Assessment of causality between breathing and heart activity signals

In physiology the analyses are often based on more than one signal, however in many cases simple linear methods do not allow solving more complicated issues. One of the relatively rarely used method for analysis of relationships between physiological signals, commonly applied in econometrics, is Granger causality test. It is not symmetrical feature in comparison with cross-correlation and it describes that as one has two time series X and Y, X one is said to be Granger cause on Y, when Y could be better predicted using both X and Y than only with Y. Among physiological signals breathing and heart activity function are strongly related each other. However this correspondence was not considered or analyzed mainly using simple mathematical methods. Therefore, the purpose of this work is to assess the causality between respiratory signal and cardiac activity. The motivation of the study is to explore the method of analyzing and predicting two times series in physiological studies, taking into account the causal link between them. In the study 13 healthy students were performed deep breathing at 10 breaths per minute rate, in sitting body posture. Impedance pneumography signal was stored in order to registered volumes of respiration (in terms of shape). R-waves amplitudes from electrocardiography signal after removing baseline (AmpR), and heart rate variability curve (HRV), were calculated. Due to the fact that the signals were quasi-sinusoidal, we combined both linear and nonlinear analysis. For all subjects the causality between breathing and Ramp or HRV was statistically confirmed. AmpR was delayed to the respiratory volume signal by 1.07 +/- 1.03 seconds, and HRV curve was ahead by 0.72 +/- 0.33 seconds.

Keywords: heart rate variability, breathing, impedance pneumography, and Granger causality