Search

Autism spectrum disorder (ASD) is a relatively common and heterogeneous, neurodevelopmental condition that comes at high cost to both the individual and the society. ASD affects approximately 1.6% of the population; it is pervasive, chronic, and associated with increased morbidity and reduced life expectancy. Nonetheless, there are no effective pharmacological treatments for the core symptoms of ASD. One reason for this is that the underpinning mechanisms that drives these symptoms symptoms are poorly characterized. Genetic factors are recognized to play an important role in ASD. The estimated heritability of ASD is 64-90%, and several genetic ASD risk factors have been identified. Collectively, these discoveries highlighted the role of genes that support different aspects of brain development such as synaptic formation and organization. However, the genetic - and (neuro)biological architecture of ASD is complex and currently identified 'risk' genes do not exhaustively explain it. Hence, it is crucial to investigate the genetics of ASD further. Particularly, attention should be drawn to genes that regulate the immune response. This is because immune genes are known to sustain neurodevelopment through diverse processes, including neuroprotection, neurogenesis, and synaptic formation and pruning. Although there is proof of the dysregulation of several immune functions, from inflammation to autoimmunity to ASD, the study of immune genes in ASD has been limited; and mainly restricted to the major histocompatibility complex genes, key regulators of adaptive immunity. The genetic regulators of the immune system are, however, multiple and they may exert different influences on brain development and thus ASD, which have yet to be examined. Therefore, this thesis aims to explore the relationship between ASD and genetic factors involved in different immune mechanisms. To address this aim, I carried out a number of studies using complementary analytic strategies. First, I conducted a systematic literature review to gain a comprehensive overview of known immune genes that have been associated to ASD (described in Chapter 3). This study laid the groundwork for my investigation since it supported a role of immune genes in ASD and their relevance to general neurodevelopment. Also, this review highlighted the need for a deeper investigation of immune genes in ASD, considering genetic factors involved in several types of immune responses. Second, given the increasing evidence to support a genetic association between ASD and autistic-like traits in the general population, I explored the genetic underpinnings of these population-based traits. Specifically, I performed a hypothesis-free genome-wide association study meta-analysis of four autistic-like traits across multiple international cohorts (presented in Chapter 4). Strikingly, the top results from this hypothesis-free approach all pointed to immune-related genes and I discovered an association between immunogenetic factors and specific autistic features (i.e., rigidity and attention to details) in the general population. Taken together, these studies provided preliminary evidence for a role of immune genes in ASD -both at a diagnostic level and at the level of population-based quantitative traits. Therefore, in my third study I further explored the genetic relationship between ASD and both innate and adaptive forms of immune dysregulations. In brief, I investigated 1) if ASD and diverse immune-related conditions, including inflammatory, allergic, and autoimmune diseases, share genetic factors; and 2) if individual variations in genetic factors linked to these immune conditions are also associated with variation in autistic-like traits in individuals of the general populations and across sex groups. In this study (described in Chapter 5), I identified significant genetic correlations between ASD and allergic and systemic autoimmunity, at both the genome-wide and local genomic level. The genetic factors underlying these correlations show increased expression in both immune and brain tissues, and an association with epigenetic changes during neurodevelopment. Also, I demonstrated that individual variations in autoimmunity-related genetic factors are associated with rigidity in the general population. Driven by the population-based findings presented in Chapter 5, in my fourth study I investigated if - and how - individual variations in identified immunogenetic factors are linked to different symptoms and cognitive domains in a clinical sample of ASD. To do this, I leveraged genetic and deep phenotypic data from the Longitudinal European Autism Project (LEAP). In this study, described in Chapter 6, I identified an association between genetic regulators linked to lymphocytic count and the worsening of repetitive and restricted behaviors over time. In summary, my thesis supports a neurodevelopmental role of immune genes and their relevance to ASD. My findings demonstrate a role of genes that regulate autoimmune and allergic responses in ASD. Also, my work provides preliminary evidence that genetic factors involved in autoimmunity are linked specifically to the rigid aspect of the autistic phenotype, in the general population. This association was also observed in a clinical sample of ASD. Taken together my findings encourage the study of immune genes with respect to specific (i.e., rigid) features of the complex autistic phenotype. Taking a step further, this work invites to extend immunogenetic investigation to other neurodevelopmental diagnostic categories, and in relation to several neurodevelopmental windows and neural systems.

A small group of men are responsible for most violent crimes, resulting in considerable interpersonal and societal costs. They meet diagnostic criteria for Conduct Disorder (CD) in childhood and Antisocial Personality Disorder (ASPD) in adulthood. Approximately one-third meet criteria for an additional diagnosis of psychopathy (ASPD+P), exhibiting callous unemotional (CU) traits from early in life, engaging in a broader range of offending behaviours, and responding less well to treatments than those without psychopathy (ASPD-P). Those with ASPD+P show particular deficits in empathic processing, while both ASPD+P and ASPD-P share deficits in decision-making. Evidence demonstrates that these clinical and neurocognitive differences may be explained by some shared, but some differential neurobiological underpinnings. However, the neurochemical underpinnings remain relatively under-researched. One neurochemical of particular interest is oxytocin- a neuropeptide shown to modulate empathic processing and decision-making in healthy populations. However, no study to date has examined its effect on neural processing of these functions in ASPD+/-P. The neurochemical underpinnings of shared decision-making deficits in ASPD+/-P also remain under-researched. One important consideration is regulation of striatal input into decision-making by Glutamate and GABA, which has not yet been examined in ASPD+/-P. In this thesis, I sought to address these gaps in the literature. Firstly, using a double-blinded, placebo controlled design, I explored similarities and differences in i) an important component of empathic processing (neural response to facial emotional expressions) and ii) decision-making in violent men with ASPD+/-P (n= 26 across studies), compared to healthy non-offenders (NO; n= 21 across studies), with placebo and after acute application of intranasal oxytocin, using functional Magnetic Resonance Imaging (fMRI). Secondly, I investigated differences in striatal excitatory/inhibitory (E/I) balance between groups, as measured by Glutamate:GABA ratio, using Magnetic Resonance Spectroscopy (MRS). Subjects were classified using Psychopathy Checklist PCL-R scores. In the first fMRI study, in the placebo condition, the ASPD+P group showed reduced responsivity to fearful faces in in bilateral cingulate cortex, compared to ASPD-P. Further, oxytocin abolished these differences and enhanced activity in ASPD+P, in these regions and also in bilateral insula. In the second fMRI study, there were no significant behavioural findings. While the task elicited an overall effect across the sample in bilateral cerebellum (offer phase) and anterior cingulate cortex (decision phase), there were no effects of group or oxytocin. In the MRS study, both ASPD+P and ASPD-P had lower Glutamate:GABA ratios than NO, though there was no significant difference between ASPD+P and ASPD-P. Glutamate:GABA ratio correlated significantly with a reward-based decision metric from a neuropsychological task ('Now or Later' task). Taken together, my findings provide further support for a model of some shared, but some differential neurocognitive deficits in ASPD+P and ASPD-P. They build on previous research supporting stratification of ASPD into subcategories of ASPD+P and ASPD-P. Importantly, for the first time, I have also demonstrated normalising effects of oxytocin in empathic processing in ASPD+P. This finding is a particularly important step forward towards developing targeted and effective interventions for this especially challenging patient population.

At least 20-50% of individuals with Autism Spectrum Disorder (ASD) experience clinically relevant anxiety and depression symptoms, as compared to approximately 7-14% of individuals from the general population. Moreover, anxiety and depression symptoms have a significant impact on the wellbeing of autistic people and their families, contributing to increased risk of suicidal ideation and attempts - one of the leading causes of premature mortality in ASD. Consequently, understanding the characteristics and mechanisms associated with anxiety and depression to inform more effective interventions for managing these symptoms has been declared the number one priority for autism research (Autistica, 2017; Pellicano, Dinsmore, & Charman, 2014). To address this research priority, a multimodal and mixed methods approach was utilised in this thesis to investigate two primary research aims. Firstly, aim one was to elucidate the impact of anxiety and depression symptoms (relative to core autism traits) on quality of life. Second, aim two was to investigate psychological and neurobiological factors contributing to anxiety and depression symptom severity in ASD. Both aims were predominantly addressed using both cross-sectional and longitudinal data from a cohort of almost 600 males and females with and without autism, aged 6-30 years and including individuals with mild intellectual disability from the EU-AIMS Longitudinal European Autism Project. In relation to aim one, individual variability in subjective quality of life was assessed, with results showing that depression symptoms were the factor most strongly associated with reduced physical and psychological wellbeing in autistic adults and children/adolescents (Chapter Three). In relation to aim two, relationships between difficulties identifying and describing emotion ('alexithymia'; Chapter Four) and neurofunctional/physiological arousal indices (amygdala response to emotional faces and resting heart rate/variability; Chapters Five and Six) with anxiety and depression symptoms in autism were investigated. These investigations revealed that difficulties identifying emotion (a facet of alexithymia) were significantly associated with both anxiety and depression symptoms at a single timepoint and predicted anxiety 12-24 months later. Autistic individuals with both high anxiety and difficulties identifying emotion also had the most severe self-reported social-communication difficulties and poorest QoL outcomes (Chapter Seven). In contrast, neither amygdala response to emotional faces, nor resting heart rate/variability were robustly associated with anxiety and depression symptoms. Instead, both arousal indices were associated with irritability, which has been posited as an 'atypical' manifestation of, or marker for, emotional problems in ASD. Last, complementary to these quantitative studies, a qualitative approach was taken, using semi-structured interviews to explore autistic adult's perceptions of QoL and the factors that they felt contributed to their anxiety and depression symptoms (Chapter Eight). The most frequently discussed contributing factor for anxiety and depression symptoms was exhaustion following environmental overload, which also related to feelings of loneliness and a desire for control. Other contributing factors for mental health expressed by autistic adults included societal issues (prejudice, barriers to accessing services), establishing an identity as an autistic person, exposure to bullying and other stressful life events, negative cognitions about the self and the world and emotion awareness/regulation difficulties. Overall, the work presented in this thesis emphasises the substantial and persistent impact of anxiety and depression symptoms on autistic individuals, both over time and across developmental stages. Difficulties with emotion awareness/regulation and reduced threshold for overload appear to be particularly prominent associates of anxiety and depression symptoms in ASD. This has implications for clinical practice, supporting that intervention approaches targeting emotion awareness/regulation and support for developing coping strategies to manage stress and environmental overload may be effective for improving mental health for some autistic people. Furthermore, from a policy perspective, better support in education and employment and improving social awareness and understanding of autism may not only target risk factors, but also bolster protective factors for mental health in ASD.

Autism Spectrum Disorder (ASD) is a pervasive, neurodevelopmental disorder characterised by impairment in reciprocal social communication, and restricted, repetitive patterns of behaviours or interests. The presence of co-occurring conditions such as anxiety and depression are common and impart additional functional impairment. However, there are no validated screening tools for co-occurring conditions in ASD and no consensus guidelines for the diagnosis and treatment of core and co-occurring conditions. Therefore, in the first part of this thesis (Study 1), I conducted a pilot investigation to examine the validity and reliability of the Strengths and Difficulties Questionnaire (SDQ). My results suggest that the SDQ is a valid and reliable instrument for assessing co-occurring conditions in ASD. Next, I worked with a multidisciplinary team, established by the British Association of Psychopharmacology, to conduct a systematic review and co-author consensus guidelines on the assessment, treatment and research for ASD and its most common co-occurring conditions (Study 2). The high co-occurrence of anxiety and depression in ASD may indicate a shared neurobiological underpinning of 'co-occurring and 'core' symptoms in this condition. One such candidate is the (im)balance between excitatory (E) glutamate and inhibitory (I) gamma-aminobutyric acid (GABA), but this has yet to be investigated directly in the living human brain. Therefore, I conducted a pharmacological MRI study to investigate the relationship between E-I and both core and co-occurring symptoms and to investigate whether E-I can be 'shifted' acutely by targeting the serotonergic system and/or μ-opioid receptor (Studies 3-5). To do this, I administered citalopram (a selective serotonin re-uptake inhibitor) and tianeptine (a μ-opioid receptor agonist) to adult males with and without ASD. I used two neuroimaging approaches to examine E-I at two different levels, namely, i) [1H]MRS to assess 'bulk' levels of glutamate+glutamine (Glx) and GABA in a pre-defined region of interest and ii) regional homogeneity (ReHo) analysis of whole-brain resting-state fMRI to quantify regional synchrony of BOLD signals, as a proxy measure of E-I influence on brain activity. In Study 3, I found that core symptoms and symptoms of anxiety and depression were respectively related to Glx and GABA levels in individuals with ASD. I also found that Glx response to citalopram was significantly different between ASD and controls. Glx decreased in controls but was unchanged in ASD. However, in both groups, there was marked individual variation in [1H]MRS response to citalopram. Individual variation in [1H]MRS response was also found for tianeptine (Study 4), and there were no main effects of group or drug or interactions. Finally, in the ReHo analyses reported in Study 5, I found that citalopram increased ReHo in the left occipital middle lobe in ASD but not in controls; tianeptine had no significant impact on this measure. My results suggest that E-I systems are related to core ASD symptoms and symptoms of co-occurring anxiety and depression. I found evidence that E-I could be modulated using citalopram and tianeptine; but that this response may differ between ASD and controls. The considerable individual variation in pharmacological response in men with and without ASD evident across the studies carried out here suggests future work to examine whether these responsivity differences are predicative of treatment response.

Many neurodevelopmental disorders (NDD) including autism spectrum (ASD) are diagnosed in early childhood. However, the pathophysiological mechanisms underpinning vulnerability to these conditions operate from much earlier, with foetal and early postnatal development representing a sensitive period to adverse environmental exposures and genetic risks. Unfortunately, very little is known about normal human brain development during this period, let along the impact of NDD risk factors on brain maturation. Most of the current knowledge of typical and atypical early brain development comes from animal and post-mortem work. Preclinical studies argue that genetic risk factors for ASD and related NDD act upon fundamental biological processes governing early brain maturation, including dendritic development. How these findings translate to the living human brain remains poorly understood. Advances in diffusion magnetic resonance imaging (MRI) and their application to human neonatal cohorts have provided a unique opportunity to probe the developing cortical microstructure in vivo. However, such research to date has mainly focused on preterm born infants. Preterm infants are known to be at risk of atypical development including a wide-range of brain alterations and behavioural deficits, and thus findings might not necessary generalise to the term-born brain of either typical infants, or those with genetic risk factors for NDD. This work aimed to advance current understanding about typical and atypical cortical microstructural maturation during early human brain development (term birth to 6 months of age). Specifically, (i) to describe typical neurite development in the term-born brain and (ii) to test the hypothesis that vulnerability to NDD is associated with altered dendritic maturation during this period. This was achieved by applying advanced diffusion MRI analyses to characterise neurite maturation in neonates and 6-month-old infants with and without NDD risk factors, namely familial history of ASD/ADHD. First, cortical microstructural maturation within the first month of life was examined in a sample of term-born neonates. The relative spatial distribution of diffusion metrics previously reported across the adult cortex was already evident within the first month of life, although quantitative values were consistent with the more immature state of the newborn cortex. Lower-level sensory cortices were better 'developed' than higher order regions, characterised by lower anisotropy and higher measures of neurite density and orientation dispersion. Age-related changes in neurite composition were observed in a fall in cortical anisotropy and tissue-water content together with an increase in neurite orientation dispersion and density, likely to be associated with rapid dendritic outgrowth. However, the pattern of cortical development observed in term-born 4 neonates was distinct from that recorded in studies of preterm cohorts of the same postmenstrual age. Second, consistent with findings from preclinical studies, a familial risk of NDD was associated with atypical cortical microstructure in early postnatal life. Neonates vulnerable to NDD had altered neurite geometry in cortical regions involved in lower-level sensory processing and multisensory integration, abilities known to be altered in high risk infants and individuals with a diagnosis of NDD. This abnormality was present in the absence of any atypical neurite density or altered macrostructural features, including cortical volume, thickness and curvature. Last, cortical microstructure was examined in the typical and atypical infant brain. For the first time, the spatial pattern of cortical diffusion was described in 6-month-old term-born infants. Consistent with findings from the neonatal study in this thesis, the primary sensory areas continued to show a more 'mature' state than higher-order cortices. Furthermore, vulnerability to NDD was also associated with altered neurite geometry in cortical regions governing lower-level sensory processing and multi-sensory integration, showing a substantial overlap with observations in at-risk neonates. Taken together, the work presented here shows that early postnatal life is characterised by dynamic changes in cortical microstructural composition. Furthermore, building on preclinical research it suggests that dendritic pathology represents a primary mechanism underlying vulnerability to ASD and related NDD.

Background: There is an association between schizophrenia and violence, although the specific drivers for this link remain unclear. This study aimed to explore the relationships between childhood adversity, conduct disorder, substance misuse and violence among men with schizophrenia. It has been suggested that there may be different pathways to violence in schizophrenia, one primarily linked with pre-morbid conduct disorder, so patients were specifically grouped on the basis of pre-morbid conduct disorder. Methods: Ninety-three male participants were recruited for the study, fifty-four with schizophrenia and thirty-nine healthy controls. Participants underwent a range of clinical assessments, including symptoms of conduct disorder, exposure to childhood adversities and history of substance use disorders. Adult propensity to violence was measured using the Gunn Robertson Violence Scale. Those participants who consented also had a structural magnetic resonance imaging scan. Results: The Gunn Robertson Violence Scale has good validity in this population. Conduct disorder was associated with an increased propensity to violence. Exposure to domestic violence during childhood and the cumulative number of childhood adversities were both associated with adult propensity to violence and attenuation of the association with cumulative adversities suggested that conduct disorder may be a mediator of the relationship. Patients with pre-morbid conduct disorder began using alcohol and cannabis earlier and more frequently, and had higher rates of lifetime substance use disorders which were associated with an increased propensity to violence. An increase in grey matter volume in the caudate was correlated with an increased lifetime propensity to violence. Conclusions: Conduct disorder, substance use disorders, childhood adversity, schizophrenia and violence are all associated with each other. Hence there is a complex interplay of factors, with their origin in childhood, which increase the risk of violent behaviour in schizophrenia.

Individuals who have autism spectrum disorders (ASD) commonly experience social anxiety. Whether this manifests as part of, or is comorbid to, ASD, continues to prompt clinical and academic discussion. To date, empirical studies have predominantly investigated social anxiety in children and adolescents. We know relatively little about prevalence rates, demographic and clinical correlates, causal and maintaining mechanisms, and best ways to provide psychological interventions for adults who have both ASD and social anxiety. This thesis focuses on two principal areas: understanding social anxiety and treating this (and associated characteristics) in adults with ASD. A systematic review of literature on social anxiety in individuals with ASD sets the scene for the empirical studies reported in the thesis. Rates and levels of social anxiety and a range of demographic and clinical correlates are examined cross-sectionally, in community and clinical samples. Using a qualitative study design, the opinions of multidisciplinary clinicians and researchers about assessment, formulation and interventions for individuals with ASD and social anxiety, are explored. The empirical evidence for psychological interventions for adults, in particular group social skills interventions (GSSI), cognitive behaviour therapy (CBT) and mindfulness, are synthesised in three systematic reviews. Finally, the design, delivery and evaluation of two novel pilot group CBT interventions - one addressing social interaction anxiety, and the second addressing low self-esteem - are described. Overall, study findings indicate that at least 45% of males and females with ASD have clinically significant cognitive and affective symptoms and behavioural responses indicative of social anxiety. In line with previous research, social anxiety, in these samples, was significantly associated with age, self-reported ASD traits, depression and general anxiety, but not with clinician-rated ASD measures, IQ or theory of mind. Causal and maintaining mechanisms likely comprise bio-psycho-social factors, underpinned by the presence and impact of core ASD symptomatology. A conceptual framework is put forward to outline putative relationships between these mechanisms. CBT interventions, adapted to suit the needs of adults with ASD, appear to be feasible and effective for reducing social anxiety.

There are currently no effective pharmacologic treatments for the core symptoms of Autism Spectrum Disorder (ASD); and candidate treatments are often assumed to act in a similar way in people with and without ASD. This may not be the case, as there is recent evidence that the excitatory (E) glutamate and inhibitory (I) GABA systems, which are crucial to brain development and function, are altered in ASD. Abnormalities of E-I balance at rest in ASD are described in the literature, however findings often contradict and mostly stem from cross-sectional studies. To date, no prior studies of ASD have examined the 'responsivity' of the E-I system to pharmacologic challenge. No single technique can fully capture E-I dynamics in the living brain, however E-I balance can be measured at a number of 'levels', using different species-specific modalities. Hence, in the present thesis, I aimed to explore the effect of pharmacologically modulating E-I balance at the i) intracellular, ii) functional, and iii) extracellular level, in both humans, and an animal model of ASD (the Neurexin1a knock out rat), after a pharmacological challenge with the E-I acting drug, riluzole. To first capture E-I dynamics at the intracellular level, I investigated riluzole-evoked changes in bulk tissue levels of E-I neurotransmitters in the medial prefrontal cortex (mPFC) and basal ganglia (BG)- regions known to be important in ASD- using Magnetic Resonance Spectroscopy and ex vivo tissue analyses in humans and rats, respectively. In humans, I compared the change in 'Inhibitory Index' - the GABA fraction within the pool of glutamate plus GABA metabolites - post riluzole challenge in adult men with (n = 17) and without (n = 20) ASD. Despite comparable baseline measures of GABA and Glx (glutamate + glutamine) in the mPFC, I found the response of the Inhibitory Index to riluzole to be diametrically opposite in men with and without ASD. In contrast, in the BG, I reported a significant group difference in baseline Glx, in which the ASD cohort had lower Glx levels compared to controls, yet both groups responded in the same direction to riluzole challenge, with an increase in the Inhibitory Index. These findings were partially replicated in the Neurexin1a knock out rat- with lower subcortical glutamate also reported in an equivalent basal ganglia region (the caudate putamen), compared to wild type controls. However, no differences in E-I responsivity were observed in the rodent in either strain or brain region. Given the reported group differences in E-I responsivity profile of the human mPFC, and that E-I imbalances are likely to affect the whole brain, I next explored the responsivity of functional connectivity of the mPFC with the rest of the brain, using resting-state functional Magnetic Resonance Imaging (rs-fMRI). In the ASD group, functional connectivity was abnormal at baseline, but restored to control levels after riluzole administration. Conversely, riluzole had no effect on the functional connectivity in the control group, thus highlighting again a group difference in responsivity. Finally, to examine whether extracellular changes in E-I dynamics were driving differences in functional connectivity, I explored neurotransmitter efflux into the extracellular space in the living rat brain, using in vivo microdialysis. However, riluzole had no measurable effect on glutamate nor GABA efflux in either wild type, or Neurexin1a knock out rats. Overall, I have, for the first time, identified that E-I balance at both the intracellular and functional level may be shifted in adult men with ASD, and that the brain in ASD is pharmacologically different to that of controls. This has implications for future drug discovery in ASD and the results of this thesis could be translated forwards to enhance future clinical trial design. Further to this, I found the Neurexin1a knock out rat to have a comparable baseline biology to the human condition, which may support its use in future studies of back-translate these findings. Such work will be necessary to further understand E-I pharmacology in ASD, and further validate E-I balance as a tractable therapeutic treatment target for this increasingly prevalent condition.

In the recent years our understanding of the frontal lobe functions has greatly advanced. The advances in the field of the neuropsychology, neuroimaging and neurosciences all contributed to a rapidly changing perspective on the role of the frontal lobes in behaviour and cognition and they also changed clinical approaches to the evaluation of patients with frontal lobe disorders. The structure of the brain can be non-invasively assessed in vivo using magnetic resonance imaging (MRI). The MRI diffusion tensor imaging (DTI) can be used to reconstruct the human brain white matter and to quantify their microstructural integrity. The aim of this thesis is to investigate the anatomy of the frontal networks underlying cognitive and behavioural functions and includes three studies: an investigation of the association between extra-motor white matter tracts and cognitive and behaviour symptoms in Motor Neuron Disease (MND); a study of the ventral fronto-temporal network and its association with behavioural symptoms in Primary Progressive Aphasia (PPA); a study to investigate the association between white matter abnormalities in several long association tracts and deficits in non-verbal and verbal communication in Autism Spectrum Disorders (ASD). Overall, the findings of this thesis indicate that both uncinate fasciculus and cingulum are frontal lobe structures particularly vulnerable to disease regardless the nature of the underlying pathology. Damage to these tracts could manifest with abnormalities in several aspects of social behaviour and cognition. These considerations will help to broaden our understanding of the frontal lobe function beyond motor and language functions.

Autism Spectrum Disorder (ASD) and Obsessive-Compulsive Disorder (OCD) are frequently comorbid and share deficits in executive function, highlighting a need to understand the shared and/or disorder-specific neurofunctional abnormalities underlying these behaviours. First, a comparative, multimodal meta-analysis between ASD and OCD was conducted on whole-brain voxel-based morphometry structural magnetic resonance imaging (MRI) studies and functional MRI studies of cognitive control. Second, functional MRI (fMRI) was used to scan adolescent boys with ASD or OCD, and control boys while they performed tasks measuring sustained attention and reward-based decision-making, including temporal discounting and gambling. Shared abnormalities were observed in the meta-analysis, where both clinical groups had reduced structure and function during cognitive control in medial prefrontal and anterior cingulate regions. During fMRI, shared abnormalities were also observed during executive function tasks of reward-based decision-making, where both clinical groups had reduced activation in ventromedial, inferior frontal and orbitofronto-striatal as well as temporo-parietal regions compared to controls. Disorder-specific abnormalities, on the other hand, were seen predominantly during tasks of non-emotional executive function. OCD patients had disorder-specific increases in striato-insular structure and function, whereas ASD individuals had increased structure but decreased function in dorsolateral prefrontal cortex during cognitive control. Temporo-parietal underactivation during sustained attention was uniquely associated with OCD compared to ASD and controls. These results present novel evidence that neurofunctional abnormalities, including temporo-parietal underactivation and striato-insular overactivation during non-emotional tasks of executive function may be mostly disorder-specific to OCD compared with ASD, whereas abnormalities during emotionally-driven tasks of reward-based decision-making are predominantly shared between ASD and OCD, particularly in ventromedial, inferior and orbitofronto-striatal regions. These studies provide preliminary indication that both shared and disorder-specific neurostructural and neurofunctional biomarkers underpin cognitive dysfunction in these disorders that may have implications for future diagnosis and treatment.

There is currently very little known about early brain development; even less about how normal development may be disrupted in individuals genetically predisposed to neurodevelopmental conditions, including autism spectrum disorder (ASD). Hence, by using magnetic resonance imaging (MRI) to examine the structure and chemistry of the fetal, neonatal, and infant human brain, my intention was to provide further insight into early neurodevelopment. The additional focus on individuals with a familial risk of ASD, also aimed to detect any potential deviations from typical brain development that may be associated with an ASD risk status. The first experimental study of this thesis used MRI at 1.5T to reveal that the early environment – specifically, mother-infant interactions – is associated with variations in brain biology, within a sample of assumed-to-be typically developing individuals. Therefore, the next study, which compared brain volume in infants with and without a familial risk of ASD also incorporated measures of mother-infant interactions. In this second study, subcortical and cerebellar enlargements were identified in 4-6 month-old infants at risk of ASD, and larger volumes were associated with more autistic symptoms at 36 months. Within the high-risk group, a higher measure of maternal sensitivity was correlated with lower subcortical brain volumes. Next, to examine even earlier development and to estimate when differences between risk groups might first appear, fetuses and neonates with and without a familial risk of ASD were scanned using advanced MRI protocols at 3T. At both these timepoints, the cortex and cerebellum were identified as the fastest growing brain regions. In addition, fetuses at risk of ASD had smaller cortical volumes when compared to low-risk controls. Postnatally, within the first month of life, neonates at risk of ASD also had smaller intracranial and total brain volumes, as well as a smaller lentiform nucleus. Placed together with the infant findings, these studies suggest that an ASD genetic risk constrains brain growth in the perinatal period, but is associated with volumetric expansion of both subcortical and cerebellar regions in infancy. The subcortical region also appeared to be particularly affected by the abnormal growth trajectory of high-risk participants, as differences in this region were observed amongst high-risk individuals at both neonatal and infant timepoints. The subcortex was therefore selected as a region-of-interest in the final study, which mapped metabolic maturation from fetal to early postnatal life using magnetic resonance spectroscopy. In both low and high-risk groups, choline and myo-inositol decreased significantly with age, whilst N-acetylaspartate and creatine increased. In contrast, glutamate (measured as glutamate and glutamine; Glx) decreased from fetal to neonatal life, with a marked ‘dip’ around the time of birth before increasing in early infancy. The sample size at the 4-6 month timepoint also permitted an explicit comparison of low and high-risk infants, and because previous studies of older cohorts have linked glutamate abnormalities to ASD, Glx levels were compared between groups. The results showed, for the first time, that high-risk infants have significantly elevated levels of Glx at 4-6 months of age, when compared to low-risk controls. In conclusion, both the volume and biochemistry of the human brain undergoes rapid change in the late prenatal and early postnatal periods. Moreover, individuals with a familial risk of ASD already show differences in brain maturation from fetal life, including neurochemical pathways modulating the balance of neuronal excitation and inhibition. This is well before the onset of autistic symptoms, and therefore, has important implications for ASD in terms of risk (and resilience) pathways.

Background: Stroke-induced aphasia is associated with adverse effects on quality of life and the ability to return to work. However, the predictors of recovery are still poorly understood. Anatomical variability of the arcuate fasciculus, connecting Broca’s and Wernicke’s areas, has been reported in the healthy population using diffusion tensor imaging tractography. In about 40% of the population the arcuate fasciculus is bilateral and this pattern is advantageous for certain language related functions, such as auditory verbal learning (Catani et al. 2007). Methods: In this prospective longitudinal study, anatomical predictors of post-stroke aphasia recovery were investigated using diffusion tractography and arterial spin labelling. Patients An 18-subject strong aphasia cohort with first-ever unilateral left hemispheric middle cerebral artery infarcts underwent post stroke language (mean 5±5 days) and neuroimaging (mean 10±6 days) assessments and neuropsychological follow-up at six months. Ten of these patients were available for reassessment one year after symptom onset. Aphasia was assessed with the Western Aphasia Battery, which provides a global measure of severity (Aphasia Quotient, AQ). Results: Better recover from aphasia was observed in patients with a right arcuate fasciculus [beta = .730, t(2.732), p = .020] (tractography) and increased fractional anisotropy in the right hemisphere (p < 0.05) (Tract-based spatial statistics). Further, an increase in left hemisphere perfusion was observed after one year (p < 0.01) (perfusion). Lesion analysis identified maximal overlay in the periinsular white matter (WM). Lesion-symptom mapping identified damage to periinsular structure as predictive for overall aphasia severity and damage to frontal lobe white matter as predictive of repetition deficits. Conclusion: These findings suggest an important role for the right hemisphere language network in recovery from aphasia after left hemispheric stroke.

Conduct disorder (CD) is a serious disruptive behaviour disorder that is diagnosed in children who display repetitive and persistent antisocial behaviour, such as violence, robbery and vandalism. Children with CD present substantial costs to society, and are the group of children most commonly referred to mental health services. Further, CD is a strong predictor of adult Antisocial Personality Disorder and psychopathy. Research to date on the biological associates of CD has mostly compared the anatomy and function of specific brain regions in people with CD to controls. However, there is increasing recognition that brain regions do not act in isolation. Rather, they form part of integrated neural systems. Nevertheless, to date, there have been no studies on anatomical ‘connectivity’ in CD. Also there are few studies of how prenatal environment modulates the development of human limbic ‘social brain’ regions that are implicated in CD, and other abnormalities in social development. For example, prior studies reported behaviour problems in babies and children of mothers with elevated levels of stress or anxiety during pregnancy. Preliminary evidence suggests that these maternal emotional factors modulate intrauterine environment (e.g. through the stress hormone cortisol); and so may alter the development of limbic brain structures (such as the amygdala and orbitofrontal cortex) that are crucial to emotion processing and social cognition. However, to date, only one human study has examined the association between prenatal maternal mood and altered development of neural systems in children.