Design And Implementation Of An FMRI-based Alzheimer's Disease Diagnosis System

An auxiliary diagnosis system for Alzheimer's Disease(AD)is designed and implemented in this paper.The system uses a front-end and back-end separated architecture.The front-end uses Vue and the back-end uses the Django framework.With a functional magnetic resonance imaging(fMRI)data matrix from user input,the system uses machine learning methods to determine whether the sample owner should be in AD state or mild cognitive impairment(MCI)state or health(HC)state.The main contributions of this article are in three aspects:(1)Regarding the data of different brain regions in the sample fMRI data as sampling signals of some data distributions,Kolmogorov-Smirnov(KS)test and Kruskal-Wallis(KW)test are used to test data distributions of all 90 brain regions in this paper.The results show that the data distributions of AD population,MCI population and HC population on four brain areas are the same with in same class but different in different classes.At the same time,we also find that data distributions of other brain regions in different categories of people have similarity characteristics.(2)Using the data distribution characteristics of the four brain areas,logistic regression,support vector machine,and artificial neural network are used to construct classifiers to classify AD,MCI,and HC respectively.A voting-based aggregation method is employed to make the final conclusion for the patient from the results of the three classifiers.Experiments show that this method achieves good performances.(3)A convenient online AD disease diagnosis system is designed and implemented in this paper.This platform can diagnose both AD and MCI.It can prevent patients to be AD state before AD state happens.It improves the cure rate,and contributes to the early detection of patients which can improve quality of their life.Generally,based on the different characteristics of brain area data distribution,an aggregation method based on three classifiers is used to judge the state of patient,and a WEB-based online system is implemented to provide effective support for the diagnosis of AD patients.