학술논문
Towards high-performance differentiation between Narcolepsy and Idiopathic Hypersomnia in 10 minute EEG recordings using a Novel Machine Learning Approach
Document Type
Conference
Source
2019 IEEE International Conference on E-health Networking, Application & Services (HealthCom) E-health Networking, Application & Services (HealthCom), 2019 IEEE International Conference on. :1-7 Oct, 2019
Subject
Language
Abstract
While time and cost intensive tests are the current standard for diagnosing and classifying sleep disorders, we present results of using ten minute electroencephalographic recordings to differentiate between narcolepsy and idiopathic hypersomnia. Using a novel and fast machine learning approach, we reach an accuracy of almost 75 percent and moreover we show that there are systematic differences in the delta and beta-1 frequencies. For personalized treatments, it is important to differentiate between sleep disorders as they not only have side effects, but are also not equally effective for each respective sleep disorder.