功能连接偏侧化在AD早期诊断中的应用研究

The Applied Research of Asymmetric Functional Connectivity in AD Early Diagnosis

借助脑功能连接方法,研究AD的偏侧化现象,并将其用于AD的辅助诊断中。实验采用ADNI数据集,首先制作可用于偏侧化研究的脑膜板,接着构建半球功能脑网络,计算网络连接强度,并计算偏侧化指数。利用统计分析的方法,筛选可用于AD辅助诊断的特征,并使用SVM(support vector machine)分类器训练分类模型。结果显示,加入偏侧化特征后的分类准确率为89.17%,敏感度为90.28%,特异度为88.24%,证明偏侧化指数的加入对于AD的分类准确率有提高作用。

n this study,functional connection strength was used to explore the hemispheric asymmetry in AD so that it can serve the AD aided diagnosis and improve the classification accuracy.ADNI database was used to verify the idea we proposed.The symmetric brain template was made to construct hemisphere brain networks,then the functional connection strength and laterality index were calculated.By using statistical analysis,the features for AD aided diagnosis were screened.According to the features,the feature space was made and the classification model was trained with SVM classifier.Importantly,the classification accuracy was improved to 89.17%,sensitivity to 90.28%and specificity to 88.24%.In conclusion,the laterality index was helpful for classification.