Automatic sleep staging of EEG signals: recent development, challenges, and future directions

Phan, Huy; Mikkelsen, Kaare B.

doi:10.1088/1361-6579/ac6049

Cited by 78 publications

(69 citation statements)

References 176 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When interpreting the performance of automated sleep staging algorithms it is important to keep in mind that manual scoring by humans is highly subjective [27]. Inter-rater agreement for PSG labeled by human scorers is reported as a κ of 0.76 (95% confidence interval, 0.71-0.81) [28] for 5-class sleep staging.…”

Section: Discussionmentioning

confidence: 99%

SleepPPG-Net: a deep learning algorithm for robust sleep staging from continuous photoplethysmography

Kotzen¹,

Charlton²,

Salabi³

et al. 2022

Preprint

View full text Add to dashboard Cite

Introduction: Sleep staging is an essential component in the diagnosis of sleep disorders and management of sleep health. It is traditionally measured in a clinical setting and requires a labor-intensive labeling process. We hypothesize that it is possible to perform robust 4-class sleep staging using the raw photoplethysmography (PPG) time series and modern advances in deep learning (DL). Methods: We used two publicly available sleep databases that included raw PPG recordings, totalling 2,374 patients and 23,055 hours. We developed SleepPPG-Net, a DL model for 4-class sleep staging from the raw PPG time series. SleepPPG-Net was trained end-to-end and consists of a residual convolutional network for automatic feature extraction and a temporal convolutional network to capture long-range contextual information. We benchmarked the performance of SleepPPG-Net against models based on the best-reported state-of-the-art (SOTA) algorithms. Results: When benchmarked on a held-out test set, SleepPPG-Net obtained a median Cohen's Kappa (κ) score of 0.75 against 0.69 for the best SOTA approach. SleepPPG-Net showed good generalization performance to an external database, obtaining a κ score of 0.74 after transfer learning. Perspective Overall, SleepPPG-Net provides new SOTA performance. In addition, performance is high enough to open the path to the development of wearables that meet the requirements for usage in clinical applications such as the diagnosis and monitoring of obstructive sleep apnea.

show abstract

Section: Discussionmentioning

confidence: 99%

SleepPPG-Net: a deep learning algorithm for robust sleep staging from continuous photoplethysmography

Kotzen¹,

Charlton²,

Salabi³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Here, C = 5 as we are dealing with 5-stage sleep staging. Similar to a sequence-to-sequence sleep staging model [1], given a sequence (S 1 , S 2 , . .…”

Section: L-seqsleepnet With Long Sequence Modelling Capabilitymentioning

confidence: 99%

“…Capturing this sequential information has been shown to be crucial for automatic sleep staging systems to achieve good performance. In fact, the capacity of sequential modelling has been the driving force behind existing deeplearning-based sleep staging models, bringing the machine scoring performance on par with that of human experts [1].…”

Section: Introductionmentioning

confidence: 99%

“…For this investigation, we employed two models, SeqSleep-Net [4] and SleepTransformer [7], and the SHHS database [10], [11], a large database consisting of recordings from 5,791 subjects. Both models conform to the state-of-theart sequence-to-sequence sleep staging framework [1] but the former uses LSTM and the latter uses Transformer for sequential modelling purposes. The obtained overall staging accuracy in Table I indeed attest the negligible possible impact of a long sequence length in case of SeqSleepNet, marginally improving the accuracy by 0.2 − 0.3% when the sequence length is 5 and 10 times longer than the typical value (i.e., 20 epochs).…”

Section: Introductionmentioning

confidence: 99%

“…We argue that the negative effects of long sequences observed so far in the literature are due to model deficiency. That is, considering a long sequence as a "flat" sequence per se, the typical sequential modelling method in existing models [1] is incapable of handling long sequences, e.g., up to one whole sleep cycle. We hypothesize that an appropriate method for equipping a sleep staging system with long sequential modelling capability would benefit its performance.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

L-SeqSleepNet: Whole-cycle Long Sequence Modelling for Automatic Sleep Staging

Phan¹,

Lorenzen²,

Heremans³

et al. 2023

Preprint

View full text Add to dashboard Cite

Human sleep is cyclical with a period of approximately 90 minutes, implying long temporal dependency in the sleep data. Yet, exploring this long-term dependency when developing sleep staging models has remained untouched. In this work, we show that while encoding the logic of a whole sleep cycle is crucial to improve sleep staging performance, the sequential modelling approach in existing state-of-the-art deep learning models are inefficient for that purpose. We thus introduce a method for efficient long sequence modelling and propose a new deep learning model, L-SeqSleepNet, which takes into account whole-cycle sleep information for sleep staging. Evaluating L-SeqSleepNet on four distinct databases of various sizes, we demonstrate state-of-the-art performance obtained by the model over three different EEG setups, including scalp EEG in conventional Polysomnography (PSG), in-ear EEG, and around-the-ear EEG (cEEGrid), even with a single EEG channel input. Our analyses also show that L-SeqSleepNet is able to alleviate the predominance of N2 sleep (the major class in terms of classification) to bring down errors in other sleep stages. Moreover the network becomes much more robust, meaning that for all subjects where the baseline method had exceptionally poor performance, their performance are improved significantly. Finally, the computation time only grows at a sub-linear rate when the sequence length increases.

show abstract

Temporal Sequences of EEG Covariance Matrices for Automated Sleep Stage Scoring with Attention Mechanisms

Seraphim,

Dequidt,

Lechervy

et al. 2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Automatic sleep staging of EEG signals: recent development, challenges, and future directions

Cited by 78 publications

References 176 publications

SleepPPG-Net: a deep learning algorithm for robust sleep staging from continuous photoplethysmography

SleepPPG-Net: a deep learning algorithm for robust sleep staging from continuous photoplethysmography

L-SeqSleepNet: Whole-cycle Long Sequence Modelling for Automatic Sleep Staging

Temporal Sequences of EEG Covariance Matrices for Automated Sleep Stage Scoring with Attention Mechanisms

Contact Info

Product

Resources

About