. In this study, we aimed to use machine learning methods to establish the quantitative value of MRI alone in the prediction of changes between disease states such as from normal cognitive function (NC) to mild cognitive impairment (MCI), and MCI to AD, and compare with the combined predictive power of MRI, PET, neuropsychological evaluations and CSF analysis. Very high overall accuracy can be achieved using both RF and DNN methods. Interestingly, predictive power of MRI features is very close to all features combined, suggesting MRI might contain much of the information provided by neuropsychological evaluations, PET scans among others combined. The methodology adopted in this study also provides a framework for evaluating the value of different imaging techniques in a quantitative manner.
This abstract and the presentation materials are available to members only; a login is required.