The SIESTA project polygraphic and clinical database

Transfer Learning for Automatic Sleep Staging Using a Pre-Gelled Electrode Grid

Novel sensor solutions for sleep monitoring at home could alleviate bottlenecks in sleep medical care as well as enable selective or continuous observation over long periods of time and contribute to new insights in sleep medicine and beyond. Since especially in the latter case the sensor data differ strongly in signal, number and extent of sensors from the classical polysomnography (PSG) sensor technology, an automatic evaluation is essential for the application. However, the training of an automatic algorithm is complicated by the fact that the development phase of the new sensor technology, extensive comparative measurements with standardized reference systems, is often not possible and therefore only small datasets are available. In order to circumvent high system-specific training data requirements, we employ pre-training on large datasets with finetuning on small datasets of new sensor technology to enable automatic sleep phase detection for small test series. By pre-training on publicly available PSG datasets and finetuning on 12 nights recorded with new sensor technology based on a pre-gelled electrode grid to capture electroencephalography (EEG), electrooculography (EOG) and electromyography (EMG), an F1 score across all sleep phases of 0.81 is achieved (wake 0.84, N1 0.62, N2 0.81, N3 0.87, REM 0.88), using only EEG and EOG. The analysis additionally considers the spatial distribution of the channels and an approach to approximate classical electrode positions based on specific linear combinations of the new sensor grid channels.

Research and application of deep learning-based sleep staging: Data, modeling, validation, and clinical practice

Over the past few decades, researchers have attempted to simplify and accelerate the process of sleep stage classification through various approaches; however, only a few such approaches have gained widespread acceptance. Artificial intelligence technology, particularly deep learning, is promising for earning the trust of the sleep medicine community in automated sleep-staging systems, thus facilitating its application in clinical practice and integration into daily life. We aimed to comprehensively review the latest methods that are applying deep learning for enhancing sleep staging efficiency and accuracy. Starting from the requisite "data" for constructing deep learning algorithms, we elucidated the current landscape of this domain and summarized the fundamental modeling process, encompassing signal selection, data pre-processing, model architecture, classification tasks, and performance metrics. Furthermore, we reviewed the applications of automated sleep staging in scenarios such as sleep-disorder screening, diagnostic procedures, and health monitoring and management. Finally, we conducted an in-depth analysis and discussion of the challenges and future in intelligent sleep staging, particularly focusing on large-scale sleep datasets, interdisciplinary collaborations, and human-computer interactions.

Quantifying the Suitability of Biosignals Acquired During Surgery for Multimodal Analysis

Goal: Recently, large datasets of biosignals acquired during surgery became available. As they offer multiple physiological signals measured in parallel, multimodal analysis - which involves their joint analysis - can be conducted and could provide deeper insights than unimodal analysis based on a single signal. However, it is unclear what percentage of intraoperatively acquired data is suitable for multimodal analysis. Due to the large amount of data, manual inspection and labelling into suitable and unsuitable segments are not feasible. Nevertheless, multimodal analysis is performed successfully in sleep studies since many years as their signals have proven suitable. Hence, this study evaluates the suitability to perform multimodal analysis on a surgery dataset (VitalDB) using a multi-center sleep dataset (SIESTA) as reference. Methods: We applied widely known algorithms entitled "signal quality indicators" to the common biosignals in both datasets, namely electrocardiography, electroencephalography, and respiratory signals split in segments of 10 s duration. As there are no multimodal methods available, we used only unimodal signal quality indicators. In case, all three signals were determined as being adequate by the indicators, we assumed that the whole signal segment was suitable for multimodal analysis. Results: 82% of SIESTA and 72% of VitalDB are suitable for multimodal analysis. Unsuitable signal segments exhibit constant or physiologically unreasonable values. Histogram examination indicated similar signal quality distributions between the datasets, albeit with potential statistical biases due to different measurement setups. Conclusions: The majority of data within VitalDB is suitable for multimodal analysis.

Dynamic networks of cortico-muscular interactions in sleep and neurodegenerative disorders

The brain plays central role in regulating physiological systems, including the skeleto-muscular and locomotor system. Studies of cortico-muscular coordination have primarily focused on associations between movement tasks and dynamics of specific brain waves. However, the brain-muscle functional networks of synchronous coordination among brain waves and muscle activity rhythms that underlie locomotor control remain unknown. Here we address the following fundamental questions: what are the structure and dynamics of cortico-muscular networks; whether specific brain waves are main network mediators in locomotor control; how the hierarchical network organization relates to distinct physiological states under autonomic regulation such as wake, sleep, sleep stages; and how network dynamics are altered with neurodegenerative disorders. We study the interactions between all physiologically relevant brain waves across cortical locations with distinct rhythms in leg and chin muscle activity in healthy and Parkinson's disease (PD) subjects. Utilizing Network Physiology framework and time delay stability approach, we find that 1) each physiological state is characterized by a unique network of cortico-muscular interactions with specific hierarchical organization and profile of links strength; 2) particular brain waves play role as main mediators in cortico-muscular interactions during each state; 3) PD leads to muscle-specific breakdown of cortico-muscular networks, altering the sleep-stage stratification pattern in network connectivity and links strength. In healthy subjects cortico-muscular networks exhibit a pronounced stratification with stronger links during wake and light sleep, and weaker links during REM and deep sleep. In contrast, network interactions reorganize in PD with decline in connectivity and links strength during wake and non-REM sleep, and increase during REM, leading to markedly different stratification with gradual decline in network links strength from wake to REM, light and deep sleep. Further, we find that wake and sleep stages are characterized by specific links strength profiles, which are altered with PD, indicating disruption in the synchronous activity and network communication among brain waves and muscle rhythms. Our findings demonstrate the presence of previously unrecognized functional networks and basic principles of brain control of locomotion, with potential clinical implications for novel network-based biomarkers for early detection of Parkinson's and neurodegenerative disorders, movement, and sleep disorders.

A computationally efficient algorithm for wearable sleep staging in clinical populations

This study describes a computationally efficient algorithm for 4-class sleep staging based on cardiac activity and body movements. Using an accelerometer to calculate gross body movements and a reflective photoplethysmographic (PPG) sensor to determine interbeat intervals and a corresponding instantaneous heart rate signal, a neural network was trained to classify between wake, combined N1 and N2, N3 and REM sleep in epochs of 30 s. The classifier was validated on a hold-out set by comparing the output against manually scored sleep stages based on polysomnography (PSG). In addition, the execution time was compared with that of a previously developed heart rate variability (HRV) feature-based sleep staging algorithm. With a median epoch-per-epoch κ of 0.638 and accuracy of 77.8% the algorithm achieved an equivalent performance when compared to the previously developed HRV-based approach, but with a 50-times faster execution time. This shows how a neural network, without leveraging any a priori knowledge of the domain, can automatically "discover" a suitable mapping between cardiac activity and body movements, and sleep stages, even in patients with different sleep pathologies. In addition to the high performance, the reduced complexity of the algorithm makes practical implementation feasible, opening up new avenues in sleep diagnostics.

Compensatory mechanisms of reduced interhemispheric EEG connectivity during sleep in patients with apnea

We performed a mathematical analysis of functional connectivity in electroencephalography (EEG) of patients with obstructive sleep apnea (OSA) (N = 10; age: 52.8 ± 13 years; median age: 49 years; male/female ratio: 7/3), compared with a group of apparently healthy participants (N = 15; age: 51.5 ± 29.5 years; median age: 42 years; male/female ratio: 8/7), based on the calculation of wavelet bicoherence from nighttime polysomnograms. Having observed the previously known phenomenon of interhemispheric synchronization deterioration, we demonstrated a compensatory increase in intrahemispheric connectivity, as well as a slight increase in the connectivity of the central and occipital areas for high-frequency EEG activity. Significant changes in functional connectivity were extremely stable in groups of apparently healthy participants and OSA patients, maintaining the overall pattern when comparing different recording nights and various sleep stages. The maximum variability of the connectivity was observed at fast oscillatory processes during REM sleep. The possibility of observing some changes in functional connectivity of brain activity in OSA patients in a state of passive wakefulness opens up prospects for further research. Developing the methods of hypnogram evaluation that are independent of functional connectivity may be useful for implementing a medical decision support system.

A novel automated robust dual-channel EEG-based sleep scoring system using optimal half-band pair linear-phase biorthogonal wavelet filter bank

Nowadays, the hectic work life of people has led to sleep deprivation. This may further result in sleep-related disorders and adverse physiological conditions. Therefore, sleep study has become an active research area. Sleep scoring is crucial for detecting sleep-related disorders like sleep apnea, insomnia, narcolepsy, periodic leg movement (PLM), and restless leg syndrome (RLS). Sleep is conventionally monitored in a sleep laboratory using polysomnography (PSG) which is the recording of various physiological signals. The traditional sleep stage scoring (SSG) done by professional sleep scorers is a tedious, strenuous, and time-consuming process as it is manual. Hence, developing a machine-learning model for automatic SSG is essential. In this study, we propose an automated SSG approach based on the biorthogonal wavelet filter bank's (BWFB) novel least squares (LS) design. We have utilized a huge Wisconsin sleep cohort (WSC) database in this study. The proposed study is a pioneering work on automatic sleep stage classification using the WSC database, which includes good sleepers and patients suffering from various sleep-related disorders, including apnea, insomnia, hypertension, diabetes, and asthma. To investigate the generalization of the proposed system, we evaluated the proposed model with the following publicly available databases: cyclic alternating pattern (CAP), sleep EDF, ISRUC, MIT-BIH, and the sleep apnea database from St. Vincent's University. This study uses only two unipolar EEG channels, namely O1-M2 and C3-M2, for the scoring. The Hjorth parameters (HP) are extracted from the wavelet subbands (SBS) that are obtained from the optimal BWFB. To classify sleep stages, the HP features are fed to several supervised machine learning classifiers. 12 different datasets have been created to develop a robust model. A total of 12 classification tasks (CT) have been conducted employing various classification algorithms. Our developed model achieved the best accuracy of 83.2% and Cohen's Kappa of 0.7345 to reliably distinguish five sleep stages, using an ensemble bagged tree classifier with 10-fold cross-validation using WSC data. We also observed that our system is either better or competitive with existing state-of-art systems when we tested with the above-mentioned five databases other than WSC. This method yielded promising results using only two EEG channels using a huge WSC database. Our approach is simple and hence, the developed model can be installed in home-based clinical systems and wearable devices for sleep scoring.

Scoring sleep with artificial intelligence enables quantification of sleep stage ambiguity: hypnodensity based on multiple expert scorers and auto-scoring

STUDY OBJECTIVES

To quantify the amount of sleep stage ambiguity across expert scorers and to validate a new auto-scoring platform against sleep staging performed by multiple scorers.

METHODS

We applied a new auto-scoring system to three datasets containing 95 PSGs scored by 6-12 scorers, to compare sleep stage probabilities (hypnodensity; i.e. the probability of each sleep stage being assigned to a given epoch) as the primary output, as well as a single sleep stage per epoch assigned by hierarchical majority rule.

RESULTS

The percentage of epochs with 100% agreement across scorers was 46 ± 9%, 38 ± 10% and 32 ± 9% for the datasets with 6, 9, and 12 scorers, respectively. The mean intra-class correlation coefficient between sleep stage probabilities from auto- and manual-scoring was 0.91, representing excellent reliability. Within each dataset, agreement between auto-scoring and consensus manual-scoring was significantly higher than agreement between manual-scoring and consensus manual-scoring (0.78 vs. 0.69; 0.74 vs. 0.67; and 0.75 vs. 0.67; all p < 0.01).

CONCLUSIONS

Analysis of scoring performed by multiple scorers reveals that sleep stage ambiguity is the rule rather than the exception. Probabilities of the sleep stages determined by artificial intelligence auto-scoring provide an excellent estimate of this ambiguity. Compared to consensus manual-scoring, sleep staging derived from auto-scoring is for each individual PSG noninferior to manual-scoring meaning that auto-scoring output is ready for interpretation without the need for manual adjustment.

Contactless Camera-Based Sleep Staging: The HealthBed Study

Polysomnography (PSG) remains the gold standard for sleep monitoring but is obtrusive in nature. Advances in camera sensor technology and data analysis techniques enable contactless monitoring of heart rate variability (HRV). In turn, this may allow remote assessment of sleep stages, as different HRV metrics indirectly reflect the expression of sleep stages. We evaluated a camera-based remote photoplethysmography (PPG) setup to perform automated classification of sleep stages in near darkness. Based on the contactless measurement of pulse rate variability, we use a previously developed HRV-based algorithm for 3 and 4-class sleep stage classification. Performance was evaluated on data of 46 healthy participants obtained from simultaneous overnight recording of PSG and camera-based remote PPG. To validate the results and for benchmarking purposes, the same algorithm was used to classify sleep stages based on the corresponding ECG data. Compared to manually scored PSG, the remote PPG-based algorithm achieved moderate agreement on both 3 class (Wake-N1/N2/N3-REM) and 4 class (Wake-N1/N2-N3-REM) classification, with average κ of 0.58 and 0.49 and accuracy of 81% and 68%, respectively. This is in range with other performance metrics reported on sensing technologies for wearable sleep staging, showing the potential of video-based non-contact sleep staging.

The Future of Sleep Staging, Revisited

In 1996, I published a paper entitled "The Future of Sleep Staging". At this time, paper and ink records were the standard way of recording sleep records. Computerised systems had only recently become commercially available. The original article was a response to those initial computer-based systems, pointing out the potential limitations of the systems. Now, digital sleep recording is ubiquitous and software and hardware capabilities have improved immeasurably. However, I will argue that despite 50 years of progress, there has not been an increase in the accuracy of sleep staging. I will propose that this is due to the limitations of the task that we have set the automatic analysis methods.

Sleep Quality Evaluation Based on Single-Lead Wearable Cardiac Cycle Acquisition Device

In clinical conditions, polysomnography (PSG) is regarded as the "golden standard" for detecting sleep disease and offering a reference of objective sleep quality. For healthy adults, scores from sleep questionnaires are more reliable than other methods in obtaining knowledge of subjective sleep quality. In practice, the need to simplify PSG to obtain subjective sleep quality by recording a few channels of physiological signals such as single-lead electrocardiogram (ECG) or photoplethysmography (PPG) signal is still very urgent. This study provided a two-step method to differentiate sleep quality into "good sleep" and "poor sleep" based on the single-lead wearable cardiac cycle data, with the comparison of the subjective sleep questionnaire score. First, heart rate variability (HRV) features and ECG-derived respiration features were extracted to construct a sleep staging model (wakefulness (W), rapid eye movement (REM), light sleep (N1&N2) and deep sleep (N3)) using the multi-classifier fusion method. Then, features extracted from the sleep staging results were used to construct a sleep quality evaluation model, i.e., classifying the sleep quality as good and poor. The accuracy of the sleep staging model, tested on the international public database, was 0.661 and 0.659 in Cardiology Challenge 2018 training database and Sleep Heart Health Study Visit 1 database, respectively. The accuracy of the sleep quality evaluation model was 0.786 for our recording subjects, with an average F₁-score of 0.771. The proposed sleep staging model and sleep quality evaluation model only requires one channel of wearable cardiac cycle signal. It is very easy to transplant to portable devices, which facilitates daily sleep health monitoring.

A systematic review of deep learning methods for modeling electrocardiograms during sleep

Sleep is one of the most important human physiological activities and plays an essential role in human health. Polysomnography (PSG) is the gold standard for measuring sleep quality and disorders, but it is time-consuming, labor-intensive, and prone to errors. Current research has confirmed the correlations between sleep and the respiratory/circulatory system. Electrocardiography (ECG) is convenient to perform, and ECG data are rich in breathing information. Therefore, sleep research based on ECG data has become popular. Currently, deep learning (DL) methods have achieved promising results on predictive health care tasks using ECG signals. Therefore, in this review, we systematically identify recent research studies and analyze them from the perspectives of data, model, and task. We discuss the shortcomings, summarize the findings, and highlight the potential opportunities. For sleep-related tasks, many ECG-based DL methods produce more accurate results than traditional approaches by combining multiple signal features and model structures. Methods that are more interpretable, scalable, and transferable will become ubiquitous in the daily practice of medicine and ambient-assisted-living applications. This paper is the first systematic review of ECG-based DL methods for sleep tasks.

Dynamic networks of physiologic interactions of brain waves and rhythms in muscle activity

The brain plays a central role in facilitating vital body functions and in regulating physiological and organ systems, including the skeleto-muscular and locomotor system. While neural control is essential to synchronize and coordinate activation of various muscle groups and muscle fibers within muscle groups in relation to body movements and distinct physiologic states, the dynamic networks of brain-muscle interactions have not been explored and the complex regulatory mechanism of brain-muscle control remains unknown. Here we present a first study of network interactions between brain waves at different cortical locations and peripheral muscle activity across key physiologic states - wake, sleep and distinct sleep stages. Utilizing a novel approach based on the Network Physiology framework and the concept of time delay stability, we find that for each physiologic state the network of cortico-muscular interactions is characterized by a specific hierarchical organization of network topology and network links strength, where particular brain waves are main mediators of interaction and control of muscular activity. Further, we uncover that with transition from one physiological state to another, the brain-muscle interaction network undergoes marked reorganization in the profile of network links strength, indicating a direct association between network structure and physiological state and function. The pronounced stratification in brain-muscle network characteristics across sleep stages is consistent for chin and leg muscle groups and persists across subjects, indicating a remarkable universality and a previously unrecognized basic physiologic mechanism that regulates muscle activity even during rest and in the absence targeted direct movement. Our findings demonstrate previously unrecognized coordination between brain waves and activation of different muscle fiber types within muscle groups, laws of brain-muscle cross-communication and principles of network integration and control. These investigations demonstrate the potential of network-based biomarkers for classification of distinct physiological states and conditions, for the diagnosis and prognosis of neurodegenerative, movement and sleep disorders, and for developing efficient treatment strategies.

Investigating the relationship between objective measures of sleep and self-report sleep quality in healthy adults: a review

STUDY OBJECTIVES

Sleep is one of the most common factors related to health, yet a standard definition of sleep quality has not been identified. Polysomnography provides important information about objective sleep variables. However, the relationship between objective sleep variables and perception of sleep quality remains unclear. The purpose of this review was to (1) summarize the current methods of measuring objective sleep macrostructure and microstructure, including electroencephalography arousals, spectral frequency, cyclic alternating pattern, and self-report sleep quality, and (2) investigate the relationship between objective measures of sleep physiology and self-report sleep quality in healthy adults.

METHODS

A literature search was conducted using Medline, PubMed, and PsycInfo databases and cited reference searches. Eligible studies included a comparison between self-report sleep quality and polysomnography sleep measures in healthy adults.

RESULTS

Thirteen studies were identified. Measurement of self-report sleep quality varied widely across studies. Total sleep time and sleep efficiency were most consistently related to sleep quality, while other objective sleep variables, including electroencephalography spectral analysis, were not reliably predictive of self-report sleep quality in healthy adults. There is preliminary support that microstructural sleep analysis with cyclic alternating pattern may be related to self-reported sleep quality.

CONCLUSIONS

Further research is needed to define and standardize self-report measures of sleep quality and investigate the microstructure of sleep. Objective measures of sleep and experiences of "quality" sleep are not as closely related as one may expect in healthy individuals, and understanding this relationship further is necessary to improve the clinical utility of sleep physiology.

CITATION

Cudney LE, Frey BN, McCabe RE, Green SM. Investigating the relationship between objective measures of sleep and self-report sleep quality in healthy adults: a review. J Clin Sleep Med. 2022;18(3):927-936.

Automated Scoring of Sleep and Associated Events

Conventionally, sleep and associated events are scored visually by trained technologists according to the rules summarized in the American Academy of Sleep Medicine Manual. Since its first publication in 2007, the manual was continuously updated; the most recent version as of this writing was published in 2020. Human expert scoring is considered as gold standard, even though there is increasing evidence of limited interrater reliability between human scorers. Significant advances in machine learning have resulted in powerful methods for addressing complex classification problems such as automated scoring of sleep and associated events. Evidence is increasing that these autoscoring systems deliver performance comparable to manual scoring and offer several advantages to visual scoring: (1) avoidance of the rather expensive, time-consuming, and difficult visual scoring task that can be performed only by well-trained and experienced human scorers, (2) attainment of consistent scoring results, and (3) proposition of added value such as scoring in real time, sleep stage probabilities per epoch (hypnodensity), estimates of signal quality and sleep/wake-related features, identifications of periods with clinically relevant ambiguities (confidence trends), configurable sensitivity and rule settings, as well as cardiorespiratory sleep staging for home sleep apnea testing. This chapter describes the development of autoscoring systems since the first attempts in the 1970s up to the most recent solutions based on deep neural network approaches which achieve an accuracy that allows to use the autoscoring results directly for review and interpretation by a physician.

The Reconstruction of Causal Networks in Physiology

We systematically compare strengths and weaknesses of two methods that can be used to quantify causal links between time series: Granger-causality and Bivariate Phase Rectified Signal Averaging (BPRSA). While a statistical test method for Granger-causality has already been established, we show that BPRSA causality can also be probed with existing statistical tests. Our results indicate that more data or stronger interactions are required for the BPRSA method than for the Granger-causality method to detect an existing link. Furthermore, the Granger-causality method can distinguish direct causal links from indirect links as well as links that arise from a common source, while BPRSA cannot. However, in contrast to Granger-causality, BPRSA is suited for the analysis of non-stationary data. We demonstrate the practicability of the Granger-causality method by applying it to polysomnography data from sleep laboratories. An algorithm is presented, which addresses the stationarity condition of Granger-causality by splitting non-stationary data into shorter segments until they pass a stationarity test. We reconstruct causal networks of heart rate, breathing rate, and EEG amplitude from young healthy subjects, elderly healthy subjects, and subjects with obstructive sleep apnea, a condition that leads to disruption of normal respiration during sleep. These networks exhibit differences not only between different sleep stages, but also between young and elderly healthy subjects on the one hand and subjects with sleep apnea on the other hand. Among these differences are 1) weaker interactions in all groups between heart rate, breathing rate and EEG amplitude during deep sleep, compared to light and REM sleep, 2) a stronger causal link from heart rate to breathing rate but disturbances in respiratory sinus arrhythmia (breathing to heart rate coupling) in subjects with sleep apnea, 3) a stronger causal link from EEG amplitude to breathing rate during REM sleep in subjects with sleep apnea. The Granger-causality method, although initially developed for econometric purposes, can provide a quantitative, testable measure for causality in physiological networks.

Directional couplings between the respiration and parasympathetic control of the heart rate during sleep and wakefulness in healthy subjects at different ages

Cardiorespiratory interactions are important, both for understanding the fundamental processes of functioning of the human body and for development of methods for diagnostics of various pathologies. The properties of cardiorespiratory interaction are determined by the processes of autonomic control of blood circulation, which are modulated by the higher nervous activity. We study the directional couplings between the respiration and the process of parasympathetic control of the heart rate in the awake state and different stages of sleep in 96 healthy subjects from different age groups. The detection of directional couplings is carried out using the method of phase dynamics modeling applied to experimental RR-intervals and the signal of respiration. We reveal the presence of bidirectional couplings between the studied processes in all age groups. Our results show that the coupling from respiration to the process of parasympathetic control of the heart rate is stronger than the coupling in the opposite direction. The difference in the strength of bidirectional couplings between the considered processes is most pronounced in deep sleep.

Directional Couplings Between Electroencephalogram and Interbeat Intervals Signals in Awake State and Different Stages of Sleep

Purpose of the work is to identify the directional coupling between the structures of the brain and the autonomic control of the heart rate variability, to analyze the changes in these coupling in sleep and in wakefulness. Infra-slow oscillations of the electroencephalograms potential and low-frequency components (0.04-0.15 Hz) of the interbeat intervals signal where analyzed using a sensitive method for identifying the directional coupling. The technique, based on modeling the dynamics of instantaneous phases of oscillations, made it possible to reveal the presence and quantify the directional couplings between the structures of the brain and the autonomic control of the heart rate variability. It was shown that the coupling coefficients in the frequency band of 0.04-0.15 Hz (associated mainly with sympathetic control of blood circulation), on average, decrease with falling asleep. We have also shown the asymmetry of coupling. At the same time, stronger connections were revealed in the direction from the autonomic control of the heart rate variability to the brain structures than in the opposite direction. It has been shown that the strength of such couplings decreases with increasing of sleep depth.

A deep transfer learning approach for wearable sleep stage classification with photoplethysmography

Unobtrusive home sleep monitoring using wrist-worn wearable photoplethysmography (PPG) could open the way for better sleep disorder screening and health monitoring. However, PPG is rarely included in large sleep studies with gold-standard sleep annotation from polysomnography. Therefore, training data-intensive state-of-the-art deep neural networks is challenging. In this work a deep recurrent neural network is first trained using a large sleep data set with electrocardiogram (ECG) data (292 participants, 584 recordings) to perform 4-class sleep stage classification (wake, rapid-eye-movement, N1/N2, and N3). A small part of its weights is adapted to a smaller, newer PPG data set (60 healthy participants, 101 recordings) through three variations of transfer learning. Best results (Cohen's kappa of 0.65 ± 0.11, accuracy of 76.36 ± 7.57%) were achieved with the domain and decision combined transfer learning strategy, significantly outperforming the PPG-trained and ECG-trained baselines. This performance for PPG-based 4-class sleep stage classification is unprecedented in literature, bringing home sleep stage monitoring closer to clinical use. The work demonstrates the merit of transfer learning in developing reliable methods for new sensor technologies by reusing similar, older non-wearable data sets. Further study should evaluate our approach in patients with sleep disorders such as insomnia and sleep apnoea.

Automatic sleep staging by cardiorespiratory signals: a systematic review

BACKGROUND

Because of problems with the recording and analysis of the EEG signal, automatic sleep staging using cardiorespiratory signals has been employed as an alternative. This study reports on certain critical points which hold considerable promise for the improvement of the results of the automatic sleep staging using cardiorespiratory signals.

METHODS

A systematic review.

RESULTS

The review and analysis of the literature in this area revealed four outstanding points: (1) the feature extraction epoch length, denoting that the standard 30-s segments of cardiorespiratory signals do not carry enough information for automatic sleep staging and that a 4.5-min length segment centering on each 30-s segment is proper for staging, (2) the time delay between the EEG signal extracted from the central nervous system activity and the cardiorespiratory signals extracted from the autonomic nervous system activity should be considered in the automatic sleep staging using cardiorespiratory signals, (3) the information in the morphology of ECG signals can contribute to the improvement of sleep staging, and (4) applying convolutional neural network (CNN) and long short-term memory network (LSTM) deep structures simultaneously to a large PSG recording database can lead to more reliable automatic sleep staging results.

CONCLUSIONS

Considering the above-mentioned points simultaneously can improve automatic sleep staging by cardiorespiratory signals. It is hoped that by considering the points, staging sleep automatically using cardiorespiratory signals, which does not have problems with the recording and analysis of EEG signals, yields results acceptably close to the results of automatic sleep staging by EEG signals.

Decrease of coherence between the respiration and parasympathetic control of the heart rate with aging

The study of coordinated behavior between different systems of the human body provides useful information on the functioning of the body. The peculiarities of interaction and coordinated dynamics of the heart rate and respiration are of particular interest. We investigated the coherence of the processes of respiration and autonomic control of the heart rate for people of different ages in the awake state, in sleep with rapid eye movement, and in deep sleep. Our analysis revealed a monotonic decrease in the coherence of these processes with increasing age. This can be explained by age-related changes in the system of autonomic control of circulation. For all age groups, we found a qualitatively similar dynamics of the coherence between the studied processes during a transition from the awake state to sleep.

Estimating sleep stages using cardiorespiratory signals: validation of a novel algorithm across a wide range of sleep-disordered breathing severity

STUDY OBJECTIVES

We have developed the CardioRespiratory Sleep Staging (CReSS) algorithm for estimating sleep stages using heart rate variability and respiration, allowing for estimation of sleep staging during home sleep apnea tests. Our objective was to undertake an epoch-by-epoch validation of algorithm performance against the gold standard of manual polysomnography sleep staging.

METHODS

Using 296 polysomnographs, we created a limited montage of airflow and heart rate and deployed CReSS to identify each 30-second epoch as wake, light sleep (N1 + N2), deep sleep (N3), or rapid eye movement (REM) sleep. We calculated Cohen's kappa and the percentage of accurately identified epochs. We repeated our analyses after stratification by sleep-disordered breathing (SDB) severity, and after adding thoracic respiratory effort as a backup signal for periods of invalid airflow.

RESULTS

CReSS discriminated wake/light sleep/deep sleep/REM sleep with 78% accuracy; the kappa value was 0.643 (95% confidence interval, 0.641-0.645). Discrimination of wake/sleep demonstrated a kappa value of 0.711 and accuracy of 89%, non-REM sleep/REM sleep demonstrated a kappa of 0.790 and accuracy of 94%, and light sleep/deep sleep demonstrated a kappa of 0.469 and accuracy of 87%. Kappa values did not vary by more than 0.07 across subgroups of no SDB, mild SDB, moderate SDB, and severe SDB. Accuracy increased to 80%, with a kappa value of 0.680 (95% confidence interval, 0.678-0.682), when CReSS additionally utilized the thoracic respiratory effort signal.

CONCLUSIONS

We observed substantial agreement between CReSS and the gold-standard comparator of manual sleep staging of polysomnographic signals, which was consistent across the full range of SDB severity. Future research should focus on the extent to which CReSS reduces the discrepancy between the apnea-hypopnea index and the respiratory event index, and the ability of CReSS to identify REM sleep-related obstructive sleep apnea.

Automatic analysis of single-channel sleep EEG in a large spectrum of sleep disorders

STUDY OBJECTIVES

To assess the performance of the single-channel automatic sleep staging (AS) software ASEEGA in adult patients diagnosed with various sleep disorders.

METHODS

Sleep recordings were included of 95 patients (38 women, 40.5 ± 13.7 years) diagnosed with insomnia (n = 23), idiopathic hypersomnia (n = 24), narcolepsy (n = 24), and obstructive sleep apnea (n = 24). Visual staging (VS) was performed by two experts (VS1 and VS2) according to the American Academy of Sleep Medicine rules. AS was based on the analysis of a single electroencephalogram channel (Cz-Pz), without any information from electro-oculography nor electromyography. The epoch-by-epoch agreement (concordance and Conger's coefficient [κ]) was compared pairwise (VS1-VS2, AS-VS1, AS-VS2) and between AS and consensual VS. Sleep parameters were also compared.

RESULTS

The pairwise agreements were: between AS and VS1, 78.6% (κ = 0.70); AS and VS2, 75.0% (0.65); and VS1 and VS2, 79.5% (0.72). Agreement between AS and consensual VS was 85.6% (0.80), with the following distribution: insomnia 85.5% (0.80), narcolepsy 83.8% (0.78), idiopathic hypersomnia 86.1% (0.68), and obstructive sleep disorder 87.2% (0.82). A significant low-amplitude scorer effect was observed for most sleep parameters, not always driven by the same scorer. Hypnograms obtained with AS and VS exhibited very close sleep organization, except for 80% of rapid eye movement sleep onset in the group diagnosed with narcolepsy missed by AS.

CONCLUSIONS

Agreement between AS and VS in sleep disorders is comparable to that reported in healthy individuals and to interexpert agreement in patients. ASEEGA could therefore be considered as a complementary sleep stage scoring tool in clinical practice, after improvement of rapid eye movement sleep onset detection.

U-Sleep: resilient high-frequency sleep staging

Sleep disorders affect a large portion of the global population and are strong predictors of morbidity and all-cause mortality. Sleep staging segments a period of sleep into a sequence of phases providing the basis for most clinical decisions in sleep medicine. Manual sleep staging is difficult and time-consuming as experts must evaluate hours of polysomnography (PSG) recordings with electroencephalography (EEG) and electrooculography (EOG) data for each patient. Here, we present U-Sleep, a publicly available, ready-to-use deep-learning-based system for automated sleep staging ( sleep.ai.ku.dk ). U-Sleep is a fully convolutional neural network, which was trained and evaluated on PSG recordings from 15,660 participants of 16 clinical studies. It provides accurate segmentations across a wide range of patient cohorts and PSG protocols not considered when building the system. U-Sleep works for arbitrary combinations of typical EEG and EOG channels, and its special deep learning architecture can label sleep stages at shorter intervals than the typical 30 s periods used during training. We show that these labels can provide additional diagnostic information and lead to new ways of analyzing sleep. U-Sleep performs on par with state-of-the-art automatic sleep staging systems on multiple clinical datasets, even if the other systems were built specifically for the particular data. A comparison with consensus-scores from a previously unseen clinic shows that U-Sleep performs as accurately as the best of the human experts. U-Sleep can support the sleep staging workflow of medical experts, which decreases healthcare costs, and can provide highly accurate segmentations when human expertize is lacking.

Automatic sleep staging using heart rate variability, body movements, and recurrent neural networks in a sleep disordered population

STUDY OBJECTIVES

To validate a previously developed sleep staging algorithm using heart rate variability (HRV) and body movements in an independent broad cohort of unselected sleep disordered patients.

METHODS

We applied a previously designed algorithm for automatic sleep staging using long short-term memory recurrent neural networks to model sleep architecture. The classifier uses 132 HRV features computed from electrocardiography and activity counts from accelerometry. We retrained our algorithm using two public datasets containing both healthy sleepers and sleep disordered patients. We then tested the performance of the algorithm on an independent hold-out validation set of sleep recordings from a wide range of sleep disorders collected in a tertiary sleep medicine center.

RESULTS

The classifier achieved substantial agreement on four-class sleep staging (wake/N1-N2/N3/rapid eye movement [REM]), with an average κ of 0.60 and accuracy of 75.9%. The performance of the sleep staging algorithm was significantly higher in insomnia patients (κ = 0.62, accuracy = 77.3%). Only in REM parasomnias, the performance was significantly lower (κ = 0.47, accuracy = 70.5%). For two-class wake/sleep classification, the classifier achieved a κ of 0.65, with a sensitivity (to wake) of 72.9% and specificity of 94.0%.

CONCLUSIONS

This study shows that the combination of HRV, body movements, and a state-of-the-art deep neural network can reach substantial agreement in automatic sleep staging compared with polysomnography, even in patients suffering from a multitude of sleep disorders. The physiological signals required can be obtained in various ways, including non-obtrusive wrist-worn sensors, opening up new avenues for clinical diagnostics.

Simulating Dynamics of Circulation in the Awake State and Different Stages of Sleep Using Non-autonomous Mathematical Model With Time Delay

We propose a mathematical model of the human cardiovascular system. The model allows one to simulate the main heart rate, its variability under the influence of the autonomic nervous system, breathing process, and oscillations of blood pressure. For the first time, the model takes into account the activity of the cerebral cortex structures that modulate the autonomic control loops of blood circulation in the awake state and in various stages of sleep. The adequacy of the model is demonstrated by comparing its time series with experimental records of healthy subjects in the SIESTA database. The proposed model can become a useful tool for studying the characteristics of the cardiovascular system dynamics during sleep.

Automated Detection of Sleep Stages Using Deep Learning Techniques: A Systematic Review of the Last Decade (2010–2020)

Sleep is vital for one’s general well-being, but it is often neglected, which has led to an increase in sleep disorders worldwide. Indicators of sleep disorders, such as sleep interruptions, extreme daytime drowsiness, or snoring, can be detected with sleep analysis. However, sleep analysis relies on visuals conducted by experts, and is susceptible to inter- and intra-observer variabilities. One way to overcome these limitations is to support experts with a programmed diagnostic tool (PDT) based on artificial intelligence for timely detection of sleep disturbances. Artificial intelligence technology, such as deep learning (DL), ensures that data are fully utilized with low to no information loss during training. This paper provides a comprehensive review of 36 studies, published between March 2013 and August 2020, which employed DL models to analyze overnight polysomnogram (PSG) recordings for the classification of sleep stages. Our analysis shows that more than half of the studies employed convolutional neural networks (CNNs) on electroencephalography (EEG) recordings for sleep stage classification and achieved high performance. Our study also underscores that CNN models, particularly one-dimensional CNN models, are advantageous in yielding higher accuracies for classification. More importantly, we noticed that EEG alone is not sufficient to achieve robust classification results. Future automated detection systems should consider other PSG recordings, such as electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG) signals, along with input from human experts, to achieve the required sleep stage classification robustness. Hence, for DL methods to be fully realized as a practical PDT for sleep stage scoring in clinical applications, inclusion of other PSG recordings, besides EEG recordings, is necessary. In this respect, our report includes methods published in the last decade, underscoring the use of DL models with other PSG recordings, for scoring of sleep stages.

Network Physiology of Cortico-Muscular Interactions

Skeletal muscle activity is continuously modulated across physiologic states to provide coordination, flexibility and responsiveness to body tasks and external inputs. Despite the central role the muscular system plays in facilitating vital body functions, the network of brain-muscle interactions required to control hundreds of muscles and synchronize their activation in relation to distinct physiologic states has not been investigated. Recent approaches have focused on general associations between individual brain rhythms and muscle activation during movement tasks. However, the specific forms of coupling, the functional network of cortico-muscular coordination, and how network structure and dynamics are modulated by autonomic regulation across physiologic states remains unknown. To identify and quantify the cortico-muscular interaction network and uncover basic features of neuro-autonomic control of muscle function, we investigate the coupling between synchronous bursts in cortical rhythms and peripheral muscle activation during sleep and wake. Utilizing the concept of time delay stability and a novel network physiology approach, we find that the brain-muscle network exhibits complex dynamic patterns of communication involving multiple brain rhythms across cortical locations and different electromyographic frequency bands. Moreover, our results show that during each physiologic state the cortico-muscular network is characterized by a specific profile of network links strength, where particular brain rhythms play role of main mediators of interaction and control. Further, we discover a hierarchical reorganization in network structure across physiologic states, with high connectivity and network link strength during wake, intermediate during REM and light sleep, and low during deep sleep, a sleep-stage stratification that demonstrates a unique association between physiologic states and cortico-muscular network structure. The reported empirical observations are consistent across individual subjects, indicating universal behavior in network structure and dynamics, and high sensitivity of cortico-muscular control to changes in autonomic regulation, even at low levels of physical activity and muscle tone during sleep. Our findings demonstrate previously unrecognized basic principles of brain-muscle network communication and control, and provide new perspectives on the regulatory mechanisms of brain dynamics and locomotor activation, with potential clinical implications for neurodegenerative, movement and sleep disorders, and for developing efficient treatment strategies.

Reinventing polysomnography in the age of precision medicine

For almost 50 years, sleep laboratories around the world have been collecting massive amounts of polysomnographic (PSG) physiological data to diagnose sleep disorders, the majority of which are not utilized in the clinical setting. Only a small fraction of the information available within these signals is utilized to generate indices. For example, the apnea-hypopnea index (AHI) remains the primary tool for diagnostic and therapeutic decision-making for obstructive sleep apnea (OSA) despite repeated studies showing it to be inadequate in predicting clinical consequences. Today, there are many novel approaches to PSG signals, making it possible to extract more complex metrics and analyses that are potentially more clinically relevant for individual patients. However, the pathway to implement novel PSG metrics/analyses into routine clinical practice is unclear. Our goal with this review is to highlight some of the novel PSG metrics/analyses that are becoming available. We suggest that stronger academic-industry relationships would facilitate the development of state-of-the-art clinical research to establish the value of novel PSG metrics/analyses in clinical sleep medicine. Collectively, as a sleep community, it is time to reinvent how we utilize the polysomnography to move us towards Precision Sleep Medicine.

Dynamic network interactions among distinct brain rhythms as a hallmark of physiologic state and function

Brain rhythms are associated with a range of physiologic states, and thus, studies have traditionally focused on neuronal origin, temporal dynamics and fundamental role of individual brain rhythms, and more recently on specific pair-wise interactions. Here, we aim to understand integrated physiologic function as an emergent phenomenon of dynamic network interactions among brain rhythms. We hypothesize that brain rhythms continuously coordinate their activations to facilitate physiologic states and functions. We analyze healthy subjects during sleep, and we demonstrate the presence of stable interaction patterns among brain rhythms. Probing transient modulations in brain wave activation, we discover three classes of interaction patterns that form an ensemble representative for each sleep stage, indicating an association of each state with a specific network of brain-rhythm communications. The observations are universal across subjects and identify networks of brain-rhythm interactions as a hallmark of physiologic state and function, providing new insights on neurophysiological regulation with broad clinical implications.

Central Sleep Apnea Alters Neuronal Excitability and Increases the Randomness in Sleep-Wake Transitions

OBJECTIVE

While most studies on Central Sleep Apnea (CSA) have focused on breathing and metabolic disorders, the neuronal dysfunction that causes CSA remains largely unknown. Here, we investigate the underlying neuronal mechanism of CSA by studying the sleep-wake dynamics as derived from hypnograms.

METHODS

We analyze sleep data of seven groups of subjects: healthy adults (n = 48), adults with obstructive sleep apnea (OSA) (n = 29), adults with CSA (n = 25), healthy children (n = 40), children with OSA (n = 18), children with CSA (n = 73) and CSA children treated with CPAP (n = 10). We calculate sleep-wake parameters based on the probability distributions of wake-bout durations and sleep-bout durations. We compare these parameters with results obtained from a neuronal model that simulates the interplay between sleep- and wake-promoting neurons.

RESULTS

We find that sleep arousals of CSA patients show a characteristic time scale (i.e., exponential distribution) in contrast to the scale-invariant (i.e., power-law) distribution that has been reported for arousals in healthy sleep. Furthermore, we show that this change in arousal statistics is caused by triggering more arousals of similar durations, which through our model can be related to a higher excitability threshold in sleep-promoting neurons in CSA patients.

CONCLUSIONS

We propose a neuronal mechanism to shed light on CSA pathophysiology and a method to discriminate between CSA and OSA. We show that higher neuronal excitability thresholds can lead to complex reorganization of sleep-wake dynamics.

SIGNIFICANCE

The derived sleep parameters enable a more specific evaluation of CSA severity and can be used for CSA diagnosis and monitor CSA treatment.

Lying Awake at Night: Cardiac Autonomic Activity in Relation to Sleep Onset and Maintenance

Insomnia, i.e., difficulties initiating and/or maintaining sleep, is one of the most common sleep disorders. To study underlying mechanisms for insomnia, we studied autonomic activity changes around sleep onset in participants without clinical insomnia but with varying problems with initiating or maintaining sleep quantified as increased sleep onset latency (SOL) and wake after sleep onset (WASO), respectively. Polysomnography and electrocardiography were simultaneously recorded in 176 participants during a single night. Cardiac autonomic activity was assessed using frequency domain analysis of RR intervals and results show that the normalized spectral power in the low frequency band (LF_nu) after sleep onset was significantly higher in participants with long SOL compared to participants with short SOL. Furthermore, the normalized spectral power in the high frequency band (HF_nu) was significantly lower in participants with long SOL as compared to participants with short SOL over 3 time periods (first 10 min in bed intending to sleep, 10 min before, and 10 min after sleep onset). These results suggest that participants with long SOL are more aroused in all three examined time periods when compared to participants with short SOL, especially for young adults (20–40 years). As there is no clear consensus on the cutoff for an increased WASO, we used a data-driven approach to explore different cutoffs to define short WASO and long WASO groups. LF_nu, HF_nu, and LF/HF differed between the long and the short WASO groups. A higher LF_nu and LF/HF and a lower HF_nu was observed in participants with long WASO for most cutoffs. The highest effect size was found using the cutoff of 66 min. Our findings suggest that autonomic cardiac activity has predictive value with respect to sleep characteristics pertaining to sleep onset and maintenance.

Network physiology in insomnia patients: Assessment of relevant changes in network topology with interpretable machine learning models

Network physiology describes the human body as a complex network of interacting organ systems. It has been applied successfully to determine topological changes in different sleep stages. However, the number of network links can quickly grow above the number of parameters that are typically analyzed with standard statistical methods. Artificial Neural Networks (ANNs) are a promising approach as they are successful in large parameter spaces, such as in digital imaging. On the other hand, ANN models do not provide an intrinsic approach to interpret their predictions, and they typically require large training data sets. Both aspects are critical in biomedical research. Medical decisions need to be explainable, and large data sets of quality assured patient and control data are rare. In this paper, different models for the classification of insomnia-a common sleep disorder-have been trained with 59 patients and age and gender matched controls, based on their physiological networks. Feature relevance evaluation is employed for all methods. For ANNs, the extrinsic interpretation method DeepLift is applied. The results are not identical across methods, but certain network links have been rated as relevant by all or most of the models. While ANNs show less classification accuracy (0.89) than advanced tree-based models (0.92 and 0.93), DeepLift provides an in-depth ANN interpretation with feature relevance scores for individual data samples. The analysis revealed modifications in the pulmonar, ocular, and cerebral subnetworks that have not been described before but are consistent with known findings on the physiological impact of insomnia.

Protocol of the SOMNIA project: an observational study to create a neurophysiological database for advanced clinical sleep monitoring

INTRODUCTION

Polysomnography (PSG) is the primary tool for sleep monitoring and the diagnosis of sleep disorders. Recent advances in signal analysis make it possible to reveal more information from this rich data source. Furthermore, many innovative sleep monitoring techniques are being developed that are less obtrusive, easier to use over long time periods and in the home situation. Here, we describe the methods of the Sleep and Obstructive Sleep Apnoea Monitoring with Non-Invasive Applications (SOMNIA) project, yielding a database combining clinical PSG with advanced unobtrusive sleep monitoring modalities in a large cohort of patients with various sleep disorders. The SOMNIA database will facilitate the validation and assessment of the diagnostic value of the new techniques, as well as the development of additional indices and biomarkers derived from new and/or traditional sleep monitoring methods.

METHODS AND ANALYSIS

We aim to include at least 2100 subjects (both adults and children) with a variety of sleep disorders who undergo a PSG as part of standard clinical care in a dedicated sleep centre. Full-video PSG will be performed according to the standards of the American Academy of Sleep Medicine. Each recording will be supplemented with one or more new monitoring systems, including wrist-worn photoplethysmography and actigraphy, pressure sensing mattresses, multimicrophone recording of respiratory sounds including snoring, suprasternal pressure monitoring and multielectrode electromyography of the diaphragm.

ETHICS AND DISSEMINATION

The study was reviewed by the medical ethical committee of the Maxima Medical Center (Eindhoven, the Netherlands, File no: N16.074). All subjects provide informed consent before participation.The SOMNIA database is built to facilitate future research in sleep medicine. Data from the completed SOMNIA database will be made available for collaboration with researchers outside the institute.

Coupling and Decoupling between Brain and Body Oscillations

Cross frequency coupling is used to study the cross talk between brain oscillations. In this paper we focus on a special type of frequency coupling between brain and body oscillations, which is reflected by the numerical ratio (r) between two frequencies (m and n; n > m). This approach is motivated by theoretical considerations, indicating that during alert wakefulness brain-body oscillations form a coupled hierarchy of frequencies with integer relationships that are binary multiples (r = n:m = 1:2, 1:4, 1:8…..). During sleep we expect an irrational relationship (r = n/m = irrational number) between brain and body oscillations that reflects decoupling. We analyzed alpha frequency, heart rate, breathing frequency during performance of a memory tasks and in addition spindle frequency from data collected by the SIESTA sleep research group. As predicted, our results show a binary multiple frequency relationship between alpha, heart rate and breathing frequency during task performance but an irrational relationship between spindle frequency, heart rate and breathing frequency during sleep.

Sleep stage classification from heart-rate variability using long short-term memory neural networks

Automated sleep stage classification using heart rate variability (HRV) may provide an ergonomic and low-cost alternative to gold standard polysomnography, creating possibilities for unobtrusive home-based sleep monitoring. Current methods however are limited in their ability to take into account long-term sleep architectural patterns. A long short-term memory (LSTM) network is proposed as a solution to model long-term cardiac sleep architecture information and validated on a comprehensive data set (292 participants, 584 nights, 541.214 annotated 30 s sleep segments) comprising a wide range of ages and pathological profiles, annotated according to the Rechtschaffen and Kales (R&K) annotation standard. It is shown that the model outperforms state-of-the-art approaches which were often limited to non-temporal or short-term recurrent classifiers. The model achieves a Cohen’s k of 0.61 ± 0.15 and accuracy of 77.00 ± 8.90% across the entire database. Further analysis revealed that the performance for individuals aged 50 years and older may decline. These results demonstrate the merit of deep temporal modelling using a diverse data set and advance the state-of-the-art for HRV-based sleep stage classification. Further research is warranted into individuals over the age of 50 as performance tends to worsen in this sub-population.

Feature relevance in physiological networks for classification of obstructive sleep apnea

OBJECTIVE

Physiological networks (PN) model couplings between organs in a high-dimensional parameter space. Machine learning methods, in particular artifical neural networks (ANNs), are powerful on high-dimensional classification tasks. However, lack of interpretability of the resulting models has been a drawback in research. We assess relevant PN topology changes in obstructive sleep apnea (OSA) by novel ANN interpretation techniques.

APPROACH

ANNs are trained to classify OSA based on the PNs of 48 patients and 48 age and gender matched healthy controls. The PNs consisting of 2812 links are derived from overnight biosignal recordings. The interpretation technique DeepLift is applied to the resulting ANN models, enabling the determination of the relevant features for classification decisions on individual subjects. The mean relevance scores of the features are compared to other machine learning methods (decision tree and random forests) and statistical tests on group differences.

MAIN RESULTS

The ANN interpretation results show good agreement with the compared methods and 87% of the samples could be correctly classified. OSA patients show a significantly higher coupling (p [Formula: see text] 0.001) in light sleep (N2) between breathing rate and EEG [Formula: see text] power in all electrode locations and to chin and leg muscular tone. In deep sleep (N3), OSA leads to significantly lower coupling (p [Formula: see text] 0.01) in lateral connections of EEG [Formula: see text] and [Formula: see text] power in central and frontal positions. Misclassified OSA patients had all mild/moderate AHIs and did not show PN topology changes. Both nights of these patients have been consistently misclassified as healthy. This may indicate, that the impact of respiratory events differs in subjects, thus forming different phenotypes.

SIGNIFICANCE

The proposed PN analysis provides a powerful and robust method to quantify a broad range of physiological interactions. Interpretability of the ANN make them a promising tool to identify new diagnostic markers in data-driven approaches.

Autonomic cardiac activity in adults with short and long sleep onset latency

Autonomic cardiac activity during sleep has been widely studied. Research has mostly focused on cardiac activity between different sleep stages and wakefulness as well as between normal and pathological sleep. This work investigates autonomic activity changes during sleep onset in healthy subjects with long and short sleep onset latency (SOL). Polysomnography (PSG) and electrocardiography (ECG) were simultaneously recorded in 186 healthy subjects during a single night. Autonomic activity was assessed based on frequency domain analysis of RR intervals and results show that the analysis of RR intervals differs significantly between the short SOL and the long SOL groups. We found that the spectral power in the low frequency band (LF) was significantly higher in the long SOL group compared to the short SOL group in the first 10 minutes in bed intended to sleep. There was no significant difference for LF and the spectral power in the high frequency band (HF) 10 minutes before and after sleep onset between the two groups. Only in the short SOL group there was a significant increase in HF from the first 10 minutes in bed intended to sleep to 10 minutes before SO, while LF decreased significantly in both groups. The effect of time (5.5-min bin) on the heart rate variability (HRV) features around sleep onset showed that both LF and HF differed significantly during the period surrounding sleep onset only in the short SOL group.

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.

Recurrent Deep Neural Networks for Real-Time Sleep Stage Classification From Single Channel EEG

Objective: We investigate the design of deep recurrent neural networks for detecting sleep stages from single channel EEG signals recorded at home by non-expert users. We report the effect of data set size, architecture choices, regularization, and personalization on the classification performance. Methods: We evaluated 58 different architectures and training configurations using three-fold cross validation. Results: A network consisting of convolutional (CONV) layers and long short term memory (LSTM) layers can achieve an agreement with a human annotator of Cohen's Kappa of ~0.73 using a training data set of 19 subjects. Regularization and personalization do not lead to a performance gain. Conclusion: The optimal neural network architecture achieves a performance that is very close to the previously reported human inter-expert agreement of Kappa 0.75. Significance: We give the first detailed account of CONV/LSTM network design process for EEG sleep staging in single channel home based setting.

A comparison of probabilistic classifiers for sleep stage classification

OBJECTIVE

To compare conditional random fields (CRF), hidden Markov models (HMMs) and Bayesian linear discriminants (LDs) for cardiorespiratory sleep stage classification on a five-class sleep staging task (wake/N1/N2/N3/REM), to explore the benefits of incorporating time information in the classification and to evaluate the feasibility of sleep staging on obstructive sleep apnea (OSA) patients.

APPROACH

The classifiers with and without time information were evaluated with 10-fold cross-validation on five-, four- (wake/N1  +  N2/N3/REM) and three-class (wake/NREM/REM) classification tasks using a data set comprising 443 night-time polysomnography (PSG) recordings of 231 participants (180 healthy participants, 100 of which had a 'regular' sleep architecture, and 51 participants previously diagnosed with OSA).

MAIN RESULTS

CRF with time information (CRFt) outperforms all other classifiers on all tasks, achieving a median accuracy and Cohen's κ for all participants of 62.8% and 0.44 for five classes, 68.8% and 0.47 for four classes, and 77.6% and 0.55 for three classes. An advantage was found in training classifiers, specifically for 'regular' and 'OSA' participants, achieving an improvement in classification performance for these groups. For 'regular' participants, CRFt achieved a median accuracy and Cohen's κ of 67.0% and 0.51, 70.8% and 0.53 and 81.3% and 0.62 for five-, four- and three-classes respectively, and for 'OSA' patients, of 59.9% and 0.40, 69.7% and 0.45, and 75.8% and 0.51 for five-, four- and three-classes respectively.

SIGNIFICANCE

The results suggest that CRFt is not only better at learning and predicting more complex and irregular sleep architectures, but that it also performs reasonably well in five-class classification-the standard for sleep scoring used in clinical PSG. Additionally, and albeit with a decrease in performance when compared with healthy participants, sleep stage classification in OSA patients using cardiorespiratory features and CRFt seems feasible with reasonable accuracy.

Neuronal noise as an origin of sleep arousals and its role in sudden infant death syndrome

In addition to regular sleep/wake cycles, humans and animals exhibit brief arousals from sleep. Although much is known about consolidated sleep and wakefulness, the mechanism that triggers arousals remains enigmatic. Here, we argue that arousals are caused by the intrinsic neuronal noise of wake-promoting neurons. We propose a model that simulates the superposition of the noise from a group of neurons, and show that, occasionally, the superposed noise exceeds the excitability threshold and provokes an arousal. Because neuronal noise decreases with increasing temperature, our model predicts arousal frequency to decrease as well. To test this prediction, we perform experiments on the sleep/wake behavior of zebrafish larvae and find that increasing water temperatures lead to fewer and shorter arousals, as predicted by our analytic derivations and model simulations. Our findings indicate a previously unrecognized neurophysiological mechanism that links sleep arousals with temperature regulation, and may explain the origin of the clinically observed higher risk for sudden infant death syndrome with increased ambient temperature.

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.

Validation of Photoplethysmography-Based Sleep Staging Compared With Polysomnography in Healthy Middle-Aged Adults

Study Objectives

To compare the accuracy of automatic sleep staging based on heart rate variability measured from photoplethysmography (PPG) combined with body movements measured with an accelerometer, with polysomnography (PSG) and actigraphy.

Methods

Using wrist-worn PPG to analyze heart rate variability and an accelerometer to measure body movements, sleep stages and sleep statistics were automatically computed from overnight recordings. Sleep-wake, 4-class (wake/N1 + N2/N3/REM) and 3-class (wake/NREM/REM) classifiers were trained on 135 simultaneously recorded PSG and PPG recordings of 101 healthy participants and validated on 80 recordings of 51 healthy middle-aged adults. Epoch-by-epoch agreement and sleep statistics were compared with actigraphy for a subset of the validation set.

Results

The sleep-wake classifier obtained an epoch-by-epoch Cohen's κ between PPG and PSG sleep stages of 0.55 ± 0.14, sensitivity to wake of 58.2 ± 17.3%, and accuracy of 91.5 ± 5.1%. κ and sensitivity were significantly higher than with actigraphy (0.40 ± 0.15 and 45.5 ± 19.3%, respectively). The 3-class classifier achieved a κ of 0.46 ± 0.15 and accuracy of 72.9 ± 8.3%, and the 4-class classifier, a κ of 0.42 ± 0.12 and accuracy of 59.3 ± 8.5%.

Conclusions

The moderate epoch-by-epoch agreement and, in particular, the good agreement in terms of sleep statistics suggest that this technique is promising for long-term sleep monitoring, although more evidence is needed to understand whether it can complement PSG in clinical practice. It also offers an improvement in sleep/wake detection over actigraphy for healthy individuals, although this must be confirmed on a larger, clinical population.

Large-Scale Automated Sleep Staging

Study Objectives

Automated sleep staging has been previously limited by a combination of clinical and physiological heterogeneity. Both factors are in principle addressable with large data sets that enable robust calibration. However, the impact of sample size remains uncertain. The objectives are to investigate the extent to which machine learning methods can approximate the performance of human scorers when supplied with sufficient training cases and to investigate how staging performance depends on the number of training patients, contextual information, model complexity, and imbalance between sleep stage proportions.

Methods

A total of 102 features were extracted from six electroencephalography (EEG) channels in routine polysomnography. Two thousand nights were partitioned into equal (n = 1000) training and testing sets for validation. We used epoch-by-epoch Cohen's kappa statistics to measure the agreement between classifier output and human scorer according to American Academy of Sleep Medicine scoring criteria.

Results

Epoch-by-epoch Cohen's kappa improved with increasing training EEG recordings until saturation occurred (n = ~300). The kappa value was further improved by accounting for contextual (temporal) information, increasing model complexity, and adjusting the model training procedure to account for the imbalance of stage proportions. The final kappa on the testing set was 0.68. Testing on more EEG recordings leads to kappa estimates with lower variance.

Conclusion

Training with a large data set enables automated sleep staging that compares favorably with human scorers. Because testing was performed on a large and heterogeneous data set, the performance estimate has low variance and is likely to generalize broadly.