Optimal frequency ranges for extracting information on autonomic activity from the heart rate spectrogram

Heart rate variability spectrum analysis provides useful quantitative indices of neural control of the SA node. This method is attractive both for its simplicity and the lack of invasive instrumentation, particularly for human investigation. The differing spectral characteristics of parasympathetic and sympathetic control of heart rate allows separate measurement. However, there are widely varying opinions as to the appropriate frequency bands to represent these two inputs. We compared the heart rate variability spectra of 16 humans in supine and upright positions. Adequate measures of parasympathetic or sympathetic activity change should correlate respectively inversely or directly with heart rate change. Frequently used spectral measures of sympathetic activation did not correlate with heart rate changes. With optimization of frequency bands, we found that restricting the sympathetic band to frequencies below 0.1 Hz and above 0.05 Hz (0.055 to either 0.086-0.098 Hz), and dividing by total spectral amplitude 0.004-0.5 Hz (to account for parasympathetic fluctuations within the sympathetic band) produced the best results. The parasympathetic band was best from 0.1 Hz to a frequency greater than that of the respiratory sinus arrhythmia. The optimization method detailed here is easily applied to circumstances other than active orthostasis, and should provide a means of empirically determining useful frequency limits.