Focal epilepsy is a common serious neurologic disorder. One out of three patients is medication refractory and epilepsy surgery may be the best treatment option. Neuroimaging and electroencephalography (EEG) techniques are critical tools to localise the ictal onset zone and for performing functional mapping to identify the eloquent cortex in order to minimise functional deficits following resection. Diffusion tensor magnetic resonance imaging (DTI) informs about amplitude (diffusivity) and directionality (anisotropy) of diffusional motion of water molecules in tissue.This allows inferring information of microstructure within the brain and reconstructing major white matter tracts (diffusion tensor tractography, DTT), providing in vivo insights into connectivity. The contribution of DTI to the evaluation of candidates for epilepsy surgery was examined: 1. Structure function relationships were explored particularly correlates of memory and language dysfunction often associated with intractable temporal lobe epilepsy (TLE; chapters 3 and 4). Abnormal diffusion measures were found in both the left and right uncinate fasciculus (UF), correlating in the expected directions in the left UF with auditory memory and in the right UF with delayed visual memory performance. Examining the arcuate fasciculus (AF), bilateral diffusion changes were found with correlations between left AF DTI measures and language scores. 2. The second aim of this thesis was to validate DTT results and test the hypothesis that cortical language areas determined by cortical stimulation serve as anchor points for the tractography defined AF (chapter 5). Subdural grid contacts overlying anterior language cortex co-localised in 84.2% with the AF, and in 55.8% in posterior language areas. This provides some validation that the AF reconstructed using DTT subserves language function, but further study is needed. 3. Lastly, seizure propagation was investigated in a case series of patients with cortical dysplasia (chapter 6). Reduced connectivity with reduced arborization and thinning of the fibre bundles between subcortical WM and the dysplastic cortex was demonstrated. Fibre tracts reconstructed from regions underlying the ictal onset zone showed abnormal connectivity.

Epilepsy is the most common serious disease of the brain. Magnetic Resonance Spectroscopy (MRS) is a novel imaging technique that offers the opportunity for co-localising biochemical information relating to metabolites specific to the study of epilepsy with high resolution MRI. Aims: The work included in this thesis was undertaken with two fundamental aims. The first was to apply a standardised MRS methodology in order to gain reproducible semi-quantitative information about the variation of relevant neuro-metabolites such as gamma amino butyric acid (GABA), glutamate (as glutamate plus glutamine [GLX]), N acetyl aspartate (NAA), myo-inositol (Ins) and creatine plus phosphocreatine (Cr) within epilepsy syndromes or pathological groups. The second main aim was to test a series of hypotheses relating to the regulation of the concentrations of these metabolites in the region of epileptic seizures, immediately following seizures and associated with particular medical and surgical treatment interventions. Methods: Seven experiments were performed in this thesis. In all seven studies the findings in the patient groups were compared against results from an acquired control group made up of healthy volunteers. In the first experiment [3.1] twenty patients with temporal lobe epilepsy, with (10), and without hippocampal sclerosis were studied using multi voxel magnetic resonance spectroscopic imaging (MRSI) sequences in order to examine for differences in the obtained metabolites N acetyl aspartate (NAA), creatine plus phosphocreatine (Cr), choline containing compounds (Cho), GLX and myo-inositol (Ins) across the pathological groups and against a control population. In experiments [3.2], [3.3], [3.4] and [3.6] an MRS protocol that incorporated a double quantum filter acquisition sequence was applied in order to allow measurement of GABA+ (a combined measure of GABA plus homocarnosine) in addition to measurement of the metabolites examined in [3.1]. Studies were performed in the occipital lobes in patients with idiopathic generalised epilepsy (IGE) (n =10) or occipital lobe epilepsy (n = 10) [3.2], in the frontal lobes in patients with IGE (n = 21) and within regions of the MRI visible pathology in patients with large focal malformations of cortical development (MCD, n =10) [3.4]. In the last experiment using this technique patients with hippocampal sclerosis and temporal lobe epilepsy (n = 16) were studied in the ipsilateral and also in the contralateral temporal lobes and following temporal lobe surgery (n = 10) [3.6]. In experiment [3.5] ten patients were examined whilst taking and when not taking sodium valproate in order to further examine for an effect of this medication on the measured metabolite concentrations. In experiment [3.7] ten patients were studied immediately after an epileptic seizure and then again during a subsequent inter-ictal period in order to examine for an influence of the recent seizure on the measured concentrations of the main metabolites. Results: MRSI in the temporal lobes in patients with temporal lobe epilepsy identified low NAA in the anterior hippocampus that was most severe in those patients with hippocampal sclerosis. GLX elevation was a feature in the patients without hippocampal sclerosis. Metabolic abnormality was most marked in the anterior compared to the posterior hippocampal regions. GABA+ levels were elevated in patients with MCD and in the ipsilateral temporal lobe in temporal lobe epilepsy associated with hippocampal sclerosis but levels were not altered in patients with IGE or OLE. GLX was also elevated in MCD in the region of MRI visible abnormality and in IGE patients when measured in the frontal lobes. Low NAA was a feature of TLE and MCD. Patients with IGE showed normal NAA levels in the occipital lobes but reduced frontal lobe concentrations. Cr concentrations were abnormal in the immediate post ictal period but normalised within 120 minutes. NAA was not altered and no significant change in lactate concentrations was observed. Finally sodium valproate treatment was associated with a reduction in the levels of Ins and with unchanged NAA and GLX levels. Main Conclusions: MRS techniques demonstrate metabolite abnormalities in epileptic patients. NAA is the most sensitive metabolite marker of chronic pathology but levels are insensitive to recent seizure history. These findings repeat earlier observations of the usefulness of NAA measurement in the assessment of chronic epilepsy whilst illustrating ongoing uncertainty as to the correct patho-physiological interpretation of reduced NAA levels. Measurable changes in the combined Cr signal are detectable whilst elevated lactate is not reliably observed following brief epileptic seizures at 1.5T. This finding indicates a potential role for MRS in functional activation studies. Malformations of cortical development have abnormal levels of both GABA+ and GLX and MCD sub-types may well demonstrate different metabolite profiles. This finding suggests that MRS could be a useful tool in the MRI classification of MCD and in the pre-surgical assessment of patients with focal malformations. Following successful temporal lobe surgery levels of NAA remain unchanged but NAA/Cr levels appear to normalise in the contralateral temporal lobe. NAA and GLX/NAA levels were altered in the frontal lobes but not in the occipital lobes in Idiopathic Generalised Epilepsy. This finding provides imaging support for frontal lobe dysfunction as a cause or consequence of IGE. Metabolite levels are affected by administered antiepileptic drugs. Sodium valproate reduces the levels of MRS visible Ins levels whilst topiramate and gabapentin appear to be associated with higher GABA+ levels. These findings may be of major importance in the assessment of treatment effect or in the investigation of patients with possible drug resistance. The effect of valproate on Ins levels may become particularly interesting in the light of a growing understanding of the role of astrocyte dysfunction in a range of neurological conditions which include migraine, epilepsy, Alzheimer’s disease, motor neurone disease and in ischaemic lesions.

This thesis describes a study of 70 patients with epilepsy and a normal control group. Subjects were studied with Transcranial Magnetic Brain Stimulation using a variety of parameters including measurement of motor evoked potentials, active and resting motor threshold, paired pulse inhibition and facilitation and cortical silent period. Data from normal subjects, 40 patients with temporal lobe epilepsy (TLE) and 17 patients with neocortical epilepsy was collected. The following data were acquired from some or all of each group: resting motor threshold (RMT), active motor threshold (AMT), cortical silent period (CSP), and conditioning-test stimulation data from which were derived measures of intracortical inhibition (SICI) and facilitation (SICF). Data were also obtained regarding drug treatment and seizure timing. Groups of subjects were compared. TLE patients were studied serially to examine the effects of seizures on TMS parameters. The main findings were that the groups differed regarding RMT and AMT, probably reflecting drug treatment; the patients groups differed in SICF; serial studies of the TLE patients showed changes in SICI and SICF which preceded seizures. RMT and AMT were elevated following seizures. Comparing the TLE and neocortical patient groups, there was higher SICF in the mTLE group. The difference between the patient groups demonstrates that epilepsies arising from distinct areas of the brain effect cortical excitability measured from the motor cortex in different ways. Repetitive TMS (rTMS) was undertaken in 11 subjects with refractory localisation related epilepsy. No effect of rTMS on the number of EEG spikes was seen as a result of a single session of rTMS, suggesting a variety of possibilities; rTMS may have no beneficial effect on focal seizures; the stimulus parameters may have been unsuitable; the chosen outcome measure – number of spikes – may be too inherently variable to show an effect after a relatively brief period of rTMS.

Synaptic neurotransmitter release is mediated by presynaptic voltage-gated calcium channels (VGCCs) via a tightly controlled mechanism. Naturally occurring mutations in the CACNA1A gene, which encodes the pore forming α1-subunit of P/Q-type VGCCs, can disrupt synaptic neurotransmission and lead to neurological disorders such as migraine, ataxia and epilepsy. Here, we aimed to understand how mutations linked to a severe form of migraine – familial hemiplegic migraine type 1 (FHM1) – affects neurotransmitter release in small central synapses. The effects of FHM1 mutations on both VGCC function and transmitter release are currently controversial. It is widely agreed that FHM1 mutations cause a shift to more negative potentials in the voltage activation threshold of P/Q-type channels. However, it is less understood whether this leads to a gain- or loss- of function of neurotransmitter release. We studied the effects of two particular FHM1 mutations using a combination of genetic manipulations and fluorescent imaging methods. First, in order to overcome species-specific differences in subunit expression, we attempted to express human cDNA encoding mutant VGCCs in the human neuronal-like cell line, NT2. Second, we compared synaptic neurotransmission in neuronal hippocampal cultures prepared from mice harbouring FHM1 mutations, with wild-type controls. Our results revealed that synaptic transmission was highly heterogeneous among individual synapses. However, FHM1 mutations did not have a significant effect on synaptic vesicular release. We also investigated the role of VGCCs in triggering spontaneous neurotransmission. We found that approximately 25 % of miniature excitatory post-synaptic currents were dependent on the stochastic opening of presynaptic VGCCs. Further analysis showed that the slow Ca2+ buffer EGTA blocks VGCC-dependent spontaneous and evoked release to a similar extent, suggesting that the two types of neurotransmission may share the same release machinery. Taken together these data provide new insights into the regulation of evoked and spontaneous release by presynaptic VGCCs.

Epileptic activity in the brain arises from dysfunctional neuronal networks involving cortical and subcortical grey matter as well as their connections via white matter fibres. Physiological brain networks can be affected by the structural abnormalities causing the epileptic activity, or by the epileptic activity itself. A better knowledge of physiological and pathological brain networks in patients with epilepsy is critical for a better understanding the patterns of seizure generation, propagation and termination as well as the alteration of physiological brain networks by a chronic neurological disorder. Moreover, the identification of pathological and physiological networks in an individual subject is critical for the planning of epilepsy surgery aiming at resection or at least interruption of the epileptic network while sparing physiological networks which have potentially been remodelled by the disease. This work describes the combination of neuroimaging methods to study the functional epileptic networks in the brain, structural connectivity changes of the motor networks in patients with localisation-related or generalised epilepsy and finally structural connectivity of the epileptic network. The combination between EEG source imaging and simultaneous EEG-fMRI recordings allowed to distinguish between regions of onset and propagation of interictal epileptic activity and to better map the epileptic network using the continuous activity of the epileptic source. These results are complemented by the first recordings of simultaneous intracranial EEG and fMRI in human. This whole-brain imaging technique revealed regional as well as distant haemodynamic changes related to very focal epileptic activity. The combination of fMRI and DTI tractography showed subtle changes in the structural connectivity of patients with Juvenile Myoclonic Epilepsy, a form of idiopathic generalised epilepsy. Finally, a combination of intracranial EEG and tractography was used to explore the structural connectivity of epileptic networks. Clinical relevance, methodological issues and future perspectives are discussed.

Valproic acid (VPA) is a commonly used drug for the treatment of epilepsy, bipolar disorder and migraine, yet its mechanisms of action are unknown. The neuroprotective effect of VPA has been hypothesized to be secondary to inhibition of the cAMP/protein kinase A (PKA) pathway. Here, the result show that VPA (1mM) inhibited mossy fibre long-term potentiation induced (LTP) by application of high frequency stimulation in dentate gyrus. Furthermore, VPA (1mM) inhibited enhancement of mossy fibre responses induced by application of forskolin (50 μM), consistent with an effect on the PKA pathway. Using biochemical assays, it was further demonstrated that this was not due to a direct effect on PKA, but resulted from inhibition of adenylyl cyclase. The results further show using in vitro seizure models (Pentylenetetrazole model and low- Mg2+ model) that this mechanism cannot fully explain VPA’s anti-seizure effect, but rather, by modifying synaptic plasticity, it may be more important for VPA’s antiepileptogenic and neuroprotective action. VPA therefore has distinct mechanisms of action that contribute to its diverse biological activity. In hippocampi from epileptic rats (following pilocarpine-induced status epilepticus), but not in control tissue, VPA affects short-term plasticity, indicating that VPA may have specific effects in epileptic rather than control animals. Using in vitro seizure models (Pentylenetetrazole model and low-Mg2+ model) and an in vivo status epilepticus model (the perforant pathway stimulation model), 4- methyloctanoic acid is further established that it is a more potent antiepileptic drug than VPA and provides neuroprotective effects which are similar to VPA. Furthermore, 4- methyloctanoic acid (1mM) inhibited enhancement of mossy fibre responses induced by application of forskolin (50 μM), indicating that 4-methyloctanoic acid shares the same effect as VPA on modulation of PKA.

Epilepsy, one of the most prevalent neurological diseases, is characterized by recurrent seizures, which are often poorly controlled by present treatments. Neuronal cell death, one of the complications of uncontrolled seizures, contributes substantially to the disease-burden through cognitive decline. This thesis explores mechanisms underlying neuronal death following seizure activity with a focus on mitochondria and ROS as major culprits of seizure induced neuronal cell death. By applying fluorescent dye imaging to glio-neuronal cell cultures, the first study examines mitochondrial mechanisms of seizure induced cell death and how such cell death can be reversed. The second study, using both in vitro and in vivo epilepsy models, explores sources of ROS production during seizures and in chronic epilepsy and how these can be reversed by blocking ROS producing enzymes. Experiments performed in the first part demonstrate that seizure activity leads to a sustained cyclosporine-A-sensitive depolarization of the mitochondrial membrane potential indicating mitochondrial permeability transition pore opening. This work proposes that neuronal ATP levels decrease, and correlate with the frequency of the oscillatory Ca2+ signal, indicative of activity-dependent ATP consumption. Cellular ATP levels during seizure like activity are dependent on the functioning of the mitochondrial complexes and seizure induced neuronal death is reduced when mitochondrial complex I substrate pyruvate is supplemented. In the second part I show that seizure activity induced increases in ROS production were seen both in vitro and in vivo. ROS production in vivo is elevated during periods of spontaneous brief seizures leading to a brain-region specific decrease in the major antioxidant system glutathione. Seizure-induced ROS were not dependent upon mitochondrial dysfunction but were generated in a Ca2+ -independent fashion through initially NADPH oxidase and later xanthine oxidase activity. Inhibition of either enzyme, consequently, reduced seizure-induced neuronal cell death. This works outlines strategies to treat seizure induced cell death.

Background: One of the few studies to describe therapy for phonological assembly difficulties in aphasia is a single case by Franklin, Buerk and Howard (2002). Their client improved significantly in picture naming for treated and untreated words after therapy targeting auditory awareness and self-monitoring. Aims: This thesis comprises two studies. Study one aimed to determine whether the generalised improvements reported by Franklin et al (2002) are replicable with other people with impaired phonological assembly, and to explore any differences in outcome. Study two aimed to compare the effectiveness of Franklin et al’s therapy with a production-focussed approach. The overall aim of both studies was to discover whether different subgroups of people with phonological assembly difficulties may respond differently to therapy, and whether any differences in treatment outcome may provide insight into theoretical models of phonological output processing. Method: A case series of eight participants with aphasia with mixed impairments including phonological assembly difficulties is reported. In study one, four participants received a replication of the treatment described by Franklin et al. In study two, four further participants received a novel production therapy in addition to Franklin et al’s therapy. Outcomes: No participant responded in the same way as Franklin et al’s original client. All post-therapy naming improvements were item-specific, except for one participant, who also showed signs of spontaneous recovery. Two participants showed no significant naming improvements after either treatment. Conclusions: Whereas Franklin et al’s original client had a relatively pure post-lexical phonological assembly impairment, six of the eight participants in the current study had phonological assembly difficulties combined with either lexical retrieval or motor speech impairments. The item-specific naming gains were proposed to reflect improved mapping between semantics and lexical phonology, rather than improved phonological assembly. These results support a model of speech production containing both lexical and post-lexical levels of phonological processing.

This research looked at the ability of children with autism to understand theory of mind. This is the ability to attribute mental states (e.g. believing, thinking, knowing etc.) to oneself and to others. The main aim of the study was to provide evidence, contrary to a large body of previous research, that individuals with autism can exhibit a theory of mind, when the standard tests used in the past are simplified for this population. A further aim of the study was to show that language ability is significant in terms of theory of mind task performance. It was hypothesised, because of the nature of the theory of mind tasks, that matching participants in terms of their understanding of grammar, rather than single word understanding (as in past research), would be more appropriate. Three groups took part in the study; autistic, learning disabled and normally developing children. The learning disabled and normally developing participants were selected to match the subjects with autism, in terms of receptive verbal age, on either the British Picture Vocabulary Scale (BPVS) or the Test for Reception of Grammar (TROG) or both. All the participants were given two first order theory of mind tasks; the standard 'Sally-Anne' task, which has been used in past research, and a simplified cartoon version of this task designed by the author. These first order tasks test the ability to consider another person's thoughts about an objective event. Those participants who passed one of the first order tasks were then given three second order theory of mind tasks. These test the ability to consider another person's thoughts about a third person's thoughts regarding an objective event. The second order tasks consisted of the standard 'Ice-Cream Man task' (used in past research), Sullivan, Zaitchik and Tager-Flusberg's (1994) simplified 'Puppy task' and a simplified cartoon version of the task designed by the author. A significant difference in performance was found between the three participant groups (matched on the BPVS) on the standard first order task, but not on the simplified first order task. A significant difference in performance was found between the participant groups on the standard Ice-Cream Man task and the Puppy task, when matched on the BPVS, but not when matched on the TROG. In addition no significant difference in performance was found between the autistic and learning disabled participants on any of the theory of mind tasks. These findings are discussed in relation to other explanations of autism such as the salient object hypothesis and executive function.

It is gradually becoming acknowledged that fathers do suffer from post-natal depression, but very little is known about their experience or how many fathers are actually affected. The factors that may increase a father's susceptibility to post-natal depression, or those that might indeed protect him, have also been given little consideration in research to date. This study examined the prevalence and comorbidity of post-natal depression in 100 first time parents, using the Edinburgh Post-natal Depression Scale (EPDS). The psychological factors of infant temperament, perception of own parenting and social support were investigated in a smaller sample of 30 fathers subsequently interviewed. These were measured using the Neonatal Perception Inventory (NPI), the Parental Bonding Instrument (PBI) and the Significant Others Scale (SOS) respectively. The findings suggest that approximately 12 per cent of first time fathers may suffer from post-natal depression and fathers are significantly more likely to experience this if their partner is also depressed. Depression amongst fathers was found to be associated with having little social support, perceiving one's own baby as more difficult than the average baby and perceiving one's own father as having been uncaring. The findings are discussed together with their clinical implications and areas for future research.