S03 Implementing auto-videosomnography in pediatric sleep research
Utilization of objective assessment methods has been a long tradition in pediatric sleep research. While parent-reports of their children’s sleep provide important information, they are also subjected to social desirability bias and substantial imprecision. Assessment tools such as actigraphy, polysomnography and videosomnography enable objective measurement of child sleep based on movement or EEG monitoring. However, these methods require attachment of expensive wearable apparatuses to the infant’s body, and the latter two methods require laborious scoring procedures, thus limiting the ability to collect data from large samples. The recent advent of auto-videosomnography technologies addresses these limitations, allowing for naturalistic non-intrusive algorithm-based assessment of sleep-wake patterns, as well as parental nighttime behaviors. These computer-vision algorithms automatically yield nightly sleep metrics derived from camera-detected movement in the crib.
All symposium speakers will demonstrate recent implementations of this technology in pediatric sleep research.
- Dr Tikotzky will present results from a validation study of 50 infants which focuses on the validation of the Nanit algorithms that discriminate between sleep-wake states in infants. The Nanit data were validated against actigraphy, using the Sadeh algorithm, and manual video visual scoring.
- Dr Kahn will present auto-videosomnography data of 595 infants, comparing sleep metrics of infants whose mothers were in home confinement throughout the COVID-19 pandemic to those whose mothers were working as usual.
- Dr Ordway will present findings from a study of over 1500 toddlers 12-30 months of age that examined the associations among parenting behaviors, including parent nighttime interactions scored using videosomnography, and toddler sleep health. This presentation will highlight the use of videosomnogrpahy to examine sleep health as a multidimensional construct and include detailed examination of night-to-night sleep patterns. Findings from this study have important implications for the development of tailored sleep training programs.
- Prof Gradisar will present findings from a sample of 1,074 infants, showing that age moderates the link between screen exposure and sleep, particularly for touchscreen devices.
- Dr Berger will demonstrate possible underlying causal mechanisms between the well-established link between locomotor skill onset and co-occurring disruptions to infant sleep using Nanit-derived nightly heat maps of infant movement in the crib together with over 100 hours of frame-by-frame video coding for skill-relevant movements and posture.
Upon completion of this CME activity, participants should be able to:
- Recognize auto-videosomnography as a validated objective measure of sleep in infancy and toddlerhood
- Compare the benefits of auto-videosomnography compared to actigraphy or manual video scoring
- Recognize the opportunities for the sleep research community of auto-videosomnography including the large-scale database that is potentially available
The pediatric sleep research community
Thomas Anders (United States)
Automatic scoring of infant sleep-wake states with Nanit - A validation study
Liat Tikotzky (Israel)
Infant sleep during COVID-19: Longitudinal analysis of infants of US mothers in home confinement vs working as usual
Michal Kahn (Australia)
Associations between multidimensional sleep health variables and parent nighttime interactions and perceptions
Monica Roosa Ordway (United States)
Sleep and screen exposure across the beginning of life: Does age matter?
Michael Gradisar (Australia)
Locomotor milestone acquisition impacts movement and posture during infant sleep
Sarah E. Berger (United States)