lower well-being. Importantly, however, outcomes are heterogeneous, and most adolescents do not develop mental health problems. In order to understand how these differences emerge, both environmental and biological factors have been examined in the literature. Evidence indicates that parenting behaviour, socioeconomic status and neurobiology may contribute and, further, that individual differences in brain development may moderate the extent to which contextual factors influence adolescents. That is, individual differences in brain development may confer ‘responsivity’ to context. Several developmental and evolutionary-developmental models provide frameworks with which to interpret such brain-by-environment interactions, and to describe biological responsivity – these are broadly associated with diathesis-stress and differential susceptibility/biological sensitivity frameworks. Broad PhD aim This thesis aims to investigate whether structural brain development moderates adolescent sensitivity to maternal parenting behaviour in the prediction of adolescent depressive symptoms and psychological well-being. Two empirical studies were completed to address this aim. Study 1 examined whether longitudinal change in brain structure modified adolescent vulnerability or susceptibility to aggressive and positive maternal parenting. Effects were assessed to infer evidence in support of either diathesis-stress or differential susceptibility frameworks. Study 2, in light of evidence that positive parenting protects against adversity, considered whether brain development moderated adolescent sensitivity to positive parenting, and whether this association was more pronounced for adolescents with low-socioeconomic status (SES). Methodology During early adolescence (age 13 years), participants completed observed interactions with their mothers, and the frequency of positive maternal behaviour was coded. At three time points (mean ages 13, 17 and 19 years), participants completed structural magnetic resonance imaging (MRI) scans. During late adolescence (age 19 years), participants completed self-report measures of depressive symptoms and psychological well-being. Two separate analyses (studies) were conducted in predicting late adolescent (age 19) outcomes (depressive symptoms and psychological well-being). For both studies, longitudinal brain development was indexed by changes in cortical thickness of structures within the frontal lobe, and volumetric changes of subcortical structures, from early to late adolescence. Study 1: Regression models analysed interactions between maternal behaviour and longitudinal brain development in the prediction of adolescent outcomes. Indices designed to distinguish between diathesis-stress and differential susceptibility effects were employed. Study 2: Regression models were used to investigate interactions between SES (parental occupation), positive maternal behaviour, and longitudinal brain development in the prediction of adolescent outcomes. Results Study 1: Results supported differential susceptibility, whereby less thinning of frontal regions (the left medial orbitofrontal, rostral middle frontal and superior frontal cortices, and the right pars opercularis) was associated with higher well-being in the context of low levels of aggressive maternal behaviour, and lower well-being in the context of high levels of aggressive maternal behaviour. Study 2: Results indicated that individual differences in structural brain development moderated the extent to which positive parenting impacted adolescents dependent on SES. High levels of positive parenting were associated with reduced depressive symptoms for low-SES adolescents with greater volumetric reduction of the right putamen. Further, low positive parenting was associated with reduced psychological well-being for individuals with greater neurobiological sensitivity, however, patterns of brain development that were associated with sensitivity differed by SES. Specifically, low positive parenting was associated with reduced psychological well-being for individuals with more thinning in the context of low-SES, but for individuals with less thinning in the context of high-SES. Significance Results across studies suggested that structural brain development may be associated with individual differences in how sensitive adolescents are to context. Study 1 indicated that reduced frontal cortical thinning during adolescence increased susceptibility to maternal aggressive behaviour in the prediction of well-being, for better and for worse. This finding is significant because it suggests that neither more or less cortical thinning is consistently good or bad for mental health. Results from Study 2 indicated that, although brain change was associated with responsivity to parenting behaviour, patterns of brain development associated with heightened responsivity to parenting were different for high- and low-SES. These results suggested that responsivity functions in a context dependent fashion and highlights the complex interactions that may occur across biological and multilevel environment factors. Results from these studies suggest that structural brain development may be a marker of responsivity to environmental influence. They also emphasise the importance of examining how brain development moderates the impact of multilevel environmental factors on mental health outcomes. Such study designs may better reflect the social settings in which adolescents develop.

The transition from childhood to adolescence is a particularly vulnerable period for the development of internalising symptoms and disorders. Hormonal changes as well as changes in brain structure and function may play a role in this increased vulnerability. Most of the research to date has focused on the hormonal and brain changes during gonadarche, whereas the literature is much more limited for adrenarche, an earlier pubertal phase that takes place prior to gonadarche. Therefore, this thesis aimed to examine how (changes in) adrenarcheal hormones relate to the development of brain structural and functional connectivity, and how that in turn affects internalising symptoms in late childhood to early adolescence. Data were used from two longitudinal community-based samples with two time points each (sample 1 M ages 9.5 and 12.2 years; sample 2 M ages 8.5 and 10 years). At each time point in each study, levels of dehydroepiandrosterone (DHEA), its sulfate (DHEAS) and testosterone were measured and averaged from morning saliva samples collected across several days. Participants also underwent Magnetic Resonance Imaging (MRI) scans, and completed self-report questionnaires, at both time points. Diffusion-weighted imaging scans were analysed to examine white matter structural connectivity. Functional Magnetic Resonance Imaging (fMRI) scans during an affective face processing paradigm were analysed to examine functional connectivity. Levels of internalising symptoms were based on self-report questionnaires: the Spence Children’s Anxiety Scale, the Children’s Depression Inventory, and the Positive And Negative Affect scale. Analyses were conducted to investigate associations between hormone levels (initial levels and changes in levels over time) and brain structural and functional connectivity (baseline and change over time). Analyses were also conducted to investigate whether hormone-related changes in structural/functional connectivity were associated with symptoms. The results showed that children with high DHEA levels at age 9 had higher mean diffusivity (cross-sectionally) in a wide range of white matter tracts, suggesting that relatively early exposure to DHEA might be negatively associated with white matter microstructure. Changes in testosterone from age 8.5 to 10 years were negatively associated with the development of white matter structure as quantified by fibre cross-section in posterior white matter tracts. Higher levels of testosterone at age 8.5 years, however, were related to stronger development of fibre cross-section from age 8.5 to age 10 years. These hormone-related changes in white matter structure were not significantly associated with levels of internalising symptoms. Analyses of functional connectivity during affective face processing focused on connectivity of the amygdala to the rest of the brain, because of the crucial role of the amygdala in emotion processing and consistent findings of its involvement in internalising disorders. Indirect effects were found of adrenarcheal hormone levels (controlled for age, potentially indicating a timing effect, i.e. maturation relative to same-age peers) on anxiety symptoms at age 9 years, mediated by amygdala connectivity to visual and limbic areas. Timing of adrenarcheal hormone exposure was also found to have indirect effects on anxiety symptoms longitudinally. Specifically, higher DHEA at age 9 years was indirectly related to more anxiety symptoms at age 12 years, controlling for symptoms at age 9 years, via more positive amygdala to inferior frontal gyrus connectivity. Thus, the findings in this thesis have demonstrated that elevated adrenarceal hormone levels (potentially reflecting early timing of adrenarche) are both cross-sectionally and longitudinally associated with anxiety symptoms through an effect on amygdala functional connectivity. We also showed that the hormonal processes of adrenarche have an impact on white matter microstructure development. These findings have implications for the understanding of how individual variation in adrenarcheal processes can impact children’s brain development and mental health. Future studies should examine whether effects of variation in adrenarcheal processes on brain development and mental health are persistent, as well as establish whether predictors found in the current thesis are also relevant in clinical samples.

Background: Traumatic brain injury (TBI), has been diagnosed in over 355,000 US military service personnel since 2000. Epidemiological research indicates that veterans with TBI are two-to-four times more likely to develop dementia than controls; however, mechanisms contributing to this relationship are poorly understood. The aim of this study was to investigate if Vietnam war veterans with a TBI show evidence of Alzheimer’s disease (AD) pathological markers, as assessed by A-amyloid, tau and glucose metabolism using PET, as well as structural MRI, including diffusion tensor imaging, and neuropsychological testing. Methods: Sixty-nine male veterans - 40 with TBI (aged 68.0±2.5 years) and 29 controls (aged 70.1±5.3 years) - underwent A-amyloid (18F-Florbetaben), tau (18F-AV1451) and 18F-FDG PET, MRI, psychiatric and neuropsychological assessment. The TBI cohort included 15 participants with mild, 16 with moderate, and 9 with severe injury. Fractional Anisotropy (FA) was employed as a measure of white matter tissue integrity. PET Standardized Uptake Value Ratios (SUVR) were calculated using the cerebellar cortex as reference region. Analyses were adjusted for IQ, age, ApoE status and psychiatric comorbidities. Results: Veterans with moderate-to-severe TBI performed significantly worse than controls on composite measures of memory and learning (M = -0.55  0.69, t(67) = 2.86, p=0.006, d=0.70) and attention and processing speed (M = -0.71  1.08, t(52) = 2.53, p=0.014, d=0.69). The moderate-to-severe TBI group had significantly lower FA than controls in the genu (F(3,36)=8.81, p<0.05, partial 2 = 0.17), and body (F(3,36)=4.39, p <0.05, partial 2=0.14) of the corpus callosum, as well as in global white matter (F(3,36)=5.35, p <0.05, partial 2=0.13). There were no significant differences in 18F-Florbetaben or 18F-AV1451 uptake amongst the groups, however the moderate-to-severe TBI group had significantly lower 18F-FDG retention than controls in the mesial temporal region (F(8,44) = 2.21, p <0.05, partial 2 = 0.13). No differences were found between the mTBI group and controls on any of the outcome measures. Conclusions: These findings indicate that moderate-to-severe TBI, but not mTBI, is associated with later-life cognitive deficits, and diminished global white matter integrity, specifically in the corpus callosum. However, these deficits are not associated with AD pathology. These results are consistent with current evidence of white matter axonal damage as the primary source of cognitive impairment in TBI, and are not reflective of a neurodegenerative process.