Function mapped trajectory estimation for ECG sets
The main purpose of this paper is to present two significant contributions in characterizing electrocardiogram (ECG) signal sets. The first contribution of the paper is the introduction of a functional map representation of atrial fibrillation ECG signal sets using chaotic maps. Second contribution of this paper is to demonstrate the effectiveness of the recurrence period density entropy (RPDE) index as an effective tool in tracking changes in rhythmic behavior of ECG signals.
For the first contribution, the ECG time series is tested for chaotic nature using Lyapunov exponents. If chaotic pattern is found, phase embedded plots are extracted to compare the plots with ones of chaotic mapping functions. Kernel density estimator is used for calculating the probability density function for markovian trajectory estimation model. For the second contribution recurrence period indexes and plots are used to track changes in rhythmic behavior.
The effectiveness of the proposed trajectory estimation model in estimating future probabilities of ECG signal sets, is tested using mean-squared error metric. RPDE indexes and plots provide good results in tracking the onset of changes in the condition of the patient form the state of normal ECG rhythm to the state of atrial fibrillation and sudden cardiac arrest.
Hence by using chaotic signals to functionally represent ECG signal sets, we attempt to reduce the ambiguity associated with the trajectory estimation and with RPDE indexes/plots, the transitions in rhythmic behavior become more prominent for analysis.
KeywordsElectrocardiogram Chaotic map Trajectory estimation Recurrence period density entropy
Unable to display preview. Download preview PDF.
- Wang K, Zhao Y, Sun X, Weng T. A simple way of distinguishing chaotic characteristics in ECG signals. Conf Proc Int Conf Biomed Eng Inf. 2010; 2:713–6.Google Scholar
- Jovic A, Bogunovic N. Feature extraction for ECG time-series mining based on chaos theory. Conf Proc Int Conf Inf Tech. 2007; 63–8.Google Scholar
- Casaleggio A, Braiotta S, Corana A. Study of the Lyapunov exponents of ECG signals from MIT-BIH database. Conf Proc Comput Cardiol. 1995; 697–700.Google Scholar
- Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, Mietus JE, Moody GB, Peng C-K, Stanley HE. PhysioBank, PhysioToolkit, and PhysioNet: The MIT-BIH Normal Sinus Rhythm Database. 2000. http://www.physionet.org/physiobank/database/afdb. Accessed 15 Feb 2014.Google Scholar
- Goldberger AL, Amaral LAN, Glass L, Hausdorff JM, Ivanov PCh, Mark RG, Mietus JE, Moody GB, Peng C-K, Stanley HE. PhysioBank, PhysioToolkit, and PhysioNet: The MIT-BIH Normal Sinus Rhythm Database. 2000. http://www.physionet.org/physiobank/database/nsrdb. Accessed 10 Feb 2014.Google Scholar
- Chetan M, Miller M. Chaos analysis for EKG time series data. Dartmouth College, Department of Mathematics. 2007; 1:1–9.Google Scholar