A high degree of interest continues in studying the central effects of ketamine, an NMDA receptor (NMDAR) antagonist, due to its potential relevance to schizophrenia, antidepressant efficacy and analgesia. This thesis investigates the brain response to ketamine to (1) determine whether it is a valuable pharmacological tool to investigate glutamatergic dysfunction and (2) further our knowledge on the role of the glutamatergic system in human cognition. The administration of low-dose ketamine to healthy volunteers was combined with arterial spin labelling (ASL) and blood oxygen level dependent (BOLD) imaging in both resting state and cognitive task conditions. The test-retest reliability of the ketamine response was assessed together with its effects on working memory (N-BACK) and paired associative learning (PAL) task networks. Modulatory effects of lamotrigine and risperidone, treatments thought to reduce ketamine-induced glutamate release, were also evaluated. Ketamine induced robust and reliable effects in predicted cerebral networks with ASL and BOLD. However, risperidone and lamotrigine only attenuated the BOLD changes. Ketamine altered the load-response profile during the N-Back within the dorsolateral prefrontal cortex and subtly decreased activation in brain areas including the cingulate and thalamus. A trend towards ketamine-induced attenuation of learning-related increases in activation was found in the PAL task. Pretreatments had differential effects depending on the cognitive task, with risperidone having a more prominent role in working memory and lamotrigine in associative learning. Finally, ketamine altered associative learning-related changes in fronto-temporal connectivity. Results suggest (1) the resting brain response to ketamine is a suitable tool to test and validate pharmacological interventions; (2) fMRI demonstrates that the function of NMDARs is implicated in both region-specific and inter-regional brain communication in working memory and associative learning contexts respectively and (3) the distinct downstream neural substrates of NMDAR antagonism are evidenced by the context-dependent reversal of its effects. The findings have implications for future ketamine-related research and its use as a model of glutamatergic dysfunction relevant to psychiatric disorders.

The subventricular zone (SVZ) is one of the main neurogenic niches in the postnatal brain. Neural precursors derived from SVZ stem cells migrate in chains to the olfactory bulb (OB) via the rostral migratory stream (RMS) through channels (glial tubes) formed by the processes of astrocytes. Importantly, these precursors have the capacity to migrate away from their native route to areas of pathological damage in the adult brain. Therefore, studying the migratory properties of these cells is essential, not only to understand basic aspects of adult neurogenesis, but also to exploit the potential of adult neural stem cells in neuroregenerative strategies. Whilst considerable progress has been made in the field of SVZ neural precursor migration, the exact molecular mechanisms regulating this process remain to be fully elucidated. The endocannabinoid system is known to play an important role in the regulation of neural stem cell proliferation. Here, we show that CB signalling also regulates the migration of SVZ-derived neural precursors. In addition, stimulation of G-protein coupled cannabinoid receptors, CB1 and CB2, leads to significant activation of RalA, a Ras-like GTPase involved in the control of neuronal morphology and polarity. siRNA-mediated knockdown of RalA or the expression of dominant negative RalA abolished cannabinoid-induced stimulation of migration. Time-lapse imaging revealed that depletion of RalA strongly impaired nucleokinesis: a crucial step for efficient migration. Analysis of RalA function in vivo, using wild type and mutant constructs electroporated into the SVZ, showed that the loss of RalA function results in both altered morphology and direction of migration. Finally, selective deletion of RalA in RMS neuroblasts in vivo further confirms that RalA is required for correct polarised morphology of migrating neuroblasts in the RMS. In summary, RalA is activated by CB agonists, and is required for CB-promoted migration of RMS neuroblasts. Furthermore, RalA expression is necessary for polarised morphology and efficient migration of RMS neuroblasts both in vitro and in vivo.

Homeostatic plasticity is thought to refer to the capacity of neurons to regulate their own excitability relative to network activity and occurs in a timescale of several days through modifications to ion channels and synaptic strength. Here, we present evidence showing that homeostatic plasticity is a cell autonomous phenomenon in that a single cell can modify its synaptic strength according to the activity that cell exhibits and not that of the network. This was accomplished using optogenetic tools. Moreover, we transfected single cells with Channelrhodopsin-2 allowing us to control the activity of these cells using light to excite and induce action potentials firing. In addition, we reduced the activity of the surrounding network by applying glutamatergic blockers. This allowed us to re-introduce normal levels of activity into the transfected cells using light, effectively rescuing them from the homeostatic effects induced by drug application. Immunohistochemical and electrophysiological evidence is provided showing that transfected neurons which exhibited normal levels of activity through photo stimulation did not undergo the synaptic homeostatic plasticity observed in the surrounding network. This strongly suggests that cells primarily regulate their homeostatic strength according to their own activity rather than that of the surrounding network. This however, does not exclude the possibility that the network does not have some effect on homeostatic plasticity.

Numerous human central nervous system disorders, particularly those involving higher cognitive functions, are poorly understood. The use of rodent models of such disorders has greatly advanced knowledge of the pathophysiological mechanisms underlying these diseases, but many questions remain unanswered. The present study examines the neurophysiology of neurons in the hippocampal formation in transgenic mouse models of two human CNS disorders, namely Alzheimer's disease and schizophrenia. The rTg4S1 0 mouse model over-expresses a mutant form of the microtubule associated protein tau, a protein involved in the formation of neurofibrillary tangles, which are a major pathological hallmark of Alzheimer's disease and a number of other dementi as collectively known as tauopathies. Profound age-dependent alterations in the intrinsic and synaptic properties of hippocampal CAl pyramidal neurons were observed in rTg4Sl0 mice. These mice also displayed alterations in the intrinsic properties of stellate neurons in layer IIIIII of the medial entorhinal c0l1ex. Both the hippocampus and the entorhinal c0l1ex are intimately involved in memory functions, so these results point to potential cellular correlates for the memory deficits observed in rTg4S1 0 mice, and may have implications for the memory impairments in Alzheimer's disease and other tauopathies. The DISCl mouse model expresses a truncated form of the Disrupted-in Schizophrenia- l gene, which has been implicated in psychiatric illnesses including schizophrenia. In DISCl mice, only subtle changes in the intrinsic properties of hippocampal CAl pyramidal neurons were observed, but distinct alterations in plasticity of synaptic transmission onto these neurons were revealed. These findings may have implications for the working memory deficits observed in schizophrenia.

Sporadic inclusion body myositis (IBM) is the commonest myopathy acquired after 50 years of age and causes significant disability. No effective treatment exists despite several clinical trials of immunotherapies. This reflects a lack of pre-clinical disease models. The work described in this thesis began with development of such a system, in which disease relevant pathological outcome measures were characterised in vitro using primary satellite cell cultures. Through over-expression of β-Amyloid Precursor Protein or exposure to inflammatory mediators IL1β and TNFα, IBM-like pathology was induced. Myotubes demonstrated ubiquinated intracellular aggregates, increased expression of MHC Class I, cytoplasmic translocation of TDP-43, ER stress, calcium dyshomeostasis and activation of NFκB. As some of these are proposed to be central pathogenic mechanisms in IBM, they provide a panel of tools on which to assess new IBM treatment strategies. The effects of heat shock response augmentation were examined using Arimoclomol, a co-inducer of the transcription factor HSF-1 that drives expression of key endogenous Heat Shock Proteins. Arimoclomol treatment ameliorated several IBM-relevant features, represented by improved cell viability, reduced ER stress, inhibition of NFκB and reduced cytoplasmic translocation of TDP-43. These data supported further evaluation of HSR manipulation as a potential therapy for IBM. Therefore, Arimoclomol was advanced into a randomised, placebo-controlled clinical trial involving 24 patients with IBM over one year. The primary outcome measure was safety and tolerability. Muscle biopsy was performed before and after the treatment phase to evaluate HSP expression and histopathological changes. Together, this in vitro model and clinical trial represent a novel translational research pathway in IBM, the lack of which has hampered the development of effective treatments for this disease.

The vascular supply to the human spinal cord has been studied in a series of observations made in the cadaver during routine necropsies performed at the Radcliffe Infirmary, Oxford. In addition to conventional anatomical dissection of the larger vessels concerned, injections with either indian ink or radio-opaque material were made into the spinal vessels. The examination of the larger spinal vessels was aided by X-rays or by rendering the spinal cord transparent with oil of wintergreen. The origin of the spinal vessels was studied and a division suggested between a aubolavian-vertebral supply to the cephalic part of the spinal cord and a direct aortic supply to the more caudal part of the spinal oord. From the evidence of both anatomical and pathological observations, the point of transition between the subolavian-vertebral supply to the direct aortic supply was considered to be in the region of the T1 spinal-cord segment. The anterior radicular tributary arteries to the anterior spinal artery were studied in a series of 40 injected spinal cords. The number of tributaries ranged from 3 to 9 with a mean of 5.6. More tributaries joined the anterior spinal artery from the left (127) than the right (97) side. The largest anterior tributary (the artery of Adamkiewicz) was always found accompanying one of the nerve roots from T7 to L3. It was three times commoner on the left than on the right side and in 12 out of the 40 specimens was accompanying the left T10 spinal nerve root. The anterior spinal artery was confirmed as the most important single vessel and was always found to run longitudinally for the whole length of the spinal cord. The posterior spinal arteries were much smaller and were more irregular vessels than the anterior spinal artery. A good anastomosis between the anterior and posterior spinal arteries was always demonstrable around the conus medullaris and was often seen in the cervical region. This was seen in the arterial infections which, when made into the anterior arterial system, flowed easily into the posterior spinal arteries. The sulcal arteries arose from the anterior spinal artery and passed backwards in the anterior median sulcus. They entered the grey commissure and by turning either left or right supplied the grey matter and central white matter of one side. These vessels and their branches formed the centrifugal arterial system. The centripetal arterial system, of lesser importance, was formed by radial arteries proceeding inwardly as branches from the coronal arterial plexus surrounding the spinal cord. The spinal veins, with some differences, were arranged on the same general plan as the spinal arteries. The differences were that the spinal veins were more numerous and that the anastomoses between large veins were more common. The veins were most prominent on the posterior aspect of the spinal cord and the largest single vessel was usually a single median posterior vein. The pathological observations were made on cases selected for the information they provided on disorders of the human spinal vascular supply. These cases showed that obstruction or disturbance of the spinal cord vessels could occur at different points of the pathway of arterial supply and venous drainage. Aorta. Aortic atherosclerosis and aortic thrombosis were shown to cause a syndrome of chronic ischaemia of the spinal cord. Aortic trauma with mural haematoma caused acute infarction of the spinal cord, a similar sequel to that more commonly seen following spontaneous medial dissection of the aortic wall. Vertebral arteries. Two cases illustrated spinal cord infarction and ischaemia caused by obstruction to both vertebral arteries. This unusual syndrome arose from trauma to the cervical spine causing a tear through one intervertebral disc. The resultant angulation at the point of the intervertebral disc tear kinked and obstructed both vertebral arteries. Interpostal arteries. Surgical damage to an intercostal artery giving rise to an important radicular tributary to the anterior spinal artery was shown to cause extensive infarction of the spinal cord in the territory of the anterior spinal artery. These paired intercostal and lumbar spinal arteries supply the spinal cord below the T1 spinal-cord segment. The majority of these spinal arteries may be obstructed with impunity but when, as in this case, the occluded artery is an intercostal artery giving rise to a major radicular tributary, serious infarction of the spinal cord will be caused. Radicular tributaries. Obstruction to these vessels in the intervertebral foramina was shown to be caused acutely by herpes soster and more gradually in cases of malignant tumour Infiltration. The sequel to this obstruction of the radicular tributaries was a state of haemorrhagic infarction affecting most of the cross-sectional area of the spinal cord and extending through several spinal-cord segments. Anterior spinal artery. Three cases of infarction of the spinal cord due to obstruction of the anterior spinal artery by thrombus were described. In two of these cases the cervical part of the anterior spinal artery was obstructed whilst in the third case the spinal cord below T11 segmental level was affected. In one case there was a relationship between the site of the thrombosis in the anterior spinal artery and the position of cervical spondylotic protrusions which were adjacent to the thrombosed artery. Ho cause for the arterial lesion was found in the other two cases. The longitudinal extent of the infarction in the two cervical cases was similar and its caudal extent was at the T1 segmental level. Below this level the spinal cord was supplied by a radicular tributary to the upper thoracic region and the blood flow was unaffected by the thrombus in the anterior spinal artery above. In its cross-sectional area the infarction always involved the anterior horns and the grey commissure but not the most posterior part of the posterior horns. The white matter of those parts of the anterior and lateral white columns adjacent to the grey matter was also infarcted. The region of infarction was often asymmetrical and might differ markedly at slightly different levels. The asymmetry was considered to be an effect of the various suloal arteries which entered to supply either the left or the right side of the spinal cord. Small spinal vessels. Three cases were selected to describe the effects of disease affecting the small spinal vessels. In one of these cases meningo-vascular syphilis had caused chronic atrophy of the spinal cord affecting both grey and white matter. The spinal vessels showed connective tissue thickening due to former syphilitic arteritis. Another case studied was a case of tuberculous meningitis causing, by arteritis, a severe haemorrhagic infarction of the spinal cord. In this case the subarachnoid space was distended with tuberculous exudate which had caused an endarteritis affecting all the spinal arteries surrounding the spinal cord. The third case considered in this group was a case of sarcoidosis affecting the spinal cord. Sarcoid granulomata were present around small spinal vessels in the leptomeninges and entering the spinal cord along perivascular spaces. In contrast to the effects of syphilis and tuberculosis, the vessels in contact with the sarcoid granulomata were not obstructed and there was no gross infarction evident in the spinal cord. Veins. Five cases of a chronic necrotising spinal cord disorder associated with grossly abnormal spinal vessels were described. The necropsies showed a complicated abnormality of the spinal vessels with large abnormal veins in the caudal and posterior regions of the spinal cord.

To understand the computations performed by the neurons within cortical structures it is essential to determine the relationship between sensory-evoked synaptic input and the resulting pattern of output spikes. In the cerebellum, granule cells constitute the input layer, translating mossy fibre signals into parallel fibre input to Purkinje cells. Until now their small size and dense packing have precluded recordings from individual granule cells in vivo. Here I use whole-cell patch-clamp recordings to show the relationship between mossy fibre synaptic currents evoked by somatosensory stimulation and the resulting granule cell output patterns. Granule cells exhibited low ongoing firing rates, due in part to dampening of excitability by a tonic inhibitory conductance mediated by gamma-aminobutyric acid type A (GABAA) receptors. Sensory stimulation produced bursts of mossy fibre excitatory postsynaptic currents that summate and trigger bursts of spikes. Remarkably, burst responses were evoked by only a few quantal excitatory postsynaptic currents. Recordings from putative mossy fibre terminals suggest that sensory-evoked granule cell spiking may result from high frequency activation of single mossy fibres. These results reveal that the input layer of the cerebellum balances exquisite sensitivity with a high signal-to-noise ratio. Granule cell bursts are optimally suited to trigger glutamate receptor activation and plasticity at downstream synapses, providing a link between input representation and memory storage in the cerebellum. Purkinje cells integrate inputs from other neurons in the cerebellar cortex to provide the sole output of the cerebellar cortex. Using whole-cell and cell-attached recordings, I describe a profound and robust bistability of spike output and membrane potential in these cells. Membrane potential toggles between a hyperpolarised down state and an active up state, in a complex spike-dependent manner. Sensory-evoked complex spikes reliably induced state transitions in both directions, indicating that bistability may be relevant for sensory processing in the cerebellum.

Generation of new axonal sprouts and the process of axonal elongation play a vital role in neural regeneration and repair. The facial nerve axotomy model is a well-established, prototypical experimental paradigm for the systematic study of nerve regeneration and degeneration, providing insights into molecular signals that determine axonal regeneration and neuronal cell death. Interestingly, this model of peripheral injury induces a delayed appearance of galanin+ and calcitonin gene-related peptide+ (CGRP) neuropeptide-immunoreactive growth cones in the facial nucleus which peak at 14 days following axotomy and surprisingly, increase in number if recut within this time window. Furthermore, application of the retrograde tracer mini-ruby to the distal portion of the cut nerve demonstrates the motoneuron origin of these sprouting neurites. To gain an insight into the molecular mechanisms inducing the sprouting response we examined how neuronal cell death and the inflammatory response of various transgenics affected sprouting and regeneration. Deletion of the a7 integrin, which demonstrated a moderate reduction in regeneration, showed enhanced sprouting neural c-jun blocked regeneration, abolished regeneration-associated neuronal proteins and neuronal cell death, also completely eliminated central axonal sprouting. Absence of TNFR1&2 which displayed reduced neuronal cell death and inflammation, showed a tendency toward enhanced sprouting TGF(31 deletion, which showed an elevated inflammatory response and a 4-fold increase in neuronal cell death, resulted in decreased central sprouting. Similarly, enhanced neural inflammation following systemic injection of E.coli lipopolysaccharide (LPS) also reduced central sprouting. Finally, neuronally expressed constitutively active Ras (Ras+), dominant-negative MEK1 (MEKIdn) and Ras+xMEK1dn double mutant (DM) all demonstrated reduced neuronal cell death as well as substantially enhanced central sprouting, particularly in the MEKIdn mutant, suggesting that the sprouting response in these mutants may be beneficial for improving regeneration in the CNS. Sprouting and regeneration studies in Ras+ and MEKIdn mutants were therefore extended to the injured corticospinal tract (CST) and rubrospinal tract. These mutants showed extensive collateral sprouting of corticospinal tract (CST) axons, in the grey and white matter on the ipsilateral side in Ras+, MEKIdn and DM animals compared with wild-type (WT) controls when the injury spared the dorsolateral CST, enhanced green fluorescent protein (EGFP) labelled rubrospinal axons showed increased sprouting below the site of injury following a C4 injury in Ras+, MEKIdn and DM mice but this was not statistically significant compared with wild-type controls. To determine functional recovery rearing and grid-walk tests over 28 days following a unilateral left dorsal hemisection (DH) at C4 were used. Ras+, MEKIdn and DM groups performed significantly better in left forepaw use than WT in the rearing test at day 28 (25.0% 3.0% 32% 1% 50.0% 9.0%) compared with WT (13.0% 5.0%). Similarly, Ras, MEKIdn and DM animals showed significantly less footslips on the left forepaw compared with WT at day 28 (11.0% 2.0% 10.0% 1.0%+/- 12.0% 1.0% 19.0% 2.0% respectively). Overall, data from facial nucleus studies suggest central axonal sprouting is an injury but not a reinnervation-driven response that it is not directly connected to neuronal cell death, that excessive inflammation is detrimental, and that jun-, Ras-, and MEK1-mediated changes in regeneration-associated gene and protein expression play a vital part in shaping the growth cone response. Following spinal injury, expression of MEKIdn enhanced CST sprouting below the injury site. Furthermore, the combination with Ras+ also enhanced functional recovery following C4 DH. These data suggest that neuronal expression of active Ras and MEKIdn might serve as a promising biochemical strategy for regrowth in the injured spinal cord.

The contributions of human medial temporal lobe (MTL) structures to perceiving and remembering visuospatial features of the environment were investigated from the perspective of human cognitive neuropsychology. The experimental chapters focussed on three main issues: the contribution of the right-sided MTL to spatial awareness the roles of the right MTL and hippocampus in spatial and non-spatial scene perception and memory and the role of the hippocampus in anterograde memory for different types of memoranda. Visual neglect (a failure of spatial awareness), was investigated after lesions of the right medial occipitotemporal lobe. Lesion analyses and diffusion tensor imaging demonstrated that the lesions associated with neglect interrupted a white matter pathway connecting the parahippocampal gyrus with the parietal lobe. The interaction of these areas may be critical to subserve normal spatial awareness. Perception and short-term retention of spatial and non-spatial aspects of visual scenes were investigated using a new task - the 4 Mountains Test. Right MTL damage impaired both spatial perception and memory. Hippocampal damage impaired spatial memory but had little impact on scene perception. Damage to these areas did not affect non-spatial perception or memory, suggesting a role for the hippocampus in the short-term retention of allocentric spatial information. A rather similar pattern of performance was documented in patients with possible Alzheimer's disease. The role of the hippocampus in anterograde memory was investigated in two patients using recognition tests for words, scenes and faces. Receiver operating characteristics analyses were used to assess the contribution of recollection and familiarity to performance. Hippocampal damage impacted upon scene and word recognition but not face recognition. In terms of scene recognition, the hippocampus may be critical to make recognition judgements based on both recollection and highly confident feelings of familiarity.

The perception of visual forms is crucial for humans for successful interactions with the environment. This process occurs automatically, and its outcome is reflected in the inferences and decisions we constantly make. The focus of this thesis is on how the brain handles different aspects of the perception of forms. To study this in normal human individuals, experiments were performed using functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG) and psychophysical methods. This thesis first discusses experiments designed to unravel the mechanisms of form construction, i.e. those from which all the component parts of a single form are assembled. Results suggest that the construction of very simple forms occurs in intermediate visual areas in a parallel and recursive process, with an increase in brain activity with increments in form complexity. A further experiment was performed to study how regularities or known characteristics of images, and the brain responses they elicit, will contribute to explain current percepts. Results from this experiment are consistent with a model where images with learnt attributes activate more strongly anterior visual areas and images with random patterns cause higher activations in earlier visual areas, probably due to top-down signals that reduce activity when it is possible to explain the causes of the sensory stimulation. Finally, it shows differences in the evoked neural activity when forms are either detected or classified, relating these processes to the activity generated in early visual areas. Based on the results of these experiments, a mechanism of top-down and bottom-up interactions between visual areas in the human brain is discussed in the context of the perception of forms.