This thesis investigates the application of physiological measures made using magnetic resonance imaging (MRI) to study cerebral haemodynamics and the pharmacological modulation of brain activity. Blood Oxygen Level Dependent (BOLD) functional MRI (fMRI), together with Arterial Spin Labelling (ASL), were used to study of the effect of beta-blockers on the brain’s response to emotional visual stimuli at 3 T. The study aimed to test the James-Lange theory, which states that emotions result from the perception of bodily arousal. Autonomic nervous system responses to emotional stimuli can be predicted by the level of activity in the limbic system (including amygdala, brainstem and salience network). This thesis assesses the action of the peripherally acting beta-blocker, nadolol, on the fMRI response to neutral, pleasant and unpleasant visual stimuli, and during rest. 80 mg nadolol and placebo tablets were administered to twenty healthy human subjects in a double blinded, randomised, placebo controlled crossover design. Drug induced reduction in anterior insula response to emotional stimuli supports results from previously published studies, and the James-Lange theory. The preliminary finding of drug induced increase in amygdala response is contradictory. Territorial ASL (TASL) and multi-phase ASL were used in combination for advanced investigation of the vascular territories, and quantitative perfusion and blood transit times. These measures were used for the assessment of the location and haemodynamic properties of the cerebral watershed regions. Watershed region masks formed from TASL and multi-phase ASL data exhibited reduced perfusion and lengthened transit times compared to other cortical regions. The accuracy of watershed delineation was shown to be enhanced by the use of both techniques in combination. Furthermore, TASL is developed and implemented at 7 T. Alternative labelling schemes were compared and parameters optimised for control condition efficiency, and TASL assessment was performed in three healthy volunteers.

Functional magnetic resonance imaging (fMRI) offers a number of opportunities to non-invasively study the properties of the human visual system. Advances in scanner technology, particularly the development of high-field scanners, allow improvements in fMRI such as higher resolution and higher signal to noise ratio (SNR). We aimed to examine what these advances in scanner technology, combined with novel analysis techniques, can tell us about the processing of motion stimuli in the human visual cortex. In Chapter 3 we investigated whether high-resolution fMRI allows us to directly study motion-selective responses in MT+. We used event-related and adaptation methods to examine selectivity for coherent motion and selectivity for direction of motion, and examined the potential limitations of these techniques. One particular analysis technique that has been developed in recent years uses multivariate methods to classify patterns of activity from visual cortex. In Chapter 4 we investigated these methods for classifying direction of motion, particularly whether successful classification responses are based on fine-scale information such as the arrangement of direction-selective columns, or a global signal at a coarser scale. In Chapter 5 we investigated multivariate classification of non-translational motion (e.g. rotation) to see how this compared to the classification of translational motion. The processing of such stimuli have been suggested to be free from the large-scale signals that may be involved in other stimuli, and therefore a more powerful tool for studying the neural architecture of visual cortex. Chapter 6 investigated the processing of plaid motion stimuli, specifically ’pattern’ motion selectivity in MT+ as opposed to ’component’ motion selectivity. These experiments highlight the usefulness of multivariate methods even if the scale of the signal is unknown.

The work presented in this thesis describes the development and implementation of a number of ideas and methods that allow fMRI to be carried out using echo-planar imaging at ultra high field strength, despite the significant problems associated with this. In the first study, EPI is used to probe how the gradient echo (GE) and spin echo (SE) BOLD responses relate to the underlying neurological processes, whilst the brain is in both its active and resting states. These finding show that SE BOLD contrast is harder to detect but less localised to areas around large draining veins than GE BOLD contrast and thus potentially more localised to sites that represent true functional areas of activation. The second study describes how dynamic delta B0 mapping can be performed during fMRI experiments with a hyperoxic challenge in order to assess the magnitude and extent of delta B0 effects that arise due to susceptibility differences between air and tissue. Developing on this, this work describes the steps involved in the design and implementation of a dual echo GE/SE EPI sequence and how it can be used to enable off-resonance effects, such as image distortion and signal concentration/dilution, to be corrected on a dynamic basis for, simultaneously acquired, GE and SE data. The final study demonstrates how such a sequence can be used to detect resting state networks. Showing that the correspondingly low temporal separation of the GE and SE data allows GE and SE BOLD contrast mechanisms to be compared in a number of novels ways in different resting state networks.

MRI has become an invaluable tool for diagnostic medicine. Its operation is based on the principles of electromagnetism that are dictated by Maxwell's equations. MRI relies on the existence of well defined, spatially and temporally controlled magnetic fields, which are usually generated by coils of wire. Human exposure to these fields has become a safety concern, especially with the increase in the strength of the magnetic fields used. In this thesis, problems in electromagnetism relevant to different areas in MRI and involving the calculation of solutions to both forward and inverse problems are investigated using techniques derived for computational mechanics. The first section of the work focuses on the development of an accurate technique for the solution of magnetostatic inverse problems using boundary element methods (BEM) with the aim of designing optimised gradient coils. This approach was found to be an extremely effective method which can be applied to a wide range of coil geometries and is particularly valuable for designs where the coil surface has low symmetry. BEM-based approaches to designing gradient coils that reduce the likelihood of peripheral nerve stimulation due to rapidly switched magnetic fields are also considered. In the second section of the work, a novel BEM tool to allow the calculation of solutions to quasi-static forward problems has been developed, and used for the evaluation of the electric fields induced in the human body by temporally varying magnetic fields, due to either gradient switching or body movements in strong static magnetic fields. This approach has been tested by comparison with analytic solutions for simply shaped objects, exposed to switched gradients or moving in large static fields, showing good agreement between the results of numerical and analytical approaches. The BEM approach has also been applied to the evaluation of the electric fields induced in human body models. This work involved the development of an appropriate theoretical framework for the study of conducting systems moving in magnetic fields. This involved correcting some misconceptions that had propagated in the literature and allowed the development of an effcient implementation of a BEM suited to this problem.

Phase images of the human brain acquired using gradient echo based Magnetic Resonance Imaging techniques show excellent contrast at 7T. This contrast is attributed to small variations in magnetic susceptibility that perturb the main magnetic field and thus yield a spatial variation of the NMR frequency. The work described in this thesis is primarily concerned with mapping the distribution of magnetic susceptibility within the human brain using these phase images. The main technical challenges of the project were first to extract accurate field maps based on phase data, and then to solve the ill-posed problem of inverting these field maps to reconstruct susceptibility (ϰ) maps. In initial work, simulations of field shifts based on known ϰ -distributions are compared to field maps acquired in vivo to highlight the non-local relationship between measured field offsets and the underlying susceptibility. These simulations were carried out using a recently derived Fourier method. The bulk of the thesis is then devoted to a detailed study of the process of inverting field maps generated from phase data using the Fourier relationship to yield quantitative 3D ϰ -maps. Unfortunately, the inversion problem is ill-posed and requires careful conditioning, either through rotation of the sample being imaged or through regularisation. A simple k-space threshold is introduced to condition the inversion and the preliminary results of applying this method to brain data from healthy subjects and patients with Parkinson's disease and multiple sclerosis are presented. The results suggest that susceptibility mapping is sensitive to iron deposition and could be a useful tool in investigating the progression of neurodegeneratived diseases. Iterative inversion algorithms, which deal with noise more robustly and allow more sophisticated filtering techniques to be employed, are then presented. These powerful regularisation methods are compared to previously described techniques, and are shown to yield high quality whole-brain ϰ -maps.

The push toward higher magnetic fields in MRI has consistently thrown up new challenges in hardware development. The recent development of a new generation of ultra-high field scanners for human imaging is no exception. The earch presented in this thesis aims to provide solutions to new technical challenges in radio-frequency probe design. All probe designs were developed for use at 7T on a Philips Acheiva full body scanner.

This thesis addresses the problem of radiofrequency probe design for Ultra High Frequency Magnetic Resonance Imaging (7T). The signal-to-noise ratio available in Magnetic Resonance Imaging (MRI) is determined by the static magnetic field strength, causing a continued drive toward higher fields to enable faster image acquisition at finer spatial resolution. The resonant frequency increases linearly with static field strength. At 7T the proton resonant frequency is 300MHz, with a wavelength of approximately 13cm in tissue. As this is smaller than the dimensions of the human head, the phase of the radiofrequency (RF) signal varies considerably across the sample, producing field cancellation due to interference. A full wave electromagnetic simulator, using the Transmission Line Matrix (TLM) method, was developed to investigate RF probes at high frequency. A Birdcage probe operating at 64, 128 and 300MHz (corresponding to 1.5, 3 and 7T) was simulated, loaded with an anatomically detailed human head model. A half-wave microstrip was investigated for use as a high frequency probe element. Magnetic and electric fields produced by a single microstrip were simulated, and the strip dimensions varied to investigate the effect on field penetration into the head and Specific Absorption Rate (SAR). A transmit-receive array probe using four microstrip elements was then developed. Bloch simulations were run, using TLM generated magnetic fields, to investigate imaging at short wavelength. Parallel receive probes are demonstrated to offer considerable advantage over volume probes, as signals from receive elements can be combined without interference. There is no transmit equivalent to parallel reception; simultaneous excitation of independent probe elements causes interference in exactly the same manner as a volume probe. A new imaging sequence was developed using a Burst-like encoding to allow sequential excitation of probe elements, without interference, which can be recalled in a single readout. An improvement in image homogeneity was demonstrated, and SENSE acceleration of the new imaging sequence is shown. The sequence was implemented at 3T using a purpose built four element microstrip probe. An RF multiplexer was also built to enable transmit element switching during the imaging sequence. It was demonstrated that images due to different RF excitations, acquired in a single EPI readout, can be separated.

BURST is a fast single-shot imaging technique used in magnetic resonance imaging. Most previous implementations of BURST on whole body systems have been carried out at fields of 1.5T and lower. In this work BURST has been implemented on a 3T whole body system. The signal and attenuation characteristics are discussed, leading to an approximate expression for the optimum echo time. A novel method for controlling the pulse amplitude envelope is described. It is shown that this can lead to a modest gain in signal with little loss of resolution, or to a much greater gain in signal with more severe blurring of the image. Frequency modulated RF pulses have been introduced in order to reduce the peak RF power required, which was found to be beyond the range of a 2kW amplifier. The resulting images show a quadratic phase roll in the readout direction, with no reduction in magnitude or image artefacts. A reduction in peak power of more than ten-fold is demonstrated. The maximum reduction factor possible is shown to be approximately equal to the number of pulses applied in the RF train. The total RF power has been reduced by superimposing a sinusoidal oscillation onto the constant excitation gradient. Each pulse is applied at a gradient minimum, and hence need be of lower bandwidth, while the overall excitation is unchanged. The RF pulses are transformed using VERSE. A reduction in SAR of 32% is demonstrated. Greater reductions can be achieved by increasing the amplitude of the oscillation. However, this technique introduces a moderate amount of acoustic noise into the sequence.

Ever since the inception of Diffusion Tensor Imaging (DTI), unabated advancements in its capabilities and applications have been spearheaded by a vibrant research effort to devise dedicated acquisition sequences, protocols and hardware. In translational research, however, the transition of these innovations into the arenas of biomedical research, and ultimately clinical practice is frequently hampered by practical considerations. These include the availability of appropriate expertise, time and resources for their implementation, and considerations of compatibility with established techniques and results reported in literature. Such concerns provide the impetus to maximize the utility of existing protocols before attempting the development of novel dedicated techniques. In this thesis, three investigations, each targeting a different DTI application, are presented. The strategy implemented throughout involves assessing the suitability of existing sequences for the intended task, and determining any limiting factors, evaluating whether appropriate modifications of the acquisition protocols used are capable of alleviating limitations, and developing novel, dedicated protocols wherever necessary. The value and, importantly, the wide scope of this approach in answering important research questions is exemplified through the breadth of the studies presented. The first study presents, for the first time, a quantitative evaluation of the effects of cardiac pulsation on prevalent DTI metrics acquired with a specific acquisition protocol used routinely in clinical practice. Findings inform the on-going debate on whether the investment in cardiac gating is merited by improvements in data quality. Effects were observed during only 6 % of the cardiac cycle, and not 20 % as previously reported. The impact of cardiac pulsation on selected diffusion Tensor indices was minimal in group studies, but of potential practical relevance in individual cases. Methods to predict which individuals may benefit from gating have also been suggested. Secondly, the feasibility of post-mortem DTI was established through the successful acquisition, also for the first time, of DTI data on a chemically fixed whole human post-mortem brain using a clinical sequence. Previous failed attempts have been attributed to insufficient SNR. In this study scanner stability and distortion are found to be the main limiting factors, and mitigated using appropriate averaging and co-registration strategies. The third study assessed the potential of ultra-high field strength DTI by identifying and optimizing the potential strengths of DTI at 7T. Subsequent to optimization with respect to SNR, the main sources of artefact were found to be B1 inhomogeneity and inadequate fat suppression. Both were alleviated by modification of the available acquisition protocol, resulting in higher SNR and data quality than previously reported. Finally, in developing appropriate data quality measures, the ‘Difference method’, commonly used for the quantification of SNR, was found to be unsuitable for in vivo DTI acquisitions at 7 T, leading to the proposal, and successful implementation and validation of an alternative.

This thesis addresses the much needed quantitative assessment of the physiological effects the magnetic fields used in MRI. MRI has a well-earned reputation for being a safe, non-ionising alternative imaging modality for both the patient and MR practioner alike. Although the operating environment can prove hazardous for inexperienced and untrained personnel, appropriate training and adoption of safe working practices can prevent adverse incidents.