Effect of window length on the analysis of cardiorespiratory synchronization

The cardiac and respiratory rhythms interact with each other and may exhibit a synchronous behaviour in the human body. The mutually coupled relationship is investigated by analyzing the biological electrocardiogram (ECG) and respiration signals. In general, these two types of signals are non-stationary, irregular and somewhat chaotic. Consequently, it is hard to use the widely applied Fourier transform or any other stationary processes to analyze the data. Respiratory and cardiac rhythms alike are periodical in themselves. Signals of these two physiological systems exhibit similar patterns and the change in individual rhythm may affect on one another. Comparing the phase variations of the selected signatures between the two signals periods of cardiorespiratory synchronization can be identified. The continuous `R´ peaks of heartbeat are picked out and the phase difference of two adjacent peaks is set to 2π. The starting points of respiratory signal are used as the zero-phase points and the phases within two continuous signals are linearly interpolated in the range of 2π. This derived phase information is related to the instantaneous radian frequency for each one. By inspecting the existence of R peaks and different respiratory cycles, the cardio-respiratory synchrogram is presented. Gamma evaluation method (γ evaluation) is used to quantify the strength of cardiorespiratory synchronization. This method processes a selected window length of data and calculates the degree of interaction. The analysis of the synchronization between cardiac and respiratory rhythms is presented and the effects of using different window lengths on γ, n, m are discussed.