Robust, automated sleep scoring by a compact neural network with distributional shift correction

• Published in: PloS one 2019-12, Vol.14 (12), p.e0224642-e0224642
• Main Authors: Barger, Zeke, Frye, Charles G, Liu, Danqian, Dan, Yang, Bouchard, Kristofer E
• Format: Article
• Language: English
• Subjects: 60 APPLIED LIFE SCIENCES; Algorithms; Animal experimentation; Animals; Arousal; Artificial intelligence; Artificial neural networks; Automation; Biology and Life Sciences; Cable television broadcasting industry; Cellular biology; Classification; Collection (Accounting); Computer and Information Sciences; Computer applications; Computer simulation; Data mining; Datasets; Deep Learning; EEG; Electrodes; Electroencephalography; Electroencephalography - methods; Electromyography; Electromyography - methods; Engineering; Engineering and Technology; Fourier transforms; Humans; In vivo methods and tests; Learning algorithms; Machine Learning; Medicine and Health Sciences; Mice; Neural networks; Neural Networks, Computer; Neurosciences; Physical Sciences; Polysomnography - methods; Reproducibility of Results; Research and Analysis Methods; Signal processing; Sleep; Sleep - physiology; Sleep Stages - physiology; Software; Standardization; User-Computer Interface
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0224642

Studying the biology of sleep requires the accurate assessment of the state of experimental subjects, and manual analysis of relevant data is a major bottleneck. Recently, deep learning applied to electroencephalogram and electromyogram data has shown great promise as a sleep scoring method, approaching the limits of inter-rater reliability. As with any machine learning algorithm, the inputs to a sleep scoring classifier are typically standardized in order to remove distributional shift caused by variability in the signal collection process. However, in scientific data, experimental manipulations introduce variability that should not be removed. For example, in sleep scoring, the fraction of time spent in each arousal state can vary between control and experimental subjects. We introduce a standardization method, mixture z-scoring, that preserves this crucial form of distributional shift. Using both a simulated experiment and mouse in vivo data, we demonstrate that a common standardization method used by state-of-the-art sleep scoring algorithms introduces systematic bias, but that mixture z-scoring does not. We present a free, open-source user interface that uses a compact neural network and mixture z-scoring to allow for rapid sleep scoring with accuracy that compares well to contemporary methods. This work provides a set of computational tools for the robust automation of sleep scoring.