摘要
识别和描述房颤有可能自发终止或持续,不仅可以更好地了解心律不齐终止的机制,还可以更有效地治疗持续房颤。本研究从两种非线性分析的角度提取房颤信号RR间期的特征,并预测房颤是否能自发终止。一是计算心电信号RR间期序列的Lempel-Ziv复杂度,二是基于符号动力学将RR间期序列转换成符号串,对符号串编码得到符号码,计算各符号码的发生概率,取符号码110的发生概率和符号码发生概率的香农熵作为RR间期序列的特征参数。基于从RR间期提取的上述三个非线性参数,建立模糊分类器来预测房颤是否能终止。实验研究了一个由Holter心电信号组成的房颤数据库,它包括一个训练集和两个测试集(A和B)。结果表明:本方法可正确分类90%的测试集A和80%的测试集B,和以前方法相比,预测房颤终止的准确率提高了约10%。可见,本方法对Holter心电信号预测房颤的自发终止是有效的。
It may lead to not only the better understanding of the termination mechanism of arrhythmia but also the improved treatment of the sustained atrial fibrillation （AF） by recognizing and characterizing whether the AF is likely to terminate spontaneously or sustained. In this paper, RR intervals of the AF were characterized with two kinds of nonlinear analysis methods to predict the termination of the AF. One feature was the Lempel-Ziv complexity calculated from RR intervals. Other two came from the symbolic dynamics. Firstly RR intervals were transformed to symbol strings. Then symbol codes were obtained by coding symbol strings. After the occurrence probability of all possible symbol codes were calculated, the occurrence probability of the symbol code 110 and the Shannon entropy of the occurrence probability of symbol codes were taken as two parameters of RR intervals. Based on the above three nonlinear parameters which were extracted from RR intervals, the fuzzy classifier was set up to predict the termination of the AF. An AF database which includes a training set and two testing sets （A and B） of Hoher Electrocardiograph （ECG） recordings was used to verify the proposed method. Results showed that 90% of the testing set A and 80% of the testing set B were correctly classified. Compared to previous methods, the proposed method improved the prediction accuracy for the AF termination by approximately 10% . h is demonstrated that the algorithm has the ability to predict spontaneous termination of the AF from Hoher ECG recordings effectively.
出处
《中国生物医学工程学报》
CAS
CSCD
北大核心
2008年第2期213-217,254,共6页
Chinese Journal of Biomedical Engineering
基金
国家基础研究项目(2005CB724303)
国家自然科学基金(30570488)
上海市科技计划(054119612)
关键词
房颤
终止
预测
RR间期
非线性分析
atrial fibrillation
termination
prediction
RR interval
nonlinear analysis
