Natural auditory scenes consist of a rich variety of temporally overlapping sounds that originate from multiple sources and locations and are characterized by distinct acoustic features. It is an important biological task to analyze such complex scenes and extract sounds of interest. The thesis addresses this question, also known as the “cocktail party problem” by developing an approach based on analysis of a novel stochastic signal contrary to deterministic narrowband signals used in previous work. This low-level signal, known as the Stochastic Figure-Ground (SFG) stimulus captures the spectrotemporal complexity of natural sound scenes and enables parametric control of stimulus features. In a series of experiments based on this stimulus, I have investigated specific behavioural and neural correlates of human auditory figure-ground segregation. This thesis is presented in seven sections. Chapter 1 reviews key aspects of auditory processing and existing models of auditory segregation. Chapter 2 presents the principles of the techniques used including psychophysics, modeling, functional Magnetic Resonance Imaging (fMRI) and Magnetoencephalography (MEG). Experimental work is presented in the following chapters and covers figure-ground segregation behaviour (Chapter 3), modeling of the SFG stimulus based on a temporal coherence model of auditory perceptual organization (Chapter 4), analysis of brain activity related to detection of salient targets in the SFG stimulus using fMRI (Chapter 5), and MEG respectively (Chapter 6). Finally, Chapter 7 concludes with a general discussion of the results and future directions for research. Overall, this body of work emphasizes the use of stochastic signals for auditory scene analysis and demonstrates an automatic, highly robust segregation mechanism in the auditory system that is sensitive to temporal correlations across frequency channels.

Huntington’s Disease is a progressive, adult onset, neurodegenerative disease. It is inherited in an autosomal dominant fashion and is caused by a trinucleotide repeat expansion in huntingtin, which encodes the protein huntingtin. The length of the expanded trinucleotide repeats accounts for some, but not all of the age of onset of the condition. Despite the monogenic basis of Huntington’s disease, the clinical features display marked variability within families and between those who carry the same length expansion. The variability in age of onset and in clinical features is likely to be due to a number of environmental and other genetic factors. Identification of these factors may lead to novel therapeutic approaches. A number of potential disease modifying agents have been elucidated and are approaching clinical trials. Robust ‘biomarkers’ of disease onset and progression are essential for developing a framework for future clinical trials that have the ability to judge the efficacy of any therapeutic intervention. In this thesis I will present work arising from TRACK-HD, an international observational biomarker study of Huntington’s Disease which has identified a panel of biomarkers for use in future clinical trials. A number of subs-studies arising from TRACK-HD are also presented here. Firstly, a study investigating the use of Positron Emission Tomography and peripheral immune markers as biomarkers of disease progression followed by a candidate genetic modifier study focusing on immune pathways as genetic modifiers of age of onset in Huntington’s Disease.

This thesis explores how a hierarchical Bayesian model of the brain can explain aspects of schizophrenia and ‘functional’ motor and sensory symptoms (FMS). Bayesian computations prescribe the optimal integration of prior expectations with sensory evidence using the relative uncertainty (precision) of each source of information. The accurate representation of precision is crucial; its loss can lead to false inference by overweighting prior expectations or sensory evidence. Numerous phenomena in schizophrenia have been thought of as due to a loss of the brain’s predictive ability. This could be the result of a loss of precision of prior beliefs, through a reduction in synaptic gain at higher hierarchical levels, and I use this alteration to model the reduced mismatch negativity and three characteristic smooth pursuit eye movement (SPEM) abnormalities found in schizophrenia. My ultimate goal is to use the pursuit (and other) models to make inferences about the cortical encoding of precision. To establish the face validity of this approach, I tried to manipulate the precision estimates in normal subjects’ internal models, using a target moving with ‘smooth’ or ‘noisy’ (imprecise) velocity. I used subjects’ eye movements and DCM to invert the pursuit model and estimate the parameters of subjects’ internal models. I showed that noisy target velocity caused subjects to attend more to the sensory aspects of the stimulus (i.e. increased sensory precision). To demonstrate the construct validity of this pursuit DCM, I used magnetoencephalography (MEG) and DCM to test its prediction that noisy target motion increases sensory precision: corresponding to a decrease in the self-inhibition of superficial pyramidal cells in early visual cortex. Noisy motion decreased self-inhibition in central V1 at the group level, and in V2, changes in sensory precision in the pursuit DCM correlated with changes in V2 self-inhibition in the MEG DCM on an individual subject basis.

Mutations in LRRK2 have been established as a common genetic cause of Parkinson’s disease (PD). Variation in gene expression of PARK loci has previously been demonstrated in PD pathogenesis, although it has not been described in detail for LRRK2 expression in the human brain. This study further elucidates the role of LRRK2 in development of PD by describing an investigation into the role of LRRK2 genetics and expression in the human brain. The G2019S mutation is a common LRRK2 mutation that exhibits a clinical and pathological phenotype indistinguishable from idiopathic PD. Thus, the study of G2019S mutation is a recurrent theme. The frequency of G2019S was estimated in unaffected subjects that lived or shared a cultural heritage to the predicted founding populations of the mutation, and was found not to be common in these populations. Morphological analysis revealed a ubiquitous expression for LRRK2 mRNA and protein in the human brain. In-situ hybridisation data suggests that LRRK2 mRNA is present as a low copy number mRNA in the human brain. A semi-quantitative analysis of LRRK2 immunohistochemistry revealed extensive regional variation in the LRRK2 protein levels, although the weakest immunoreactivity was consistently identified in the nigrostriatal dopamine region. No difference was observed in the morphological localisation of LRRK2 mRNA and protein in unaffected, IPD or G2019S positive PD subjects. Dysregulation of LRRK2 mRNA expression and the effects of cis- acting genetic variation on these levels were demonstrated. A widespread decrease of LRRK2 mRNA was observed in IPD and G2019S positive PD subjects in comparison to unaffected controls. Furthermore, non-coding genetic variation was also demonstrated to have an effect on the LRRK2 transcriptional activity in PD subjects. Collectively, these findings suggest that LRRK2 has an important physiological role, and a dysregulation in its levels could affect auxiliary mechanisms that contribute to PD pathogenesis. This data also supports the possibility of a shared mechanism contributing to the identical phenotype of IPD and G2019S linked PD.

NaV1.1 is a voltage-gated sodium channel encoded by the gene SCN1A. Mutations in SCN1A cause dominantly inherited epilepsy syndromes in humans and NaV1.1 is an important target of several anti-epileptic drugs (AEDs). A common polymorphism in this gene has been shown to alter the expression of two splice variants of the channel, NaV1.1-5N (containing exon 5N) and NaV1.1-5A (containing exon 5A). Although the splicing is highly conserved and the polymorphism that modifies it has been associated with altered AED dosage, the functional impact of the splicing on NaV1.1 is unknown. This project used whole cell voltage clamp of heterologously-expressed NaV1.1-5A and 5N to compare the intrinsic properties of the splice variants, their modulation by AEDs, their interaction with a published epilepsy mutation (R1648H), their modulation by G-proteins and how they responded to co-expression of sodium channel β subunits. The main finding was that, although when recorded at physiological temperatures the splice variants produced macroscopic currents that were similar for many parameters, they differed in the rate at which they recovered from inactivation, with NaV1.1-5N recovering more rapidly than NaV1.1-5A. This difference in recovery was conferred by a single amino acid substitution that is conserved in several sodium channels that are alternatively spliced at this site, and the difference was obscured in the presence of the common AED, phenytoin. Inclusion of the mutation R1648H also eradicated the difference by disproportionately slowing the recovery of NaV1.1-5N. Although several other subtle differences were seen, no consistent differences were found in interactions between the splice variants and G-proteins or β subunits. By converging on a single parameter, recovery from inactivation, the data presented here suggest this parameter is modulated by splicing, and may play a role in development and treatment of seizures.

Disorders affecting the basal ganglia can have a severe effect on speech motor control. The effect can vary depending on the pathophysiology of the basal ganglia disease but in general terms it can be classified as hypokinetic or hyperkinetic dysarthria. Despite the role of basal ganglia on speech, there is a marked discrepancy between the effect of medical and surgical treatments on limb and speech motor control. This is compounded by the complex nature of speech and communication in general, and the lack of animal models of speech motor control. The emergence of deep brain stimulation of basal ganglia structures gives us the opportunity to record systematically the effects on speech and attempt some assumptions on the role of basal ganglia on speech motor control. The aim of the present work was to examine the impact of bilateral subthalamic nucleus deep brain stimulation (STN-DBS) for Parkinson’s disease (PD) and globus pallidus internus (GPi-DBS) for dystonia on speech motor control. A consecutive series of PD and dystonia patients who underwent DBS was evaluated. Patients were studied in a prospective longitudinal manner with both clinical assessment of their speech intelligibility and acoustical analysis of their speech. The role of pre-operative clinical factors and electrical parameters of stimulation, mainly electrode positioning and voltage amplitude was systematically examined. In addition, for selected patients, tongue movements were studied using electropalatography. Aerodynamic aspects of speech were also studied. The impact of speech therapy was assessed in a subgroup of patients. The clinical evaluation of speech intelligibility one and three years post STN-DBS in PD patients showed a deterioration of speech, partly related to medially placed electrodes and high amplitude of stimulation. Pre-operative predictive factors included low speech intelligibility before surgery and longer disease duration. Articulation rather than voice was most frequently affected with a distinct dysarthria type emerging, mainly hyperkinetic-dystonic, rather than hypokinetic. Traditionally effective therapy for PD dysarthria had little to no benefit following STN-DBS. Speech following GPi-DBS for dystonia did not significantly change after one year of stimulation. A subgroup of patients showed hypokinetic features, mainly reduced voice volume and fast rate of speech more typical of Parkinsonian speech. Speech changes in both STN-DBS and GPi-DBS were apparent after six months of stimulation. This progressive deterioration of speech and the critical role of the electrical parameters of stimulation suggest a long-term effect of electrical stimulation of basal ganglia on speech motor control.

A wealth of evidence exists to suggest that the cerebellum has an important role in the integration of vestibular, proprioceptive and visual sensory signals. Human bipedal balance depends on sensory integration and balance impairment is a common feature of cerebellar disease. I test the hypothesis that disrupted sensori-motor processing is responsible for balance impairment in cerebellar disease. Balance control in subjects with pure cerebellar disease (SCA6) was compared with matched healthy subjects using a mix of traditional clinical and laboratory-based tests. Sensory processing was explored using a novel combination of tools designed to deliver single-sensory channel balance perturbations. The vestibular, proprioceptive and visual channels were stimulated with galvanic vestibular stimulation, vibration and visual scene motion respectively. Standing balance was explored using 3D whole body motion analysis. Sway speed when standing quietly with eyes open was significantly increased in those with SCA6 and strongly correlated with disease severity scores. Responses to isolated vestibular stimulation suggest largely normal vestibulo-motor processing in SCA6 subjects. Responses had normal latency and magnitude. Response direction followed head position in the normal way suggesting intact vestibulo-proprioceptive integration. Vision had a normal attenuating effect on response magnitude suggesting intact vestibulo-visual integration. Responses to isolated vestibular, proprioceptive and visual stimuli responses were compared to investigate whether there might be a predominant deficit in any one channel. Vestibular and proprioceptive stimuli evoked largely normal responses. In contrast, visual stimuli consistently evoked abnormally large responses with significant timing delays. Increases in SCA6 response magnitudes to moving visual stimuli strongly correlated with disease severity scores. This finding is the first to point to a specific change in sensori-motor processing in cerebellar disease. This finding could contribute to balance impairments but is unlikely to explain balance impairment observed with the eyes closed. Overall sensory processing for balance control in SCA6 is largely intact.

Developing neurones require a dynamic cytoskeleton for newly forming processes to respond to guidance cues. The Rho GTPases, including Rac1, are very important in cytoskeleton remodelling. A GTPase activating protein (GAP) for Rac1, α2-chimaerin, participates in Semaphorin 3A (Sema 3A) and ephrin/EphA4 growth cone collapse pathways, and the SH2 domain is essential, suggesting phosphotyrosine interactions are involved. The activation of α2-chimaerin in collapse pathways may be through diacylglycerol binding to the cysteine rich C1 domain or via phosphotyrosine signalling. This investigation explores the interactions of α2-chimaerin and the role of tyrosine phosphorylation in retraction signalling pathways. The Src family tyrosine kinase, Fyn, is involved in Sema 3A and ephrin A1/EphA4 signalling. Results showed that Fyn can bind to the GAP domain of α2-chimaerin and can phosphorylate α2-chimaerin activated by phorbol 12-myristate 13-acetate, Sema 3A or ephrin A1. Fyn was found to phosphorylate tyrosines 143 and 303 of α2-chimaerin. Overexpression of α2-chimaerin led to increased axon number in hippocampal neurones. Selected point mutations of α2-chimaerin were sufficient to affect axon branching in developing hippocampal neurones. In COS-7 cells, the receptor EphA4 interacted with α2-chimaerin GAP domain when unstimulated and with the SH2 domain when stimulated with the ligand ephrin A1. In N1E- 115 cells permanently expressing EphA4, α2-chimaerin shRNA inhibited ephrin A1- induced cell rounding. The SH2 and C1 domains of α2-chimaerin were both necessary for ephrin A1/EphA4 cell rounding. Another partner of α2-chimaerin is adaptor protein Nck1. The α2-chimaerin Nck1 binding region has been mapped and the binding may be sufficient to activate α2- chimaerin, allowing other partners to interact. A novel α2-chimaerin interacting protein Vps28, a component of the endosomal sorting complex required for transport-I, has been investigated in endocytosis using monoclonal antibodies and shRNA. These results together suggest an important regulatory role for α2-chimaerin and its partners in axonal guidance pathways.

Idiopathic normal pressure hydrocephalus (iNPH) is a condition affecting a small percentage of the elderly population; however it is the only known treatable cause of dementia. Surgical cerebrospinal fluid (CSF) diversion is the only known treatment for the condition today. However, such a procedure is not to be offered lightly and any expected benefit has to balance the associated surgical risks. The prognosis of a favourable surgical outcome has been problematic since the conception of the syndrome. None of current prognostic tests reaches 100% sensitivity or specificity and it is felt that there might be a need for a combination of tests, rather than a single one to maximize the chances of selecting the right patients to offer a surgical CSF diversion procedure. Biomarkers are biological substances that may act as surrogate markers of response to a treatment or to characterise a disease’s progression over time. The aim of this study was to identify CSF markers of favourable surgical outcome in patients with iNPH undergoing the insertion of a ventriculoperitoneal shunt (VPS). We first describe the effects of external lumbar drainage (ELD) on the CSF biochemistry of these patients. Correlations are made with imaging data obtained from volumetric analysis and neuropsychological tests in order to obtain a complete profile of these patients. The rostrocaudal gradients of the CSF markers examined are reported showcasing the need to understand that commonly reported values from lumbar CSF do not necessarily reflect pathological changes occurring at cerebral level. Finally, we report on the individual as well as combined prognostic value of 7 CSF markers on surgical outcomes at 6 months. The pathophysiological significance of these markers is discussed individually. It is concluded that the combined power of total tau and Aβ 1-42 may be useful in predicting favourable surgical outcomes at 6 months; further studies applying the findings in a larger cohort and correlating findings with longer outcomes are warranted to enhance the clinical application. The biochemical profile of patients with iNPH appears unique and different than patients with Alzheimer’s dementia or control subjects.

This thesis concerns the statistical characterisation of neuroimaging data; namely the detection of statistically surprising (significant) responses in measured brain activity that is continuous in space and time. It illustrates the application of these precedures developed using functional magnetic resonance time series of blood oxygenation level dependent (BOLD) signals. The spatio-temporal dynamics of BOLD reflects brain metabolism via the neurovascular coupling, and is therefore of interest to neuroscience. This coupling ensures that local use of oxygen relates to local supply, largely via the inflow of blood. As a physical measure of these fluid dynamics, BOLD unfolds in continuous space-time. In practice however, assessing experimental effects on BOLD requires statistical inference, which is not practical for continuous four-dimensional representations. This thesis contributes four ideas for statistical inference on the neuronal causes of BOLD. Each retains some reduced notion of continuity (in either space or time). The first three Chapters consider ‘spatial’ and ‘topological’ inference on atemporal transformations of the original four-dimensional data. Here, we assume measurements have already been collapsed over time at each sampled point of space via inversion of a General Linear Model; the resulting ‘SPM’ estimates the true field of parameters governing experimentally-induced BOLD. Chapters 2-3 concern topological inference on this true underlying field. The aim is to decide on the occurrence of discrete topological features (e.g. the existence of local maxima) using procedures whose error-rate is known/controlled. Chapter 4 aims to infer experimentally induced patterning in the spatial organisation of topological events (clusters or peaks) in an SPM. Chapter 5 then revisits the temporal dynamics between neuronal populations, whose spatial locations have been determined in advance, and attempts to understand their structure. Chapter 6 concludes this thesis. It summarizes the limitations of our proposed methods and explores their future extension and generalisation.