A Multi-Class Automatic Sleep Staging Method Based on Photoplethysmography Signals

Zhao, Xiangfa; Sun, Guobing

doi:10.3390/e23010116

Cited by 19 publications

(14 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As recently summarized in a comprehensive systematic review ( Imtiaz, 2021 ), PPG-based solutions are on average easier to use and better suited for wearable/nearable monitoring but often unable to reliably characterize the full spectrum of sleep stages. A recent manuscript by Zhao and Sun (2021) , proposed an approach similar to our work despite limited to a significantly smaller sample size. The reported results for accuracy and Cohen’s kappa of the 3-level and 4-level models are aligned with our findings.…”

Section: Discussionmentioning

confidence: 93%

An automated heart rate-based algorithm for sleep stage classification: Validation using conventional polysomnography and an innovative wearable electrocardiogram device

et al. 2022

View full text Add to dashboard Cite

BackgroundThe rapid advancement in wearable solutions to monitor and score sleep staging has enabled monitoring outside of the conventional clinical settings. However, most of the devices and algorithms lack extensive and independent validation, a fundamental step to ensure robustness, stability, and replicability of the results beyond the training and testing phases. These systems are thought not to be feasible and reliable alternatives to the gold standard, polysomnography (PSG).Materials and methodsThis validation study highlights the accuracy and precision of the proposed heart rate (HR)-based deep-learning algorithm for sleep staging. The illustrated solution can perform classification at 2-levels (Wake; Sleep), 3-levels (Wake; NREM; REM) or 4- levels (Wake; Light; Deep; REM) in 30-s epochs. The algorithm was validated using an open-source dataset of PSG recordings (Physionet CinC dataset, n = 994 participants, 994 recordings) and a proprietary dataset of ECG recordings (Z3Pulse, n = 52 participants, 112 recordings) collected with a chest-worn, wireless sensor and simultaneous PSG collection using SOMNOtouch.ResultsWe evaluated the performance of the models in both datasets in terms of Accuracy (A), Cohen’s kappa (K), Sensitivity (SE), Specificity (SP), Positive Predictive Value (PPV), and Negative Predicted Value (NPV). In the CinC dataset, the highest value of accuracy was achieved by the 2-levels model (0.8797), while the 3-levels model obtained the best value of K (0.6025). The 4-levels model obtained the lowest SE (0.3812) and the highest SP (0.9744) for the classification of Deep sleep segments. AHI and biological sex did not affect scoring, while a significant decrease of performance by age was reported across the models. In the Z3Pulse dataset, the highest value of accuracy was achieved by the 2-levels model (0.8812), whereas the 3-levels model obtained the best value of K (0.611). For classification of the sleep states, the lowest SE (0.6163) and the highest SP (0.9606) were obtained for the classification of Deep sleep segment.ConclusionThe results of the validation procedure demonstrated the feasibility of accurate HR-based sleep staging. The combination of the proposed sleep staging algorithm with an inexpensive HR device, provides a cost-effective and non-invasive solution deployable in the home environment and robust across age, sex, and AHI scores.

show abstract

Section: Discussionmentioning

confidence: 93%

An automated heart rate-based algorithm for sleep stage classification: Validation using conventional polysomnography and an innovative wearable electrocardiogram device

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Our validation study demonstrated that our SVM-based method was consistent with the AMI actigraph (accuracy = 94.4% ± 3.8%, specificity = 94.2% ± 5.2%, sensitivity = 94.8% ± 3.9%, and F1-score = 92.0 ± 4.5) ( 21 , 22 ). Note that while we used a classical machine learning approach for the sleep–wake classification, state-of-the-art methods, such as ensemble tree-based algorithms [e.g., extreme gradient boosting (XGBoost) ( 35 ), or light gradient boosting machine (LightGBM) ( 36 , 37 )], or deep neural networks [e.g., long short-term memory ( 38 – 40 )], may improve classification performance significantly.…”

Section: Methodsmentioning

confidence: 99%

Seasonal Sleep Variations and Their Association With Meteorological Factors: A Japanese Population Study Using Large-Scale Body Acceleration Data

Nakamura

Hayano

et al. 2021

Front. Digit. Health

View full text Add to dashboard Cite

Seasonal changes in meteorological factors [e.g., ambient temperature (Ta), humidity, and sunlight] could significantly influence a person's sleep, possibly resulting in the seasonality of sleep properties (timing and quality). However, population-based studies on sleep seasonality or its association with meteorological factors remain limited, especially those using objective sleep data. Japan has clear seasonality with distinctive changes in meteorological variables among seasons, thereby suitable for examining sleep seasonality and the effects of meteorological factors. This study aimed to investigate seasonal variations in sleep properties in a Japanese population (68,604 individuals) and further identify meteorological factors contributing to sleep seasonality. Here we used large-scale objective sleep data estimated from body accelerations by machine learning. Sleep parameters such as total sleep time, sleep latency, sleep efficiency, and wake time after sleep onset demonstrated significant seasonal variations, showing that sleep quality in summer was worse than that in other seasons. While bedtime did not show clear seasonality, get-up time varied seasonally, with a nadir during summer, and positively correlated with the sunrise time. Estimated by the abovementioned sleep parameters, Ta had a practically meaningful association with sleep quality, indicating that sleep quality worsened with the increase of Ta. This association would partly explain seasonal variations in sleep quality among seasons. In conclusion, Ta had a principal role for seasonality in sleep quality, and the sunrise time chiefly determined the get-up time.

show abstract

“…By incorporating time information and replacing the classifier by conditional random fields, Cohen's kappa increased to 0.53 in 100 healthy subjects (Fonseca et al, 2018). In 2017, Tataraidze et al (2017) reported a kappa of 0.56 based on respiratory inductance plethysmography (RIP) signals using an extreme gradient boosting classifier in 658 healthy subjects and Beattie et al (2017) reported a kappa of 0.52 based on photoplethysmography (PPG) and actigraphy signals using a linear (Li et al, 2018;Radha et al, 2019;Wei et al, 2019;Sridhar et al, 2020;Huttunen et al, 2021;Zhao and Sun, 2021;Garcia-Molina and Jiang, 2022) and 0.47 ± 0.15 for 10 datasets with external testing (Fonseca et al, 2020;Sridhar et al, 2020;Sun et al, 2020b;Bakker et al, 2021;Garcia-Molina and Jiang, 2022).…”

Section: Agreement Between Sleep Parameters Derived From Cardiorespir...mentioning

confidence: 99%

Overview of the hypnodensity approach to scoring sleep for polysomnography and home sleep testing

Anderer¹,

Ross²,

Cerny³

et al. 2023

Front. Sleep

View full text Add to dashboard Cite

Human experts scoring sleep according to the American Academy of Sleep Medicine (AASM) rules are forced to select, for every 30-second epoch, one out of five stages, even if the characteristics of the neurological signals are ambiguous, a very common occurrence in clinical studies. Moreover, experts cannot score sleep in studies where these signals have not been recorded, such as in home sleep apnea testing (HSAT). In this topic review we describe how artificial intelligence can provide consistent and reliable scoring of sleep stages based on neurological signals recorded in polysomnography (PSG) and on cardiorespiratory signals recorded in HSAT. We also show how estimates of sleep stage probabilities, usually displayed as hypnodensity graph, can be used to quantify sleep stage ambiguity and stability. As an example of the application of hypnodensity in the characterization of sleep disordered breathing (SDB), we compared 49 patients with sleep apnea to healthy controls and revealed a severity-depending increase in ambiguity and decrease in stability during non-rapid eye movement (NREM) sleep. Moreover, using autoscoring of cardiorespiratory signals, we show how HSAT-derived apnea-hypopnea index and hypoxic burden are well correlated with the PSG indices in 80 patients, showing how using this technology can truly enable HSATs as alternatives to PSG to diagnose SDB.

show abstract

A Multi-Class Automatic Sleep Staging Method Based on Photoplethysmography Signals

Cited by 19 publications

References 33 publications

An automated heart rate-based algorithm for sleep stage classification: Validation using conventional polysomnography and an innovative wearable electrocardiogram device

An automated heart rate-based algorithm for sleep stage classification: Validation using conventional polysomnography and an innovative wearable electrocardiogram device

Seasonal Sleep Variations and Their Association With Meteorological Factors: A Japanese Population Study Using Large-Scale Body Acceleration Data

Overview of the hypnodensity approach to scoring sleep for polysomnography and home sleep testing

Contact Info

Product

Resources

About