Discriminant analysis in the study of Alzheimer’s disease using feature extractions and support vector machines in positron emission tomography with 18F-FDG

With more successful applications of advanced medical imaging technologies in clinical diagnosis,various analytic discriminant approaches, by seeking the imaging based characteristics of a given disease to achieve automatic diagnosis, gain greater attention in the medical community. However the existing computer-aided discriminant procedures for Alzheimer's disease(AD) are yet to be improved for better identifying patients with mild cognitive impairment(MCI) from those with AD and those who are cognitively normal. In this work we present a computer assisted diagnosis approach by first statistically extracting characteristics from whole brain2-deoxy-2-(18F)fluoro-D-glucose positron emission tomography(18F-FDG PET) images, and then using support vector machines for classification. Evaluations of the proposed procedure with patient data exhibit satisfactory accuracies in distinguishing AD from its early stage MCI, and normal controls.