HEART-RATE CONTROL IN NORMAL AND ABORTED SIDS INFANTS

Approximate entropy (ApEn), a mathematical formula quantifying regularity in data, was applied to heart rate data from normal and aborted-sudden infant death syndrome (SIDS) infants. We distinguished quiet from rapid-eye-movement (REM) sleep via the following three criteria, refining the notion of REM as more ''variable'': 1) REM sleep has greater overall variability (0.0374 +/- 0.0138 vs. 0.0205 +/- 0.0090 s, P < 0.005); 2) REM sleep is less stationary (StatAv = 0.742 +/- 0.110) than quiet sleep (StatAv = 0.599 +/- 0.159, P < 0.03); 3) after normalization to overall variability, REM sleep is more regular (ApEn(sub) = 1.224 +/- 0.092) than quiet sleep (ApEn(sub) = 1.448 +/- 0.071, P < 0.0001). Fifty percent of aborted-SIDS infants showed greater ApEn instability across quiet sleep than any normal infant exhibited, suggesting that autonomic regulation of heart rate occasionally becomes abnormal in a high-risk subject. There was an association between low ApEn values and aborted-SIDS events; 5 of 14 aborted-SIDS infants had at least one quiet sleep epoch with an ApEn value below the minimum of 45 normal-infant ApEn values.