Evaluation of Automated Pediatric Sleep Stage Classification Using U-Sleep: a Convolutional Neural Network

Overview

Journal J Clin Sleep Med

Specialties Neurology
Psychiatry

Date 2024 Sep 26

PMID 39324691

Authors

Ajay Kevat

Rylan Steinkey

Sadasivam Suresh

Warren R Ruehland

Jasneek Chawla

Philip I Terrill

Andrew Collaro

Kartik Iyer

Affiliations

Soon will be listed here.

Abstract

Study Objectives: U-Sleep is a publicly available automated sleep stager, but has not been independently validated using pediatric data. We aimed to (1) test the hypothesis that U-Sleep performance is equivalent to trained humans, using a concordance dataset of 50 pediatric polysomnogram excerpts scored by multiple trained scorers, and (2) identify clinical and demographic characteristics that impact U-Sleep accuracy, using a clinical dataset of 3,114 polysomnograms from a tertiary center.

Methods: Agreement between U-Sleep and "gold" 30-second epoch sleep staging was determined across both datasets. Utilizing the concordance dataset, the hypothesis of equivalence between human scorers and U-Sleep was tested using a Wilcoxon 2 1-sided test. Multivariable regression and generalized additive modeling were used on the clinical dataset to estimate the effects of age, comorbidities, and polysomnographic findings on U-Sleep performance.

Results: The median (interquartile range) Cohen's kappa agreement of U-Sleep and individual trained humans relative to "gold" scoring for 5-stage sleep staging in the concordance dataset were similar, kappa = 0.79 (0.19) vs 0.78 (0.13), respectively, and satisfied statistical equivalence (2 1-sided test < .01). Median (interquartile range) kappa agreement between U-Sleep 2.0 and clinical sleep-staging was kappa = 0.69 (0.22). Modeling indicated lower performance for children < 2 years, those with medical comorbidities possibly altering sleep electroencephalography (kappa reduction = 0.07-0.15) and those with decreased sleep efficiency or sleep-disordered breathing (kappa reduction = 0.1).

Conclusions: While U-Sleep algorithms showed statistically equivalent performance to trained scorers, accuracy was lower in children < 2 years and those with sleep-disordered breathing or comorbidities affecting electroencephalography. U-Sleep is suitable for pediatric clinical utilization provided automated staging is followed by expert clinician review.

Citation: Kevat A, Steinkey R, Suresh S, et al. Evaluation of automated pediatric sleep stage classification using U-Sleep: a convolutional neural network. . 2025;21(2):277-285.

References

Huang X, Shirahama K, Li F, Grzegorzek M . Sleep stage classification for child patients using DeConvolutional Neural Network. Artif Intell Med. 2020; 110:101981. DOI: 10.1016/j.artmed.2020.101981. View

Perslev M, Darkner S, Kempfner L, Nikolic M, Jennum P, Igel C . U-Sleep: resilient high-frequency sleep staging. NPJ Digit Med. 2021; 4(1):72. PMC: 8050216. DOI: 10.1038/s41746-021-00440-5. View

Mi S, Kelly N, Brychta R, Grammer A, Jaramillo M, Chen K . Associations of sleep patterns with metabolic syndrome indices, body composition, and energy intake in children and adolescents. Pediatr Obes. 2019; 14(6):e12507. PMC: 6504608. DOI: 10.1111/ijpo.12507. View

Kelly C, Karthikesalingam A, Suleyman M, Corrado G, King D . Key challenges for delivering clinical impact with artificial intelligence. BMC Med. 2019; 17(1):195. PMC: 6821018. DOI: 10.1186/s12916-019-1426-2. View

Wang H, Lin G, Li Y, Zhang X, Xu W, Wang X . Automatic Sleep Stage Classification of Children with Sleep-Disordered Breathing Using the Modularized Network. Nat Sci Sleep. 2021; 13:2101-2112. PMC: 8643215. DOI: 10.2147/NSS.S336344. View

Hochadel J, Frolich J, Wiater A, Lehmkuhl G, Fricke-Oerkermann L . Prevalence of sleep problems and relationship between sleep problems and school refusal behavior in school-aged children in children's and parents' ratings. Psychopathology. 2013; 47(2):119-26. DOI: 10.1159/000345403. View

Ruehland W, Rochford P, Pierce R, Singh P, Thornton A . External proficiency testing improves inter-scorer reliability of polysomnography scoring. Sleep Breath. 2022; 27(3):923-932. DOI: 10.1007/s11325-022-02673-4. View

Di Mauro P, Cocuzza S, Maniaci A, Ferlito S, Rasa D, Anzivino R . The Effect of Adenotonsillectomy on Children's Behavior and Cognitive Performance with Obstructive Sleep Apnea Syndrome: State of the Art. Children (Basel). 2021; 8(10). PMC: 8535044. DOI: 10.3390/children8100921. View

Phan H, Mertins A, Baumert M . Pediatric Automatic Sleep Staging: A Comparative Study of State-of-the-Art Deep Learning Methods. IEEE Trans Biomed Eng. 2022; 69(12):3612-3622. DOI: 10.1109/TBME.2022.3174680. View

10.

Zhang X, Dong X, Kantelhardt J, Li J, Zhao L, Garcia C . Process and outcome for international reliability in sleep scoring. Sleep Breath. 2014; 19(1):191-5. DOI: 10.1007/s11325-014-0990-0. View

11.

Gomez R, Edgin J . Sleep as a window into early neural development: Shifts in sleep-dependent learning effects across early childhood. Child Dev Perspect. 2015; 9(3):183-189. PMC: 4636128. DOI: 10.1111/cdep.12130. View

12.

Lee H, Li B, DeForte S, Splaingard M, Huang Y, Chi Y . A large collection of real-world pediatric sleep studies. Sci Data. 2022; 9(1):421. PMC: 9296671. DOI: 10.1038/s41597-022-01545-6. View

13.

Grigg-Damberger M, Gozal D, Marcus C, Quan S, Rosen C, Chervin R . The visual scoring of sleep and arousal in infants and children. J Clin Sleep Med. 2007; 3(2):201-40. View

14.

Dutil C, Chaput J . Inadequate sleep as a contributor to type 2 diabetes in children and adolescents. Nutr Diabetes. 2017; 7(5):e266. PMC: 5518801. DOI: 10.1038/nutd.2017.19. View

15.

Proost R, Heremans E, Lagae L, Van Paesschen W, De Vos M, Jansen K . Automated sleep staging on reduced channels in children with epilepsy. Front Neurol. 2024; 15:1390465. PMC: 11116721. DOI: 10.3389/fneur.2024.1390465. View

16.

Somaskandhan P, Leppanen T, Terrill P, Sigurdardottir S, Arnardottir E, Olafsdottir K . Deep learning-based algorithm accurately classifies sleep stages in preadolescent children with sleep-disordered breathing symptoms and age-matched controls. Front Neurol. 2023; 14:1162998. PMC: 10140398. DOI: 10.3389/fneur.2023.1162998. View

17.

van Twist E, Hiemstra F, Cramer A, Verbruggen S, Tax D, Joosten K . An electroencephalography-based sleep index and supervised machine learning as a suitable tool for automated sleep classification in children. J Clin Sleep Med. 2023; 20(3):389-397. PMC: 11019221. DOI: 10.5664/jcsm.10880. View

18.

Baumert M, Hartmann S, Phan H . Automatic sleep staging for the young and the old - Evaluating age bias in deep learning. Sleep Med. 2023; 107:18-25. DOI: 10.1016/j.sleep.2023.04.002. View

19.

de Bruin E, van Run C, Staaks J, Meijer A . Effects of sleep manipulation on cognitive functioning of adolescents: A systematic review. Sleep Med Rev. 2016; 32:45-57. DOI: 10.1016/j.smrv.2016.02.006. View

20.

Abramoff M, Tarver M, Loyo-Berrios N, Trujillo S, Char D, Obermeyer Z . Considerations for addressing bias in artificial intelligence for health equity. NPJ Digit Med. 2023; 6(1):170. PMC: 10497548. DOI: 10.1038/s41746-023-00913-9. View