Computer-assisted sleep staging

Sleep Apnea Detection Using EEG: A Systematic Review of Datasets, Methods, Challenges, and Future Directions

PURPOSE

Sleep Apnea (SA) affects an estimated 936 million adults globally, posing a significant public health concern. The gold standard for diagnosing SA, polysomnography, is costly and uncomfortable. Electroencephalogram (EEG)-based SA detection is promising due to its ability to capture distinctive sleep stage-related characteristics across different sub-band frequencies. This study aims to review and analyze research from the past decade on the potential of EEG signals in SA detection and classification focusing on various deep learning and machine learning techniques, including signal decomposition, feature extraction, feature selection, and classification methodologies.

METHOD

A systematic literature review using the preferred reporting items for systematic reviews and meta-Analysis (PRISMA) and PICO guidelines was conducted across 5 databases for publications from January 2010 to December 2024.

RESULTS

The review involved screening a total of 402 papers, with 63 selected for in-depth analysis to provide valuable insights into the application of EEG signals for SA detection. The findings underscore the potential of EEG-based methods in improving SA diagnosis.

CONCLUSION

This study provides valuable insights, showcasing significant advancements while identifying key areas for further exploration, thereby laying a strong foundation for future research in EEG-based SA detection.

Optimized Prescreen Survey Tool for Predicting Sleep Apnea Based on Deep Neural Network: Pilot Study

Obstructive sleep apnea (OSA) is one of the common sleep disorders related to breathing. It is important to identify an optimal set of questions among the existing questionnaires, using a data-driven approach, that can prescreen OSA with high sensitivity and specificity. The current study proposes reliable models that are based on machine learning techniques to predict the severity of OSA. A total of 66 participants consisted of 45 males and 21 females (average age = 52.4 years old; standard deviation ± 14.6). Participants were asked to fill out the questionnaire items. If the value of the Respiratory Disturbance Index (RDI) was more than 30, the participant was diagnosed with severe OSA. Several different modeling techniques were applied, including deep neural networks with a scaled principal component analysis (DNN-PCA), random forest (RF), Adaptive Boosting Classifier (ABC), Decision Tree Classifier (DTC), K-nearest neighbors classifier (KNC), and support vector machine classifier (SVMC). Among the participants, 27 participants were diagnosed with severe OSA (RDI > 30). The area under the receiver operating characteristic curve (AUROC) was used to evaluate the developed models. As a result, the AUROC values of DNN-PCA, RF, ABC, DTC, KNC, and SVMC models were 0.95, 0.62, 0.53, 0.53, 0.51, and 0.78, respectively. The highest AUROC value was found in the DNN-PCA model with a sensitivity of 0.95, a specificity of 0.75, a positive predictivity of 0.95, an F1 score of 0.95, and an accuracy of 0.95. The DNN-PCA model outperforms the existing screening questionnaires, scores, and other models.

A Multi-Level Interpretable Sleep Stage Scoring System by Infusing Experts' Knowledge Into a Deep Network Architecture

In recent years, deep learning has shown potential and efficiency in a wide area including computer vision, image and signal processing. Yet, translational challenges remain for user applications due to a lack of interpretability of algorithmic decisions and results. This black box problem is particularly problematic for high-risk applications such as medical-related decision-making. The current study goal was to design an interpretable deep learning system for time series classification of electroencephalogram (EEG) for sleep stage scoring as a step toward designing a transparent system. We have developed an interpretable deep neural network that includes a kernel-based layer guided by a set of principles used for sleep scoring by human experts in the visual analysis of polysomnographic records. A kernel-based convolutional layer was defined and used as the first layer of the system and made available for user interpretation. The trained system and its results were interpreted in four levels from microstructure of EEG signals, such as trained kernels and effect of each kernel on the detected stages, to macrostructures, such as transitions between stages. The proposed system demonstrated greater performance than prior studies and the system learned information consistent with expert knowledge.

A comprehensive wireless neurological and cardiopulmonary monitoring platform for pediatrics

Neurodevelopment in the first 10 years of life is a critical time window during which milestones that define an individual's functional potential are achieved. Comprehensive multimodal neurodevelopmental monitoring is particularly crucial for socioeconomically disadvantaged, marginalized, historically underserved and underrepresented communities as well as medically underserved areas. Solutions designed for use outside the traditional clinical environment represent an opportunity for addressing such health inequalities. In this work, we present an experimental platform, ANNE EEG, which adds 16-channel cerebral activity monitoring to the existing, USA FDA-cleared ANNE wireless monitoring platform which provides continuous electrocardiography, respiratory rate, pulse oximetry, motion, and temperature measurements. The system features low-cost consumables, real-time control and streaming with widely available mobile devices, and fully wearable operation to allow a child to remain in their naturalistic environment. This multi-center pilot study successfully collected ANNE EEG recordings from 91 neonatal and pediatric patients at academic quaternary pediatric care centers and in LMIC settings. We demonstrate the practicality and feasibility to conduct electroencephalography studies with high levels of accuracy, validated via both quantitative and qualitative metrics, compared against gold standard systems. An overwhelming majority of parents surveyed during studies indicated not only an overall preference for the wireless system, but also that its use would improve their children's physical and emotional health. Our findings demonstrate the potential for the ANNE system to perform multimodal monitoring to screen for a variety of neurologic diseases that have the potential to negatively impact neurodevelopment.

Spike patterns surrounding sleep and seizures localize the seizure-onset zone in focal epilepsy

OBJECTIVE

Interictal spikes help localize seizure generators as part of surgical planning for drug-resistant epilepsy. However, there are often multiple spike populations whose frequencies change over time, influenced by brain state. Understanding state changes in spike rates will improve our ability to use spikes for surgical planning. Our goal was to determine the effect of sleep and seizures on interictal spikes, and to use sleep and seizure-related changes in spikes to localize the seizure-onset zone (SOZ).

METHODS

We performed a retrospective analysis of intracranial electroencephalography (EEG) data from patients with focal epilepsy. We automatically detected interictal spikes and we classified different time periods as awake or asleep based on the ratio of alpha to delta power, with a secondary analysis using the recently published SleepSEEG algorithm. We analyzed spike rates surrounding sleep and seizures. We developed a model to localize the SOZ using state-dependent spike rates.

RESULTS

We analyzed data from 101 patients (54 women, age range 16-69). The normalized alpha-delta power ratio accurately classified wake from sleep periods (area under the curve = .90). Spikes were more frequent in sleep than wakefulness and in the post-ictal compared to the pre-ictal state. Patients with temporal lobe epilepsy had a greater wake-to-sleep and pre- to post-ictal spike rate increase compared to patients with extra-temporal epilepsy. A machine-learning classifier incorporating state-dependent spike rates accurately identified the SOZ (area under the curve = .83). Spike rates tended to be higher and better localize the seizure-onset zone in non-rapid eye movement (NREM) sleep than in wake or REM sleep.

SIGNIFICANCE

The change in spike rates surrounding sleep and seizures differs between temporal and extra-temporal lobe epilepsy. Spikes are more frequent and better localize the SOZ in sleep, particularly in NREM sleep. Quantitative analysis of spikes may provide useful ancillary data to localize the SOZ and improve surgical planning.

Methodologies and Wearable Devices to Monitor Biophysical Parameters Related to Sleep Dysfunctions: An Overview

Sleep is crucial for human health from metabolic, mental, emotional, and social points of view; obtaining good sleep in terms of quality and duration is fundamental for maintaining a good life quality. Over the years, several systems have been proposed in the scientific literature and on the market to derive metrics used to quantify sleep quality as well as detect sleep disturbances and disorders. In this field, wearable systems have an important role in the discreet, accurate, and long-term detection of biophysical markers useful to determine sleep quality. This paper presents the current state-of-the-art wearable systems and software tools for sleep staging and detecting sleep disorders and dysfunctions. At first, the paper discusses sleep's functions and the importance of monitoring sleep to detect eventual sleep disturbance and disorders. Afterward, an overview of prototype and commercial headband-like wearable devices to monitor sleep is presented, both reported in the scientific literature and on the market, allowing unobtrusive and accurate detection of sleep quality markers. Furthermore, a survey of scientific works related the effect of the COVID-19 pandemic on sleep functions, attributable to both infection and lifestyle changes. In addition, a survey of algorithms for sleep staging and detecting sleep disorders is introduced based on an analysis of single or multiple biosignals (EEG-electroencephalography, ECG-electrocardiography, EMG-electromyography, EOG-electrooculography, etc.). Lastly, comparative analyses and insights are provided to determine the future trends related to sleep monitoring systems.

Multi-scale ResNet and BiGRU automatic sleep staging based on attention mechanism

Sleep staging is the basis of sleep evaluation and a key step in the diagnosis of sleep-related diseases. Despite being useful, the existing sleep staging methods have several disadvantages, such as relying on artificial feature extraction, failing to recognize temporal sequence patterns in the long-term associated data, and reaching the accuracy upper limit of sleep staging. Hence, this paper proposes an automatic Electroencephalogram (EEG) sleep signal staging model, which based on Multi-scale Attention Residual Nets (MAResnet) and Bidirectional Gated Recurrent Unit (BiGRU). The proposed model is based on the residual neural network in deep learning. Compared with the traditional residual learning module, the proposed model additionally uses the improved channel and spatial feature attention units and convolution kernels of different sizes in parallel at the same position. Thus, multiscale feature extraction of the EEG sleep signals and residual learning of the neural networks is performed to avoid network degradation. Finally, BiGRU is used to determine the dependence between the sleep stages and to realize the automatic learning of sleep data staging features and sleep cycle extraction. According to the experiment, the classification accuracy and kappa coefficient of the proposed method on sleep-EDF data set are 84.24% and 0.78, which are respectively 0.24% and 0.21 higher than the traditional residual net. At the same time, this paper also verified the proposed method on UCD and SHHS data sets, and the figure of classification accuracy is 79.34% and 81.6%, respectively. Compared to related existing studies, the recognition accuracy is significantly improved, which validates the effectiveness and generalization performance of the proposed method.

Validation of PiezoSleep Scoring Against EEG/EMG Sleep Scoring in Rats

INTRODUCTION

Current methods of sleep research in rodents involve invasive surgical procedures of EEG and EMG electrodes implantation. Recently, a new method of measuring sleep, PiezoSleep, has been validated against implanted electrodes in mice and rats. PiezoSleep uses a piezoelectric film transducer to detect the rodent's movements and respiration and employs an algorithm to automatically score sleep. Here, we validate PiezoSleep scoring versus EEG/EMG implanted electrodes sleep scoring in rats.

METHODS

Adult male Brown Norway and Wistar Kyoto rats were implanted with bilateral stainless-steel screws into the skull for EEG recording and bilateral wire electrodes into the nuchal muscles for EMG assessment. In Brown Norway rats, the EEG/EMG electrode leads were soldered to a miniature connector plug and fixed to the skull. In Wistar Kyoto rats, the EEG/EMG leads were tunneled subcutaneously to a telemetry transmitter implanted in the flank. Rats were allowed to recover from surgery for one week. Brown Norway rats were placed in PiezoSleep cages, and had their headsets connected to cable for recording EEG/EMG signals, which were then manually scored by a human scorer in 10-sec epochs. Wistar Kyoto rats were placed in PiezoSleep cages, and EEG/EMG signals were recorded using a telemetry system (DSI). Sleep was scored automatically in 4-sec epochs using NeuroScore software. PiezoSleep software recorded and scored sleep in the rats.

RESULTS

Rats implanted with corded EEG/EMG headsets had 85.6% concurrence of sleep-wake scoring with PiezoSleep. Rats implanted with EEG/EMG telemetry had 80.8% concurrence sleep-wake scoring with PiezoSleep. Sensitivity and specificity rates were similar between the EEG/EMG recording systems. Total sleep time and hourly sleep times did not differ in all three systems. However, automatic sleep detection by NeuroScore classified more sleep during the light period compared to the PiezoSleep.

CONCLUSION

We showed that PiezoSleep system can be a reliable alternative to both automatic and visual EEG/EMG- based sleep-wake scoring in rat.

A lightweight automatic sleep staging method for children using single-channel EEG based on edge artificial intelligence

With the development of telemedicine and edge computing, edge artificial intelligence (AI) will become a new development trend for smart medicine. On the other hand, nearly one-third of children suffer from sleep disorders. However, all existing sleep staging methods are for adults. Therefore, we adapted edge AI to develop a lightweight automatic sleep staging method for children using single-channel EEG. The trained sleep staging model will be deployed to edge smart devices so that the sleep staging can be implemented on edge devices which will greatly save network resources and improving the performance and privacy of sleep staging application. Then the results and hypnogram will be uploaded to the cloud server for further analysis by the physicians to get sleep disease diagnosis reports and treatment opinions. We utilized 1D convolutional neural networks (1D-CNN) and long short term memory (LSTM) to build our sleep staging model, named CSleepNet. We tested the model on our childrens sleep (CS) dataset and sleep-EDFX dataset. For the CS dataset, we experimented with F4-M1 channel EEG using four different loss functions, and the logcosh performed best with overall accuracy of 83.06% and F1-score of 76.50%. We used Fpz-Cz and Pz-Oz channel EEG to train our model in Sleep-EDFX dataset, and achieved an accuracy of 86.41% without manual feature extraction. The experimental results show that our method has great potential. It not only plays an important role in sleep-related research, but also can be widely used in the classification of other time sequences physiological signals.

On effective cognitive state classification using novel feature extraction strategies

Investigating new features for human cognitive state classification is an intiguing area of research with Electroencephalography (EEG) based signal analysis. We plan to develop a cost-effective system for cognitive state classification using ambulatory EEG signals. A novel event driven environment is created using external stimuli for capturing EEG data using a 14-channel Emotiv neuro-headset. A new feature extraction method, Gammatone Cepstrum Coefficients (GTCC) is introduced for ambulatory EEG signal analysis. The efficacy of this technique is compared with other feature extraction methods such as Discrete Wavelet Transformation (DWT) and Mel-Frequency Cepstral Coefficients (MFCC) using statistical metrics such as Fisher Discriminant Ratio (FDR) and Logistic Regression (LR). We obtain higher values for GTCC features, demonstrating its discriminative power during classification. A superior performance is achieved for the EEG dataset with a novel ensemble feature space comprising of GTCC and MFCC. Furthermore, the ensemble feature sets are passed through a proposed 1D Convolution Neural Networks (CNN) model to extract novel features. Various classification models like Probabilistic neural network (P-NN), Linear Discriminant Analysis (LDA), Multi-Class Support Vector Machine (MCSVM), Decision Tree (DT), Random Forest (RF) and Deep Convolutional Generative Adversarial Network (DCGAN) are employed to observe best accuracy on extracted features. The proposed GTCC, (GTCC+MFCC) & (GTCC +MFCC +CNN) features outperform the state-of-the-art techniques for all cases in our work. With GTCC+MFCC feature space and GTCC+MFCC+CNN features, accuracies of 96.42% and 96.14% are attained with the DCGAN classifier. Higher classification accuracies of the proposed system makes it a cynosure in the field of cognitive science.

Attention based convolutional network for automatic sleep stage classification

Sleep staging is an important basis for diagnosing sleep-related problems. In this paper, an attention based convolutional network for automatic sleep staging is proposed. The network takes time-frequency image as input and predict sleep stage for each 30-s epoch as output. For each CNN feature maps, our model generate attention maps along two separate dimensions, time and filter, and then multiplied to form the final attention map. Residual-like fusion structure is used to append the attention map to the input feature map for adaptive feature refinement. In addition, to get the global feature representation with less information loss, the generalized mean pooling is introduced. To prove the efficacy of the proposed method, we have compared with two baseline method on sleep-EDF data set with different setting of the framework and input channel type, the experimental results show that the paper model has achieved significant improvements in terms of overall accuracy, Cohen's kappa, MF1, sensitivity and specificity. The performance of the proposed network is compared with that of the state-of-the-art algorithms with an overall accuracy of 83.4%, a macro F1-score of 77.3%, κ = 0.77, sensitivity = 77.1% and specificity = 95.4%, respectively. The experimental results demonstrate the superiority of the proposed network.

Inter-database validation of a deep learning approach for automatic sleep scoring

STUDY OBJECTIVES

Development of inter-database generalizable sleep staging algorithms represents a challenge due to increased data variability across different datasets. Sharing data between different centers is also a problem due to potential restrictions due to patient privacy protection. In this work, we describe a new deep learning approach for automatic sleep staging, and address its generalization capabilities on a wide range of public sleep staging databases. We also examine the suitability of a novel approach that uses an ensemble of individual local models and evaluate its impact on the resulting inter-database generalization performance.

METHODS

A general deep learning network architecture for automatic sleep staging is presented. Different preprocessing and architectural variant options are tested. The resulting prediction capabilities are evaluated and compared on a heterogeneous collection of six public sleep staging datasets. Validation is carried out in the context of independent local and external dataset generalization scenarios.

RESULTS

Best results were achieved using the CNN_LSTM_5 neural network variant. Average prediction capabilities on independent local testing sets achieved 0.80 kappa score. When individual local models predict data from external datasets, average kappa score decreases to 0.54. Using the proposed ensemble-based approach, average kappa performance on the external dataset prediction scenario increases to 0.62. To our knowledge this is the largest study by the number of datasets so far on validating the generalization capabilities of an automatic sleep staging algorithm using external databases.

CONCLUSIONS

Validation results show good general performance of our method, as compared with the expected levels of human agreement, as well as to state-of-the-art automatic sleep staging methods. The proposed ensemble-based approach enables flexible and scalable design, allowing dynamic integration of local models into the final ensemble, preserving data locality, and increasing generalization capabilities of the resulting system at the same time.

A method based on cardiopulmonary coupling analysis for sleep quality assessment with FPGA implementation

The relevance of sleep quality examination for clinical diagnosis is increasing with the discovery of new relationships with several diseases and the overall wellness. This assessment is commonly performed by conducting interviews with the subjects, evaluating the self-report and psychological variables. However, this approach has a major constraint since the subject is a poor self-observer of sleep behaviors. To address this issue, a method based on the examination of a physiological signal was developed. Specifically, the single-lead electrocardiogram signal was examined to estimate the cardiopulmonary coupling between the electrocardiogram derived respiration signal and the normal-to-normal sinus interbeat interval series. A one dimensional array was created from the coupling signal and was fed to a convolutional neural network to estimate the sleep quality. The age-related cyclic alternating pattern rate percentages in healthy subjects was considered as the classification reference. An accuracy of 91 % was attained by the developed model, with an area under the receiver operating characteristic curve of 97 %. The performance is in the upper range of the reported performance by the works presented in the state of the art, advocating the relevance of the proposed method. The model was implemented in a small field programmable gate array board. Hence, a home monitoring device was created, composed of a processing unit, a sensing module and a display unit. The device is resilient, easy to self-assemble and operate, and can conceivably be employed for clinical analysis.

Greedy based convolutional neural network optimization for detecting apnea

BACKGROUND AND OBJECTIVE

Sleep apnea is a common sleep disorder, usually diagnosed using an expensive, highly specialized, and inconvenient test called polysomnography. A single SpO2 sensor based on an automated classification system can be developed to simplify the apnea detection. The main objective of this work is to develop a classifier based on a convolution neural network with the capability of detecting apnea events from one dimensional SpO2 signal. However, to find an optimum convolution neural network structure is a daunting task is usually done by a trial-and-error method. To solve this problem, a method is proposed to save time and simplify the process of searching for an optimum convolution neural network structure.

METHODS

Greedy based optimization is proposed to search for an optimized convolution neural network structure. Three different variants of greedy based optimization are proposed: the topology transfer, the weighted-topology transfer with rough estimation, and the weighted-topology transfer with fine tuning. The subject independent and the cross-database test are performed for the analysis.

RESULTS

Considering the balance between the execution time and the performance, the weighted-topology transfer with rough estimation is the best. An accuracy of 88.49% for the HuGCDN2008 database and 95.14% for the Apnea-ECG database are obtained for apnea events detection per minute. Regarding the apnea patient detection, also referred to as global classification, an accuracy of 95.71% is achieved for the HuGCDN2008 database, and 100% is achieved for the AED database without removing any subjects from both databases.

CONCLUSIONS

The proposed one-dimensional convolution neural network performs better in a similar situation than those presented in the literature. The greedy based methods, mainly the weighted-topology transfer with rough estimation, is an alternative method to extensive trial and error method.

Gaussian Mixture Models for Detecting Sleep Apnea Events Using Single Oronasal Airflow Record

Sleep apnea is a common sleep-related disorder that significantly affects the population. It is characterized by repeated breathing interruption during sleep. Such events can induce hypoxia, which is a risk factor for multiple cardiovascular and cerebrovascular diseases. Polysomnography, the gold standard, is expensive, inaccessible, uncomfortable and an expert technician is needed to score sleep-related events. To address these limitations, many previous studies have proposed and implemented automatic scoring processes based on fewer sensors and machine learning classification algorithms. However, alternative device technologies developed for both home and hospital still have limited diagnostic accuracy for detecting apnea events even though many of the previous investigational algorithms are based on multiple physiological channel inputs. In this paper, we propose a new probabilistic algorithm based on (only) oronasal respiration signal for automated detection of apnea events during sleep. The proposed model leverages AASM recommendations for characterizing apnea events with respect to dynamic changes in the local respiratory airflow baseline. Unlike classical threshold-based classification models, we use a Gaussian mixture probability model for detecting sleep apnea based on the posterior probabilities of the respective events. Our results show significant improvement in the ability to detect sleep apnea events compared to a rule-based classifier that uses the same classification features and also compared to two previously published studies for automated apnea detection using the same respiratory flow signal. We use 96 sleep patients with different apnea severity levels as reflected by their Apnea-Hypopnea Index (AHI) levels. The performance was not only analyzed over obstructive sleep apnea (OSA) but also over other types of sleep apnea events including central and mixed sleep apnea (CSA, MSA). Also the performance was comprehensively analyzed and evaluated over patients with varying disease severity conditions, where it achieved an overall performance of TPR=88.5%, TNR=82.5%, and AUC=86.7%. The proposed approach contributes a new probabilistic framework for detecting sleep apnea events using a single airflow record with an improved capability to generalize over different apnea severity conditions

Automated sleep stage scoring of the Sleep Heart Health Study using deep neural networks

STUDY OBJECTIVES

Polysomnography (PSG) scoring is labor intensive and suffers from variability in inter- and intra-rater reliability. Automated PSG scoring has the potential to reduce the human labor costs and the variability inherent to this task. Deep learning is a form of machine learning that uses neural networks to recognize data patterns by inspecting many examples rather than by following explicit programming.

METHODS

A sleep staging classifier trained using deep learning methods scored PSG data from the Sleep Heart Health Study (SHHS). The training set was composed of 42 560 hours of PSG data from 5213 patients. To capture higher-order data, spectrograms were generated from electroencephalography, electrooculography, and electromyography data and then passed to the neural network. A holdout set of 580 PSGs not included in the training set was used to assess model accuracy and discrimination via weighted F1-score, per-stage accuracy, and Cohen's kappa (K).

RESULTS

The optimal neural network model was composed of spectrograms in the input layer feeding into convolutional neural network layers and a long short-term memory layer to achieve a weighted F1-score of 0.87 and K = 0.82.

CONCLUSIONS

The deep learning sleep stage classifier demonstrates excellent accuracy and agreement with expert sleep stage scoring, outperforming human agreement on sleep staging. It achieves comparable or better F1-scores, accuracy, and Cohen's kappa compared to literature for automated sleep stage scoring of PSG epochs. Accurate automated scoring of other PSG events may eventually allow for fully automated PSG scoring.

Sleep staging algorithm based on multichannel data adding and multifeature screening

BACKGROUND

Sleep staging is an important basis of sleep research, which is closely related to both normal sleep physiology and sleep disorders. Many studies have reported various sleep staging algorithms of which the framework generally consists of three parts: signal preprocessing, feature extraction and classification. However, there are few studies on the superposition of signals and feature screening for sleep staging.

OBJECTIVE

The objectives were to (1) Analyze the effective signal enhancement based on the superposition of homologous and heterogeneous signals, (2) Find a better way to use multichannel signals, (3) Study a systematic method of feature screening for sleep staging, and (4) Improve the performance of automatic sleep staging.

METHODS

In this paper, a novel method of signal preprocessing and feature screening was proposed. In the signal preprocessing, multi-channel signal superposition was applied to improve the effective information contained in the original signal. In the feature screening, 62 features were initially selected including the time-domain features, frequency-domain features and nonlinear features, and a ReliefF algorithm was employed to select 14 features highly correlated to sleep stages from the former 62 features. Then, Pearson correlation coefficients were used to remove 2 redundant features from the 14 features to eventually obtain 12 features. Next, with the aforementioned signal preprocessing method, the 12 selected features and a support vector machine (SVM) classifier were used for sleep staging based on thirty recordings.

RESULTS

Comparing the performance of sleep staging using different single-channel signals and different multi-channel superposition signals, we found that the best performance was obtained while using the superposition of two electroencephalogram (EEG) signals. The overall accuracies of sleep staging with 2-6 classes obtained by superposing the two EEG signals reach 98.28%, 95.50%, 94.28%, 93.08% and 92.34%, respectively, and the kappa coefficient of sleep staging with 6 classes reaches 84.07%.

CONCLUSIONS

Among the proposed sleep staging methods of using single-channel signal and multi-channel signal superposition, the best performance and consistency were obtained while using the superposition of two electroencephalogram (EEG) signals. The multichannel signal superposition method pointed out a valuable direction for improving the performance of automatic sleep staging in both theoretical research and engineering applications, and the proposed systematical feature screening method opened up a reasonable pathway for better selecting type and number of features for sleep staging.

An Oximetry Based Wireless Device for Sleep Apnea Detection

Sleep related disorders can severely disturb the quality of sleep. Among these disorders, obstructive sleep apnea (OSA) is highly prevalent and commonly undiagnosed. Polysomnography is considered to be the gold standard exam for OSA diagnosis. Even though this multi-parametric test provides highly accurate results, it is time consuming, labor-intensive, and expensive. A non-invasive and easy to self-assemble home monitoring device was developed to address these issues. The device can perform the OSA diagnosis at the patient’s home and a specialized technician is not required to supervise the process. An automatic scoring algorithm was developed to examine the blood oxygen saturation signal for a minute-by-minute OSA assessment. It was performed by analyzing statistical and frequency-based features that were fed to a classifier. Afterward, the ratio of the number of minutes classified as OSA to the time in bed in minutes was compared with a threshold for the global (subject-based) OSA diagnosis. The average accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve for the minute-by-minute assessment were, respectively, 88%, 80%, 91%, and 0.86. The subject-based accuracy was 95%. The performance is in the same range as the best state of the art methods for the models based only on the blood oxygen saturation analysis. Therefore, the developed model has the potential to be employed in clinical analysis.

Spatial distribution of interictal spikes fluctuates over time and localizes seizure onset

The location of interictal spikes is used to aid surgical planning in patients with medically refractory epilepsy; however, their spatial and temporal dynamics are poorly understood. In this study, we analysed the spatial distribution of interictal spikes over time in 20 adult and paediatric patients (12 females, mean age = 34.5 years, range = 5-58) who underwent intracranial EEG evaluation for epilepsy surgery. Interictal spikes were detected in the 24 h surrounding each seizure and spikes were clustered based on spatial location. The temporal dynamics of spike spatial distribution were calculated for each patient and the effects of sleep and seizures on these dynamics were evaluated. Finally, spike location was assessed in relation to seizure onset location. We found that spike spatial distribution fluctuated significantly over time in 14/20 patients (with a significant aggregate effect across patients, Fisher's method: P < 0.001). A median of 12 sequential hours were required to capture 80% of the variability in spike spatial distribution. Sleep and postictal state affected the spike spatial distribution in 8/20 and 4/20 patients, respectively, with a significant aggregate effect (Fisher's method: P < 0.001 for each). There was no evidence of pre-ictal change in the spike spatial distribution for any patient or in aggregate (Fisher's method: P = 0.99). The electrode with the highest spike frequency and the electrode with the largest area of downstream spike propagation both localized the seizure onset zone better than predicted by chance (Wilcoxon signed-rank test: P = 0.005 and P = 0.002, respectively). In conclusion, spikes localize seizure onset. However, temporal fluctuations in spike spatial distribution, particularly in relation to sleep and post-ictal state, can confound localization. An adequate duration of intracranial recording-ideally at least 12 sequential hours-capturing both sleep and wakefulness should be obtained to sufficiently sample the interictal network.

Automated detection and localization system of myocardial infarction in single-beat ECG using Dual-Q TQWT and wavelet packet tensor decomposition

BACKGROUND AND OBJECTIVE

It is challenging to conduct real-time identification of myocardial infarction (MI) due to artifact corruption and high dimensionality of multi-lead electrocardiogram (ECG). In the present study, we proposed an automated single-beat MI detection and localization system using dual-Q tunable Q-factor wavelet transformation (Dual-Q TQWT) denoising algorithm.

METHODS

After denoising and segmentation of ECG, a fourth-order wavelet tensor (leads × subbands × samples × beats) was constructed based on the discrete wavelet packet transform (DWPT), to represent the features considering the information of inter-beat, intra-beat, inter-frequency, and inter-lead. To reduce the tensor dimension and preserve the intrinsic information, the multilinear principal component analysis (MPCA) was employed. Afterward, 84 discriminate features were fed into a classifier of bootstrap-aggregated decision trees (Treebagger). A total of 78 healthy and 328 MI (6 types) records including 57557 beats were chosen from PTB diagnostic ECG database for evaluation.

RESULTS

The validation results demonstrated that our proposed MI detection and localization system embedded with Dual-Q TQWT and wavelet packet tensor decomposition outperformed commonly used discrete wavelet transform (DWT), empirical mode decomposition (EMD) denoising methods and vector-based PCA method. With the Treebagger classifier, we obtained an accuracy of 99.98% in beat level and an accuracy of 97.46% in record level training/testing for MI detection. We also achieved an accuracy of 99.87% in beat level and an accuracy of 90.39% in record level for MI localization.

CONCLUSION

Altogether, the automated system brings potential improvement in automated detection and localization of MI in clinical practice.

Ravelo-García A, Morgado-Dias F. A Systematic Review of Detecting Sleep Apnea Using Deep Learning

Sleep apnea is a sleep related disorder that significantly affects the population. Polysomnography, the gold standard, is expensive, inaccessible, uncomfortable and an expert technician is needed to score. Numerous researchers have proposed and implemented automatic scoring processes to address these issues, based on fewer sensors and automatic classification algorithms. Deep learning is gaining higher interest due to database availability, newly developed techniques, the possibility of producing machine created features and higher computing power that allows the algorithms to achieve better performance than the shallow classifiers. Therefore, the sleep apnea research has currently gained significant interest in deep learning. The goal of this work is to analyze the published research in the last decade, providing an answer to the research questions such as how to implement the different deep networks, what kind of pre-processing or feature extraction is needed, and the advantages and disadvantages of different kinds of networks. The employed signals, sensors, databases and implementation challenges were also considered. A systematic search was conducted on five indexing services from 2008-2018. A total of 255 papers were found and 21 were selected by considering the inclusion and exclusion criteria, using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) approach.

Multichannel Sleep Stage Classification and Transfer Learning using Convolutional Neural Networks

Current sleep medicine relies on the supervised analysis of polysomnographic measurements, comprising amongst others electroencephalogram (EEG), electromyogram (EMG), and electrooculogram (EOG) signals. Convolutional neural networks (CNN) provide an interesting framework to automated classification of sleep based on these raw waveforms. In this study, we compare existing CNN approaches to four databases of pathological and physiological subjects. The best performing model resulted in Cohen's Kappa of $\kappa = 0 .75$ on healthy subjects and $\kappa = 0 .64$ on patients suffering from a variety of sleep disorders. Further, we show the advantages of additional sensor data (i.e., EOG and EMG). Deep learning approaches require a lot of data which is scarce for less prevalent diseases. For this, we propose a transfer learning procedure by pretraining a model on large public data and fine-tune this on each subject from a smaller dataset. This procedure is demonstrated using a private REM Behaviour Disorder database, improving sleep classification by 24.4%.

Hearables: Automatic Overnight Sleep Monitoring With Standardized In-Ear EEG Sensor

OBJECTIVE

Advances in sensor miniaturization and computational power have served as enabling technologies for monitoring human physiological conditions in real-world scenarios. Sleep disruption may impact neural function, and can be a symptom of both physical and mental disorders. This study proposes wearable in-ear electroencephalography (ear-EEG) for overnight sleep monitoring as a 24/7 continuous and unobtrusive technology for sleep quality assessment in the community.

METHODS

A total of 22 healthy participants took part in overnight sleep monitoring with simultaneous ear-EEG and conventional full polysomnography recordings. The ear-EEG data were analyzed in the both structural complexity and spectral domains. The extracted features were used for automatic sleep stage prediction through supervized machine learning, whereby the PSG data were manually scored by a sleep clinician.

RESULTS

The agreement between automatic sleep stage prediction based on ear-EEG from a single in-ear sensor and the hypnogram based on the full PSG was 74.1% in the accuracy over five sleep stage classification. This is supported by a substantial agreement in the kappa metric (0.61).

CONCLUSION

The in-ear sensor is feasible for monitoring overnight sleep outside the sleep laboratory and also mitigates technical difficulties associated with PSG. It, therefore, represents a 24/7 continuously wearable alternative to conventional cumbersome and expensive sleep monitoring.

SIGNIFICANCE

The "standardized" one-size-fits-all viscoelastic in-ear sensor is a next generation solution to monitor sleep-this technology promises to be a viable method for readily wearable sleep monitoring in the community, a key to affordable healthcare and future eHealth.

A Two Stage Approach for the Automatic Detection of Insomnia

Chronic insomnia can significantly impair an individual's quality of life leading to a high societal cost. Unfortunately, limited automated tools exist that can assist clinicians in the timely detection of insomnia. In this paper, we propose a two stage approach to automatically detect insomnia from an overnight EEG recording. In the first stage we trained a sleep stage scoring model and an epoch-level insomnia detection model. Both models are deep neural network (DNN)- based which are fed by a set of temporal and spectral features derived from 2 EEG channels. In the second stage we computed two subject-level feature sets. One is computed using the output of the sleep stage scoring model and consists of the sleep stage ratios, the stage pair ratios and the stage transition ratios. The second feature set is derived from the output of the epoch-level insomnia detection model and represents the ratio of detected insomniac epochs in each stage and their average posterior probability. These features are then used to train a final binary classifier to classify each subject as control, i.e., with no sleep complaints, or insomniac. We compared 5 different binary classifiers, namely the linear discriminant analysis (LDA), the classification and regression trees (CART) and the support vector machine (SVM) with linear, Gaussian and sigmoid kernels. The system was evaluated against data collected from 115 participants, 61 control and 54 with insomnia, and achieved $F1$ score, sensitivity and specificity of 0.88, 84% and 91% respectively.

Automatic Human Sleep Stage Scoring Using Deep Neural Networks

The classification of sleep stages is the first and an important step in the quantitative analysis of polysomnographic recordings. Sleep stage scoring relies heavily on visual pattern recognition by a human expert and is time consuming and subjective. Thus, there is a need for automatic classification. In this work we developed machine learning algorithms for sleep classification: random forest (RF) classification based on features and artificial neural networks (ANNs) working both with features and raw data. We tested our methods in healthy subjects and in patients. Most algorithms yielded good results comparable to human interrater agreement. Our study revealed that deep neural networks (DNNs) working with raw data performed better than feature-based methods. We also demonstrated that taking the local temporal structure of sleep into account a priori is important. Our results demonstrate the utility of neural network architectures for the classification of sleep.

EEG-Based Automatic Sleep Staging Using Ontology and Weighting Feature Analysis

Sleep staging is considered as an effective indicator for auxiliary diagnosis of sleep diseases and related psychiatric diseases, so it attracts a lot of attention from sleep researchers. Nevertheless, sleep staging based on visual inspection of tradition is subjective, time-consuming, and error-prone due to the large bulk of data which have to be processed. Therefore, automatic sleep staging is essential in order to solve these problems. In this article, an electroencephalogram- (EEG-) based scheme that is able to automatically classify sleep stages is proposed. Firstly, EEG data are preprocessed to remove artifacts, extract features, and normalization. Secondly, the normalized features and other context information are stored using an ontology-based model (OBM). Thirdly, an improved method of self-adaptive correlation analysis is designed to select the most effective EEG features. Based on these EEG features and weighting features analysis, the improved random forest (RF) is considered as the classifier to achieve the classification of sleep stages. To investigate the classification ability of the proposed method, several sets of experiments are designed and conducted to classify the sleep stages into two, three, four, and five states. The accuracy of five-state classification is 89.37%, which is improved compared to the accuracy using unimproved RF (84.37%) or previously reported classifiers. In addition, a set of controlled experiments is executed to verify the effect of the number of sleep segments (epochs) on the classification, and the results demonstrate that the proposed scheme is less affected by the sleep segments.

Achieving Accurate Automatic Sleep Staging on Manually Pre-processed EEG Data Through Synchronization Feature Extraction and Graph Metrics

Sleep staging, the process of assigning labels to epochs of sleep, depending on the stage of sleep they belong, is an arduous, time consuming and error prone process as the initial recordings are quite often polluted by noise from different sources. To properly analyze such data and extract clinical knowledge, noise components must be removed or alleviated. In this paper a pre-processing and subsequent sleep staging pipeline for the sleep analysis of electroencephalographic signals is described. Two novel methods of functional connectivity estimation (Synchronization Likelihood/SL and Relative Wavelet Entropy/RWE) are comparatively investigated for automatic sleep staging through manually pre-processed electroencephalographic recordings. A multi-step process that renders signals suitable for further analysis is initially described. Then, two methods that rely on extracting synchronization features from electroencephalographic recordings to achieve computerized sleep staging are proposed, based on bivariate features which provide a functional overview of the brain network, contrary to most proposed methods that rely on extracting univariate time and frequency features. Annotation of sleep epochs is achieved through the presented feature extraction methods by training classifiers, which are in turn able to accurately classify new epochs. Analysis of data from sleep experiments on a randomized, controlled bed-rest study, which was organized by the European Space Agency and was conducted in the "ENVIHAB" facility of the Institute of Aerospace Medicine at the German Aerospace Center (DLR) in Cologne, Germany attains high accuracy rates, over 90% based on ground truth that resulted from manual sleep staging by two experienced sleep experts. Therefore, it can be concluded that the above feature extraction methods are suitable for semi-automatic sleep staging.

Promises and Challenges in the Use of Consumer-Grade Devices for Sleep Monitoring

The market for smartphones, smartwatches, and wearable devices is booming. In recent years, individuals and researchers have used these devices as additional tools to monitor and track sleep, physical activity, and behavior. Their use in sleep research and clinical applications could address the difficulties in scaling up studies that rely on polysomnography, the gold-standard. However, the use of commercial devices for large-scale sleep studies is not without challenges. With this in mind, this paper presents an extensive review of sleep monitoring systems and the techniques used in their development. We also discuss their performance in terms of reliability and validity, and consider the needs and expectations of users, whether they are experts, patients, or the general public. Through this review, we highlight a number of challenges with current studies: a lack of standard evaluation methods for consumer-grade devices (e.g., reliability and validity assessment); limitations in the populations studied; consumer expectations of monitoring devices; constraints on the resources of consumer-grade devices (e.g., power consumption).

Automatic two-channel sleep staging using a predictor-corrector method

OBJECTIVE

We developed and implemented two predictor-corrector methods for the classification of two-channel EEG data into sleep stages.

APPROACH

The sequence of sleep stages over the night is modeled by a Markov chain of first and second order, resulting in an informative prior distribution for the new state, given the distribution of the current one. The correction step is realized by applying a Bayes classifier using the (preprocessed) data and this prior. The preprocessing step consists of a frequency analysis, a log transformation and a dimensionality reduction via principal component analysis.

MAIN RESULTS

The software automatically generates sleep profiles in which it detects wakeful phases as well as the different sleep stages with error rates of 16.5%-31.9% (n  =  8, healthy subjects, mean age  ±  SD: 39  ±  8.1 years, five females), where we compared our results to those of a certified polysomnographic technologist, who used a full polysomnograph and rated according to the American Academy of Sleep Medicine (AASM) criteria.

SIGNIFICANCE

The method presented relies on considerably less information than visual scoring and is done automatically. Furthermore, the error is comparable to visual scoring, where the inter-rater variability lies around 82%. Therefore, it has the potential to lessen the overheads associated with sleep diagnostics.

Comparing two insomnia detection models of clinical diagnosis techniques

Sleep disorders are becoming increasingly prevalent in society. However most of the burgeoning research on automated sleep analysis has been in the realm of sleep stage classification with limited focus on accurately diagnosing these disorders. In this paper, we explore two different models to discriminate between control and insomnia patients using support vector machine (SVM) classifiers. We validated the models using data collected from 124 participants, 70 control and 54 with insomnia. The first model uses 57 features derived from two channels of EEG data and achieved an accuracy of 81%. The second model uses 15 features from each participant's hypnogram and achieved an accuracy of 74%. The impetus behind using these two models is to follow the clinician's diagnostic decision-making process where both the EEG signals and the hypnograms are used. These results demonstrate that there is potential for further experimentation and improvement of the predictive capability of the models to help in diagnosing sleep disorders like insomnia.

A State Space and Density Estimation Framework for Sleep Staging in Obstructive Sleep Apnea

OBJECTIVE

Although the importance of sleep is increasingly recognized, the lack of robust and efficient algorithms hinders scalable sleep assessment in healthy persons and those with sleep disorders. Polysomnography (PSG) and visual/manual scoring remain the gold standard in sleep evaluation, but more efficient/automated systems are needed. Most previous works have demonstrated algorithms in high agreement with the gold standard in healthy/normal (HN) individuals-not those with sleep disorders.

METHODS

This paper presents a statistical framework that automatically estimates whole-night sleep architecture in patients with obstructive sleep apnea (OSA)-the most common sleep disorder. Single-channel frontal electroencephalography was extracted from 65 HN/OSA sleep studies, and decomposed into 11 spectral features in 60 903 30 s sleep epochs. The algorithm leveraged kernel density estimation to generate stage-specific likelihoods, and a 5-state hidden Markov model to estimate per-night sleep architecture.

RESULTS

Comparisons to full PSG expert scoring revealed the algorithm was in fair agreement with the gold standard (median Cohen's kappa = 0.53). Further, analysis revealed modest decreases in median scoring agreement as OSA severity increased from HN (kappa = 0.63) to severe (kappa = 0.47). A separate implementation on HN data from the Physionet Sleep-EDF Database resulted in a median kappa = 0.65, further indicating the algorithm's broad applicability.

CONCLUSION

Results of this work indicate the proposed single-channel framework can emulate expert-level scoring of sleep architecture in OSA.

SIGNIFICANCE

Algorithms constructed to more accurately model physiological variability during sleep may help advance automated sleep assessment, for practical and general use in sleep medicine.

A comparative review on sleep stage classification methods in patients and healthy individuals

BACKGROUND AND OBJECTIVE

Proper scoring of sleep stages can give clinical information on diagnosing patients with sleep disorders. Since traditional visual scoring of the entire sleep is highly time-consuming and dependent to experts' experience, automatic schemes based on electroencephalogram (EEG) analysis are broadly developed to solve these problems. This review presents an overview on the most suitable methods in terms of preprocessing, feature extraction, feature selection and classifier adopted to precisely discriminate the sleep stages.

METHODS

This study round up a wide range of research findings concerning the application of the sleep stage classification. The fundamental qualitative methods along with the state-of-the-art quantitative techniques for sleep stage scoring are comprehensively introduced. Moreover, according to the results of the investigated studies, five research papers are chosen and practically implemented on a well-known public available sleep EEG dataset. They are applied to single-channel EEG of 40 subjects containing equal number of healthy and patient individuals. Feature extraction and classification schemes are assessed in terms of accuracy and robustness against noise. Furthermore, an additional implementation phase is added to this research in which all combinations of the implemented features and classifiers are considered to find the best combination for sleep analysis.

RESULTS

According to our achieved results on both groups, entropy of wavelet coefficients along with random forest classifier are chosen as the best feature and classifier, respectively. The mentioned feature and classifier provide 87.06% accuracy on healthy subjects and 69.05% on patient group.

CONCLUSIONS

In this paper, the road map of EEG-base sleep stage scoring methods is clearly sketched. Implementing the state-of-the-art methods and even their combination on both healthy and patient datasets indicates that although the accuracy on healthy subjects are remarkable, the results for the main community (patient group) by the quantitative methods are not promising yet. The reasons rise from adopting non-matched sleep EEG features from other signal processing fields such as communication. As a conclusion, developing sleep pattern-related features deem necessary to enhance the performance of this process.