Narcolepsy is a rare life-long disease that exists in two forms, narcolepsy type-1 (NT1) or type-2 (NT2), but only NT1 is accepted as clearly defined entity. Both types of narcolepsies belong to the group of central hypersomnias (CH), a spectrum of poorly defined diseases with excessive daytime sleepiness as a core feature. Due to the considerable overlap of symptoms and the rarity of the diseases, it is difficult to identify distinct phenotypes of CH. Machine learning (ML) can help to identify phenotypes as it learns to recognize clinical features invisible for humans. Here we apply ML to data from the huge European Narcolepsy Network (EU-NN) that contains hundreds of mixed features of narcolepsy making it difficult to analyze with classical statistics. Stochastic gradient boosting, a supervised learning model with built-in feature selection, results in high performances in testing set. While cataplexy features are recognized as the most influential predictors, machine find additional features, e.g. mean rapid-eye-movement sleep latency of multiple sleep latency test contributes to classify NT1 and NT2 as confirmed by classical statistical analysis. Our results suggest ML can identify features of CH on machine scale from complex databases, thus providing 'ideas' and promising candidates for future diagnostic classifications.

AP HP Pediatric Sleep Center CHU Robert Debré Paris France

Center for Investigation and Research in Sleep Lausanne University Hospital Lausanne Switzerland

Center for Sleep Medicine Sleep Research and Epileptology Klinik Barmelweid AG Barmelweid Switzerland

Centre de Reference Nationale Maladies Rares Narcolepsie et Hypersomnie Idiopathique Service Neurologie Hôpital Gui de Chauliac INSERM U1061 Montpellier France

Centre Neuchatelois de Psychiatrie Neuchatel Switzerland

Department of Biomedical and Neuromotor Sciences Alma Mater Studiorum University of Bologna Bologna Italy

Department of Clinical Neurophysiology Institute of Psychiatry and Neurology Warsaw Poland

Department of Neurology and Clinical Neurophysiology Leiden University Medical Center Leiden The Netherlands

Department of Neurology Inselspital Bern University Hospital and University of Bern Bern Switzerland

Department of Sleep Medicine and Neuromuscular Disorders University of Münster Münster Germany

Eindhoven University of Technology Eindhoven The Netherlands

Fundació Institut d'Investigació Sanitària Illes Balears Hospital Universitari Son Espases Palma de Mallorca Spain

Helsinki Sleep Clinic Vitalmed Research Center Helsinki Finland

Institute of Molecular Medicine Portugal Medical Faculty Lisbon University Lisbon Portugal

IRCCS Istituto delle Scienze Neurologiche di Bologna ASL di Bologna Bologna Italy

National Reference Centre for Orphan Diseases Narcolepsy Idiopathic Hypersomnia and Kleine Levin Syndrome Paris France

Neurology Department and Centre of Clinical Neurosciences 1st Faculty of Medicine Charles University and General University Hospital Prague Czech Republic

Neurology Department Hephata Klinik Schwalmstadt Germany

Neurology Department Medical Faculty of P J Safarik University University Hospital of L Pasteur Kosice Kosice Slovak Republic

Neurology Department Sleep Disorders Clinic Medical University of Innsbruck Innsbruck Austria

Neurology Department University Hospital Zurich Zurich Switzerland

Neurology Service Multidisciplinary Sleep Unit Hospital Clínic of Barcelona IDIBAPS CIBERNED Barcelona Spain

Serviço de Neurofisiologia Hospital Santo António Centro Hospitalar do Porto and Instituto Ciências Biomédicas Abel Salazar Universidade do Porto Porto Portugal

Sleep and Epilepsy Center Neurocenter of Southern Switzerland Lugano Switzerland

Sleep and Epilepsy Unit Clinical Neurophysiology Department Gregorio Marañón University Hospital Complutense University of Madrid Madrid Spain

Sleep Medicine Center Kempenhaeghe Heeze The Netherlands

Sleep Wake Center SEIN Heemstede Stichting Epilepsie Instellingen Nederland Heemstede The Netherlands

Unidad de Neurofisiología y Trastornos del Sueño Hospital Vithas Internacional Madrid Madrid Spain

See more in PubMed

Khatami R, et al. The European Narcolepsy Network (EU-NN) database. J. Sleep Res. 2016;25:356–364. doi: 10.1111/jsr.12374. PubMed DOI

Scammell TE. Narcolepsy. N. Engl. J. Med. 2015;373:2654–2662. doi: 10.1056/NEJMra1500587. PubMed DOI

Dauvilliers, Y., PubMed

Nishino S, Ripley B, Overeem S, Lammers GJ, Mignot E. Hypocretin (orexin) deficiency in human narcolepsy. Lancet. 2000;355:39–40. doi: 10.1016/S0140-6736(99)05582-8. PubMed DOI

Baumann CR, et al. Challenges in Diagnosing Narcolepsy without Cataplexy: A Consensus Statement. Sleep. 2014;37:1035–1042. PubMed PMC

Schapire RE. The strength of weak learnability. Machine Learning. 1990;5:197–227.

Hastie, T., Tibshirani, R. & Friedman, J. Boosting and Additive Trees in

Kuhn M. Building Predictive Models in R Using the caret Package. Journal of Statistical Software. 2008;28:26. doi: 10.18637/jss.v028.i05. DOI

Blagus R, Lusa L. Boosting for high-dimensional two-class prediction. BMC Bioinformatics. 2015;16:300. doi: 10.1186/s12859-015-0723-9. PubMed DOI PMC

Friedman JH. Stochastic gradient boosting. Computational Statistics & Data Analysis. 2002;38:367–378. doi: 10.1016/S0167-9473(01)00065-2. DOI

Natekin A, Knoll A. Gradient boosting machines, a tutorial. Front Neurorobot. 2013;7:21. doi: 10.3389/fnbot.2013.00021. PubMed DOI PMC

Kuhn, M. The caret Package. (2017).

Aldrich MS, Chervin RD, Malow BA. Value of the multiple sleep latency test (MSLT) for the diagnosis of narcolepsy. Sleep. 1997;20:620–629. PubMed

Dietterich TG. An Experimental Comparison of Three Methods for Constructing Ensembles of Decision Trees: Bagging, Boosting, and Randomization. Machine Learning. 2000;40:139–157. doi: 10.1023/A:1007607513941. DOI

Opitz D, Maclin R. Popular Ensemble Methods: An Empirical Study. Journal of Artificial Intelligence Research. 1999;11:169–198.

Takei Y, et al. Differences in findings of nocturnal polysomnography and multiple sleep latency test between narcolepsy and idiopathic hypersomnia. Clin Neurophysiol. 2012;123:137–141. doi: 10.1016/j.clinph.2011.05.024. PubMed DOI

Mignot E, Hayduk R, Black J, Grumet FC, Guilleminault C. HLA DQB1*0602 is associated with cataplexy in 509 narcoleptic patients. Sleep. 1997;20:1012–1020. PubMed

Altman DG, Bland JM. Missing data. BMJ. 2007;334:424. doi: 10.1136/bmj.38977.682025.2C. PubMed DOI PMC

Trotti LM, Staab BA, Rye DB. Test-retest reliability of the multiple sleep latency test in narcolepsy without cataplexy and idiopathic hypersomnia. J Clin Sleep Med. 2013;9:789–795. PubMed PMC

Goldbart A, et al. Narcolepsy and predictors of positive MSLTs in the Wisconsin Sleep Cohort. Sleep. 2014;37:1043–1051. doi: 10.5665/sleep.3758. PubMed DOI PMC

Data-Driven Phenotyping of Central Disorders of Hypersomnolence With Unsupervised Clustering

Idling for Decades: A European Study on Risk Factors Associated with the Delay Before a Narcolepsy Diagnosis