Abstract:

Because of the high incidence and irreversibility of Alzheimer’s Disease（AD）, the early diagnosis of AD patients is particularly important. It has been found that AD patients are related to Apolipoprotein E（APOE）, and the complexity also changes. Complexity can be used in AD diagnosis, but its classification performance needs to be further improved. Using Permutation Entropy（PE） as index, the complexity change pattern of AD with different gene types are discussed. This paper examines the possible effects of genotype on the brain complexity in the Normal Control（NC）, Early Mild Cognitive Impairment（EMCI）, Later Mild Cognitive Impairment（LMCI） and AD by Permutation Entropy（PE）. Extracting PE values from brain regions with significant differences as training features, and training different classifiers according to gene type. The results show that the accuracy of EMCI and NC is 96.67% by adding gene information, and the classification accuracy of EMCI and LMCI is increased from 40.35% to 88.24%, which significantly improves the accuracy of early diagnosis of AD.

Key words: Alzheimer’s Disease（AD）, Apolipoprotein E（APOE） ε4, permutation entropy, classifiers

摘要：

由于阿尔茨海默病（Alzheimer’s Disease，AD）发病率高且不可逆，所以AD早诊尤为重要。已有研究发现AD患者与载脂蛋白E（Apolipoprotein E，APOE）有关，复杂度也有变化。可将复杂度用于AD诊断中，但其分类性能有待进一步提高。以排列熵（Permutation Entropy,PE）为指标，探讨了不同基因型的AD患者复杂度变化模式，研究了APOE载体的正常对照组（Normal Control，NC）、早期轻度认知损害（Early Mild Cognitive Impairment，EMCI）、晚期轻度轻度认知损害（Later Mild Cognitive Impairment，LMCI）和AD与未携带者脑信号复杂度的差异，提取显著差异脑区的PE值作为特征向量，根据基因型分别训练不同的分类器。结果表明，加上基因信息后可以96.67%的准确率区分EMCI与NC，且EMCI与LMCI的分类正确率由40.35%提高到88.24%，显著提高了AD早诊的正确率。

关键词: 阿尔茨海默症, 载脂蛋白E（APOE）, 排列熵, 分类器