This thesis investigates autism as it has been understood in the cognitive and social (neuro)sciences, within the United Kingdom, since 1985. Of specific interest is how these sciences discover, construct, and experiment upon individuals who are understood as socially abnormal. Theoretically, the thesis is positioned between Foucauldian History and Philosophy of Science, Medical Sociology, and Science and Technology Studies. Empirically, two key sources of information are relied upon. Firstly, there is an extensive critical reading of the published literature from 1985 to the present. Secondly, twenty qualitative research interviews were conducted with academic researchers, based within the UK, and with core interests in psychology in general and autism in particular. It is firstly argued that the cognitive sciences rely upon a particular, historically novel, construction of the social in order to articulate autism as social disorder. It is then argued that, although autism is frequently reported as heterogeneous and illusory within the laboratory, social disorder appears self-evident when the autistic individual is required to interact with both the researcher and broader society. Following these findings it is argued that the researcher does not observe autism but, rather, that they achieve it. Finally it is argued that the language of autism is itself capable of ushering in novel conceptualisations of social conduct that may apply to all individuals and not just those diagnosed with the condition. Following these empirical findings it is argued that autism is best understood as the consequence of particular socio-historical conditions. It is asked if these socio-historical conditions may include a novel knowledge-power nexus arising in the mid-twentieth century, named here a socio-emotive politics, of which autism is just one consequence.

The remaining three manuscripts deal with issues directly related to the graft survival and the use of gene therapy and animal models of PD, looking at the dynamics of viral vector gene expression in the pathological brain, an investigation of the two-layer staining obtained on immunohistochemical stained sections, and a re-assessment of the amphetamine induced rotational response of dopamine grafted animals

Transgenic rat models were used to investigate aspects of the role and expression of the circadian clock-regulated genes, vasopressin (VP) and Period 1 (Per1), in the suprachiasmatic nucleus (SCN), and the early growth response gene-1 (egr-1), in the SCN and anterior pituitary gland. Initial studies of the JP-17 and JP-59 (rat VP promoter / human growth hormone (hGH) reporter gene) transgenic lines detected expression of the reporter gene in the supraoptic nucleus (SON), but not the SCN, as determined by in situ hybridisation (ISH). Therefore, no further analysis of circadian clock-regulated VP expression could be conducted in these lines. Four (mPer1) promoter / destabilised enhanced green fluorescent protein (d2EGFP)) (Y) transgenic lines were generated, but failed to express the EGFP reporter gene. Sequence analysis of the transgene failed to explain the absence of expression. Subsequent studies in the egr-1 d4EGFP (57C) transgenic line detected tissue-specific constitutive expression and inducible expression in the brain, as determined by ISH. Region-specific and physiologically-regulated expression in the pituitary gland was also detected, as determined by Northern blot and dual ISH / immunohistochemical (IHC) analyses. Nevertheless, direct GFP fluorescence remained undetectable in this line. Seven egr-1/d2EGFP (Z) transgenic lines were then generated using a novel construct lacking the egr-1 intron present in the 57C construct. These transgenic lines demonstrated tissue-specific constitutive expression and inducible expression in the brain, and region-specific expression in the pituitary gland, at all levels of detection: transcript, protein and direct fluorescence. These findings indicate that the egr-1 intron is not required for directing tissue-specific and inducible expression in the rat. Z line rats will provide insights into the cellular progression of clock- and light- regulated responses in the brain. Potential uses of the egr-1/d2EGFP transgenic model are discussed

The work presented in herein focuses on the optimisation and use of established animal models to study behavioural, pharmacological, histological and molecular correlates of the debilitating motor side effects of current and future treatments for Parkinson’s disease, namely L-DOPA induced dyskinesia (LID) and graft induced dyskinesia (GID). Chapter 3 optimises the 6-OHDA lesion model in mice, from surgical approaches to behavioural assessment of motor function. The neurotoxin was injected at three different regions along the nigrostriatal tract to produce unique patterns of dopaminergic cell death in the midbrain. The resulting cell loss was correlated to behavioural deficits identified through an extensive battery of motor hand tests. Fully lesioned mice from each of the three models were chosen for chronic L-DOPA treatment, described in Chapter 4, where doses were increased every 1-2 weeks. Behaviour was assessed and correlated to deficits on motor hand tests prior to L-DOPA treatment, cell loss within sub regions of the midbrain, serotonergic density levels and upregulations in ΔFosB and striatal TH cell populations. Chapter 5 uses knowledge gained in previous chapters to use the most appropriate 6-OHDA mouse model of LID for the examination of changes in the Regulators of G-protein Signalling (RGS) following an acute and chronic L-DOPA treatment. RGS2 was the only one to increase significantly following either treatment regime. In Chapter 6 a well established rat model of GID (the induction of dyskinesia in the transplanted 6-OHDA lesioned rat through the administration of amphetamine) was used to assess the use of pharmacological agents known to reduce LID. Changes in locomotor function and abnormal inhibitory movements (AIMs) could be assessed giving an insight into the mechanism and receptors involved. To further the understanding of GID, Chapter 7 examines dopamine receptor levels, RGS transcript expression, and the proportions of dopamine and serotonin cells in the transplanted, 6-OHDA lesioned rodent brain. The aim was to determine any correlation between these parameters and amphetamine induced dyskinesia. Only the number of dopaminergic and serotonin cells could be correlated to dyskinesia and not the proportion of serotonin cells. As no previous mouse model of GID has been established, Chapter 8 demonstrates that transplantation of E12 ventral mesencephalon (VM) grafts can be optimised in the lesioned mouse of C57/Bl6 and CD1 strains to give functional recovery, and amphetamine induced dyskinesia. Both strains were also used to demonstrate that transplants were also able to reduce LID.

Amnesia impairs episodic memory in humans, and can occur following damage to the medial diencephalon. This thesis examines the importance of the anterior thalamic nuclei, a diencephalic structure, for different elements (what, when, where) of episodic memory (or episodic-like memory) in rats. The overall goal was to understand why this region is so critical for normal memory in humans. An initial series of experiments investigated the ability of rats with anterior thalamic lesions to recognise objects and odours with a variety of delays. These experiments found that anterior thalamic lesions spare item recognition (what). This same ability was further explored with the use of the immediate early gene zif268, and the results again indicated that this thalamic nucleus does not have a direct role in recognition. A related series of studies explored the effects of anterior thalamic lesions on temporal order judgments (when) for objects. Two different discrimination procedures were tested, between-block recency and within-block recency. Lesions to the anterior thalamic nuclei selectively impaired performance on within-block recency but spared between-block recency. The ability of rats with anterior thalamic damage to discriminate between two locations (where) in both complex and simple environments was also tested. Anterior thalamic lesions significantly impaired place learning compared with control animals, despite the finding that rats with anterior thalamic lesions could sometimes discriminate between the two locations (i.e., could perform significantly above chance). In addition, the effects of anterior thalamic damage on biconditional learning (what-where conjunction) were examined. The rats were trained on both item-place and item-context associations. Lesions to the anterior thalamic nuclei disrupted acquisition of the former, but not the latter. The results suggest that the contributions of the rodent anterior thalamic nuclei to episodic memory, as part of the extended-hippocampal system, primarily reflect the involvement of these nuclei in allocentric spatial learning.

The thesis investigates reasons for a failure of neutrophils to clear pulmonary bacterial infection in cystic fibrosis (CF). Patients with CF experience bacterial colonization of the lungs associated with progressive lung injury and poor prognosis. Neutrophil responsiveness in vitro was determined in patients with CF at different clinical states, and compared to healthy subjects. Neutrophils from the patients were able to phagocytose and kill Pseudomonas aeruginosa effectively, but the presence of sputum sol reduced intracellular killing. Superoxide generation and elastase release in response to fMLP were shown to be reduced in neutrophils from patients with an exacerbation of respiratory symptoms. This was not observed when the cells were stimulated with PMA, which acts intracellularly, rather than through cell surface receptors. There was no difference in the down-regulation of cell surface L-selectin and up-regulation of CD11b in response to fMLP suggesting no alteration in number or function of the fMLP receptors. There was increased adherence to nylon columns by neutrophils from patients with CF. Infection resulted in a greater proportion of band neutrophils, and this correlated with the reduced superoxide generation and elastase release, although it was not possible to separate the band forms to prove this conclusively. Alterations in circulating lipid and fatty acid composition inpatients could potentially affect neutrophil membrane composition and fluidity and therefore signal processing or release of products. The reason for the observed reduced responsiveness appears to be multifactorial. Band cell number post receptor signalling and/or receptor desensitization may result in the observed reduced responsiveness, which returns towards healthy subject levels after treatment of an exacerbation, is likely to be the result of chronic infection.

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder for which there is no cure. At the neuropathological level, AD is characterized by the presence of large numbers of amyloid-beta containing plaques (Aβ-plaques), and neurofibrillary tangles comprised mostly of hyperphosphorylated aggregated protein tau. Both types of deposit are associated with neuroinflammation, synaptic and neuronal cell loss. Accumulating evidence indicates a role for metabolic dysfunction in the pathogenesis of AD. Type 2 diabetes increases the risk of developing AD and several post-mortem analyses have reported evidence of insulin resistance in Alzheimer brain tissue. Insulin-based therapies have emerged as potential strategies to slow cognitive decline in AD, these include the use of insulin sensitizers, such as rosiglitazone, which mediates its effects on insulin sensitivity via the peroxisome proliferator-activated receptors-gamma (PPAR-γ) receptor. While the results of insulin sensitizers on cognition in animal models of AD have been largely positive, the impact of these compounds on cognitive decline in AD patients has been variable. Animal experiments provide a unique opportunity to examine the specific conditions and mechanisms by which insulin sensitizer’s impact on AD-related pathology. This thesis details experiments conducted in a popular Amyloid Precursor Protein overexpressing transgenic mouse model of amyloid pathology that overproduces human Aβ. The aim of these experiments was to determine if chronic dosing with rosiglitazone ameliorated phenotypic behavioural deficits in transgenic mice, and lowered specific biomarkers associated with Aβ over-production. The results indicate that rosiglitazone largely does not reverse phenotypic behavioural alterations in these mice, nor does it reduce total Aβ levels. From this preclinical data, it is concluded that rosiglitazone is likely not a suitable therapeutic treatment for use in human patients with AD.

The pathophysiology of acute post-stroke hyperglycaemia (PSH) is important as hyperglycaemia affects the majority of stroke patients, and is consistently associated with poorer outcome in terms of survival, disability and markers of brain injury such as infarct expansion. There appears to be an interaction between brain arterial patency and hyperglycaemia that has not been fully characterised. This thesis initially reviews the literature on hyperglycaemia and stroke before focusing on animal models of PSH and clinical trials of insulin treatment for PSH in two systematic reviews with meta-analysis. The thesis then looks at the relationship between glucose profiles and clinical outcome in a historical population receiving IV thrombolysis for acute ischaemic stroke, specifically exploring alternative indices of glycaemic state to compare the optimal predictive index for functional outcome as measured by the modified Rankin scale. The main body of the thesis details a prospective observational clinical study which recruited 108 patients within 6 hours of acute ischaemic stroke. These patients had careful monitoring of blood glucose levels over a 48 hour period and detail brain imaging including CT perfusion scanning to examine the ischaemic penumbra, CT angiography on admission and at 24-48 hours to document brain arterial patency with follow-up CT brain imaging to assess outcome infarct volume. The relationship between 48 hour blood glucose profiles, clinical outcome and imaging findings is then explored. The main findings of the thesis are summarized below. · Animal models of PSH have shown that hyperglycaemia exacerbates infarct volume in MCA occlusion models but studies are heterogeneous, and do not address the common clinical problem of PSH because they have used either the streptozotocin model of type I diabetes or extremely high glucose loads. · Animal models show that insulin has a non-significant and significantly heterogeneous effect on infarct growth. 3 · Clinical trials of insulin for post stroke hyperglycaemia have shown no benefit in terms of improved functional outcomes or mortality. Insulin is associated with an increased risk of hypoglycaemia. · In a historical cohort mean capillary blood glucose over 48 hours was more predictive of clinical outcome that admission blood glucose or two consecutive elevated glucose measurements. · A high proportion of acute stroke patients have a blood glucose level above 7mmol/L within 6 hours of onset. Different patterns of blood glucose levels define different populations. · Higher admission and mean glucose levels correlate with larger infarct volumes. Larger core perfusion lesion volumes are associated with a greater risk of mortality. Admission hyperglycaemia is more harmful than hyperglycaemia after 6 hours. · In patients with angiographic evidence of an arterial occlusion infarct volume varies significantly with glycaemic status. In some populations late hyperglycaemia is associated with better imaging outcomes. · Tandem occlusions are associated with bad outcomes after ischaemic stroke.

The aim of this thesis was to explore perceptual processing in individuals with autism and Asperger's syndrome, and to assess the extent to which the theory of weak central coherence could account for any abnormalities in this area. In Experiment 3:1 we presented individuals with autism with four illusions on a computer and asked them to adjust certain parts to appear the same. The results showed just as susceptible to illusions as those without autism on a computer task contrary to previous literature (Happe, 1996). In Experiment 3:2 we presented the same illusions on card and asked participants to judge whether parts of the stimuli were the same or different as in Happe's procedure. Our results showed that autistic populations succumbed to illusions regardless of whether they verbally judged or manually made adjustments to the stimuli. This ruled out the possibility that procedural differences could account for our failure to replicate Happe's findings. These results show that coherence is intact at low levels of perceptual processing in autism. Our second study (Experiment 4:1) explored whether individual differences in coherence may be able to explain why the results of Experiments 3:1 and 3:2 were not consistent with Happe's findings. We presented a battery of visuo-spatial tasks (block design, embedded figures, Rey complex figure test) and the visual illusion computer task to participants. Performance on these tasks was unable to predict susceptibility to visual illusions, suggesting that perception of illusions may not be related to weak central coherence. Our final investigation explored whether autistic populations were more inclined to rely on visual rather than semantic properties when asked to pair atypically coloured pictures (e. g. blue banana) with colour patches (e. g. yellow or blue). Those with autism relied on background knowledge like control participants choosing the semantically related colour. We then considered whether requiring the participants to name the object before selecting a colour may have influenced them to choose the semantic alternative in Experiment 5:2. Those with autism performed similarly to comparison groups choosing the semantic rather than the visual option. This demonstrated that background knowledge was just as salient to those with autism and Asperger's syndrome as those without autism.

Schizophrenia is a complex disorder, with several genes putatively associated with the pathogenesis of the disorder. A large genome-wide association study (O’Donovan et al. 2008) identified ZNF804A as a candidate gene for schizophrenia (meta-analysis p = 1.61 x 10-7). The association of the gene with schizophrenia (and bipolar disorder) has since been successfully replicated several times, confirming the association (Riley et al., 2010; Steinberg et al., 2010; Zhang et al., 2010, Williams et al., 2011). The aim of this thesis is to create and provide preliminary assessments of a mouse model of the murine form of ZNF804A, Zfp804a. A mutagenised DNA archive derived from mice treated with N-ethyl-N-nitrosourea (ENU) held at the MRC Mammalian Genetics Unit, Harwell, was screened for mutations in Zfp804a. Two mutations (C59X and C417Y) were selected for re-derivation based upon the estimated impact upon the protein. The mutations were backcrossed onto a C57Bl/6J background for three successive generations using a panel of genetic markers to aid selection for the highest level of C57Bl/6J congenicity (and therefore speed up the backcrossing process). G4 mice were tested in the study. Preliminary assessments of the fourth generation intercross cohort revealed, most notably, that the mice breed well, have no gross physical deficits and that male Zfp804aC59X/C59X mutants appeared less anxious than other groups in the elevated plus maze and performed better than other groups on the RotaRod. Initial indications show that Zfp804a may indeed influence behaviour and cognition however further work is necessary to expand upon these findings with larger samples.