Display options
Share it on

NPJ Digit Med. 2021 Apr 15;4(1):72. doi: 10.1038/s41746-021-00440-5.

U-Sleep: resilient high-frequency sleep staging.

NPJ digital medicine

Mathias Perslev, Sune Darkner, Lykke Kempfner, Miki Nikolic, Poul Jørgen Jennum, Christian Igel

Affiliations

  1. Department of Computer Science, University of Copenhagen, Copenhagen, Denmark.
  2. Danish Center for Sleep Medicine, Rigshospitalet, Copenhagen, Denmark.
  3. Department of Computer Science, University of Copenhagen, Copenhagen, Denmark. [email protected].

PMID: 33859353 PMCID: PMC8050216 DOI: 10.1038/s41746-021-00440-5

Abstract

Sleep disorders affect a large portion of the global population and are strong predictors of morbidity and all-cause mortality. Sleep staging segments a period of sleep into a sequence of phases providing the basis for most clinical decisions in sleep medicine. Manual sleep staging is difficult and time-consuming as experts must evaluate hours of polysomnography (PSG) recordings with electroencephalography (EEG) and electrooculography (EOG) data for each patient. Here, we present U-Sleep, a publicly available, ready-to-use deep-learning-based system for automated sleep staging ( sleep.ai.ku.dk ). U-Sleep is a fully convolutional neural network, which was trained and evaluated on PSG recordings from 15,660 participants of 16 clinical studies. It provides accurate segmentations across a wide range of patient cohorts and PSG protocols not considered when building the system. U-Sleep works for arbitrary combinations of typical EEG and EOG channels, and its special deep learning architecture can label sleep stages at shorter intervals than the typical 30 s periods used during training. We show that these labels can provide additional diagnostic information and lead to new ways of analyzing sleep. U-Sleep performs on par with state-of-the-art automatic sleep staging systems on multiple clinical datasets, even if the other systems were built specifically for the particular data. A comparison with consensus-scores from a previously unseen clinic shows that U-Sleep performs as accurately as the best of the human experts. U-Sleep can support the sleep staging workflow of medical experts, which decreases healthcare costs, and can provide highly accurate segmentations when human expertize is lacking.

References

  1. BMJ Open. 2016 Feb 17;6(2):e008119 - PubMed
  2. PLoS One. 2019 May 7;14(5):e0216456 - PubMed
  3. Comput Methods Programs Biomed. 2016 Feb;124:180-92 - PubMed
  4. Sleep. 2012 Jun 01;35(6):757-67 - PubMed
  5. Int J Environ Res Public Health. 2019 Jan 21;16(2): - PubMed
  6. Sleep Breath. 2017 May;21(2):235-242 - PubMed
  7. J Neurosci Methods. 1998 Feb 20;79(2):187-93 - PubMed
  8. Eur Neuropsychopharmacol. 2011 Sep;21(9):655-79 - PubMed
  9. J Sleep Res. 2019 Apr;28(2):e12780 - PubMed
  10. Am J Respir Crit Care Med. 1995 Mar;151(3 Pt 1):682-7 - PubMed
  11. Am J Respir Crit Care Med. 2018 Apr 15;197(8):1080-1083 - PubMed
  12. Sleep. 2011 Nov 01;34(11):1509-17 - PubMed
  13. Nat Commun. 2018 Dec 6;9(1):5229 - PubMed
  14. J Am Geriatr Soc. 2011 Dec;59(12):2217-25 - PubMed
  15. Eur Respir J. 2009 Apr;33(4):907-14 - PubMed
  16. Nat Methods. 2014 Apr;11(4):385-92 - PubMed
  17. Sleep Med. 2018 Apr;44:97-105 - PubMed
  18. Sleep. 2015 Jun 01;38(6):877-88 - PubMed
  19. Circulation. 2000 Jun 13;101(23):E215-20 - PubMed
  20. IEEE Trans Neural Syst Rehabil Eng. 2019 Mar;27(3):400-410 - PubMed
  21. IEEE Trans Biomed Eng. 2000 Sep;47(9):1185-94 - PubMed
  22. J Pediatr. 2003 Apr;142(4):383-9 - PubMed
  23. J Am Med Inform Assoc. 2018 Dec 1;25(12):1643-1650 - PubMed
  24. Sleep. 2016 Nov 01;39(11):1993-2004 - PubMed
  25. Comput Methods Programs Biomed. 2018 Jul;161:1-13 - PubMed
  26. IEEE Trans Neural Syst Rehabil Eng. 2020 Sep;28(9):1955-1965 - PubMed
  27. Nature. 2015 May 28;521(7553):436-44 - PubMed
  28. Sleep Med. 2013 Aug;14(8):795-806 - PubMed
  29. J Am Med Inform Assoc. 2018 Oct 1;25(10):1351-1358 - PubMed
  30. Int J Environ Res Public Health. 2016 Aug 18;13(8): - PubMed
  31. Healthcare (Basel). 2018 Dec 20;7(1): - PubMed
  32. J Am Geriatr Soc. 2008 Jan;56(1):45-50 - PubMed
  33. Sleep. 1997 Dec;20(12):1077-85 - PubMed
  34. Sleep Med Rev. 2012 Jun;16(3):251-63 - PubMed
  35. J Sleep Res. 2014 Dec;23(6):628-635 - PubMed
  36. IEEE Trans Neural Syst Rehabil Eng. 2018 Feb;26(2):324-333 - PubMed
  37. Neuroepidemiology. 2016;46(1):57-78 - PubMed
  38. IEEE Trans Neural Syst Rehabil Eng. 2017 Nov;25(11):1998-2008 - PubMed
  39. Sleep. 2015 Mar 01;38(3):411-21 - PubMed
  40. Sleep. 2017 Nov 1;40(11): - PubMed
  41. Nature. 2020 Nov;587(7832):78-82 - PubMed
  42. Comput Methods Programs Biomed. 2019 Jul;176:81-91 - PubMed
  43. J Sleep Res. 2019 Feb;28(1):e12672 - PubMed
  44. JAMA. 1990 Feb 2;263(5):665-8 - PubMed
  45. IEEE Trans Med Imaging. 2006 Nov;25(11):1451-61 - PubMed
  46. IEEE Trans Biomed Eng. 2020 Aug 31;PP: - PubMed
  47. IEEE Eng Med Biol Mag. 2001 May-Jun;20(3):51-7 - PubMed
  48. Sleep. 2017 Oct 1;40(10): - PubMed
  49. Nat Rev Cardiol. 2019 Apr;16(4):213-224 - PubMed
  50. Nat Methods. 2019 Jan;16(1):67-70 - PubMed
  51. N Engl J Med. 2013 Jun 20;368(25):2366-76 - PubMed
  52. Neuropsychobiology. 2010;62(4):250-64 - PubMed
  53. Annu Int Conf IEEE Eng Med Biol Soc. 2018 Jul;2018:171-174 - PubMed
  54. Annu Int Conf IEEE Eng Med Biol Soc. 2019 Jul;2019:1596-1600 - PubMed

Publication Types