Abstract:

Alzheimer's disease (AD) is a common neurodegenerative disease in the elderly. Early diagnosis and prediction plays an important role in early intervention and delaying disease progression of AD. This paper focused on the principles and process of pattern classification method, and its application in the clinical study and auxiliary diagnosis of AD. The biomarkers, neuroimaging and cognitive ability scales are important features for pattern classification. Various classification algorithms including Bayesian networks, decision trees, support vector machines (SVM) and multilayer perception have been adopted to distinguish AD, mild cognitive impairment (MCI) and normal aging subjects. Besides, they can effectively trace and analyze MCI patients.

Key words: Classification, Alzheimer disease, Biological markers, Review

摘要：

阿尔茨海默病是老年人常见的神经变性病，早期诊断、转化预测对早期干预、延缓病情进展具有重要意义。本文重点阐述模式分类方法之原理和流程，以及该方法在阿尔茨海默病临床研究和辅助诊断中的应用。其中，生物学标记、神经影像学和神经心理学测验为其重要特征，通过各种分类计算方法可以区分正常老龄化、轻度认知损害和阿尔茨海默病，并对轻度认知损害的动态追踪分析具有较好作用。

关键词: 分类法, 阿尔茨海默病, 生物学标记, 综述