Structural analyses of 1/f fluctuations in heart rate variability based on their spectrograms

In this paper, time-frequency properties of 1/f fluctuations of human heart rate variability (HRV) are analyzed. For this purpose, whole day recordings for 15 normal subjects are done by using a digital holter recorder under free moving conditions. A spectrogram of HRV is computed through sleep and wakefulness, which is reduced into powers in three frequency band LF (<0.05 Hz), MF (0.05-0.15 Hz), and HF (0.15-0.45 Hz). Because these spectral powers are expected to represent hormonal and autonomic regulatory activities, distributions of balance among these spectral powers are obtained for sleep and wakefulness as an indicator of the regulatory activities on HRV. The balance distributions during wakefulness commonly show that MF and LF powers are dominant comparing with HF. During sleep, the balance distributions are extended from those similar to the wakefulness to region where HF power dominates LF and MF powers, which are shown to correspond to a rapid eye movement sleep (REM) and non-REM sleep, respectively. Analyses on overall transition tendency of the balance show distinct patterns for sleep and wakefulness: positive correlation between variations of LF power and HF/MF (balance between spectral powers in HF and MF) during wakefulness and negative during sleep. In addition, local transition tendency suggests that there are attractive points of balance which possibly correspond to the maintained physiological states such as sleep stages and postures. Through this study, the 1/f fluctuation of HRV is suggested to be constructed by the transition dynamics of the balance among the hormonal and autonomic regulatory activities which is wandering among the attractive points.