心律失常的多重分形去趋势波动分析  

Multifractal detrended fluctuation analysis for the pathological electrocardiograph signals

作　　者：赵烨南[1] 王俊[2] 

机构地区：[1]南京邮电大学通信与信息工程学院图像处理与图像通信江苏省重点实验室,南京210003 [2]南京邮电大学地理与生物信息学院图像处理与图像通信江苏省重点实验室,南京210003

出　　处：《北京生物医学工程》2012年第5期474-477,共4页Beijing Biomedical Engineering

基　　金：江苏省自然科学基金(BK2011759)资助

摘　　要：目的心律失常的及早诊断对及时救助病人具有重要意义。方法基于多重分形去趋势波动分析方法,本文对正常心电信号及窦、房性心律失常信号进行分析。结果发现三种信号都具有不同程度的长程相关性和多重分形特性,且在波动函数的阶数为正值时,三种信号的长程相关特性最为明显。通过比较三种信号的多重分形谱,发现正常心电信号的多重分形谱宽度最小,窦性心动过缓信号次之,心房颤动宽度最大。结论此研究结果对临床医学诊断区分心律失常与正常心电信号有很好的借鉴意义。Objective The early diagnosis and timely salvation for patients with arrhythmia is of great significance. Methods In this paper, multifractal detrended fluctuation analysis method was used to analyze the multifractal characteristics of the sinus bradycardia（SBR）signals, the atrial fibrillation（AF） signals and the normal electrocardiograph （ECG） signals. Results We found that the three kinds of signals were long-range correlating and multifractal. When the order of fluctuation function was positive, the three kinds of signals showed distinct long- range correlation properties. To compare with the muhifractal spectra of the three signals, the width of the normal ECG signals＇ spectrum was the smallest, the SBR signals＇ was bigger and the width of the AF signals＇ was the biggest. Conclusions This study provided good reference for clinical diagnosis of SBR and AF signals.

关 键 词：心律失常信号 标度不变性 多重分形 去趋势波动分析 

分 类 号：R318.04[医药卫生—生物医学工程]