EXPLORE PUBLICATIONS BY COUNTRIES

HRV, Interpolation, Re-sampling, Distortion, Phase-shift

The heart rate variability (HRV), refers to the beat-to-beat alterations in heart rate, is analyzed using RR-interval (RRI) series derived from the ECG signal as an interval between successive QRS complexes. For deciphering the true HRV spectrum using FFT, the RRI series should be resampled. But re-sampling often induces a noticeable distortion in the HRV power spectral estimates. Thus, the re-sampling operation should be accurate enough in reproducing the finest variation in the given signal. This paper compared three most widely used interpolation techniques: linear, cubicspline, and Berger’s, as re-sampling methods, in an attempt to propose an optimal method of interpolation for HRV analysis. The linear and cubicspline methods based PSD estimates, for artificially generated non-uniformly sampled RRI series, introduce linear phase shifting, and thus lower the HRV frequencies. On the contrary, Berger’s method efficiently reproduced the inherent frequencies in the underlying signal except some amplitude distortion. Further, similar trends in PSD estimates were obtained for real RRI series as well. Thus, it was concluded that at the expense of some increase in computational complexity, the spectral distortion has been significantly reduced using the Berger’s interpolation based re-sampling method as compared to the linear and cubicspline methods.