Effect of using ECG derived respiration (EDR) signal in linear parametric QT-RR modeling

Abstract

Linear parametric modeling techniques are widely used for modeling the short term QT-RR interaction to explore the factors (i.e. heart rate variability, Autonomic Nervous system) controlling ventricular repolarisation variability. Recent studies established that respiration also has important effect on the ventricular repolarisation process like it has on the heart rate variability. So for the clear understanding of cardiac regulations, respiration signal should be considered for modeling the QT-RR dynamics. Due to several problems in collecting original respiration signal using the traditional recording devices that measure the nasal air flow or abdominal or chest pressure, a lot of research has been done to extract respiration information from the ECG signal called the ECG derived respiration (EDR). In this study we verify the use of EDR signal as a surrogate of original respiratory signal in modeling QT-RR interaction. We collect 10 young subjects' ECG and respiration signal from Fantasia database. We developed linear parametric autoregressive model with multiple exogenous inputs with an autoregressive noise term and check the model performance by using original respiration and EDR signal and found statistically similar result. Our findings showed that EDR can be used as a surrogate of original respiration in QT-RR model for the better understanding of cardiac regulations in young healthy subjects.