A simple approach to determine loss of physiological complexity in heart rate series.

There are several ways to assess complexity, but no method has yet been developed for quantitatively calculating the "loss of fractal complexity" under pathological or physiological states. In this paper, we aimed to quantitatively evaluate fractal complexity loss using a novel approach and new variables developed from Detrended Fluctuation Analysis (DFA) log-log graphics. Three study groups were established to evaluate the new approach: one for normal sinus rhythm (NSR), one for congestive heart failure (CHF), and white noise signal (WNS). ECG recordings of the NSR and CHF groups were obtained from PhysioNET Database were used for analysis. For all groups Detrended Fluctuation Analysis scaling exponents (DFAα1, DFAα2) were determined. Scaling exponents were used to recreate the DFA log-log graph and lines. Then, the relative total logarithmic fluctuations for each sample were identified and new parameters were computed. To do this, we used a standard log-log plane to standardize the DFA log-log curves and calculated the differences between the standardized and expected areas. We quantified the total difference in standardized areas using parameters called dS1, dS2, and TdS. Our results showed that; compared to the NSR group, DFAα1 was lower in both CHF and WNS groups. However, DFAα2 was only reduced in the CHF group and not in the WNS group. Newly derived parameters: dS1, dS2, and TdS were significantly higher in the NSR group compared to CHF and WNS groups. The new parameters derived from the DFA log-log graphs are highly distinguishing for congestive heart failure and white noise signal. In addition, it may be concluded that a potential feature of our approach can be beneficial in classifying the severity of cardiac abnormalities.