Evaluation of a commercially available automated sleep staging tool in healthy infants : Evaluating an infant sleep staging tool

The ALICE5 software package provides a commercially available automated sleep staging system designed for infants. This study aims to evaluate the accuracy of this sleep staging system in healthy infants. Polysomnograms from 32 healthy infants were obtained from the CHIME dataset. Sleep staging was performed using the Pneumo mode of the ALICE5 package.

Each polysomnogram was then re-staged by an experienced pediatric sleep physician. The agreement rate and the kappa agreement rate between the manual sleep scoring and the automated sleep scoring were calculated. The mean agreement rate was 74.6% (SD = 10%), and the kappa agreement rate was k = 0.558 (SD = 0.17).

While the mean agreement rates are comparable with agreement rates between human scorers, the intra-subject variability in performance means that the ALICE5-generated automated sleep staging should be manually validated.