Journal of Computational Physics
| Hi-index | 0.00 |
The scaling (fractal) characteristics of electrocardiograms (ECG) provide information complementary to traditional linear measurements (heart rate, repolarisation rate etc.) allowing them to discriminate signal changes induced pathologically or pharmacologically. Under such interventions scaling behaviour is described by multiple local scaling exponents and the signal is termed multifractal. Exercise testing is used extensively to quantify and monitor cardiorespiratory health, yet to our knowledge there has been no previous multifractal investigation of exercise-induced changes in heart rate dynamics. Ambulatory ECGs were acquired from eight healthy participants. Linear descriptive statistics and a parameterisation of multifractal singularity spectra were determined for inter-beat (RR) and intra-beat (QT) time-series before and after exercise. Multivariate analyses of both linear and multifractal measures discriminated between pre- and post-exercise periods and proportionally more significant correlations were observed between linear than between multifractal measures. Variance was more uniformly distributed over the first three principal components for multifractal measures and the two classes of measures were uncorrelated. Order and phase randomisation of the time-series indicated that both sample distribution and correlation properties contribute to multifractalilty. This exploratory study indicates the possibility of using physical exercise in conjunction with multifractal methodology as an adjunctive description of autonomically mediated modulation of heart rate.