基于递归图网络构建的脑电信号研究

发布时间：2018-05-20 06:53

本文选题：EEG + 递归图　；参考：《南京邮电大学》2017年硕士论文

【摘要】：人脑是由100多亿神经元细胞构成的网络结构,在大脑的神经活动中表现出高度的复杂性、非平稳性和非线性行为。利用递归图和复杂网络理论能够研究大脑在进行神经活动时不同区域之间的相互作用、拓扑结构和动力学信息以及生理功能。脑电(EEG,electroencephalogram)蕴含了大量与人脑相关的生理病理信息,是人体非常重要的生理电信号之一。本文从递归图和复杂网络角度分析了睡眠脑电和清醒时脑电信号的不同,从而得出结论,为研究睡眠状态下大脑的状态提供了重要的理论依据。其中,包含的主要工作有:一、首先使用递归图对睡眠脑电和清醒脑电进行分析。对脑电信号进行相空间重构时,分别采用自相关函数法和Cao方法确定延迟时间和嵌入维数。利用递归图(RP,Recurrence Plot)来分析睡眠脑电和清醒脑电相空间重构后网络递归特性的不同,在此基础之上使用递归定量分析(RQA,Recurrence Quantitative Analysis)对其进行量化分析,实验证明清醒脑电信号非线性特性更强,不确定性更高。二、论文应用排序递归图(ORP,Order Recurrence Plot)分析睡眠脑电和清醒脑电信号。先使用排序递归图对两种脑电信号的递归特性进行定性分析,然后使用RQA进行定量分析,实验证明,清醒状态的脑电信号的非线性特性明显强于睡眠状态,即进入睡眠状态之后,大脑神经系统非线性减弱。该方法与递归图的RQA相比,两种信号之间的区分度更高,效果更佳。三、使用递归图分析时获得的递归矩阵,构建复杂网络,并对网络进行可视化,定性的分析两种信号网络拓补结构的不同,进而使用复杂网络的参数,定量的分析两种网络之间的区别。论文中使用的方法适用于分析睡眠脑电和清醒脑电网络的之间的区别,有助于了解睡眠脑电和清醒脑电神经网络的动力学行为差异。
[Abstract]:The human brain is a network consisting of more than 10 billion neuron cells. It shows a high degree of complexity, nonstationary and nonlinear behavior in the neural activities of the brain. The recursive graph and complex network theory can be used to study the interaction, topological structure, dynamic information and physiological function of different regions of the brain in the process of neural activity. EEG electroencephalograms (EEG) contains a lot of physiological and pathological information related to human brain and is one of the most important physiological electrical signals in human body. In this paper, the difference between sleep EEG and waking EEG is analyzed from the angles of recursion graph and complex network, and the conclusion is drawn, which provides an important theoretical basis for studying the state of the brain in sleep. The main contents are as follows: firstly, sleep EEG and awake EEG are analyzed by recursion graph. In the phase space reconstruction of EEG signals, the delay time and embedding dimension are determined by autocorrelation function method and Cao method respectively. The difference of network recursion characteristics after sleep EEG and waking EEG phase space reconstruction is analyzed by using Recurrence Quantitative analysis, and then the quantitative analysis is carried out by using recursive quantitative analysis (RQA / Recurrence Quantitative Analysis) to analyze the difference between sleep EEG and waking EEG phase space remodeling, based on which the RQAurrence Quantitative analysis is used to analyze the difference between sleep EEG and waking EEG phase space reconstruction. The experimental results show that the nonlinear characteristics and uncertainty of conscious EEG signals are stronger and higher. Secondly, the order Recurrence Plotde is used to analyze sleep EEG and wakefulness EEG signals. The recursive characteristics of the two kinds of EEG signals are qualitatively analyzed by using sort recursion diagram, and then quantitative analysis is carried out by using RQA. The experimental results show that the nonlinear characteristics of EEG signals in awake state are obviously stronger than those in sleep state. After entering a sleep state, the neural system of the brain is nonlinear weakened. Compared with the RQA of recursion graph, this method has higher discriminant degree and better effect. Thirdly, using the recursive matrix obtained in the analysis of recursion graph, the complex network is constructed, and the network is visualized and qualitatively analyzed for the difference between the two signal networks, and then the parameters of the complex network are used. The difference between the two networks is analyzed quantitatively. The method used in this paper is suitable for analyzing the difference between sleep EEG and awake EEG network, and is helpful to understand the difference of dynamic behavior between sleep EEG and awake EEG neural network.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R338;TN911.7

【相似文献】