基于功能磁共振影像的阿尔茨海默病分类研究

Research on AD classification based fMRI

通过功能磁共振影像研究阿尔茨海默病与正常对照组之间的特征差异.研究对象为16例阿尔茨海默病和28例正常对照组的功能磁共振影像.首先,对静息态功能磁共振影像进行预处理;接着,运用基于信息极大化的独立成分分析算法提取阿尔茨海默病和正常对照组样本的组间特征;最后,将提取出的特征表达系数送入支持向量机分类器对阿尔茨海默病和正常对照组样本分类.在阿尔茨海默病与正常对照组的分类实验中,获得的最大平均分类准确度、敏感度及特异度分别为97.82%、94.88%、99.50%.结果显示,组间的差异特征能够比较准确地将阿尔茨海默病从正常对照组中区分开来.

The paper studies the characteristic differences between Alzheimer＇s disease（AD）and the healthy controls based on functional magnetic resonance imaging（fMRI）.The study objects are the fMRI of 16Alzheimer＇s disease patients and 28 healthy controls.Firstly,the resting-state fMRI was preprocessed.Then,group features of Alzheimer＇s disease and healthy controls with information maximization algorithm based on independent component analysis.Finally,the features into the support vector machine classifier for the classification of Alzheimer＇s patients and healthy controls.In the classification experiment of Alzheimer＇s patients and healthy controls,the obtained maximum average accuracy,sensitivity and specificity are 97.82%,94.88%,99.50%.The results showed that the differences between the groups can separate Alzheimer＇s disease from the healthy controls accurately.