Decomposition of heart rate variability by adaptive filtering for estimation of cardiac vagal tone

Heart rate fluctuations resulting from respiration and other influences upon the cardiovascular system are encoded into the patterns of heart rate variability (HRV). The fluctuations due to respiration are called respiratory sinus arrhythmia (RSA). Since RSA is primarily mediated through the autonomic nervous system (ANS), it is of interest to separate RSA from other influences to assess the underlying ANS function. On the other hand, the RSA may obscure heart rate responses to external manipulations in psychophysiological tests. A method of partitioning the HRV signal which can provide quantitative estimate of RSA as well as true heart rate responses without respiratory disturbances for psychophysiological studies is developed. The analysis of HRV signal is performed using an adaptive filtering system. With the simultaneously recorded respiration signal as a reference input, the HRV signal can be separated into two components, RSA and fluctuation due to other influences. After the separation, the variance of RSA, an estimate of cardiac vagal tone (ECVT), is readily obtained. The performance of the system was evaluated using artificial test signals as well as real HR V data. As a time domain approach, the method is simple, fast and robust.