Abstract
Sleep scoring is an important step for the detection of sleep disorders and usually performed by visual analysis. Since manual sleep scoring is time consuming, machine-learning based approaches have been proposed. Though efficient, these algorithms are black-box in nature and difficult to interpret by clinicians. In this paper, we propose a deep learning architecture for multi-modal sleep scoring, investigate the model's decision making process, and compare the model's reasoning with the annotation guidelines in the AASM manual. Our architecture, called STQS, uses convolutional neural networks (CNN) to automatically extract spatio-temporal features from 3 modalities (EEG, EOG and EMG), a bidirectional long short-term memory (Bi-LSTM) to extract sequential information, and residual connections to combine spatio-temporal and sequential features. We evaluated our model on two large datasets, obtaining an accuracy of 85% and 77% and a macro F1 score of 79% and 73% on SHHS and an in-house dataset, respectively. We further quantify the contribution of various architectural components and conclude that adding LSTM layers improves performance over a spatio-temporal CNN, while adding residual connections does not. Our interpretability results show that the output of the model is well aligned with AASM guidelines, and therefore, the model's decisions correspond to domain knowledge. We also compare multi-modal models and single-channel models and suggest that future research should focus on improving multi-modal models.
Highlights
Polysomnograms (PSGs) are recordings of body activities collected during sleep to aid the diagnosis of sleep disorders
We evaluate our models on two large datasets: the Sleep Heart Health Study (SHHS)-1 bench mark dataset (5793 PSGs, ≈6 million epochs) and a dataset collected at Medisch Spectrum Twente (MST), Enschede, Netherlands (1418 PSGs, ≈1.4 million epochs)
The results show that (i) EMG and EEG are sufficient to correctly predict Rapid Eye Move ment (REM), (ii) EEG and EOG are sufficient to correctly classify Non-REM stage 1 (N1) and NonREM stage 3 (N3), (iii) EEG alone is sufficient to classify W and Non-REM stage 2 (N2), and (iv) EOG is necessary to reduce the misclassification of other stages with REM
Summary
Polysomnograms (PSGs) are recordings of body activities collected during sleep to aid the diagnosis of sleep disorders. PSGs are analyzed and annotated by sleep technologists1 [1] based on sleep annotation guidelines [2,3]. Automatic approaches aim to make the annotation process more efficient, and can be coarsely divided into traditional machine learning (e.g., [4,5,6]) and deep learning approaches (e.g., [7,8]). The latter have the advantage of automatically extracting features and have been shown to outperform traditional approaches [9]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
