Automatic scoring of apnea and hypopnea events using blood oxygen saturation signals

Prediction of the Sleep Apnea Severity Using 2D-Convolutional Neural Networks and Respiratory Effort Signals

The high prevalence of sleep apnea and the limitations of polysomnography have prompted the investigation of strategies aimed at automated diagnosis using a restricted number of physiological measures. This study aimed to demonstrate that thoracic (THO) and abdominal (ABD) movement signals are useful for accurately estimating the severity of sleep apnea, even if central respiratory events are present. Thus, we developed 2D-convolutional neural networks (CNNs) jointly using THO and ABD to automatically estimate sleep apnea severity and evaluate the central event contribution. Our proposal achieved an intraclass correlation coefficient (ICC) = 0.75 and a root mean square error (RMSE) = 10.33 events/h when estimating the apnea-hypopnea index, and ICC = 0.83 and RMSE = 0.95 events/h when estimating the central apnea index. The CNN obtained accuracies of 94.98%, 79.82%, and 81.60% for 5, 15, and 30 events/h when evaluating the complete apnea hypopnea index. The model improved when the nature of the events was central: 98.72% and 99.74% accuracy for 5 and 15 events/h. Hence, the information extracted from these signals using CNNs could be a powerful tool to diagnose sleep apnea, especially in subjects with a high density of central apnea events.

A Dual-Scale Convolutional Neural Network for Sleep Apnea Detection with Time-Delayed SpO2 Signals

Sleep apnea (SA) is a common breathing disease, with clinical manifestations of sleep snoring at night with apnea and daytime sleepiness. It could lead to ischemic heart disease, stroke, or even sudden death. SpO2 signal is highly related to SA, and many automatic SA detection methods have been proposed. However, extant work focuses on small datasets with relatively few subjects (less than 100) and is unaware of SA syndromes occurring about 5 seconds prior to the SpO2 change. This study proposes an automatic SA detector called DSCNN using a single-lead SpO2 signal with a dual-scale convolutional neural network. To solve the time-delayed problem of SpO2 changes, we enlarge the target SpO2 segment information by combining its subsequent segment information. To utilize neighbouring segments information and further facilitate the SA detection performance, a dual-scale neural network with the fusing information of the prolonged target segment and its two surrounding segments is proposed. Three datasets from multiple centres are employed to verify the generic performance of DSCNN. Here, we must point out that we use two datasets as external datasets, and one of them is collected from the First Affiliated Hospital of Sun Yat-sen University with a large sample size (450 subjects). Extensive experiment results show that DSCNN can achieve promising results which are superior to the existing state-of-the-art methods.

Detection of Respiratory Events during Sleep Based on Fusion Analysis and Entropy Features of Cardiopulmonary Signals

Sleep apnea hypopnea syndrome (SAHS) is a common sleep disorder with a high prevalence. The apnea hypopnea index (AHI) is an important indicator used to diagnose the severity of SAHS disorders. The calculation of the AHI is based on the accurate identification of various types of sleep respiratory events. In this paper, we proposed an automatic detection algorithm for respiratory events during sleep. In addition to the accurate recognition of normal breathing, hypopnea and apnea events using heart rate variability (HRV), entropy and other manual features, we also presented a fusion of ribcage and abdomen movement data combined with the long short-term memory (LSTM) framework to achieve the distinction between obstructive and central apnea events. While only using electrocardiogram (ECG) features, the accuracy, precision, sensitivity, and F1 score of the XGBoost model are 0.877, 0.877, 0.876, and 0.876, respectively, demonstrating that it performs better than other models. Moreover, the accuracy, sensitivity, and F1 score of the LSTM model for detecting obstructive and central apnea events were 0.866, 0.867, and 0.866, respectively. The research results of this paper can be used for the automatic recognition of sleep respiratory events as well as AHI calculation of polysomnography (PSG), which provide a theoretical basis and algorithm references for out-of-hospital sleep monitoring.

Identification of arterial oxygen intermittency in oximetry data

In patients with kidney failure treated by hemodialysis, intradialytic arterial oxygen saturation (SaO₂) time series present intermittent high-frequency high-amplitude oximetry patterns (IHHOP), which correlate with observed sleep-associated breathing disturbances. A new method for identifying such intermittent patterns is proposed. The method is based on the analysis of recurrence in the time series through the quantification of an optimal recurrence threshold ([Formula: see text]). New time series for the value of [Formula: see text] were constructed using a rolling window scheme, which allowed for real-time identification of the occurrence of IHHOPs. The results for the optimal recurrence threshold were confronted with standard metrics used in studies of obstructive sleep apnea, namely the oxygen desaturation index (ODI) and oxygen desaturation density (ODD). A high correlation between [Formula: see text] and the ODD was observed. Using the value of the ODI as a surrogate to the apnea-hypopnea index (AHI), it was shown that the value of [Formula: see text] distinguishes occurrences of sleep apnea with great accuracy. When subjected to binary classifiers, this newly proposed metric has great power for predicting the occurrences of sleep apnea-related events, as can be seen by the larger than 0.90 AUC observed in the ROC curve. Therefore, the optimal threshold [Formula: see text] from recurrence analysis can be used as a metric to quantify the occurrence of abnormal behaviors in the arterial oxygen saturation time series.

A method to detect sleep apnea using residual attention mechanism network from single-lead ECG signal

Sleep apnea is a sleep disorder caused by weakened or suspended breathing during sleep, which seriously affects the work and health of patients. The traditional polysomnography (PSG) detection process is complicated and expensive, which has attracted researchers to explore a rapid detection method based on single-lead ECG signals. However, existing ECG-based sleep apnea detection methods have certain limitations and complexities, mainly relying on human-crafted features. To solve the problem, the paper develops a sleep apnea detection method based on a residual attention mechanism network. The method uses the RR interval signal and the R-peak signal derived from the ECG signal as input, realizes feature extraction through the residual network (ResNet), and adds the SENet attention mechanism to deepen the mining of channel features. Experimental results show that the per-segment accuracy of the proposed method can reach 86.2%. Compared with existing works, its accuracy has increased by 1.1-8.1%. These results show that the proposed residual attention network can effectively use ECG signals to quickly detect sleep apnea. Meanwhile, compared with existing works, the proposed method overcomes the limitations and complexity of human-crafted features in sleep apnea detection research.

Objective sleep assessments for healthy people in environmental research: A literature review

To date, although many studies had focused on the impact of environmental factors on sleep, how to choose the proper assessment method for objective sleep quality was often ignored, especially for healthy subjects in bedroom environment. In order to provide methodological guidance for future research, this paper reviewed the assessments of objective sleep quality applied in environmental researches, compared them from the perspective of accuracy and interference, and statistically analyzed the impact of experimental type and subjects' information on method selection. The review results showed that, in contrast to polysomnography (PSG), the accuracy of actigraphy (ACT), respiratory monitoring-oxygen saturation monitoring (RM-OSM), and electrocardiograph (ECG) could reach up to 97%, 80.38%, and 79.95%, respectively. In terms of sleep staging, PSG and ECG performed the best, ACT the second, and RM-OSM the worst; as compared to single methods, mix methods were more accurate and better at sleep staging. PSG interfered with sleep a great deal, while ECG and ACT could be non-contact, and thus, the least interference with sleep was present. The type of experiment significantly influenced the choice of assessment method (p < 0.001), 85.3% of researchers chose PSG in laboratory study while 82.5% ACT in field study; moreover, PSG was often used in a relatively small number of young subjects, while ACT had a wide applicable population. In general, researchers need to pay more attention at selection of assessments in future studies, and this review can be used as a reliable reference for experimental design.

A generalized algorithm for the automatic diagnosis of sleep apnea from per-sample encoding of airflow and oximetry

OBJECTIVE

Sleep apnea is a common sleep breathing disorder that can significantly decrease sleep quality and have major health consequences. It is diagnosed based on the apnea hypopnea index (AHI). This study explored a novel, generalized algorithm for the automatic diagnosis of sleep apnea employing airflow (AF) and oximetry (SpO2) signals.

APPROACH

Of the 988 polysomnography records, 45 were randomly selected for developing the automatic algorithm and the remainder 943 for validating purposes. The algorithm detects apnea events by a per-sample encoding process applied to the peak excursion of AF signal. Hypopnea events were detected from the per-sample encoding of AF and SpO2 with an adjustment to time lag in SpO2. Total recording time was automatically processed and optimized for computation of total sleep time (TST). Total number of detected events and computed TST were used to estimate AHI. The estimated AHI was validated against the scored data from the Sleep Heart Health Study.

MAIN RESULTS

Intraclass correlation coefficient of 0.94 was obtained between estimated and scored AHIs. The diagnostic accuracies were 93.5%, 92.4%, and 96.6% for AHI cut-off values of ≥5, ≥15, and ≥30 respectively. The overall accuracy for the combined severity categories (normal, mild, moderate, and severe) and kappa were 83.4% and 0.77 respectively.

SIGNIFICANCE

This new automatic technique was found to be superior to the other existing methods and can be applied to any portable sleep devices especially for home sleep apnea tests.

Conventional Machine Learning Methods Applied to the Automatic Diagnosis of Sleep Apnea

The overnight polysomnography shows a range of drawbacks to diagnose obstructive sleep apnea (OSA) that have led to the search for artificial intelligence-based alternatives. Many classic machine learning methods have been already evaluated for this purpose. In this chapter, we show the main approaches found in the scientific literature along with the most used data to develop the models, useful and large easily available databases, and suitable methods to assess performances. In addition, a range of results from selected studies are presented as examples of these methods. Very high diagnostic performances are reported in these results regardless of the approaches taken. This leads us to conclude that conventional machine learning methods are useful techniques to develop new OSA diagnosis simplification proposals and to act as benchmark for other more recent methods such as deep learning.

Automatic Segmentation to Cluster Patterns of Breathing in Sleep Apnea

Annotation of polysomnography (PSG) recordings for diagnosis of obstructive sleep apnea (OSA) is a standard procedure but an expensive and time-consuming process for clinicians. To aid clinicians in this process we present a data driven unsupervised hierarchical clustering approach for detection and visual presentation of breathing patterns in PSG recordings. The aim was to develop a model independent of manual annotations to detect and visualize respiratory events related to OSA. 10 recordings from the Sleep Heart Health Study database were used, and the proposed algorithm was evaluated based on the manually annotated events for each recording. The algorithm reached an F1-score of 0.58 across the 10 recordings when detecting the presence of an event vs. no event and a 100% correct diagnosis prediction of OSA when predicting if apnea-hypopnea index (AHI) ≥ 15, which is a clinically meaningful cut-off. The F1-score may be due to imprecise placement of events, difficulty distinguishing between hypopneas and stable breathing, and variations in scoring. In conclusion the performance can be improved despite the strong agreement in diagnostics. The method is a proof of concept that a clustering method can detect and visualize breathing patterns related to OSA while maintaining a correct diagnosis.

An LSTM Network for Apnea and Hypopnea Episodes Detection in Respiratory Signals

One of the most common sleep disorders is sleep apnea. It manifests itself by episodes of shallow breathing or pauses in breathing during the night. Diagnosis of this disease involves polysomnography examination, which is expensive. Alternatively, diagnostic doctors can be supported with recordings from the in-home polygraphy sensors. Furthermore, numerous attempts for providing an automated apnea episodes annotation algorithm have been made. Most of them, however, do not distinguish between apnea and hypopnea episodes. In this work, a novel solution for epoch-based annotation problem is presented. Utilizing an architecture based on the long short-term memory (LSTM) networks, the proposed model provides locations of sleep disordered breathing episodes and identifies them as either apnea or hypopnea. To achieve this, special pre- and postprocessing steps have been designed. The obtained labels can be then used for calculation of the respiratory event index (REI), which serves as a disease severity indicator. The input for the model consists of the oronasal airflow along with the thoracic and abdominal respiratory effort signals. Performance of the proposed architecture was verified on the SHHS-1 and PhysioNet Sleep databases, obtaining mean REI classification error of 9.24/10.52 with standard deviation of 11.61/7.92 (SHHS-1/PhysioNet). Normal breathing, hypopnea and apnea differentiation accuracy is assessed on both databases, resulting in the correctly classified samples percentage of 86.42%/84.35%, 49.30%/58.28% and 68.20%/69.50% for normal breathing, hypopnea and apnea classes, respectively. Overall accuracies are 80.66%/82.04%. Additionally, the effect of wake periods is investigated. The results show that the proposed model can be successfully used for both episode classification and REI estimation tasks.

Automatic identification of respiratory events based on nasal airflow and respiratory effort of the chest and abdomen

Objective.Disease may cause changes in an individual's respiratory pattern, which can be measured as parameters for disease evaluation, usually through manually annotated polysomnographic recordings. In this study, a machine learning model based on nasal airflow and respiratory effort of the chest and abdomen is proposed to automatically identify respiratory events, including normal breathing, hypopnea and apnea.Approach.The nasal airflow and chest-abdominal respiratory effort signals were collected by polysomnography (PSG). Time/frequency domain features, fractional Fourier transform features and sample entropy were calculated to obtain feature sets. Features selected through statistical analysis were used as input variables of the machine learning model. The performance of different input combinations on different models was studied and cross-validated.Main results.The dataset included PSG sleep records of 60 patients provided by the Chinese People's Liberation Army General Hospital. The extreme gradient boosting-based model (XGBoost) performed best in several models with an accuracy of 0.807 and a F1 score of 0.807, depending on the combination of nasal airflow and two respiratory effort signals. The precision for normal breathing, hypopnea and apnea events were 0.764, 0.789 and 0.871, respectively. In addition, the recall scores were 0.833, 0.768 and 0.823 for normal breathing, hypopnea and apnea events, respectively. Moreover, it was found that the standard deviation and kurtosis of nasal airflow were the most important features of the respiratory event detection model.Significance.Since nasal airflow and respiratory effort of the chest and abdomen contain the characteristics of respiratory events, their combined use can improve the classification performance for identification of respiratory events. With this method, respiratory events can be automatically detected and labeled from the PSG records, which can be used to screen for patients with sleep apnea-hypopnea syndrome.