An automated method for coding sleep states in human infants based on respiratory rate variability

Area of Science:

• Pediatric Sleep Medicine
• Quantitative Physiology
• Infant Neurodevelopment

Background:

• Accurate infant sleep state coding is crucial for developmental research.
• Current methods, like visual inspection and polysomnography, can be time-consuming or subjective.
• A need exists for a reliable, objective, and efficient method for infant sleep state classification.

Purpose of the Study:

• To develop and validate a novel quantitative method for coding infant active and quiet sleep epochs.
• To utilize respiration patterns as the primary data source for automated sleep state identification.
• To establish the concordance of this new method with established sleep coding techniques.

Main Methods:

• The method quantifies sleep epochs based on the variance of instantaneous breathing rate.
• Breathing rate variances are normalized within subjects using the 75th percentile.
• A threshold of 0.29 normalized variance was determined to differentiate active and quiet sleep.

Main Results:

• The novel method achieved high concordance with visual respiratory variability inspection (100% for quiet sleep, 90% for active sleep).
• Independent validation showed strong agreement with polysomnography (87% quiet, 80% active) and behavioral coding (92% quiet, 78% active).
• Identified sleep states correlated with expected differences in infant heart rate variability and electrocortical activity.

Conclusions:

• This novel quantitative method provides a reliable and objective approach for coding infant sleep states using respiration.
• The method demonstrates strong validity when compared to standard polysomnography and behavioral coding.
• This technique has the potential to enhance the efficiency and accuracy of infant sleep research.