Computer Aided Diagnosis system for Alzheimer Disease using brain Diffusion Tensor Imaging features selected by Pearson's correlation

► We contribute a feature selection process based on the correlation of Fractional Anisotropy (FA) and Mean Diffusivity (MD) voxels with the subject class indicative variable. ► We define and test an image co-registration pipeline for anatomical reference of FA and MD volumes which is critical for the feature selection process. ► Voxel sites found by feature selection on a database of Alzheimer's Disease (AD) and healthy controls (HC) agree with previous findings on. ► Classification experiments discriminating AD patients from HC on the basis of FA and MD features compare or improve over results found in the literature for other image modalities (i.e. PET, SPECT). ► We conclude that the proposed features can be used to build classification-based computer aided diagnosis tools.