A behavioral approach to annotating sleep in infants: Building on the classic framework

Area of Science:

• Developmental Psychology
• Neuroscience
• Sleep Medicine

Background:

• Polysomnography (PSG) is the gold standard for infant sleep monitoring but is obtrusive and impractical for home use.
• Behavioral observation offers a non-obtrusive, infant-friendly method for assessing sleep patterns.
• Accurate infant sleep assessment is crucial for understanding sleep problems and neuro-cognitive development.

Purpose of the Study:

• To describe and validate a novel behavior-based framework for annotating infant sleep states.
• To provide a practical, non-obtrusive alternative to polysomnography for infant sleep monitoring.
• To establish the reliability of a new framework using established sleep assessment methods.

Main Methods:

• Developed a behavior-based framework for infant sleep annotation, utilizing Prechtl's method and annotator consensus.
• Integrated multi-modal data, including infant movements, breaths, and sounds, within a contextual framework.
• Validated the framework by comparing its annotations to polysomnography (PSG) data from 14 infants during napping.

Main Results:

• The behavior-based framework demonstrated results comparable to PSG for scoring wake, active sleep, and quiet sleep.
• Achieved a Cohen's Kappa agreement of ≥0.74 when comparing framework annotations to PSG.
• The framework's accuracy suggests its potential as a viable alternative to PSG in certain settings.

Conclusions:

• The developed behavior-based framework offers a reliable and non-obtrusive method for infant sleep state annotation.
• Further validation with larger cohorts and clinical populations is recommended to enhance generalizability.
• This framework has the potential to facilitate more accessible and practical infant sleep monitoring.