A Multiple stage Discriminative Event Based Model for Alzheimer’s Disease Progression Timeline Estimation

摘要

As an irreversible neurodegenerative disease, the progression of Alzheimer’s disease includes three stages: cognitively normal (CN), mild cognitive impairment (MCI) and Alzheimer’s disease (AD). Early diagnosis of the disease i.e. the stage of MCI is helpful for its early intervention. In this paper we propose a multistage discriminative event-based model (MDEBM). By modelling the discriminative events in cross-sectional data of the patients into three categories through Gaussian Mixture, then staging the patient by Bayesian classifier, the model can detect not only the patients with AD but also those with MCI. besides, the diagnosis of Alzheimer’s disease is often based on a combination of factors including the patient's medical history, the results of cognitive tests and neuroimaging. This model supports gradually improving the accuracy of diagnosis through the consecutive input of different categories of patient’s examination results. Experiments on Alzheimer's Disease Neuroimaging Initiative (ADNI) data prove the effectiveness of our model.