Phenotypization of Infant Sleep by Videosomnography

Introduction: During the first year of life, infants go through significant changes in their sleep patterns. From the irregular sleep-wake cycles of newborns, sleep gradually evolves to more predictable patterns. Establishing healthy sleep patterns in infants is crucial for their physical, cognitive, and emotional development. Parents and caregivers can help promote healthy sleep patterns, establishing consistent bedtime routines, creating a sleep-friendly environment, and responding to the infant's cues and needs. Understanding and accepting individual differences in sleep patterns can help parents adapt their caregiving strategies to meet the unique needs of their baby. In this study we propose to identify different sleep patterns in healthy infants, relying on objecti-ve sleep metrics. Materials and Methods: A total of 623 parents of infants aged 9 to 13 months (M=10.31 months ± 1.13, 52.00% females) were recruited among users of Nanit baby-monitor in the United States and were asked to comple-te the Brief Infant Sleep Questionnaire (BISQ-R). Objective infant sleep me-trics obtained from Nanit auto-videosomnography (1 week of data averaged) were: nighttime infant sleep duration, number of nighttime infant awake-nings, number of parental nighttime visits, nighttime sleep efficiency, bedti-me and wake up time. To group infants based on sleep variables, a cluster analysis was conducted using a series of hierarchical (Ward's method) and non-hierarchical (k-means) cluster analyses to determine the best represen-tative model and to test stability and replicability of clusters. The between-cluster comparison was performed using ANOVA for parametric and the χ2 test for nonparametric variables. Post hoc tests were performed. Results: Three reproduci-ble and stable sleep groups were identified: - Long Sleepers (LS, n.340), - Interrupted Sleepers (IS, n.126), - Short Sleepers (SS, n.156). All sleep metrics were signi-ficantly different in the three groups. LS had longer nighttime sleep duration than IS (0.65 ± 0.08 h, p<.001) and SS (1.71 ± 0.07 h, p<.001), also IS had lon-ger nighttime duration than SS (1.06 ± 0.09 h, p<.001), but presented more awakenings than SS (2.16 ± 0.10 h, p<.001) and LS (2.16 ± 0.09 h, p<.001), the latter did not differ in awakenings but in parent interventions, being more frequent in SS than in LS (0.65 ± 0.16, p<.001). Bedtime and wake up time were similar for LS and IS, whereas SS presented later bedtime and earlier wa-ke up time than LS (1.29 ± 0.06 h, p<.001 and -0.26 ± 0.07 h, p<.001) and IS (1.22 ± 0.07 h, p<.001 and -0.30 ± 0.08 h, p<.001). Nighttime sleep efficiency was better in LS than in IS (0.05 ± 0.01, p<.001) and SS (0.01 ± 0.01, p<.001), SS presented better sleep efficiency than IS (0.04 ± 0.01, p<.001). No age or gender difference was found between clusters. Conclusions: Cluster analysis based on objective sleep metrics offers a novel multidimensional approach to identify and under-stand infants' sleep patterns. The categorization of infant sleep pattern through a non-invasive objective method like videosom-nography might identify infants with sleep problems and might allow an early and specific inter-vention.