Classification of AD patients based on the type of visible abnormalities in Lead Investigator: Saeid Taheri Institution : University of South Florida E-Mail : taheris@health.usf.edu Proposal ID : 795 Proposal Description: The goal of this research is to build a classifier to classify cognitive impaired patients based on MR-visible cerebral abnormalities and cognitive data. Specifically we are interested to use in vivo imaging biomarkers of cerebral abnormalities that supposedly have vascular origin and investigate their impact on the severity of cognitive impairment. MR anatomical images (T1 and T2) and diffusion images (DTI raw, or calculated FA, MD) when analyzed together may offer a more detailed picture of cerebral damage with vascular origin, as seen by hyperintensity in anatomical images. We aim to use advanced non-parametric classification methods to define classes of cerebral injuries in AD patients and compare with the known pathologies such as multiple infarcts, lacunar, and microbleeds. Of course access to perfusion images will add more details on the pathology of micro bleeds.