Spectral analysis of heart rate variability revisited: comparison of the methods

Since the early 70's, interest in spectral analysis of heart rate variability (HRV) has grown considerably. The clinical significance of the HRV spectrum has been made apparent by studies which have shown it to provide predictors of mortality after myocardial infarction. Unfortunately, there are difficulties in quantitatively interpreting and comparing the results of the many published HRV studies owing to the different methods which have been used in arriving at the HRV spectrum. In particular, most of the methods introduce spectral bias. This study reports a systematic comparison of how effective each of the published methods for estimating the HRV spectrum are in faithfully retrieving the autonomic activity converging on the heart. Moreover, the use of spontaneous HRV activity to estimate the sympathovagal balance is challenged and the 'respiratory-evoked HRV spectrum' is put forward as a more reliable method of obtaining this parameter. >