Project Summary The objective of this R21 proposal is to give proof that magnetic susceptibility and volume of cerebral microbleeds quantified by an innovative MRI technique, CISSCO, and validated by pathological examination can differentiate postmortem brain samples from control and demented subjects. Microbleeds appear in varying numbers in images, but they are strongly associated with vascular cognitive impairment (VCI), small vessel diseases (SVD), and Alzheimer's disease (AD). They also appear in healthy older people at a lower prevalence. Current clinical diagnosis only counts the number of microbleeds and their mimics from images. However, the number of these “apparent” micro-objects is subject to imaging parameters (including the MRI field strength) and does not correlate well with the progression of cognitive decline. As microbleeds with hemorrhagic components show magnetic susceptibility effects in MRI, these investigators hypothesize that magnetic properties of microbleeds may help to differentiate certain clinical dementia subtypes. These novel quantitative parameters may be better markers for incipient dementia. They will advance our current understanding of the contribution of microbleeds to dementia beyond postmortem analysis and toward living patients. To design clinical imaging protocols and parameters, susceptibility values of microbleeds used to distinguish between control and demented subjects must be known first. The proposed aims are to image and pathologically examine micro-objects in postmortem samples obtained from the Michigan Brain Bank (MBB), which will blind these investigators to the clinical and neuropathological diagnosis of each subject until the late stage of this project. Formalin fixed, paraffin-embedded blocks where microvascular lesions are suspected will be obtained from subjects who died with no cognitive impairment or mixed dementia (including cases with high- and intermediate- likelihood of AD pathology). The CISSCO method will be applied to MR images of postmortem samples for accurate quantification of the magnetic susceptibility and volume of each micro-object. These micro-objects observed in MRI will then be co-registered and identified histologically in the same samples. These micro-objects, which are microbleeds and their mimics, will be categorized based on pathological results and quantified susceptibility values from MRI. They will also be compared between the diagnostic groups after the investigators are un-blinded. Cox hazard ratios will be calculated for different categorized results. If different magnetic properties can differentiate the clinical groups, then this outcome would indicate that magnetic properties of microbleeds is a potential advance in imaging biomarkers for dementia. With proof from this R21, future clinical testing plans will be proposed to NIH.