STQS: Interpretable multi-modal Spatial-Temporal-seQuential model for automatic Sleep scoring

作者：

Highlights：

• Deep learning model for multimodal multichannel signals for predicting sleep stages.

• Spatio-temporal and sequential information are promising for sleep stage prediction.

• Post-hoc interpretability shows the model’s conformance with sleep scoring rules.

• Multimodal input signals (EEG, EOG, EMG) are important for sleep stage prediction.

摘要

•Deep learning model for multimodal multichannel signals for predicting sleep stages.•Spatio-temporal and sequential information are promising for sleep stage prediction.•Post-hoc interpretability shows the model’s conformance with sleep scoring rules.•Multimodal input signals (EEG, EOG, EMG) are important for sleep stage prediction.

论文关键词：Sleep scoring,Sleep stage annotation,Deep learning,EEG, EOG, EMG signals,Post-hoc interpretability,Explainable AI

论文评审过程：Received 16 July 2020, Revised 27 January 2021, Accepted 16 February 2021, Available online 27 February 2021, Version of Record 11 March 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102038