Implementation and evaluation of an adaptive method for reduce the respiration influence on Heart Rate Variability

Abstract

In this paper it is described the implementation and evaluation of an adaptive method that has as aim to cancel the influence of the respiratory signal over the Heart Rate Variability (HRV) signal in order to enhance the power estimation of its spectral components. The method consists in an Adaptive Noise Cancellation (ANC) structure that uses a Finite Impulse Response (FIR) filter together with the Normalized Least Mean Squares (NLMS) adaptation algorithm. Respiration and electrocardiogram (ECG) signals were obtained simultaneously using 240Hz sampling frequency. After data acquisition, tachogram was derived from ECG signal to obtain its HRV signal; then ANC filtering is applied, reducing variations due to respiration from HRV signal. This method was evaluated for spontaneous and for two controlled respiration frequencies. 6-minutes registers were taken form 10 people during the 3 different scenarios giving a total of 30 registers. Power Spectral Density (PSD) was estimated from the HRV signal before and after filtering and compared. At the results, it is observed that frequency components related to respiration are cancelled in the HRVs PSD, reaching an improved estimation of the control exerted by the Autonomic Nervous System (ANS) over the heart rate.