The effect of epoch length and smoothing on infant sleep and waking state architecture for term infants at 42 to 46 weeks postconceptional age.

Study Objectives: Epoch lengths from 20 seconds to 1 minute, and smoothing strategies from zero to three minutes are encountered in the infant sleep a nd waking literature. The present study systematically examined the impact of various epoch lengths and smoothing strategies on infant sleep state arc hitecture. Design: Overnight polysomnographic recordings were visually assessed by epo ch as wake or as each of four sleep state parameters: electroencephalograph ic patterns, respiration, body movement, and eye movement. From these findi ngs, steep and waking states were assigned for each of six combinations of epoch length (30-second or 1-minute) and smoothing window length (none, 3-e poch, or 5-epoch). Setting: N/A Participants: Subjects were 91 term infants, 42-46 weeks postconceptional a ge, from the Collaborative Home Infant Monitoring Evaluation (CHIME) study. Interventions: N/A Measurements and Results: A greater epoch length resulted in more active an d less quiet sleep as a percentage of total study; however, the size of the smoothing window did not affect the percentage of sleep/waking states. In general, the greater the epoch length and the greater the smoothing window length, the fewer the number of, the greater the mean duration of, and the greater the longest continuous episode of sleep/waking states. Analysis of significant interactions indicated that a 1-minute epoch length relative to a 30-second epoch length resulted in increasingly longer episodes of quiet and especially active sleep with a greater smoothing window length. Conclusions: Smoothing strategy significantly altered sleep state architect ure in infants and may explain part of the variability in infant sleep stat e findings between laboratories.