Abstract
The behaviour of an NMR signal (the 'NMR response') from a sample depends on the method used to excite the signal, and on three characteristic NMR properties of the sample; the concentration of NMR-sensitive atomic nuclei, the long itudinal relaxation time and the transverse relaxation time. In NMR imaging cross sectional maps are produced which are related to the spatial distribution of NMR properties within the sample. This thesis forms part of a continuing programme of research to investigate the application of NMR imaging to biological systems. The relationship between the intensities on an NMR image and the NMR properties of the sample is analysed. For the medically important case of biological systems imaged by the method of reconstruction from projections a new model is described which predicts the observed behaviour. Earlier models are shown to be unsatisfactory. A review is made of published measurements of the NMR properties of excised tissues. For the soft body organs the NMR response is shown to arise mainly from tissue water, with contributions up to 20% from fat. The relaxation times for excised tissues are shown to depend on the intracellular water content. A model is described which explains this dependence as the result of water molecules exchanging rapidly between two phases of the intracellular water. An extensive series of measurements of the relaxation times of live human tissues is presented. The properties of live tissues are shown to be significantly different from excised tissues. It is suggested that the difference is due to the presence of blood in live tissues. Experiments are described which will measure the properties of the tissue intracellular water and blood supply separately. The relative intensities of different tissues on an NMR image is shown to depend in a complicated way on the imaging method, with drastic changes in relative intensities for even small changes in the imaging method. The use of earlier simple models to interpret NMR images is shown to be misleading in many cases. A procedure is proposed, using, the new models described, which allows the 'optimum' NMR imaging method to be chosen for any given application.