Related Experiment Videos
Nonlinear time series analysis of electrocardiograms
A. Bezerianos1, T. Bountis, G. Papaioannou
1University of Patras, 261 10 Patras, Greece.
Chaos (Woodbury, N.Y.)
|March 1, 1995
Summary
Nonlinear dynamics analysis of electrocardiograms (ECGs) reveals correlation dimension estimates can be unreliable. Further tests show no major dynamical differences between normal and heavy smoker subjects, suggesting increased cardiac complexity during mild exercise.
Area of Science:
- Nonlinear Dynamics
- Physiology
- Biomedical Engineering
Background:
- Increasing application of nonlinear dynamics to electrocardiogram (ECG) time series.
- Previous studies suggest ECG dynamics are deterministic and chaotic, with dimension distinguishing health status.
Purpose of the Study:
- Critically evaluate correlation dimension estimates for ECG analysis.
- Investigate dynamical differences in ECGs between normal subjects and heavy smokers.
- Assess changes in cardiac dynamics during mild exercise.
Main Methods:
- Correlation dimension calculations.
- Time differencing, smoothing, principal component analysis, surrogate data analysis.
- Analysis of ECGs from normal and heavy smoker subjects at rest and during exercise.
Main Results:
- Correlation dimension calculations require careful application for reliable "dimension" estimates.
- No significant dynamical differences observed between normal and heavy smoker ECGs.
- Estimated dynamical variables increase from 3-4 to 5-6 after removing temporal correlations.
- Mild exercise generally increases cardiac dynamics complexity.
Conclusions:
- Standard correlation dimension methods may oversimplify ECG dynamics.
- Dynamical differences between normal and heavy smoker subjects at rest are not pronounced.
- Cardiac dynamics exhibit increased complexity with mild physical exertion.