Analysis of heart period variability is a dynamic noninvasive techniqu
e to quantify the autonomic control over the heart period. We recorded
electroencephalographic, electro-oculographic, electromyographic and
electrocardiographic data from 10 normal subjects during sleep using a
n ambulatory polysomnographic monitor. R-R intervals were determined f
or 10 min segments of electrocardiographic data from wakefulness, stag
e 2 sleep, slow wave Sleep and REM sleep. Average heart period, instan
taneous changes greater than 50 msec and fractal dimension were calcul
ated and the time domain and phase plots were depicted. The R-R interv
al time domain plots were subsequently analyzed using the discrete Fou
rier transform. We found sleep stage specific, time domain and frequen
cy domain changes in heart period variability, particularly using spec
tral analysis of heart period. Increased power in the 0.2-0.4 Hz band
was associated with stage 2 sleep when compared to awake and slow wave
sleep states. Power in the 0.0-0.04 and 0.04-0.12 Hz bands was increa
sed in association with REM sleep when compared to non-REM sleep, and
slow wave sleep had diminished power in al! frequency bands. Our resul
ts support other investigations demonstrating stage 2 sleep is associa
ted with increased parasympathetic influences and REM sleep is associa
ted with increased sympathetic and neurohumoral influences. We feel th
at spectral analysis of heart period variability is an effective nonin
vasive method to quantify changes in the autonomic influences over the
heart during sleep.