Identification of healthy and pathological heartbeat dynamics based on ECG-waveform using multifractal spectrum

Human body surface electrocardiogram (ECG) is non-stationary and frequency-varying by nature that belongs to a typical nonlinear signal. Therefore, traditional linear and time–frequency​ analysis methods cannot fully disclose its nonlinear nature. Meanwhile, physiological complexity of heartbeat signal may vary with age, diseases, drug administrations, or even behavioral modifiers. To test the intrinsic relationships among them, we first put forward a theoretical model for nonlinear time series analysis and then took the generally accepted multifractal sets to verify it. Upon that, we then investigated the multifractal singularity spectrum areas of synchronous 12-lead ECG signals taken from crowds with different age stages, healthy conditions and drug medications. Our results suggested that aging and diseases can not only decrease multifractal complexity of the signals, but also increase inhomogeneity of it. With aging and deepening lesion, fractal-like structure of the heartbeat system is damaged or even structurally changed, which lead to the declination of physiological complexity and at the same time the increasement of irregularity and anisotropy of ECG signal’s propagation. In addition, the mean value of multifractal spectrum area of human 12-lead ECG signals also reflect the self-discipline regulation of human autonomic nervous system. The value descends with age growing or drug intervention to restrain sympathetic nerves. That suggest self-discipline control function weakens when people are getting old, or they are under repressed heartbeat activities with lower heart rate and lower blood pressure. Then, complexity of heartbeat signal declines and even tends to turn from multifractality to monofractality, which means drops off of human individual adaptability.