Analysis of ECG signal by chaos principle to help automatic diagnosis of myocardial infarction

Introduction Cardiovascular disease is a major cause of premature dying. If any patient is diagnosed at early stage of cardiovascular disease [(myocardial infarction (MI)], it may be possible to reverse disease process. Functional behavior of heart is represented by electrocardiogram (ECG), signals of which are quasiperiodic (Fig. 1). Various transform techniques (Fourier, wavelet etc.) are ineffective to distinguish MI signals from other class and require improvement of accuracy in differentiating MI signals from rest of ECG. Studies are available on nonlinear dynamics to describe deviation from periodicity of ECG. Phase space fractal dimension (PSFD) highly varies for two different class patterns of ECG signals. Therefore, a systematic experiment on finding a good characteristic feature based on chaos is needed for identifying quasi periodicity. This study investigates embedded diagnostic information concerning MI disease within chaos profile of a quasi-periodic ECG by extracting PSFD of ECG. A discriminant function based technique has been proposed for better identification of MI.