Multiple Sclerosis is a demyelinating disease affecting the communication in the central nervous system. Magnetic resonance diffusion imaging provides information about water diffusion in white matter and allows an early detection of abnormalities, comparing to conventional magnetic resonance techniques. The aim of this study is to find out which the brain regions that are damaged during disease progression. Tract-Based Spatial Statistics is a voxelwise multi-subject statistical analysis which performs non-linear registration of each subject's image and projects them onto an alignment-invariant tract representation where the statistical tests are accomplished. This approach improves the sensitivity, objectivity and interpretability of results. The study compares brain images of 64 healthy controls and 59 patients with different stages of Multiple Sclerosis. Some preliminary statistical tests were performed and although the results are still under study, it shows that corticospinal tracts as the most region affected by the disease.