Measurement of cerebral blood flow in humans using MRI with arterial spin labelling

Cerebral blood flow (CBF) is an important marker for tissue well-being and brain function. Traditionally, CBF has been measured with invasive methods using exogenous tracers. With the introduction of arterial spin labelling (ASL) MRI techniques, non-invasive, quantitative CBF mapping in humans became feasible. The aim of the work described in this thesis was to implement, optimise and validate an ASL MRI technique for measuring CBF in humans. Furthermore, the utility of this technique for longitudinal studies was to be evaluated. FAIR, a pulsed ASL technique, was successfully implemented. A data acquisition and analysis protocol was proposed after an extensive survey of acquisition and quantification issues such as rf pulse profiles, static subtraction error, voxel partial voluming, volunteer movement and choice of CBF quantification models. Fitting for CSF fraction resulted in an improved goodness-of-fit when compared to the standard fit model. It was also derived that it is necessary to analyse the magnitude of the FAIR data to avoid erroneous results. Furthermore, it was shown that the inflow delay has to be included in the fit model for an accurate CBF value. Two image realignment protocols were proposed and evaluated. The CBF measurement was validated by comparing baseline and parametric CBF results with literature values. Finally, the technique was applied in a study of CBF, inflow delay, BOLD signal and cerebrovascular reactivity (CR) over the course of a day. The findings were used to propose upper limits to diurnal changes in these parameters. The standard deviation of a CBF measurement was found to be satisfactory for functional experiments, but not for longitudinal studies where small CBF changes are of interest. It was argued how FAIR could be made more suitable for longitudinal studies by applying 3D-realignment to whole brain FAIR, acquired preferably at a higher magnetic field strength