Abstract
The power of the high-frequency (HF) region in the power spectral density function of heart rate variability has been recognized as an indicator of parasympathetic nervous system activity and is expected to be applicable as an evaluation parameter not only for the autonomic nervous system but also for mental and physical workload. To quantify HF power in accordance with the power spectral density (PSD) estimated by Fast Fourier Transform (FFT) or autoregressive model, several kinds of different methods have been suggested. These methods, for example, calculation of the PSD area from 0.15Hz to 0.5Hz, calculation of the area of the specific band width around the HF component and that utilizing the peak value of the HF component, do not always yield the same result in the tendency of HF power. To evaluate the performance of these methods quantitatively, simulation data were induced and the “error rate” of each method was calculated and compared. The error rate showed that the method by which PSD was estimated by FFT and the HF power calculated by the area of specific band width around the HF component was robust against noise and was effective in measuring HF power in the spontaneous breathing condition.
