A deep learning architecture to detect events in EEG signals during\n sleep

2018/07/11 by Stanislas Chambon, Valentin Thorey, Chambon, Stanislas +7
Computer Science · Neuroscience · #EEG and Brain-Computer Interfaces #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Signal Processing (eess.SP) #Sleep and Wakefulness Research #Time Series Analysis and Forecasting #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.1807.05981

openalex publication_date 2018/07/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Electroencephalography (EEG) during sleep is used by clinicians to evaluate\nvarious neurological disorders. In sleep medicine, it is relevant to detect\nmacro-events (> 10s) such as sleep stages, and micro-events (<2s) such as\nspindles and K-complexes. Annotations of such events require a trained sleep\nexpert, a time consuming and tedious process with a large inter-scorer\nvariability. Automatic algorithms have been developed to detect various types\nof events but these are event-specific. We propose a deep learning method that\njointly predicts locations, durations and types of events in EEG time series.\nIt relies on a convolutional neural network that builds a feature\nrepresentation from raw EEG signals. Numerical experiments demonstrate\nefficiency of this new approach on various event detection tasks compared to\ncurrent state-of-the-art, event specific, algorithms.\n

Related

• DOSED: a deep learning approach to detect multiple sleep micro-events in EEG signal
• An Attention-Based Deep Learning Approach for Sleep Stage Classification With Single-Channel EEG
• Multi-task deep-learning for sleep event detection and stage classification
• A multi-level interpretable sleep stage scoring system by infusing experts' knowledge into a deep network architecture
• Classification of sleep stages from EEG, EOG and EMG signals by SSNet