Because the state of alertness exerts a profound influence on autonomic cardiac control, we hypothesized, that the heart rate response to a vagal stimulus, i.e., ocular compression, may differ during different states of alertness. We studied 8 healthy infants with a postconceptional age of 35-41 weeks (mean ± SD 37.9 ± 2.1 weeks). They underwent a standardized ocular compression test during polygraphically controlled wakeful-ness, rapid eye movement (REM) sleep, and non-REM (NREM) sleep. The R-R intervals were measured (1) during the 60 s preceding the ocular compression test, to determine the mean control R-R interval, and (2) during compression. Percent R-R interval was defined as the longest R-R interval in milliseconds during the test divided by the mean control R-R interval and multiplied by 100. The longest R-R interval during the test was significantly greater in REM sleep than in wakefulness (p < 0.05) and in NREM sleep (p < 0.01): 939 ± 360, 623 ± 355, and 538 ± 60 ms, respectively. The percent R-R interval was significantly greater in REM sleep than in NREM sleep (p < 0.01): 236 ± 91 and 129 ± 16, respectively. The time from the longest R-R interval to return to mean control R-R interval, i.e., vagal escape, was significantly shorter in REM sleep than in NREM sleep and in wakefulness (p < 0.01): 843 ± 168,2,131 ± 712, and 2,078 ± 913 ms, respectively. This study indicates that the state of alertness should be defined when performing tests on autonomic reflexes in infants.