There is a need for sensitive measures of disease progression in multiple sclerosis (MS) to monitor treatment effects and understand disease evolution. MRI measures of brain atrophy have been proposed for this purpose. This thesis investigates a number of measurement techniques to assess their relative ability to monitor disease progression in clinically isolated syndromes (CIS) and early relapsing remitting MS (RRMS). Presented, is work demonstrating that measurement techniques and MR acquisitions can be optimised to give small but significant improvements in measurement sensitivity and precision, which provided greater statistical power. Direct comparison of numerous techniques demonstrated significant differences between them. Atrophy measurements from SIENA and the BBSI (registration-based techniques) were significantly more precise than segmentation and subtraction of brain volumes, although larger percentage losses were observed in grey matter fraction. Ventricular enlargement (VE) gave similar statistical power and these techniques were robust and reliable; scan-rescan measurement error was <0.01% of brain volume for BBSI and SIENA and <0.04ml for VE. Annual atrophy rates (using SIENA) were -0.78% in RRMS and -0.52% in CIS patients who progressed to MS, which were significantly greater than the rate observed in controls (-0.07%). Sample size calculations for future trials of disease-modifying treatments in RRMS, using brain atrophy as an outcome measure, are described. For SIENA, the BBSI and VE respectively, an estimated 123, 157 and 140 patients per treatment arm respectively would be required to show a 30% slowing of atrophy rate over two years. In CIS subjects brain atrophy rate was a significant prognostic factor, independent of T2 MRI lesions at baseline, for development of MS by five year follow-up. It was also the most significant MR predictor of disability in RRMS subjects. Cognitive assessment of RRMS patients at five year follow-up is described, and brain atrophy rate was a significant predictor of overall cognitive performance, and more specifically, of performance in tests of memory. The work in this thesis has identified methods for sensitively measuring progressive brain atrophy in MS. It has shown that brain atrophy changes in early MS are related to early clinical evolution, providing complementary information to clinical assessment that could be utilised to monitor disease progression.