Document Type : Full Research Paper


1 Biomedical Engineering Faculty, Amirkabir University of Technology

2 , Electrical and Computer Engineering Faculty, Tehran University



Ventricular Fibrillation (VF) is a dangerous abnormality in the heart activity. During the VF, well known shape of electrocardiogram (ECG) signal changes to a pseudo-noise waveform. Recent researches have depicted that VF is not a noisy signal. The characteristics of VF and chaotic signals are the same. In this research, these characteristics were studied and used for discriminating the VF signal from the other electrocardiogram signals. Three types of electrocardiogram signals including VF, Tachycardia and Normal ECG were used for training and testing a back propagation neural network. We used these signals in three stages. At the first stage, the power spectrum of signals was used for training and testing the neural network. Time Series signals were used in the second stage. The result of the first experience was better than the second. At the third stage, we used surrogate technique to enrich the training signals in the time domain. The surrogate technique is a method which has been used in the chaotic systems. By using these new generated signals for training the neural network, the results of classification were extremely improved. Furthermore, the results of simulations showed that the chaotic dynamic of VF signal is a time dependant one.


Main Subjects

[1]     Small M: Temporal evolution of nonlinear dynamics in ventricular arrhythmia; International Journal of Bifurcation and Chaos 2001; 11: 2531–2548.
[2]     Lombardi F: Chaos theory, heart rate variability, and arrhythmic mortality; American Heart Association- Circulation 2000; 101: 8–10.
[3]     Guyton A; Textbook of Medical Physiology: 8th Edition, Saunders; 1991.
[4]     Timo L; Correlation properties and complexity of preoperative RR interval dynamics in coronary artery bypass surgery patients; Anaesthesiology 2000; 93: 69–80.
[5]     Korpelainen J: Dynamic behavior of heart rate in ischemic stroke; American Heart Association-Stroke 1999; 30: 1008–1013.
[6]     Small M: Uncovering nonlinear structure in human ECG recordings; Chaos Solitons & Fractals 2002; 13: 1755-1762.
[7]     Haefner J; Modelling Biological Systems: Principles and Applications; Chapman and Hall; 1996.
[8]     Small M; Nonlinear analysis of human ECG during sinus rhythm and arrhythmia; Computers in Cardiology 2000; 27: 147 150.
[9]     Minami K; Real time discrimination of ventricular tachyarrhythmia with Fourier transform neural network; IEEE Trans on Biomedical Engineering 1999; 46: 179–185.
[10] Eberhart R; Chaos theory for the biomedical engineer; IEEE Eng in Med and Biology 1998; 8:41–45.
[11] Small M; Testing time series for nonlinearity; Statistic and Computing 2001; 11:257–268.
[12] Povinelli R; Are nonlinear ventricular arrhythmia characteristics lost as signal duration decreases?; Proc Computers in Cardiology 2002; 29: 221-224.
[13] Parker T; Chaos: a tutorial for engineers; Proc of IEEE 1987; 75(8): 982–1008.
[14] Small M; Surrogate test for pseudo periodic time series data; Physics Review Letters 2001; 87(18): 1–4.