Quantification of Respiratory Influence in Heart Rate Variability Through

Respiratory Sinus Arrhythmia (RSA) induces cyclic variations, synchronous with respiration, in the RR interval series. Quantification of RSA is important for the correct separation between sympathetic and parasympathetic components of HRV and for the analysis of Autonomic Nervous System (ANS) activity. In this paper we propose a Wavelet packets subband decomposition algorithm for the extraction of respiratory-related component of RR variability. Wavelet Packets are used to filter the RR interval and respiratory series and to generate a set of orthonormal sub-band signals. An Autoregressive with eXogenous input (ARX) model is used to describe the relationships between the two signals in the different sub-bands. Starting from the parameters of the different sub-band models it is possible reconstruct the respiratory related component in full-band. A few preliminary results will be presented to discuss the performance of the proposed approach and to show that the new approach is able to provide a better reconstruction of the RSA than tradition full-band methods.