By applying an appropriate magnetic field gradients, MR imaging may be able to register the random, thermally driven motion of water molecules of brain(diffusion). Diffusion is strongly anisotropic, directionally dependent in White Matter (WM) fiber tracts, because axonal membranes and myelin sheaths represent barriers to the water molecular motion, in directions not parallel to their ones. A common measure of diffusion motion is the fractional anisotropy FA (Pierpaoli, Basser 1996). In recent years, there has been a great interest in using diffusion anisotropy as a marker for WM tracts integrity, so FA is a useful quantity to compare across subjects, because it is a scalar voxelwise defined value. In our study we have analyzed a dataset composed by a group of 17 patients affected by Alzheimer diseases and a control group of 11 unities. MRI images have been acquired by means a 3 Tesla Siemens Impact Scanner, from the Neuroradiological Unit of the Ospedale Maggiore (Borgo Trento, Verona). After have been realized FA images for each subject, we have used the Tract Based Spatial Statistical method (Smith et al.,2006) for a voxelwise statistical analysis. The TBSS is a fully automated approach, investigating the whole brain with a great limitations of problems caused by alignment inaccuracies and smoothing extent. This is achieved by estimated a group mean FA skeleton, which represents the centers of fiber tracts common to the investigated subjects. After have been projected each subject FA image onto to the common skeleton with the FA values derived from the nearest tract centre, we carry out voxelwise statistics across subjects using skeleton FA data.