Classification Research For Alzheimer’s Disease Based On The Spatial Texture Features Of Cortical Thickness

Alzheimer’s disease (AD) is the main type of the old age dementia, accounts for about60%. Patients with MCI (Mild Cognitive Impairment) is considered to be a group at high risk of developing into AD. AD is a neurodegenerative and onset conceals disease of the central nervous system, which early is not easy to find. Moreover, presently no biological markers as diagnostic criteria are established before the symptoms are shown, it makes the AD early diagnosis difficult. And based on the existing biological indexes related to AD disease, study of the classification between AD/MCI and the normal controls, which can provide help for early clinical diagnosis^prevention and treatment of AD, is a very effective ways to reduce the harm of AD. Therefore, Study of automated method to accurately identify the patients from clinical subjects is a hot and difficult problem of AD based on the neuro-imaging studies. This paper intends to explore the classification of AD and MCI based on AD sMR1data, thus providing theoretical basis and method guidance for early diagnosis of the disease.The main contents and achievements of the study are as described below.1) We compares and analyzes the biological indicators and classification methods used in the classification studies of AD and MCI based on sMRI, in order to provide a reference for the follow-up classification study, it has an important significance for the classification research of AD and MCI;2) Based on the cortex thickness which is a mainstream biological indicator in classification of AD and MCI, proposed a new feature extraction algorithm. On the one hand, the higher classification performance of the proposed method verifies that the methods which extracted features by using PC A or taking the average of cortical thickness zones is not very appropriate to differentiate patients with AD or MCI from normal controls. On the other hand, it also illustrates that the cortex thickness changes caused by AD is not a simple uniform decrease pattern, but a kind of uneven atrophy models which has not been found by the current study so far. And the local spatial features of the cortex thickness changes in the process of this shrinking. This may guide a new direction for the classification of AD based on the cortical thickness.