Wavelet analysis performed almost identically to Fourier analysis for computing heart period power values in both stationary and nonstationary segments, with differences of <1%.
Observational (n=40)
Does Wavelet analysis provide significantly different heart period power values compared to Fourier analysis in stationary and nonstationary data?
Fourier and Wavelet methods perform almost identically for computing heart period power values, indicating Wavelet is only superior when time-frequency domain analyses are specifically required.
The aim of this study was to assess the error made by violating the assumption of stationarity when using Fourier analysis for spectral decomposition of heart period power. A comparison was made between using Fourier and Wavelet analysis (the latter being a relatively new method without the assumption of stationarity). Both methods were compared separately for stationary and nonstationary segments. An ambulatory device was used to measure the heart period data of 40 young and healthy participants during a psychological stress task and during periods of rest. Surprisingly small differences (<1%) were found between the results of both methods, with differences being slightly larger for the nonstationary segments. It is concluded that both methods perform almost identically for computation of heart period power values. Thus, the Wavelet method is only superior for analyzing heart period data when additional analyses in the time-frequency domain are required.
Houtveen et al. (Sat,) conducted a observational in Healthy (n=40). Wavelet analysis vs. Fourier analysis was evaluated on Spectral decomposition of heart period power. Wavelet analysis performed almost identically to Fourier analysis for computing heart period power values in both stationary and nonstationary segments, with differences of <1%.