心音的非线性时间序列分析——冠状动脉疾病无损检测的新方法  

Nonlinear time series analysis of heart sounds——New method for noninvasive detection of coronary artery disease

作　　者：杨群清[1] 周庆利[2] 刘军[1] 董筠[1] 楼正国[1] 

机构地区：[1]浙江大学生物医学工程系 [2]浙江大学医学院附属妇产科医院,浙江杭州310006

出　　处：《浙江大学学报（工学版）》2006年第8期1473-1476,共4页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(30470470)

摘　　要：为了获取冠状动脉堵塞病人的心音特征,从而为临床预诊断冠状动脉疾病提供新的无损检测方法,通过非线性时间序列分析的方法———相关维来分析心音时间序列,获得了有关冠状动脉堵塞的血液动力学特征信息.对相关病例进行分析,发现通过相关维数刻划出了大多数冠状动脉堵塞病人的心舒期心音序列内在的确定性行为;而大多数正常人的心舒期心音序列得不到具体的相关维数.利用本文提供的相关维数分析方法,在曲线D2(M;r)-ln(r)上寻找平台期,可以达到对冠状动脉疾病无损检测的目的.To acquire the feature of heart sounds of patients with coronary artery disease （CAD）, nonlinear dynamics methods, such as correlation dimension, were employed to describe heart sounds of the diastolic for noninvasive detection of CAD. The analysis showed that the patient group has complex deterministic behavior characterized by correlation dimension. However, the majority of the normal subjects could not be scaled by correlation dimension. By the method of correlation dimension analysis, flat period of D2 （M;r）-ln （r） curve can be found for noninvasive detection of CAD.

关 键 词：心舒期 相空间重构 奇异值分解 相关维 

分 类 号：R318.04[医药卫生—生物医学工程]