This thesis describes the development and application of novel techniques to reduce the EEG artefacts at source during the simultaneous acquisition of EEG and fMRI data. The work described in this thesis was carried out by the author in the Sir Peter Mansfield Magnetic Resonance Centre, School of Physics & Astronomy at the University of Nottingham, between October 2010 and January 2013. Large artefacts compromise EEG data quality during simultaneous fMRI. These artefact voltages pose heavy demands on the bandwidth and dynamic range of EEG amplifiers and mean that even small fractional variations in the artefact voltages give rise to significant residual artefacts after correction, which can easily swamp signals from brain activity. Therefore any intrinsic reduction in the magnitude of the artefacts would be highly advantageous, allowing data with a higher bandwidth to be acquired without amplifier saturation, and facilitating improved detection of brain activity. This thesis firstly explores a new method for reducing the gradient artefact (GA), which is induced in EEG data recorded during concurrent MRI, by investigating the effects of the cable configuration on the characteristics of the GA. This work showed that the GA amplitude and its sensitivity to movement of the cabling is reduced by minimising wire loop areas in the cabling between the EEG cap and amplifier. Another novel approach for reducing the magnitude and variability of the artefacts is the use of an EEG cap that incorporates electrodes embedded in a reference layer, which has a similar conductivity to tissue and is electrically isolated from the scalp. With this arrangement, the artefact voltages produced on the reference layer leads are theoretically similar to those induced in the scalp leads, but neuronal signals are not detected in the reference layer. Therefore taking the difference of the voltages in the reference and scalp channels should reduce the artefacts, without affecting sensitivity to neuronal signals. The theoretical efficacy of artefact correction that can be achieved by using this new reference layer artefact subtraction (RLAS) method was investigated. This was done through separate electromagnetic simulations of the artefacts induced in a hemispherical reference layer and a spherical volume conductor in a time-varying magnetic field and the results showed that similar artefacts are induced on the surface of both conductors. Simulations are also performed to find the optimal design for an RLAS system, by varying the geometry of the system. A simple experimental realisation of the RLAS system was implemented to investigate the degree of artefact attenuation that can be achieved via RLAS. Through a series of experiments on phantoms and human subjects, it is shown here that RLAS significantly reduces the GA, pulse (PA) and motion (MA) artefacts, while allowing accurate recording of neuronal signals. The results indicate that RLAS generally outperforms the standard artefact correction method, average artefact subtraction (AAS), in the removal of the GA and PA when motion is present, while the combination of RLAS and AAS always produces higher artefact attenuation than AAS alone. Additionally, this work demonstrates that RLAS greatly attenuates the unpredictable and highly variable MA that are very hard to remove using post-processing methods.

Hyperpolarisation encompasses a multitude of methods to increase the species' spin polarisation for nuclear magnetic resonance (NMR) and magnetic resonance imaging (MRI) applications. Hyperpolarised 129Xe is produced via spin-exchange optical-pumping (SEOP). Firstly, electronic spins of alkali metal vapour are polarised via absorption of circularly polarised light. Alkali metal polarisation is subsequently transferred to noble gas nuclei via collisions. Within this thesis, the SEOP process is examined by probing the kinetics of the 129Xe polarisation build up. A combination of diagnostic techniques are used including low field NMR to measure 129Xe polarisation (PXe) at different spatial positions, near-IR optical absorption to give a global estimate of the alkali metal polarisation, and in situ Raman spectroscopy to spatially monitor the energy transport processes by detecting the internal gas temperatures (TN2). TN2 values were found to be dramatically elevated above oven thermocouple readings, with observations of up to 1000 K for an oven heated to only 400 K. Internal gas temperatures are presented for the first time along the length of the optical cell, showing spatial temperature and PXe variations during steady state and rubidium runaway conditions. Two contrasting methods of Raman spectroscopy are examined: a conventional orthogonal arrangement of detection and excitation optics, where intrinsic spatial filtering of the probe laser is utilised; and a newly designed inline module with all components in the same optical plane. Optical filtering is used to reduce the Rayleigh scattering and the probe laser line. This new inline device is presented herein and has a 23 fold improvement in signal to noise enabling increased accuracy and precision of `real-time' temperature monitoring. Rubidium, caesium and a rubidium/caesium hybrid are compared as the alkali metal of choice in the SEOP process. Caesium has a higher spin-exchange cross-section with 129Xe, thus a system is envisaged where current Rb D1 lasers in many polarisers can be utilised with a Rb/Cs hybrid to gain improvements in polarisation rates or levels. Xenon polarisations are shown up to 50% for a hybrid cell. Finally, preparatory experiments crucial to the imminent lung imaging study are presented, including measurements of PXe at low and high magnetic fields. In addition, polariser technology is examined including the current Nottingham device and an open-source consortium polariser.

This thesis is intended to contribute to a quantitative understanding of the blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) signal in order to increase its clinical potential. Here, the vascular, neuronal and physical processes which combine to give a resulting BOLD signal are investigated using respiratory challenges. The effect of isocapnic hyperoxia on vascular responses is investigated at 7 Tesla. No significant change was found in resting-state cerebral blood flow (CBF), resting-state cerebral blood volume (CBV) and task-evoked CBF. This challenges a previously held idea that hyperoxia is vasoconstrictive. The effect of isocapnic hyperoxia on neuronal oscillations was assessed with magnetoencephalography (MEG). Whilst a significant reduction in oscillatory power is reported in the occipital lobe, the change is significantly smaller than the global reduction previously measured with hypercapnia. These findings suggest that hyperoxia is an ideal tool for calibrated BOLD fMRI. The relationship between the change in blood oxygenation and change in transverse relaxation plays a key role in calibrated BOLD fMRI. However, previous measurements have been confounded by a change in CBV. Here, the relationship was found to be sub-linear across 1.5, 3 and 7 Tesla. Previous results which suggest a supralinear relationship at 1.5/3 Tesla and a linear relationship at 7 Tesla, are attributed to the relative contribution of intravascular/extravascular signals and their dependence on field strength, blood oxygenation and echo time. Finally, a comparison of single and multiphase ASL is made at 7 Tesla, with a modified Look-locker EPI sequence presented which allows simultaneous measurement of CBF and transit time, whilst increasing the available BOLD signal. This could have important implications for hypercapnia calibrated BOLD fMRI, where choice of ASL sequence may affect the estimated change in CMRO2. Furthermore, it provides a framework for future cerebral haemodynamic studies where simultaneous measurements are required.

Foetal MRI represents a non-invasive imaging technique that allows detailed visualisation of foetus in utero and the maternal structure. This thesis outlines the quantitative imaging techniques used to investigate the effect of maternal diabetes and maternal smoking on foetal development at 1.5 Tesla. The effect of maternal diabetes on placental blood flow and foetal growth was studied. The placental images were acquired using Echo Planar Imaging and blood flow was measured using Intra Voxel Incoherent Motion. The results indicate that peak blood flow in the basal plate and chorionic plate increases across gestation in both normal and diabetic pregnancies. Conversely, diffusion in the whole placenta decreases across gestation, with a more pronounced decrease in diabetic placentae. Following this, a method was developed to use a Tl weighted fat suppressed MRI scan to quantify foetal fat images in-utero. In addition, HAlf Fourier Single-shot Turbo spin Echo (HASTE) and balanced Fast Field Echo (bFFE) were used to acquire images encompassing the whole foetus in three orthogonal planes. These scans were used to measure foetal volume, foetal length and shoulder width. The data shows that foetal fat volume and intra-abdominal fat were increased in foetuses of diabetic mothers at third trimester. The HASTE and bFFE sequences were also used to study the effect of maternal smoking on foetal development. Here, foetal organ volumes, foetal and placental volume, shoulder width and foetal length were measured using a semiautomatic approach based on the concept of edge detection and a stereological method, the Cavalieri technique. The data shows that maternal smoking has significant negative effect on foetal organ growth and foetal growth, predominantly foetal kidney and foetal volume. The work described here certainly has a great potential in non-invasive assessment of abnormal placental function and can be used to study foetal development.

Cette thèse est consacrée aux propriétés des exciton-polaritons, les particules mixtes formées à partir de la lumière et la matière dans les microcavités de semi-conducteurs dans le régime de couplage fort. D'abord, j'analyse la possibilité de condensation de Bose des exciton- polaritons à température ambiante dans les microcavités de GaN avec les équations de Boltzmann semi-classiques. Puis les effets de polarisation dans le régime d'oscillateur paramétrique sont étudiés avec les équations de Boltzmann semi-classiques avec pseudospin. La deuxième partie de la thèse est consacrée aux propriétés des condensats et modes macrooccupés des exciton-polaritons. Leur polarisation, dispersion des excitations, propagation, localisation et superfluidité sont décrits avec l'équation de Gross-Pitaevskii

Ce manuscrit présente les travaux théoriques et expérimentaux réalisés dans le cadre de la construction du laser pilote de la station LASERIX. L'objectif en début de thèse était de concevoir un amplificateur de type "saphir dopé titane" (Ti:Sa) haute énergie et haute cadence permettant d'obtenir une énergie de 40 joules à la cadence de 0.1Hz avant compression des impulsions.<br />Après une introduction générale sur l'amplification dans les chaines laser Ti:sa à amplification à dérive de fréquence (chapitre 1), je présente les deux développements particuliers réalisés durant ce travail pour l'amplification à haute énergie (chapitre 2 :<br />- la mise en forme spatiale et l'homogénéisation à l'aide de systèmes à microlentilles des huit faisceaux des lasers de pompe Nd-Verre utilisés pour le pompage du dernier étage d'amplification<br />- La suppression des phénomènes parasites d'amplification transverse de la fluorescence dans le cristal amplificateur du dernier étage d'amplification. Deux techniques différentes sont présentées.<br />Ces développements ont permis d'amplifier, dans un cristal de 10cm de diamètre, une impulsion de 2J jusqu'à une énergie de 39J.<br />Je présente ensuite le programme de simulation que j'ai réalisé (Chapitre 3) pour modéliser en trois dimensions les phénomènes de lasage parasite transverse et de fluorescence transverse amplifiée dans les cristaux de Ti:Sa de grandes dimensions. Ce code m'a permis d'étudier les principaux paramètres influant ces phénomènes parasites.<br />Finalement, la dernière partie (Chapitre 4) présente les perspectives de mon travail dans le cadre des futurs grands projets lasers ILE et ELI.

non disponible

The inflationary paradigm of standard big bang cosmology provides a mechanism to generate primordial curvature perturbations and explain the large scale homogeneity and isotropy of the observable universe. This is achieved through requiring a period of accelerated expansion during the early universe and requires a deep understanding of particle physics for its correct formulation. With the emergence of string theory as a potential description of a fundamental laws of nature provides a the natural framework in which we can construct realistic models of inflation seems plausible. A common feature of string theories is the requirement of extra dimensions and, in the absence of a complete formulation of the theory, it is necessary to dimensionally reduce the theories to give a 4d effective theory. String compactifications provide a promising approach through which this can be done. However compactifications lead to the generation of a large number of massless scalar fields (moduli) which would mediate unobserved 'fifth forces'. Methods of stabilising these fields give rise to exponentially flat potentials which provide the means of obtaining inflation quite naturally. In the introductory chapters a review of Type IIb flux compactifications gives methods to stabilise the complex structure moduli and dilaton through the use of fluxes. In order to stabilise the Kähler moduli additional non perturbative corrections to the superpotential are required. We introduce the well know class of meta stable de Sitter string vacua obtained when such corrections are included. An additional class vacua at large volume are discussed, these are found when leading order perturbative corrections to the Kähler potential are also considered. The large volume vacua are then shown to give rise to a model of inflaton using a Kähler modulus as an inflaton field. We show that there exists a large class of inflationary solutions corresponding to a constant volume V of the compactification manifold. In a second chapter on this inflationary model the existence of a basin of attraction for inflation with a constant volume is described. We also find a larger class of inflationary solutions when we evolve the axionic components of the Kähler moduli and the phenomenological aspects are discussed. We finally review the standard slow roll analysis and discuss its use in multiple field inflationary models. We introduce two multiple field extensions to the standard single field slow roll approach. We proceed with an investigation into the suitability of the multiple field slow roll approaches in predicting the slow roll footprint of Supergravity models of inflation. This is achieved through comparing the results with single field results and numerical simulation data when more complex models are considered.

The inherent quantum-mechanical nature of the proton transfer process in hydrogen bonds has been investigated through its effects on the nuclear spin-lattice relaxation rate. The fast magnetic field-cycling techniques employed allowed a direct measure of the rate characterising this dynamic process, which is closely related to the potential energy environment experienced by the mobile proton. Various heteronuclear effects from magnetic and non-magnetic nuclei outside the hydrogen bond were characterised. The contribution to proton tunnelling from the displacement of heavy atoms in the molecule is an important consideration within a complete description of the process. This interdependence was accurately measured for the carboxyl-group oxygen atoms in benzoic acid dimers through the isotope effect. Careful comparison of ¹⁶O and ¹⁸O-enriched benzoic acid relaxation allowed this relationship to be measured from the difference in low-temperature tunnelling rates. Fluctuating dipolar interactions caused by proton transfer motion couples the Zeeman states of different nuclear species. The cross-relaxation occurring through this natural coupling was explored as a function of field in 2,4,6-trifluorobenzoic acid and ¹³C-enriched pure benzoic acid. Characterising the strength of this interaction endeavoured to broaden the comprehension of heteronuclear coupling and served as confirmation of the model used. Beyond the carboxylic acid dimer, this investigation also showed dynamic disorder in intermolecular short, strong hydrogen bonds of pyridine-3,5-dicarboxylic acid. This proton transfer mechanism was found to be strongly dependent on the molecular vibrational modes creating a pathway between two potential minima. A finite change in entropy between the proton sites ensured that greatest proton mobility occurred at intermediate temperature, between relatively stable configurations at the extremes of temperature. A study of different sources of molecular dynamics within one compound showed the efficiency of field-cycling NMR at separating their contributions to relaxation. Dynamic rates from the proton transfer and methyl group rotation in 4-methylbenzoic acid were reliably extracted to the extent of identifying separate contributions from a small percentage of molecules around impurity centres.

In this thesis, a study of a range of functional surfaces formed in ultra-high vacuum (UHV) conditions using primarily scanning probe microscopy is presented. The construction of a combined scanning tunnelling and atomic force microscope, and the experiments performed using this instrument, are also detailed. Boron nitride and graphene monolayers were formed on rhodium thin films in UHV and investigated with in-situ and ex-situ (ambient conditions) scanning tunnelling microscopy. Simultaneous scanning tunnelling and atomic force microscopy images were also produced for the graphene monolayers. X-ray photoelectron spectroscopy and diffraction results for graphene monolayers on Rh(111) surfaces, as well as low energy electron diffraction data, are also included. The novel formation of monolayer and few-layer graphene on nickel thin films is also described. Graphene layers were detached from these nickel thin films and isolated on other substrates. The results of characterisation experiments using scanning probe microscopy, X-ray photoelectron spectroscopy, X-ray diffraction and electron microscopy techniques are detailed. Graphene layers with approximately 75% monolayer graphene coverage and an increased electronic quality, when compared to many other methods of graphene production, were revealed. Different organic molecules were adsorbed on both the boron nitride and graphene monolayers formed on rhodium thin films in UHV conditions. Perylene tetracarboxylic diimide (PTCDI) and di(propylthio)-PTCDI molecules were investigated on these surfaces and compared with the adsorption of PTCDI on a graphite surface. Furthermore, dibutyl-coronene tetra-carboxylic diimide was deposited on the graphene (on rhodium) surface, in UHV. Although the boron nitride and graphene surfaces were similar, it was discovered that very contrasting molecular formations were formed on the dierent surfaces. The positioning of these nanostructures was determined by the Moire superstructure formed due to the mismatch between the monolayers and the Rh(111) surface. Additionally, different hydrogen-bonded molecular junctions were formed depending on the length of the side chains of the adsorbed organic molecules.