基于可视图类方法的静息态脑电信号特征分析

Feature Analysis of Resting EEG Signals Based on Visual Graph Class Method

基于可视图类方法的脑电网络构建对于区分健康脑电与病理性脑电具有重要的意义。以10名健康志愿者(5男5女)的数据集为研究对象。首先,为脑电的每个通道创建可视图或者水平可视图;接着,构建脑电的多路可视图,并将其映射成复杂网络;进而,计算该网络的传递性、全局效率以及自定义平均权重3个特征;最后,通过T检验法分析闭、睁眼状态下特征分布的差异性。实验结果表明:相较于可视图法,基于水平可视图法获得的特征分布差异性更为显著。

The construction of Electroencephalogram(EEG)network based on visual graph class method is of great significance for distinguishing healthy EEG from pathological EEG.In this paper,the data set of 10 healthy volunteers(5 men and 5 women)are taken as the research object.Firstly,we create a visual graph or horizontal visual graph for each channel of the EEG.Secondly,we construct multi-channel visual graph of the EEG and map them into a complex network.Thirdly,we calculate three features of the network:transmissibility,global efficiency and custom average weight.Finally,we analyze the difference of features distribution between closed-eye and opened-eye state by the T-test method.The experimental results show that compared with the visual graph method,the difference of features distribution obtained by the horizontal visual graph method is more significant.