The ability of the nervous system to retain, manipulate and use visual information which is no longer present in the external environment contributes to intelligent behaviour. A new approach to studying visual working memory has led to re-evaluation of the nature of its limitations in keeping with a finite memory resource which is flexibly distributed across space according to attentional priority. Using a novel behavioural paradigm to study visual working memory precision for sequentially presented items, I demonstrate how the resolution with which healthy subjects recall simple objects changes dynamically with each new item in the sequence. Stochastic modelling of the distribution of responses suggested that memory for earlier objects in the sequence was especially prone to failure in integration of visual features, such as orientation and colour, into complete objects. Next, I examined how memory precision was affected by attentional selection according to the relative behavioural relevance of objects in a sequence, and explored the limitations in this filtering process and their relationship with performance on standard measures of memory and intelligence. The role of updating of non-spatial visual working memory across time was then examined in patients with visual neglect following right hemisphere stroke, revealing a profound non-spatial impairment in WM and its voluntary attentional control in neglect, when compared to stroke patients without neglect and healthy control subjects. Lesion analysis identified separable neural correlates of these deficits. Dopaminergic activity in the prefrontal cortex and basal ganglia has a pivotal and complex role in mediating and controlling working memory and attentional processes. In a randomised, double-blind, placebo controlled study, employing a replicated ABA N-of-1 randomised design, I tested the hypothesis that the dopamine agonist rotigotine improves visual neglect following right-hemisphere stroke. Rotigotine was associated with significant improvement in visual search, an effect that appears to have been mediated by an enhancement of selective, goal-directed attention. The medial temporal lobe (MTL) has an established role in supporting long-term memory processes, but its involvement in working memory has been debated recently. I studied visual working memory for sequentially presentated objects in four patients with MTL lesions and found that short-term memory can be compromised in such individuals. Overall, this thesis explores how visual working memory is updated dynamically across time according to attentional priority in health, how these processes are affected in patients with visual neglect following right hemisphere stroke and in those with medial temporal lesions, and how a dopamine agonist might ameliorate visual neglect by modulating selective attention. The thesis concludes with a brief discussion suggesting further research directions.

This thesis focuses on the non-motor symptoms (NMS) of Parkinson's disease (PD), with particular emphasis on impulsive-compulsive spectrum behaviours (ICBs). The relevance of NMS is investigated as presenting complaints amongst 21% of 433 patients with pathologically-proven PD. These patients had delayed diagnosis of PD, and often underwent unnecessary and potentially harmful interventions. Motor and NMS of LRRK2- associated PD are compared with idiopathic PD, and NMS of LRRK2-associated PD are more benign than in idiopathic PD. Increased Lewy body burden was shown to correlate with NMS in a subgroup of idiopathic PD patients at post-mortem. A metanalysis investigates pathological gambling and the dopamine dysregulation syndrome in PD, highlighting the importance of dopamine agonists in pathological gambling. A new ICB in PD, named "reckless generosity", occurring in association with dopamine agonist use is described. Dopamine agonists are also implicated in the development of ICBs in patients with pathologically-confirmed progressive supranuclear palsy. An increased prevalence of excessive hoarding as a NMS of PD is shown, particularly in patients with other ICBs. Hoarding is shown to relate more to impulsivity measures than obsessive-compulsive symptoms. The overlap between sleep disturbance, dreams, mood and ICBs in PD is demonstrated. Sleep disturbance is independent of dopaminergic medication use, and instead relates to anxiety, depression, and the presence of ICBs. Reward learning in PD patients with and without ICBs, and healthy controls is tested on a reaction time game, the Salience Attribution Test. PD patients with ICBs show increased impulsivity on temporal discounting tasks, but do not show increased aberrant salience to rewards. Positron emission tomography (PET) scanning is used to investigate the dopaminergic pathways involved in ICBs by comparing PD patients with ICBs to those without ICBs after a L-dopa challenge and visual reward-related cue exposure. The results differ from studies on PD patients with addictive dopamine dysregulation, or other non-PD drug addicted individuals. The role of deep brain stimulation surgery for PD, patients with ICBs is investigated in a case series of 21 operated PD patients exhibiting ICBs at some stage during the course of their disease. The mixed results suggest the need for careful selection of patients for this procedure.

Background: Inherited prion diseases (IPD) show remarkable clinical heterogeneity, posing problems for clinicians in making an early diagnosis and raising questions about genetic or environmental modifiers. The P102L prion protein gene (PRNP) mutation, one of the most frequently identified causes of IPD, has been linked to a large English kindred for three decades. A series of further smaller kindreds did not share apparent ancestry with this large kindred, raising the possibility of distant common ancestry of a rare mutation or no ancestry and relatively common novel mutations occurring. Identification of the genetic modifiers of phenotype may have implications for sporadic and variant Creutzfeldt-Jakob disease (sCJD), and neurodegeneration more widely, while novel mutation rates in PRNP inform study of the unknown aetiology of sCJD. IPD, although rare, has unique advantages in terms of looking for genetic modifiers of prion disease. Methods: Genealogical work and microsatellite haplotyping was carried out on cases of IPD P102L. Clinical information on affected patients was collected retrospectively as well as prospectively. Heritability estimates of IPD were produced by comparing parental and offspring phenotypes. The expanded collection of IPD cases with reliable clinical information (composing P102L and other IPD associated mutations), allowed for the testing of candidate genetic modifiers of phenotype. Polymorphisms previously reported to have an impact on prion disease susceptibility or phenotype were tested, along with a panel of candidate polymorphisms selected from a genome-wide association study (GWAS) in vCJD. Findings: Common ancestry between the known large P102L kindred and a number of previously unlinked P102L kindreds was identified. However, a number of apparently unrelated IPD P102L kindreds remain, suggesting that multiple separate mutational events are responsible: PRNP codon 102 may be a mutation ‘hot- spot’. Microsatellite work on other IPD subtypes in the UK and from elsewhere in Europe also finds evidence of the existence of multiple unrelated kindreds. The genotype-phenotype analysis discovered that codon 129 is a modifier of IPD P102L; in that codon 129 homozygosity is associated with an earlier age at clinical onset. Also identified was a possible effect of APOE genotype on IPD. APOE-E4 is associated with a significantly later age at onset in IPD, which has recently been identified in other categories of neurodegenerative disease. Interpretation: Overall, this work provides indirect support for the plausibility of the ‘somatic mutation hypothesis’ of sporadic CJD (sCJD), by suggesting a relatively frequent novel mutation rate in PRNP. The genetic contribution to phenotypic heterogeneity in IPD was estimated. This contribution is significant and may inform the search for genetic modifiers of susceptibility to acquired prion disease. The models established to analyse IPD as a quantitative trait will be used in future prion disease GWAS.

Parkinson’s disease (PD), affecting about 1% of the population over 60 years old, is the second most common neurodegenerative disease. Although the majority of PD cases are sporadic, mutations in a number of genes are known to cause or increase the risk of developing PD. Studying the function of these genes can shed light on our understanding of the molecular mechanisms contributing to both familial and sporadic forms of PD. This thesis is focused on two of these genes: PINK1 and LRRK2. Mutations in the PINK1 gene are the second most common cause of autosomal recessive forms of PD. PINK1 is a serine/threonine kinase with a mitochondrial targeting sequence. Studies have suggested a neuroprotective role for PINK1 and loss of PINK1 function can lead to mitochondrial dysfunction, increased oxidative stress, altered mitochondrial fission/fusion and autophagy. In this study, I have investigated the effects of PINK1 deficiency on mitochondrial and cellular function as well as brain and skeletal muscle pathology using a PINK1 knockout (KO) mouse model. Although aged PINK1 KO mice don’t develop obvious pathological changes (e.g. neurodegeneration or abnormal protein aggregation), PINK1 KO leads to impaired mitochondrial respiration in both midbrain neurons and skeletal myocytes. The difference in cellular energy metabolism between these two cell types determines their different responses to PINK1 deficiency in terms of mitochondrial membrane potential and calcium regulation, which may explain the selective vulnerability of neurons to PINK1 associated PD. In addition to this work, I have also studied the expression and distribution of putative LRRK2 substrates, ERM (ezrin, radixin and moesin) proteins, in PD brains compared to healthy controls. Mutations in the LRRK2 gene are the most common cause of autosomal dominant forms of PD. LRRK2 is a large protein kinase with multiple functional domains. ERM proteins have been demonstrated as putative LRRK2 substrates by in vitro assays. In this study, I examined the expression levels and distribution of moesin and phospho-ERM proteins in post-mortem PD brains and found that the level of phospho-ERM was increased in striatum from sporadic PD brains compared to controls. Furthermore, phospho-ERM proteins were localized to Lewy bodies, one of the main PD pathological hallmarks, in both IPD brains and brains of PD patients carrying the G2019S LRRK2 mutation. These results suggest that ERM proteins might play a role in the pathogenesis of PD.

Imaging studies have shown that response inhibition and conflict resolution, key executive functions essential for adaptive behaviour, are mediated by fronto-striatal circuits, specifically the right inferior frontal gyrus (IFG), pre-supplementary motor area (pre-SMA), and the subthalamic nucleus (STN). Overactivity of the STN and underactivation of the pre-SMA are features of Parkinson's disease (PD). The aim of this thesis was to investigate the hypothesis that the STN is involved in inhibition of action and serves to provide time for reflection and to slow down decision-making under conflict (Frank et al. 2006; 2007). The conditional stop signal reaction time was used. In Study 1, compared to controls, medicated PD patients showed impaired motor inhibition with prolonged stop signal reaction times (SSRTs), problems with conflict resolution, and impairment on tests of cognitive inhibition. In Study 2, PD patients were impaired on motor inhibition and conflict resolution compared to controls both on and off medication. Most importantly, medication status did not influence motor inhibition or conflict resolution in PD. In Study 3, bilateral deep brain stimulation of the STN impaired inhibition as indicated by significantly prolonged SSRTs, but had no significant effect on conflict resolution. However, Study 4, showed that compared to controls and unoperated PD, PD with unilateral subthalamotomies showed faster Go RTs and less conflict-induced slowing but at the expense of greater response selection errors. Finally, Study 5 investigated the contributions of the pre-SMA and IFG to motor inhibition and conflict resolution in healthy controls using transcranial magnetic stimulation. The results confirmed that the contribution of the right pre-SMA to response inhibition is essential. The results are consistent with the proposal that the STN and pre-SMA are involved in reactive inhibition and that the STN together with the pre-SMA influences the response threshold and speed-accuracy trade-offs.

Visual working memory, a limited temporary storage system for relevant information, is a critical contributor to many essential cognitive functions. In this thesis, I aimed to investigate some of the mechanisms underlying working memory in healthy humans and neurological patients, as well as its modulation by processes attributed to attention and the neurotransmitter dopamine. There currently is an important controversy regarding models of working memory. I investigated whether a resource model of memory (which argues for a limited resource distributed amongst to-be-remembered items) might be extended to the domain of visual motion. The results suggest that this is indeed be the case, supporting the utility of this model as a general conceptual framework for understanding working memory across a range of visual features and modes of presentation (Chapter 2). A comprehensive model of working memory should consider its relationship with attention. My findings point to an intimate yet highly specific relationship between these two processes, demonstrating that attention is essential for maintenance of integrated features within working memory (Chapters 2 and 4). Further, evidence for a causal role of early visual areas in maintenance of items in focus of attention, compared to the full content of working memory, is provided using transcranial magnetic stimulation (Chapter 3). Finally, I investigated neuromodulation of working memory processes by dopamine in patients with dopamine dysfunction (Parkinson’s disease) and using the dopamine agonist, Cabergoline, in healthy controls. The results demonstrate that dopamine can modulate working memory precision (Chapter 5 and 6). Furthermore, deficits in working memory were also observed in individuals with glucocerebrosidase mutations who have a significantly raised risk of developing Parkinson’s disease (Chapter 7). I discuss the possibility that specific deficits in working memory might provide a cognitive marker of risk for neurodegeneration and development of Parkinson’s disease.

Progressive supranuclear palsy (PSP) belongs to a group of neurodegenerative disorders that are characterised by hallmark pathology consisting of intra-neuronal aggregates of the microtubule-associated protein, tau. In PSP, these aggregates are almost exclusively composed of one of the two major tau protein isoform groups normally expressed at similar levels in the healthy brain, indicating a role for altered isoform regulation in PSP aetiology. Although no causal mutations have been identified, common variation within the gene encoding tau, MAPT, has been highly associated with PSP risk. The A-allele of the rs242557 single nucleotide polymorphism has been repeatedly shown to significantly increase the risk of developing PSP. Its location within a distal region of the MAPT promoter region is significant, with independent studies – including this one – demonstrating that the rs242557-A allele alters the function of a transcription regulatory domain. As transcription and alternative splicing processes have been shown to be co-regulated in some genes, it was hypothesised that the rs242557-A allele could directly affect MAPT alternative splicing through its differential effect on transcription. This project describes an investigation into the molecular mechanism linking the MAPT association with the tau isoform dysregulation characteristic of PSP. The design, construction and in vitro investigation of minigenes representing common MAPT variants will be presented in detail and will demonstrate that promoter identity plays an important role in regulating the alternative splicing of MAPT transcripts. The specific role of the rs242557 polymorphism in MAPT transcription and splicing are investigated and the two alleles of the polymorphism are shown to differentially influence these two molecular processes, providing a plausible mechanism linking the two phenomena known to be associated with PSP – a common genetic variant within the MAPT promoter region and detrimental changes to tau isoform production.

Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disorder, for which there is no effective treatment. Clinical trials in ALS are hindered by the lack of reliable biomarkers to enable an early diagnosis and monitoring of disease progression. There is therefore an urgent need to develop biomarkers for ALS. In this Thesis, I examined the possibility that plasma Neurofilament levels may be a biomarker of disease progression in mouse models and ALS patients. I first developed an enzyme-linked immunosorbent assay (ELISA) method that overcomes a well-known technical obstacle, the NfH ’Hook Effect’, to enable accurate quantification of plasma Neurofilament Heavy Chain (NfH). Using this ELISA, I next examined the longitudinal changes in plasma NfH levels in SOD1G93A mice that model ALS. I observed a significant increase in plasma NfH levels as disease progressed, which correlated with the decrease in functional and morphological read-outs of neuromuscular function and motor neuron survival. Moreover, treatment of SOD1G93A mice with different disease-modifying agents, Arimoclomol, Cogane and Riluzole, was found to not only ameliorate disease but also reduce plasma NfH levels compared to untreated controls. Plasma NfH levels may therefore be of use in monitoring drug efficacy and disease progression in clinical trials. Examination of plasma NfH levels in an alternative ALS model with a slower disease progression, the SOD1G93Adl mouse, confirmed that plasma NfH levels increase during disease progression in models other than the rapidly progressing SOD1G93A model. The relevance of these findings of increased plasma NfHs in SOD1 mice as a biomarker of disease progression in the human disease were examined in the final Results Chapter using serial samples longitudinally collected from a cohort of 136 ALS patients with follow-up up to three years. To sum up, the results presented in this Thesis detail the investigation of plasma NfH levels as a disease progression marker in SOD1 mice and in ALS patients.

Disorders of Coenzyme Q10 (CoQ10) biosynthesis represent the most treatable subgroup of mitochondrial diseases. Neurological involvement is frequently observed in CoQ10 deficiency, typically presenting as cerebellar ataxia and/or seizures. The aetiology of the neurological presentation of CoQ10 deficiency has yet to be fully elucidated and therefore in order to investigate these phenomena we have established a neuronal cell model of CoQ10 deficiency by treatment of the neuronal SH-SY5Y cell line with Para-AminoBenzoic Acid (PABA). This neuronal cell model provides insights into the effects of CoQ10 deficiency on neuronal mitochondrial function and oxidative stress. A marginal decrease in CoQ10 status (76% residual CoQ10) appears to be sufficient to impair Electron Transport Chain (ETC) function and increase mitochondrial oxidative stress, highlighting the vulnerability of neurons to a small deficit in CoQ10 status. In contrast to CoQ10 deficient fibroblasts, a CoQ10 deficiency (46% residual CoQ10) in neuronal cells appears to result in reversal of Complex V activity. This phenomenon has not been reported in previous studies of CoQ10 deficiency and may be a unique characteristic of neuronal cells. This neuronal cell model was subsequently utilised in the evaluation of candidate therapies for neurological conditions associated with CoQ10 deficiency. The efficacy of CoQ10 supplementation and methylene blue (MB) treatment were evaluated. CoQ10 supplementation proved effective at preventing mitochondrial oxidative stress and partially restoring neuronal mitochondrial function. However ETC complex activities were still compromised, suggesting an explanation for the refractory nature of neurological CoQ10 deficiency to treatment. Muscle is considered the “gold standard” for CoQ10 quantification; however neurological CoQ10 deficiency does not always present with a significant decrease in muscle CoQ10 status, despite a genetic diagnosis of CoQ10 deficiency. Cerebrospinal Fluid (CSF) CoQ10 quantification offers a more direct measurement of cerebellar CoQ10 levels. A tandem mass spectrometry (MS/MS) method capable of quantifying nanomolar (nM) levels of CoQ10 was therefore developed. In conclusion this PhD thesis has been successful in expanding our understanding of the pathophysiology of neuronal CoQ10 deficiency and subsequently suggesting why neurological CoQ10 deficiency might be refractory to CoQ10 treatment. This thesis has also led to the development of a new technique for quantification of CSF CoQ10 concentration, opening up many possibilities for future studies and applications.

MMN has been characterized as an automatic, pre-attentive, error detection mechanism that may aid switch in attention towards a salient stimulus as well as assisting with contrast enhancement of sensory data. There is interest clinically in the MMN given its abnormality in a number of neurological/neuropsychiatric disorders. This thesis explores tDCS as a practical tool to modulate MMN in health, to assess impaired somatosensory error detection as an explanation of the underlying mechanism of dystonia by utilising somatosensory mismatch negativity (MMN) and to explore the role of the cerebellum in both error detection and dystonia using tDCS. We found that tDCS is a reliable way to alter the generation of MMN perhaps through homeostatic interaction in frontal auditory MMN generator. Subsequently, we found a reliable way to generate somatosensory MMN and then by using tDCS stimulation were able to give more evidence of the relationship between the cerebellum and somatosensory MMN. Finally, we have demonstrated an abnormality in somatosensory MMN in patients with dystonia which is strongly correlated with somatosensory temporal discrimination threshold. Additionally, we showed that there was no modulation of sMMN after cerebellar tDCS in patients with dystonia. This thesis suggests that, in health, tDCS is able to modulate both auditory and somatosensory MMN. These experiments provide evidence for a new way in which MMN might be modulated for experimental purposes and potentially for clinical/therapeutic purposes. The finding of reduced somatosensory MMN (but not auditory MMN) in dystonia suggests a specific deficit in error detection in the somatosensory domain in dystonia. The correlation of somatosensory MMN with somatosensory temporal discrimination threshold, which is robustly abnormal in dystonia, implies a shared mechanism between the two phenomena. We suggest that somatosensory temporal discrimination could be subsumed within somatosensory MMN as an error detection task. The dependence we have demonstrated for somatosensory MMN on the cerebellum in healthy people fits with a role for the cerebellum in the pathophysiology of dystonia. We could not modulate somatosensory MMN in patients with dystonia with tDCS as we could in healthy people, and this provides some additional evidence that cerebellar dysfunction might be relevant both for the abnormal somatosensory MMN in dystonia but also for the clinical manifestation of dystonia itself.