功能脑网络偏侧化在AD早期诊断中的应用研究

发布时间：2018-04-08 13:07

本文选题：阿尔茨海默病　切入点：轻度认知障碍　出处：《太原理工大学》2017年硕士论文

【摘要】：阿尔茨海默病(Alzheimer’s Disease,AD)又称原发性老年痴呆病,以严重的高级认知功能和记忆功能障碍为临床表现。轻度认知障碍(Mild Cognitive Impairment,MCI)是AD与正常人间的一种过渡状态。该类疾病潜伏期长不易被发现,且病变严重,损害不可逆。我国已逐渐进入老龄化社会,防治AD成为一项严峻的工作。如能在更早的时间发现疾病的征兆,并及时进行医疗干预,可以降低病人的痛苦,节省治疗的开支,因此发现AD早期异常特征是AD辅助诊断系统的研究热点之一。目前,基于磁共振影像的AD辅助诊断虽已取得了一些成果,但使用的主要是结构特征,而一旦结构发生变化,患者已进入AD晚期,进入了不可逆阶段。如能利用认知功能异常作为辅助诊断的依据,就可以做到早发现、早干预。因此AD早期诊断的关键是找到AD认知功能的异常指标。大脑偏侧化现象是指大脑的左右半球存在结构和功能上的不对称性,本研究借助脑功能网络研究方法,以图论理论为基础,研究AD的偏侧化现象,并将其用于AD的辅助诊断中,提高AD早期诊断的分类准确度。本文主要研究工作如下:(1)不同于传统的脑功能网络研究过程,本研究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,计算网络连接强度与拓扑属性,并计算偏侧化指数。(2)利用统计分析的方法,筛选可用于AD辅助诊断的特征,并对筛选的特征进行生理意义的解释。(3)根据筛选的特征,构造特征空间,使用SVM(support vector machine)分类器训练分类模型,采用留一验证法测试分类模型。(4)采用ADNI数据集,验证本文提出的方法,结果表明,AD患者确实存在偏侧化异常现象,筛选出的偏侧化特征与其他研究方法得到较为一致的结论。加入偏侧化特征后的分类准确率为85.71%,敏感度为87.06%,特异度为84.34%。本文对比了使用功能连接、功能连接及其偏侧化指数,网络属性、网络属性及其偏侧化指数的分类结果,结合已有文献,得出了偏侧化指数的加入对于AD的分类准确率有提高作用,尤其对于轻度认识障碍与正常对照组的准确率提升作用明显。说明偏侧化指数确实能提高AD辅助诊断准确率,对AD早期辅助诊断的研究提供了一定的依据。
[Abstract]:Alzheimer's disease (Alzheimer's disease), also known as primary Alzheimer's disease (AD), is characterized by severe advanced cognitive and memory impairment.Mild cognitive impairment (mild Cognitive Impairment MCI) is a transitional state between AD and normal subjects.The long incubation period of the disease is not easy to be found, and the lesion is serious and irreversible.China has gradually entered an aging society, prevention and treatment of AD has become a serious task.If we can find the symptoms of the disease earlier and carry out medical intervention in time, we can reduce the suffering of patients and save the cost of treatment. Therefore, the discovery of early abnormal features of AD is one of the hot topics in AD auxiliary diagnosis system.At present, although some achievements have been made in AD aided diagnosis based on magnetic resonance imaging, the main structural features are used, and once the structure changes, the patients have entered the late stage of AD and have entered the irreversible stage.If cognitive dysfunction can be used as the basis of auxiliary diagnosis, early detection and early intervention can be achieved.Therefore, the key to early diagnosis of AD is to find abnormal indicators of AD cognitive function.The phenomenon of cerebral lateralization refers to the asymmetry in the structure and function of the left and right hemispheres of the brain. This study studied the lateralization of AD on the basis of graph theory and applied it to the auxiliary diagnosis of AD.To improve the classification accuracy of early diagnosis of AD.The main work of this paper is as follows: (1) different from the traditional research process of brain functional network, the meningeal plate which can be used for lateralization is first made, and then the hemispherical functional brain network is constructed to calculate the network connection strength and topological properties.Using statistical analysis method to screen the features that can be used for AD diagnosis, and to interpret the physiological meaning of the selected features. The feature space is constructed according to the characteristics of the screening.The classification model was trained by SVM(support vector machine, and the classification model was tested by a residual verification method. The ADNI data set was used to verify the method proposed in this paper. The results show that the abnormal phenomenon of partial lateralization does exist in AD patients.The selected characteristics of lateralization are in good agreement with other research methods.The classification accuracy, sensitivity and specificity were 85.71, 87.06 and 84.34, respectively.This paper compares the classification results of using functional connection, functional connection and its lateralization index, network attribute, network attribute and partial lateralization index.It is concluded that the addition of the hemilateralization index can improve the classification accuracy of AD, especially for the mild cognitive impairment and the normal control group.The results show that the hemipartialization index can improve the accuracy of AD auxiliary diagnosis and provide some basis for the study of AD early auxiliary diagnosis.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R749.16;O157.5

【参考文献】