Evaluation of two detrending techniques for application in Heart Rate Variability

The performance of two different algorithms of detrending the RR-interval before Heart Rate Variability (HRV) analysis has been evaluated using both, simulated signals and real RR-interval time series. The first algorithm is based on the smoothness prior approach (SPA) and the second algorithm is implemented using wavelet packet (WP) analysis. The calculated time and frequency domain parameters obtained from real signals after detrending and the results obtained from simulated signals suggest that the WP method performed better than the SPA. The WP method provided more attenuation of the slow varying trend and was able to preserve the other signal components better than the SPA method. Also the SPA method was computationally slower and it might be not appropriate with long signals.