Time-Frequency analysis of heart rate variability during transient segments

The spectral characteristics of heart rate variability (HRV) are related to the modulation of the autonomic nervous system. As the physiological condition is changed by such external stimuli such as drugs, postural changes, and anesthesia, or by internal deregulation such as in syncope, adjective autonomic responses could alter HRV characteristics. Time-frequency analysis is commonly used to investigate the time-related HRV characteristics. An alteration of the autonomic regulation resulting in a change in mean heart rate induces a transient component in heart rate, which, with any analysis method based on signals from multiple beats, results in the apparent spread of the spectrum of frequencies. This obscures the spectral components related to the autonomic function. In this paper we investigated the influence of the transient component in several time-frequency methods including the short-time Fourier transform, the Choi-Williams distribution, the smoothed pseudo Wigner-Ville distribution (SPWVD), the filtering SPWVD compensation, and the discrete wavelet transform. One simulated signal and two heart rate signals during general anesthesia and postural change were used for this assessment. The result demonstrates that the filtering SPWVD compensation and the discrete wavelet transform have small spectrum interference from the transient component.