Lomb-scargle periodogram applied to heart rate variability study

Frequency domain analysis techniques have been used in the heart rate variability (HRV) study, and its interpretation has been associated with both clinical investigation and diagnosis. Fast Fourier Transform (FFT) and Autoregressive (AR) based periodograms have been the most commonly used approaches. This work compares these two methods with the Lomb-Scargle (LS) periodogram, which is intended to deal with unevenly sampled data, the HRV. A simulated data with known frequency band amplitude was generated to compare these three methods. Once the HRV signal is an unevenly sampled signal, the LS periodogram seemed to be an interesting choice, presenting less noise estimates in comparison with the other methods. However, additional investigations applying this technique and comparing it with the other approaches in real data are still necessary.