Volume 29, Issue 5, 2020
DOI: 10.24205/03276716.2020.1111
Heart rate variability assessment during sleep derived from ultrashort-term and short-term windows
Abstract
Objective: Heart rate variability (HRV) is a useful measure to quantify autonomic nervous system (ANS) function. A short-term window (5 min) for HRV analysis has been adopted in many sleep studies to differentiate between rapid eye movement (REM) and non-REM (NREM) sleep. In this study, an ultrashort-term window (2 min) was proposed to overcome the problem that the short-term window cannot investigate instantaneous ANS changes during REM or NREM.
Methods: We obtained a 10-minute HRV dataset from polysomnographic data for 21 individual patients in both REM and NREM (N3 stage). This target dataset was analyzed by shifting the ultrashort-term and short-term windows forward by 2 s to create multiple power spectral densities (n=250 and n=140, respectively) with the short time Fourier transform (STFT). Three main frequency bands were investigated: very low frequency (Ln VLF), low frequency (ln LF), and high frequency (Ln HF).
Results: The standard deviation (SD) of spectral profiles obtained by ultrashort-term windows was found to be a new potential indicator to differentiate REM and NREM (p<0.0001). How many times or whether the Ln LF/Ln HF ratio and Ln HF crossed the designated threshold line could be used to detect rapid ANS changes during both REM and NREM. Ln VLF and Ln LF were higher in REM than NREM, whereas Ln HF was lower.
Conclusion: The results suggested that an ultrashort-term window based on the STFT with a time resolution of 2 s would be more useful for tracking rapid changes in ANS activity than a short-term window.
Keywords
heart rate variability; rapid eye movement; power spectrum; autonomic nervous system; sleep