Large-Scale Automated Sleep Staging

Refining sleep staging accuracy: Transfer learning coupled with scorability models

STUDY OBJECTIVES

This study aimed to 1) improve sleep staging accuracy through transfer learning, to achieve or exceede human inter-expert agreement; 2) introduce a scorability model to assess the quality and trustworthiness of automated sleep staging.

METHODS

A deep neural network (base model) was trained on a large multi-site polysomnography (PSG) dataset from the United States. Transfer learning was used to calibrate the model to a reduced montage and limited samples from the Korean Genome and Epidemiology Study (KoGES) dataset. Model performance was compared to inter-expert reliability among three human experts. A scorability assessment was developed to predict the agreement between the model and human experts.

RESULTS

Initial sleep staging by the base model showed lower agreement with experts (κ=0.55) compared to inter-expert agreement (κ=0.62). Calibration with 324 randomly sampled training cases matched expert agreement levels. Further targeted sampling improved performance, with models exceeding inter-expert agreement (κ=0.70). The scorability assessment, combining biosignal quality and model confidence features, predicted model-expert agreement moderately well (R²=0.42). Recordings with higher scorability scores demonstrated greater model-expert agreement than inter-expert agreement. Even with lower scorability scores, model performance was comparable to inter-expert agreement.

CONCLUSIONS

Fine-tuning a pre-trained neural network through targeted transfer learning significantly enhances sleep staging performance for an atypical montage, achieving and surpassing human expert agreement levels. The introduction of a scorability assessment provides a robust measure of reliability, ensuring quality control and enhancing the practical application of the system before deployment. This approach marks an important advancement in automated sleep analysis, demonstrating the potential for AI to exceed human performance in clinical settings.

Classifying Routine Clinical Electroencephalograms With Multivariate Iterative Filtering and Convolutional Neural Networks

Electroencephalogram (EEG) is widely used in basic and clinical neuroscience to explore neural states in various populations, and classifying these EEG recordings is a fundamental challenge. While machine learning shows promising results in classifying long multivariate time series, optimal prediction models and feature extraction methods for EEG classification remain elusive. Our study addressed the problem of EEG classification under the framework of brain age prediction, applying a deep learning model on EEG time series. We hypothesized that decomposing EEG signals into oscillatory modes would yield more accurate age predictions than using raw or canonically frequency-filtered EEG. Specifically, we employed multivariate intrinsic mode functions (MIMFs), an empirical mode decomposition (EMD) variant based on multivariate iterative filtering (MIF), with a convolutional neural network (CNN) model. Testing a large dataset of routine clinical EEG scans (n = 6540) from patients aged 1 to 103 years, we found that an ad-hoc CNN model without fine-tuning could reasonably predict brain age from EEGs. Crucially, MIMF decomposition significantly improved performance compared to canonical brain rhythms (from delta to lower gamma oscillations). Our approach achieved a mean absolute error (MAE) of 13.76 ± 0.33 and a correlation coefficient of 0.64 ± 0.01 in brain age prediction over the entire lifespan. Our findings indicate that CNN models applied to EEGs, preserving their original temporal structure, remains a promising framework for EEG classification, wherein the adaptive signal decompositions such as the MIF can enhance CNN models' performance in this task.

Generalized sleep decoding with basal ganglia signals in multiple movement disorders

Sleep disturbances profoundly affect the quality of life in individuals with neurological disorders. Closed-loop deep brain stimulation (DBS) holds promise for alleviating sleep symptoms, however, this technique necessitates automated sleep stage decoding from intracranial signals. We leveraged overnight data from 121 patients with movement disorders (Parkinson's disease, Essential Tremor, Dystonia, Essential Tremor, Huntington's disease, and Tourette's syndrome) in whom synchronized polysomnograms and basal ganglia local field potentials were recorded, to develop a generalized, multi-class, sleep specific decoder - BGOOSE. This generalized model achieved 85% average accuracy across patients and across disease conditions, even in the presence of recordings from different basal ganglia targets. Furthermore, we also investigated the role of electrocorticography on decoding performances and proposed an optimal decoding map, which was shown to facilitate channel selection for optimal model performances. BGOOSE emerges as a powerful tool for generalized sleep decoding, offering exciting potentials for the precision stimulation delivery of DBS and better management of sleep disturbances in movement disorders.

Interhemispheric differences of electroencephalography signal characteristics in different sleep stages

OBJECTIVE

The current electroencephalography (EEG) measurement setup is complex, laborious to set up, and uncomfortable for patients. We hypothesize that differences in EEG signal characteristics for sleep staging between the left and right hemispheres are negligible; therefore, there is potential to simplify the current measurement setup. We aimed to investigate the technical hemispheric differences in EEG signal characteristics along with electrooculography (EOG) signals during different sleep stages.

METHODS

Type II portable polysomnography (PSG) recordings of 50 patients were studied. Amplitudes and power spectral densities (PSDs) of the EEG and EOG signals were compared between the left (C3-M2, F3-M2, O1-M2, and E1-M2) and the right (C4-M1, F4-M1, O2-M1, and E2-M2) hemispheres. Regression analysis was performed to investigate the potential influence of sleep stages on the hemispheric differences in PSDs. Wilcoxon signed-rank tests were also employed to calculate the effect size of hemispheres across different frequency bands and sleep stages.

RESULTS

The results showed statistically significant differences in signal characteristics between hemispheres, but the absolute differences were minor. The median hemispheric differences in amplitudes were smaller than 3 μv with large interquartile ranges during all sleep stages. The absolute and relative PSD characteristics were highly similar between hemispheres in different sleep stages. Additionally, there were negligible differences in the effect size between hemispheres across all sleep stages.

CONCLUSIONS

Technical signal differences between hemispheres were minor across all sleep stages, indicating that both hemispheres contain similar information needed for sleep staging. A reduced measurement setup could be suitable for sleep staging without the loss of relevant information.

U-PASS: An uncertainty-guided deep learning pipeline for automated sleep staging

With the increasing prevalence of machine learning in critical fields like healthcare, ensuring the safety and reliability of these systems is crucial. Estimating uncertainty plays a vital role in enhancing reliability by identifying areas of high and low confidence and reducing the risk of errors. This study introduces U-PASS, a specialized human-centered machine learning pipeline tailored for clinical applications, which effectively communicates uncertainty to clinical experts and collaborates with them to improve predictions. U-PASS incorporates uncertainty estimation at every stage of the process, including data acquisition, training, and model deployment. Training is divided into a supervised pre-training step and a semi-supervised recording-wise finetuning step. We apply U-PASS to the challenging task of sleep staging and demonstrate that it systematically improves performance at every stage. By optimizing the training dataset, actively seeking feedback from domain experts for informative samples, and deferring the most uncertain samples to experts, U-PASS achieves an impressive expert-level accuracy of 85% on a challenging clinical dataset of elderly sleep apnea patients. This represents a significant improvement over the starting point at 75% accuracy. The largest improvement gain is due to the deferral of uncertain epochs to a sleep expert. U-PASS presents a promising AI approach to incorporating uncertainty estimation in machine learning pipelines, improving their reliability and unlocking their potential in clinical settings.

Linking brain structure, cognition, and sleep: insights from clinical data

STUDY OBJECTIVES

To use relatively noisy routinely collected clinical data (brain magnetic resonance imaging (MRI) data, clinical polysomnography (PSG) recordings, and neuropsychological testing), to investigate hypothesis-driven and data-driven relationships between brain physiology, structure, and cognition.

METHODS

We analyzed data from patients with clinical PSG, brain MRI, and neuropsychological evaluations. SynthSeg, a neural network-based tool, provided high-quality segmentations despite noise. A priori hypotheses explored associations between brain function (measured by PSG) and brain structure (measured by MRI). Associations with cognitive scores and dementia status were studied. An exploratory data-driven approach investigated age-structure-physiology-cognition links.

RESULTS

Six hundred and twenty-three patients with sleep PSG and brain MRI data were included in this study; 160 with cognitive evaluations. Three hundred and forty-two participants (55%) were female, and age interquartile range was 52 to 69 years. Thirty-six individuals were diagnosed with dementia, 71 with mild cognitive impairment, and 326 with major depression. One hundred and fifteen individuals were evaluated for insomnia and 138 participants had an apnea-hypopnea index equal to or greater than 15. Total PSG delta power correlated positively with frontal lobe/thalamic volumes, and sleep spindle density with thalamic volume. rapid eye movement (REM) duration and amygdala volume were positively associated with cognition. Patients with dementia showed significant differences in five brain structure volumes. REM duration, spindle, and slow-oscillation features had strong associations with cognition and brain structure volumes. PSG and MRI features in combination predicted chronological age (R2 = 0.67) and cognition (R2 = 0.40).

CONCLUSIONS

Routine clinical data holds extended value in understanding and even clinically using brain-sleep-cognition relationships.

Enhanced sleep staging with artificial intelligence: a validation study of new software for sleep scoring

Manual sleep staging (MSS) using polysomnography is a time-consuming task, requires significant training, and can lead to significant variability among scorers. STAGER is a software program based on machine learning algorithms that has been developed by Medibio Limited (Savage, MN, USA) to perform automatic sleep staging using only EEG signals from polysomnography. This study aimed to extensively investigate its agreement with MSS performed during clinical practice and by three additional expert sleep technicians. Forty consecutive polysomnographic recordings of patients referred to three US sleep clinics for sleep evaluation were retrospectively collected and analyzed. Three experienced technicians independently staged the recording using the electroencephalography, electromyography, and electrooculography signals according to the American Academy of Sleep Medicine guidelines. The staging initially performed during clinical practice was also considered. Several agreement statistics between the automatic sleep staging (ASS) and MSS, among the different MSSs, and their differences were calculated. Bootstrap resampling was used to calculate 95% confidence intervals and the statistical significance of the differences. STAGER's ASS was most comparable with, or statistically significantly better than the MSS, except for a partial reduction in the positive percent agreement in the wake stage. These promising results indicate that STAGER software can perform ASS of inpatient polysomnographic recordings accurately in comparison with MSS.

Decentralized Data-Privacy Preserving Deep-Learning Approaches for Enhancing Inter-Database Generalization in Automatic Sleep Staging

Automatic sleep staging has been an active field of development. Despite multiple efforts, the area remains a focus of research interest. Indeed, while promising results have reported in past literature, uptake of automatic sleep scoring in the clinical setting remains low. One of the current issues regards the difficulty to generalization performance results beyond the local testing scenario, i.e. across data from different clinics. Issues derived from data-privacy restrictions, that generally apply in the medical domain, pose additional difficulties in the successful development of these methods. We propose the use of several decentralized deep-learning approaches, namely ensemble models and federated learning, for robust inter-database performance generalization and data-privacy preservation in automatic sleep staging scenario. Specifically, we explore four ensemble combination strategies (max-voting, output averaging, size-proportional weighting, and Nelder-Mead) and present a new federated learning algorithm, so-called sub-sampled federated stochastic gradient descent (ssFedSGD). To evaluate generalization capabilities of such approaches, experimental procedures are carried out using a leaving-one-database-out direct-transfer scenario on six independent and heterogeneous public sleep staging databases. The resulting performance is compared with respect to two baseline approaches involving single-database and centralized multiple-database derived models. Our results show that proposed decentralized learning methods outperform baseline local approaches, and provide similar generalization results to centralized database-combined approaches. We conclude that these methods are more preferable choices, as they come with additional advantages concerning improved scalability, flexible design, and data-privacy preservation.

Sleep electroencephalography biomarkers of cognition in obstructive sleep apnea

Obstructive sleep apnea has been associated with cognitive impairment and may be linked to disorders of cognitive function. These associations may be a result of intermittent hypoxaemia, sleep fragmentation and changes in sleep microstructure in obstructive sleep apnea. Current clinical metrics of obstructive sleep apnea, such as the apnea-hypopnea index, are poor predictors of cognitive outcomes in obstructive sleep apnea. Sleep microstructure features, which can be identified on sleep electroencephalography of traditional overnight polysomnography, are increasingly being characterized in obstructive sleep apnea and may better predict cognitive outcomes. Here, we summarize the literature on several major sleep electroencephalography features (slow-wave activity, sleep spindles, K-complexes, cyclic alternating patterns, rapid eye movement sleep quantitative electroencephalography, odds ratio product) identified in obstructive sleep apnea. We will review the associations between these sleep electroencephalography features and cognition in obstructive sleep apnea, and examine how treatment of obstructive sleep apnea affects these associations. Lastly, evolving technologies in sleep electroencephalography analyses will also be discussed (e.g. high-density electroencephalography, machine learning) as potential predictors of cognitive function in obstructive sleep apnea.

Challenges of Applying Automated Polysomnography Scoring at Scale

Automatic polysomnography analysis can be leveraged to shorten scoring times, reduce associated costs, and ultimately improve the overall diagnosis of sleep disorders. Multiple and diverse strategies have been attempted for implementation of this technology at scale in the routine workflow of sleep centers. The field, however, is complex and presents unsolved challenges in a number of areas. Recent developments in computer science and artificial intelligence are nevertheless closing the gap. Technological advances are also opening new pathways for expanding our current understanding of the domain and its analysis.

Decoding information about cognitive health from the brainwaves of sleep

Sleep electroencephalogram (EEG) signals likely encode brain health information that may identify individuals at high risk for age-related brain diseases. Here, we evaluate the correlation of a previously proposed brain age biomarker, the "brain age index" (BAI), with cognitive test scores and use machine learning to develop and validate a series of new sleep EEG-based indices, termed "sleep cognitive indices" (SCIs), that are directly optimized to correlate with specific cognitive scores. Three overarching cognitive processes were examined: total, fluid (a measure of cognitive processes involved in reasoning-based problem solving and susceptible to aging and neuropathology), and crystallized cognition (a measure of cognitive processes involved in applying acquired knowledge toward problem-solving). We show that SCI decoded information about total cognition (Pearson's r = 0.37) and fluid cognition (Pearson's r = 0.56), while BAI correlated only with crystallized cognition (Pearson's r = - 0.25). Overall, these sleep EEG-derived biomarkers may provide accessible and clinically meaningful indicators of neurocognitive health.

Functional extreme learning machine

Introduction

Extreme learning machine (ELM) is a training algorithm for single hidden layer feedforward neural network (SLFN), which converges much faster than traditional methods and yields promising performance. However, the ELM also has some shortcomings, such as structure selection, overfitting and low generalization performance.

Methods

This article a new functional neuron (FN) model is proposed, we takes functional neurons as the basic unit, and uses functional equation solving theory to guide the modeling process of FELM, a new functional extreme learning machine (FELM) model theory is proposed.

Results

The FELM implements learning by adjusting the coefficients of the basis function in neurons. At the same time, a simple, iterative-free and high-precision fast parameter learning algorithm is proposed.

Discussion

The standard data sets UCI and StatLib are selected for regression problems, and compared with the ELM, support vector machine (SVM) and other algorithms, the experimental results show that the FELM achieves better performance.

The sleep and wake electroencephalogram over the lifespan

Both sleep and wake encephalograms (EEG) change over the lifespan. While prior studies have characterized age-related changes in the EEG, the datasets span a particular age group, or focused on sleep and wake macrostructure rather than the microstructure. Here, we present sex-stratified data from 3372 community-based or clinic-based otherwise neurologically and psychiatrically healthy participants ranging from 11 days to 80 years of age. We estimate age norms for key sleep and wake EEG parameters including absolute and relative powers in delta, theta, alpha, and sigma bands, as well as sleep spindle density, amplitude, duration, and frequency. To illustrate the potential use of the reference measures developed herein, we compare them to sleep EEG recordings from age-matched participants with Alzheimer's disease, severe sleep apnea, depression, osteoarthritis, and osteoporosis. Although the partially clinical nature of the datasets may bias the findings towards less normal and hence may underestimate pathology in practice, age-based EEG reference values enable objective screening of deviations from healthy aging among individuals with a variety of disorders that affect brain health.

Automated Sleep Stages Classification Using Convolutional Neural Network From Raw and Time-Frequency Electroencephalogram Signals: Systematic Evaluation Study

BACKGROUND

Most existing automated sleep staging methods rely on multimodal data, and scoring a specific epoch requires not only the current epoch but also a sequence of consecutive epochs that precede and follow the epoch.

OBJECTIVE

We proposed and tested a convolutional neural network called SleepInceptionNet, which allows sleep classification of a single epoch using a single-channel electroencephalogram (EEG).

METHODS

SleepInceptionNet is based on our systematic evaluation of the effects of different EEG preprocessing methods, EEG channels, and convolutional neural networks on automatic sleep staging performance. The evaluation was performed using polysomnography data of 883 participants (937,975 thirty-second epochs). Raw data of individual EEG channels (ie, frontal, central, and occipital) and 3 specific transformations of the data, including power spectral density, continuous wavelet transform, and short-time Fourier transform, were used separately as the inputs of the convolutional neural network models. To classify sleep stages, 7 sequential deep neural networks were tested for the 1D data (ie, raw EEG and power spectral density), and 16 image classifier convolutional neural networks were tested for the 2D data (ie, continuous wavelet transform and short-time Fourier transform time-frequency images).

RESULTS

The best model, SleepInceptionNet, which uses time-frequency images developed by the continuous wavelet transform method from central single-channel EEG data as input to the InceptionV3 image classifier algorithm, achieved a Cohen κ agreement of 0.705 (SD 0.077) in reference to the gold standard polysomnography.

CONCLUSIONS

SleepInceptionNet may allow real-time automated sleep staging in free-living conditions using a single-channel EEG, which may be useful for on-demand intervention or treatment during specific sleep stages.

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.

An accessible and versatile deep learning-based sleep stage classifier

Manual sleep scoring for research purposes and for the diagnosis of sleep disorders is labor-intensive and often varies significantly between scorers, which has motivated many attempts to design automatic sleep stage classifiers. With the recent introduction of large, publicly available hand-scored polysomnographic data, and concomitant advances in machine learning methods to solve complex classification problems with supervised learning, the problem has received new attention, and a number of new classifiers that provide excellent accuracy. Most of these however have non-trivial barriers to use. We introduce the Greifswald Sleep Stage Classifier (GSSC), which is free, open source, and can be relatively easily installed and used on any moderately powered computer. In addition, the GSSC has been trained to perform well on a large variety of electrode set-ups, allowing high performance sleep staging with portable systems. The GSSC can also be readily integrated into brain-computer interfaces for real-time inference. These innovations were achieved while simultaneously reaching a level of accuracy equal to, or exceeding, recent state of the art classifiers and human experts, making the GSSC an excellent choice for researchers in need of reliable, automatic sleep staging.

Feasibility of wearable devices and machine learning for sleep classification in children with Rett syndrome: A pilot study

Sleep is vital to many processes involved in the well-being and health of children; however, it is estimated that 80% of children with Rett syndrome suffer from sleep disorders. Caregiver reports and questionnaires, which are the current method of studying sleep, are prone to observer bias and missed information. Polysomnography is considered the gold standard for sleep analysis but is labor and cost-intensive and limits the frequency of data collection for sleep disorder studies. Wearable digital health technologies, such as actigraphy devices, have shown potential and feasibility as a method for sleep analysis in Rett syndrome, but have not been validated against polysomnography. Furthermore, the collected accelerometer data has limitations due to the rigidity, periodic limb movement, and involuntary muscle contractions prevalent in Rett syndrome. Heart rate and electrodermal activity, along with other physiological signals, have been linked to sleep stages and can be utilized with machine learning to provide better resistance to noise and false positives than actigraphy. This research aims to address the gap in Rett syndrome sleep analysis by comparing the performance of a machine learning model utilizing both accelerometer data and physiological data features to the gold-standard polysomnography for sleep analysis in Rett syndrome. Our analytical validation pilot study (n = 7) found that using physiological and accelerometer features, our machine learning models can differentiate between awake, non-rapid eye movement sleep, and rapid eye movement sleep in Rett syndrome children with an accuracy of 85.1% when using an individual model. Additionally, this work demonstrates that it is feasible to use digital health technologies in Rett syndrome, even at a young age, without data loss or interference from repetitive movements that are characteristic of Rett syndrome.

XSleepNet: Multi-View Sequential Model for Automatic Sleep Staging

Automating sleep staging is vital to scale up sleep assessment and diagnosis to serve millions experiencing sleep deprivation and disorders and enable longitudinal sleep monitoring in home environments. Learning from raw polysomnography signals and their derived time-frequency image representations has been prevalent. However, learning from multi-view inputs (e.g., both the raw signals and the time-frequency images) for sleep staging is difficult and not well understood. This work proposes a sequence-to-sequence sleep staging model, XSleepNet,¹ that is capable of learning a joint representation from both raw signals and time-frequency images. Since different views may generalize or overfit at different rates, the proposed network is trained such that the learning pace on each view is adapted based on their generalization/overfitting behavior. In simple terms, the learning on a particular view is speeded up when it is generalizing well and slowed down when it is overfitting. View-specific generalization/overfitting measures are computed on-the-fly during the training course and used to derive weights to blend the gradients from different views. As a result, the network is able to retain the representation power of different views in the joint features which represent the underlying distribution better than those learned by each individual view alone. Furthermore, the XSleepNet architecture is principally designed to gain robustness to the amount of training data and to increase the complementarity between the input views. Experimental results on five databases of different sizes show that XSleepNet consistently outperforms the single-view baselines and the multi-view baseline with a simple fusion strategy. Finally, XSleepNet also outperforms prior sleep staging methods and improves previous state-of-the-art results on the experimental databases.

Validation Study on Automated Sleep Stage Scoring Using a Deep Learning Algorithm

Background and Objectives: Polysomnography is manually scored by sleep experts. However, manual scoring is a time-consuming and labor-intensive task. The goal of this study was to verify the accuracy of automated sleep-stage scoring based on a deep learning algorithm compared to manual sleep-stage scoring. Materials and Methods: A total of 602 polysomnography datasets from subjects (Male:Female = 397:205) aged 19 to 65 years (mean age, 43.8, standard deviation = 12.2) were included in the study. The performance of the proposed model was evaluated based on kappa value and bootstrapped point-estimate of median percent agreement with a 95% bootstrap confidence interval and R = 1000. The proposed model was trained using 482 datasets and validated using 48 datasets. For testing, 72 datasets were selected randomly. Results: The proposed model exhibited good concordance rates with manual scoring for stages W (94%), N1 (83.9%), N2 (89%), N3 (92%), and R (93%). The average kappa value was 0.84. For the bootstrap method, high overall agreement between the automated deep learning algorithm and manual scoring was observed in stages W (98%), N1 (94%), N2 (92%), N3 (99%), and R (98%) and total (96%). Conclusions: Automated sleep-stage scoring using the proposed model may be a reliable method for sleep-stage classification.

Monitoring of sedation in mechanically ventilated patients using remote technology

PURPOSE OF REVIEW

Two years of coronavirus disease 2019 (COVID-19) pandemic highlighted that excessive sedation in the ICU leading to coma and other adverse outcomes remains pervasive. There is a need to improve monitoring and management of sedation in mechanically ventilated patients. Remote technologies that are based on automated analysis of electroencephalogram (EEG) could enhance standard care and alert clinicians real-time when severe EEG suppression or other abnormal brain states are detected.

RECENT FINDINGS

High rates of drug-induced coma as well as delirium were found in several large cohorts of mechanically ventilated patients with COVID-19 pneumonia. In patients with acute respiratory distress syndrome, high doses of sedatives comparable to general anesthesia have been commonly administered without defined EEG endpoints. Continuous limited-channel EEG can reveal pathologic brain states such as burst suppression, that cannot be diagnosed by neurological examination alone. Recent studies documented that machine learning-based analysis of continuous EEG signal is feasible and that this approach can identify burst suppression as well as delirium with high specificity.

SUMMARY

Preventing oversedation in the ICU remains a challenge. Continuous monitoring of EEG activity, automated EEG analysis, and generation of alerts to clinicians may reduce drug-induced coma and potentially improve patient outcomes.

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

Modern deep learning holds a great potential to transform clinical practice on human sleep. Teaching a machine to carry out routine tasks would be a tremendous reduction in workload for clinicians. Sleep staging, a fundamental step in sleep practice, is a suitable task for this and will be the focus in this article. Recently, automatic sleep staging systems have been trained to mimic manual scoring, leading to similar performance to human sleep experts, at least on scoring of healthy subjects. Despite tremendous progress, we have not seen automatic sleep scoring adopted widely in clinical environments. This review aims to give a shared view of the authors on the most recent state-of-the-art development in automatic sleep staging, the challenges that still need to be addressed, and the future directions for automatic sleep scoring to achieve clinical value.

SleepTransformer: Automatic Sleep Staging with Interpretability and Uncertainty Quantification

BACKGROUND

Black-box skepticism is one of the main hindrances impeding deep-learning-based automatic sleep scoring from being used in clinical environments.

METHODS

Towards interpretability, this work proposes a sequence-to-sequence sleep staging model, namely SleepTransformer. It is based on the transformer backbone and offers interpretability of the models decisions at both the epoch and sequence level. We further propose a simple yet efficient method to quantify uncertainty in the models decisions. The method, which is based on entropy, can serve as a metric for deferring low-confidence epochs to a human expert for further inspection.

RESULTS

Making sense of the transformers self-attention scores for interpretability, at the epoch level, the attention scores are encoded as a heat map to highlight sleep-relevant features captured from the input EEG signal. At the sequence level, the attention scores are visualized as the influence of different neighboring epochs in an input sequence (i.e. the context) to recognition of a target epoch, mimicking the way manual scoring is done by human experts.

CONCLUSION

Additionally, we demonstrate that SleepTransformer performs on par with existing methods on two databases of different sizes.

SIGNIFICANCE

Equipped with interpretability and the ability of uncertainty quantification, SleepTransformer holds promise for being integrated into clinical settings.

Enhancing Current Cardiorespiratory-based Approaches of Sleep Stage Classification by Temporal Feature Stacking

This paper presents a generic method to enhance performance and incorporate temporal information for cardiorespiratory-based sleep stage classification with a limited feature set and limited data. The classification algorithm relies on random forests and a feature set extracted from long-time home monitoring for sleep analysis. Employing temporal feature stacking, the system could be significantly improved in terms of Cohen's κ and accuracy. The detection performance could be improved for three classes of sleep stages (Wake, REM, Non-REM sleep), four classes (Wake, Non-REM-Light sleep, Non-REM Deep sleep, REM sleep), and five classes (Wake, N1, N2, N3/4, REM sleep) from a κ of 0.44 to 0.58, 0.33 to 0.51, and 0.28 to 0.44 respectively by stacking features before and after the epoch to be classified. Further analysis was done for the optimal length and combination method for this stacking approach. Overall, three methods and a variable duration between 30 s and 30 min have been analyzed. Overnight recordings of 36 healthy subjects from the Interdisciplinary Center for Sleep Medicine at Charité-Universitätsmedizin Berlin and Leave-One-Out-Cross-Validation on a patient-level have been used to validate the method.Clinical relevance- The method can be employed generically to feature sets for (small scale) datasets to improve classification performance for classification problems with temporal relations with random forest classifiers.

An open-source, high-performance tool for automated sleep staging

The clinical and societal measurement of human sleep has increased exponentially in recent years. However, unlike other fields of medical analysis that have become highly automated, basic and clinical sleep research still relies on human visual scoring. Such human-based evaluations are time-consuming, tedious, and can be prone to subjective bias. Here, we describe a novel algorithm trained and validated on +30,000 hr of polysomnographic sleep recordings across heterogeneous populations around the world. This tool offers high sleep-staging accuracy that matches human scoring accuracy and interscorer agreement no matter the population kind. The software is designed to be especially easy to use, computationally low-demanding, open source, and free. Our hope is that this software facilitates the broad adoption of an industry-standard automated sleep staging software package.

Statistical uncertainty quantification to augment clinical decision support: a first implementation in sleep medicine

Machine learning has the potential to change the practice of medicine, particularly in areas that require pattern recognition (e.g. radiology). Although automated classification is unlikely to be perfect, few modern machine learning tools have the ability to assess their own classification confidence to recognize uncertainty that might need human review. Using automated single-channel sleep staging as a first implementation, we demonstrated that uncertainty information (as quantified using Shannon entropy) can be utilized in a "human in the loop" methodology to promote targeted review of uncertain sleep stage classifications on an epoch-by-epoch basis. Across 20 sleep studies, this feedback methodology proved capable of improving scoring agreement with the gold standard over automated scoring alone (average improvement in Cohen's Kappa of 0.28), in a fraction of the scoring time compared to full manual review (60% reduction). In summary, our uncertainty-based clinician-in-the-loop framework promotes the improvement of medical classification accuracy/confidence in a cost-effective and economically resourceful manner.

Physiological Assessment of Delirium Severity: The Electroencephalographic Confusion Assessment Method Severity Score (E-CAM-S)

OBJECTIVES

Delirium is a common and frequently underdiagnosed complication in acutely hospitalized patients, and its severity is associated with worse clinical outcomes. We propose a physiologically based method to quantify delirium severity as a tool that can help close this diagnostic gap: the Electroencephalographic Confusion Assessment Method Severity Score (E-CAM-S).

DESIGN

Retrospective cohort study.

SETTING

Single-center tertiary academic medical center.

PATIENTS

Three-hundred seventy-three adult patients undergoing electroencephalography to evaluate altered mental status between August 2015 and December 2019.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We developed the E-CAM-S based on a learning-to-rank machine learning model of forehead electroencephalography signals. Clinical delirium severity was assessed using the Confusion Assessment Method Severity (CAM-S). We compared associations of E-CAM-S and CAM-S with hospital length of stay and inhospital mortality. E-CAM-S correlated with clinical CAM-S (R = 0.67; p < 0.0001). For the overall cohort, E-CAM-S and CAM-S were similar in their strength of association with hospital length of stay (correlation = 0.31 vs 0.41, respectively; p = 0.082) and inhospital mortality (area under the curve = 0.77 vs 0.81; p = 0.310). Even when restricted to noncomatose patients, E-CAM-S remained statistically similar to CAM-S in its association with length of stay (correlation = 0.37 vs 0.42, respectively; p = 0.188) and inhospital mortality (area under the curve = 0.83 vs 0.74; p = 0.112). In addition to previously appreciated spectral features, the machine learning framework identified variability in multiple measures over time as important features in electroencephalography-based prediction of delirium severity.

CONCLUSIONS

The E-CAM-S is an automated, physiologic measure of delirium severity that predicts clinical outcomes with a level of performance comparable to conventional interview-based clinical assessment.

Automatic Respiratory Event Scoring in Obstructive Sleep Apnea Using a Long Short-Term Memory Neural Network

The diagnosis of obstructive sleep apnea is based on daytime symptoms and the frequency of respiratory events during the night. The respiratory events are scored manually from polysomnographic recordings, which is time-consuming and expensive. Therefore, automatic scoring methods could considerably improve the efficiency of sleep apnea diagnostics and release the resources currently needed for manual scoring to other areas of sleep medicine. In this study, we trained a long short-term memory neural network for automatic scoring of respiratory events using input signals from peripheral blood oxygen saturation, thermistor-airflow, nasal pressure -airflow, and thorax respiratory effort. The signals were extracted from 887 in-lab polysomnography recordings. 787 patients with suspected sleep apnea were used to train the neural network and 100 patients were used as an independent test set. The epoch-wise agreement between manual and automatic neural network scoring was high (88.9%, κ = 0.728). In addition, the apnea-hypopnea index (AHI) calculated from the automated scoring was close to the manually determined AHI with a mean absolute error of 3.0 events/hour and an intraclass correlation coefficient of 0.985. The neural network approach for automatic scoring of respiratory events achieved high accuracy and good agreement with manual scoring. The presented neural network could be used for analysis of large research datasets that are unfeasible to score manually, and has potential for clinical use in the future In addition, since the neural network scores individual respiratory events, the automatic scoring can be easily reviewed manually if desired.

Artificial intelligence in sleep medicine: an American Academy of Sleep Medicine position statement

None

Sleep medicine is well positioned to benefit from advances that use big data to create artificially intelligent computer programs. One obvious initial application in the sleep disorders center is the assisted (or enhanced) scoring of sleep and associated events during polysomnography (PSG). This position statement outlines the potential opportunities and limitations of integrating artificial intelligence (AI) into the practice of sleep medicine. Additionally, although the most apparent and immediate application of AI in our field is the assisted scoring of PSG, we propose potential clinical use cases that transcend the sleep laboratory and are expected to deepen our understanding of sleep disorders, improve patient-centered sleep care, augment day-to-day clinical operations, and increase our knowledge of the role of sleep in health at a population level.

Towards More Accurate Automatic Sleep Staging via Deep Transfer Learning

BACKGROUND

Despite recent significant progress in the development of automatic sleep staging methods, building a good model still remains a big challenge for sleep studies with a small cohort due to the data-variability and data-inefficiency issues. This work presents a deep transfer learning approach to overcome these issues and enable transferring knowledge from a large dataset to a small cohort for automatic sleep staging.

METHODS

We start from a generic end-to-end deep learning framework for sequence-to-sequence sleep staging and derive two networks as the means for transfer learning. The networks are first trained in the source domain (i.e. the large database). The pretrained networks are then finetuned in the target domain (i.e. the small cohort) to complete knowledge transfer. We employ the Montreal Archive of Sleep Studies (MASS) database consisting of 200 subjects as the source domain and study deep transfer learning on three different target domains: the Sleep Cassette subset and the Sleep Telemetry subset of the Sleep-EDF Expanded database, and the Surrey-cEEGrid database. The target domains are purposely adopted to cover different degrees of data mismatch to the source domains.

RESULTS

Our experimental results show significant performance improvement on automatic sleep staging on the target domains achieved with the proposed deep transfer learning approach.

CONCLUSIONS

These results suggest the efficacy of the proposed approach in addressing the above-mentioned data-variability and data-inefficiency issues.

SIGNIFICANCE

As a consequence, it would enable one to improve the quality of automatic sleep staging models when the amount of data is relatively small.¹¹The source code and the pretrained models are published at https://github.com/pquochuy/sleep_transfer_learning.

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.

Deep learning enables sleep staging from photoplethysmogram for patients with suspected sleep apnea

Study Objectives

Accurate identification of sleep stages is essential in the diagnosis of sleep disorders (e.g. obstructive sleep apnea [OSA]) but relies on labor-intensive electroencephalogram (EEG)-based manual scoring. Furthermore, long-term assessment of sleep relies on actigraphy differentiating only between wake and sleep periods without identifying specific sleep stages and having low reliability in identifying wake periods after sleep onset. To address these issues, we aimed to develop an automatic method for identifying the sleep stages from the photoplethysmogram (PPG) signal obtained with a simple finger pulse oximeter.

Methods

PPG signals from the diagnostic polysomnographies of susptected OSA patients (n = 894) were utilized to develop a combined convolutional and recurrent neural network. The deep learning model was trained individually for three-stage (wake/NREM/REM), four-stage (wake/N1+N2/N3/REM), and five-stage (wake/N1/N2/N3/REM) classification of sleep.

Results

The three-stage model achieved an epoch-by-epoch accuracy of 80.1% with Cohen’s κ of 0.65. The four- and five-stage models achieved 68.5% (κ = 0.54), and 64.1% (κ = 0.51) accuracies, respectively. With the five-stage model, the total sleep time was underestimated with a mean bias error (SD) of of 7.5 (55.2) minutes.

Conclusion

The PPG-based deep learning model enabled accurate estimation of sleep time and differentiation between sleep stages with a moderate agreement to manual EEG-based scoring. As PPG is already included in ambulatory polygraphic recordings, applying the PPG-based sleep staging could improve their diagnostic value by enabling simple, low-cost, and reliable monitoring of sleep and help assess otherwise overlooked conditions such as REM-related OSA.

The Dreem Headband compared to polysomnography for electroencephalographic signal acquisition and sleep staging

Study Objectives

The development of ambulatory technologies capable of monitoring brain activity during sleep longitudinally is critical for advancing sleep science. The aim of this study was to assess the signal acquisition and the performance of the automatic sleep staging algorithms of a reduced-montage dry-electroencephalographic (EEG) device (Dreem headband, DH) compared to the gold-standard polysomnography (PSG) scored by five sleep experts.

Methods

A total of 25 subjects who completed an overnight sleep study at a sleep center while wearing both a PSG and the DH simultaneously have been included in the analysis. We assessed (1) similarity of measured EEG brain waves between the DH and the PSG; (2) the heart rate, breathing frequency, and respiration rate variability (RRV) agreement between the DH and the PSG; and (3) the performance of the DH’s automatic sleep staging according to American Academy of Sleep Medicine guidelines versus PSG sleep experts manual scoring.

Results

The mean percentage error between the EEG signals acquired by the DH and those from the PSG for the monitoring of α was 15 ± 3.5%, 16 ± 4.3% for β, 16 ± 6.1% for λ, and 10 ± 1.4% for θ frequencies during sleep. The mean absolute error for heart rate, breathing frequency, and RRV was 1.2 ± 0.5 bpm, 0.3 ± 0.2 cpm, and 3.2 ± 0.6%, respectively. Automatic sleep staging reached an overall accuracy of 83.5 ± 6.4% (F1 score: 83.8 ± 6.3) for the DH to be compared with an average of 86.4 ± 8.0% (F1 score: 86.3 ± 7.4) for the 5 sleep experts.

Conclusions

These results demonstrate the capacity of the DH to both monitor sleep-related physiological signals and process them accurately into sleep stages. This device paves the way for, large-scale, longitudinal sleep studies.

Clinical Trial Registration

NCT03725943.

Gated temporal convolutional neural network and expert features for diagnosing and explaining physiological time series: A case study on heart rates

BACKGROUND AND OBJECTIVE

Physiological time series are common data sources in many health applications. Mining data from physiological time series is crucial for promoting healthy living and reducing governmental medical expenditure. Recently, research and applications of deep learning methods on physiological time series have developed rapidly because such data can be continuously recorded by smart wristbands or smartwatches. However, existing deep learning methods suffer from excessive model complexity and a lack of explanation. This paper aims to handle these issues.

METHODS

We propose TEG-net, which is a novel deep learning method for accurately diagnosing and explaining physiological time series. TEG-net constructs T-net (a multi-scale bi-directional temporal convolutional neural network) to model physiological time series directly, E-net (personalized linear model) to model expert features extracted from physiological time series, and G-net (gating neural network) to combine T-net and E-net for diagnosis. The combination of T-net and E-net through G-net improves diagnosis accuracy and E-net can be utilized for explanation.

RESULTS

Experimental results demonstrate that TEG-net outperforms the second-best baseline by 13.68% in terms of area under the receiver operating characteristic curve and 11.49% in terms of area under the precision-recall curve. Additionally, intuitive justifications can be provided to explain model predictions.

CONCLUSIONS

This paper develops an ensemble method to combine expert features and deep learning method for modeling physiological time series. Improvements in diagnostic accuracy and explanation make TEG-net applicable to many real-world health applications.

DeepSleep convolutional neural network allows accurate and fast detection of sleep arousal

Sleep arousals are transient periods of wakefulness punctuated into sleep. Excessive sleep arousals are associated with symptoms such as sympathetic activation, non-restorative sleep, and daytime sleepiness. Currently, sleep arousals are mainly annotated by human experts through looking at 30-second epochs (recorded pages) manually, which requires considerable time and effort. Here we present a deep learning approach for automatically segmenting sleep arousal regions based on polysomnographic recordings. Leveraging a specific architecture that 'translates' input polysomnographic signals to sleep arousal labels, this algorithm ranked first in the "You Snooze, You Win" PhysioNet Challenge. We created an augmentation strategy by randomly swapping similar physiological channels, which notably improved the prediction accuracy. Our algorithm enables fast and accurate delineation of sleep arousal events at the speed of 10 seconds per sleep recording. This computational tool would greatly empower the scoring process in clinical settings and accelerate studies on the impact of arousals.

Night-to-night variability of sleep electroencephalography-based brain age measurements

OBJECTIVE

Brain Age Index (BAI), calculated from sleep electroencephalography (EEG), has been proposed as a biomarker of brain health. This study quantifies night-to-night variability of BAI and establishes probability thresholds for inferring underlying brain pathology based on a patient's BAI.

METHODS

86 patients with multiple nights of consecutive EEG recordings were selected from Epilepsy Monitoring Unit patients whose EEGs reported as within normal limits. While EEGs with epileptiform activity were excluded, the majority of patients included in the study had a diagnosis of chronic epilepsy. BAI was calculated for each 12-hour segment of patient data using a previously established algorithm, and the night-to-night variability in BAI was measured.

RESULTS

The within-patient night-to-night standard deviation in BAI was 7.5 years. Estimates of BAI derived by averaging over 2, 3, and 4 nights had standard deviations of 4.7, 3.7, and 3.0 years, respectively.

CONCLUSIONS

Averaging BAI over n nights reduces night-to-night variability of BAI by a factor of n, rendering BAI a more suitable biomarker of brain health at the individual level. A brain age risk lookup table of results provides thresholds above which a patient has a high probability of excess BAI.

SIGNIFICANCE

With increasing ease of EEG acquisition, including wearable technology, BAI has the potential to track brain health and detect deviations from normal physiologic function. The measure of night-to-night variability and how this is reduced by averaging across multiple nights provides a basis for using BAI in patients' homes to identify patients who should undergo further investigation or monitoring.

Machine and Deep Learning in Molecular and Genetic Aspects of Sleep Research

Epidemiological sleep research strives to identify the interactions and causal mechanisms by which sleep affects human health, and to design intervention strategies for improving sleep throughout the lifespan. These goals can be advanced by further focusing on the environmental and genetic etiology of sleep disorders, and by development of risk stratification algorithms, to identify people who are at risk or are affected by, sleep disorders. These studies rely on comprehensive sleep-related data which often contains complex multi-dimensional physiological and molecular measurements across multiple timepoints. Thus, sleep research is well-suited for the application of computational approaches that can handle high-dimensional data. Here, we survey recent advances in machine and deep learning together with the availability of large human cohort studies with sleep data that can jointly drive the next breakthroughs in the sleep-research field. We describe sleep-related data types and datasets, and present some of the tasks in the field that can be targets for algorithmic approaches, as well as the challenges and opportunities in pursuing them.

Estimating daytime sleepiness with previous night electroencephalography, electrooculography, and electromyography spectrograms in patients with suspected sleep apnea using a convolutional neural network

A common symptom of obstructive sleep apnea (OSA) is excessive daytime sleepiness (EDS). The gold standard test for EDS is the multiple sleep latency test (MSLT). However, due to its high cost, MSLT is not routinely conducted for OSA patients and EDS is instead evaluated using sleep questionnaires. This is problematic however, since sleep questionnaires are subjective and correlate poorly with the MSLT. Therefore, new objective tools are needed for reliable evaluation of EDS. The aim of this study was to test our hypothesis that EDS can be estimated with neural network analysis of previous night polysomnographic signals. We trained a convolutional neural network (CNN) classifier using electroencephalography, electrooculography, and chin electromyography signals from 2,014 patients with suspected OSA. The CNN was trained to classify the patients into four sleepiness categories based on their mean sleep latency (MSL); severe (MSL < 5min), moderate (5 ≤ MSL < 10), mild (10 ≤ MSL < 15), and normal (MSL ≥ 15). The CNN classified patients to the four sleepiness categories with an overall accuracy of 60.6% and Cohen's kappa value of 0.464. In two-group classification scheme with sleepy (MSL < 10 min) and non-sleepy (MSL ≥ 10) patients, the CNN achieved an accuracy of 77.2%, with sensitivity of 76.5%, and specificity of 77.9%. Our results show that previous night's polysomnographic signals can be used for objective estimation of EDS with at least moderate accuracy. Since the diagnosis of OSA is currently confirmed by polysomnography, the classifier could be used simultaneously to get an objective estimate of the daytime sleepiness with minimal extra workload.

Rapid annotation of seizures and interictal-ictal-injury continuum EEG patterns

BACKGROUND

Manual annotation of seizures and interictal-ictal-injury continuum (IIIC) patterns in continuous EEG (cEEG) recorded from critically ill patients is a time-intensive process for clinicians and researchers. In this study, we evaluated the accuracy and efficiency of an automated clustering method to accelerate expert annotation of cEEG.

NEW METHOD

We learned a local dictionary from 97 ICU patients by applying k-medoids clustering to 592 features in the time and frequency domains. We utilized changepoint detection (CPD) to segment the cEEG recordings. We then computed a bag-of-words (BoW) representation for each segment. We further clustered the segments by affinity propagation. EEG experts scored the resulting clusters for each patient by labeling only the cluster medoids. We trained a random forest classifier to assess validity of the clusters.

RESULTS

Mean pairwise agreement of 62.6% using this automated method was not significantly different from interrater agreements using manual labeling (63.8%), demonstrating the validity of the method. We also found that it takes experts using our method 5.31 ± 4.44 min to label the 30.19 ± 3.84 h of cEEG data, more than 45 times faster than unaided manual review, demonstrating efficiency.

COMPARISON WITH EXISTING METHODS

Previous studies of EEG data labeling have generally yielded similar human expert interrater agreements, and lower agreements with automated methods.

CONCLUSIONS

Our results suggest that long EEG recordings can be rapidly annotated by experts many times faster than unaided manual review through the use of an advanced clustering method.

Unsupervised Sleep and Wake State Identification in Long-Term Electrocorticography Recordings

Studying the neural correlates of sleep can lead to revelations in our understanding of sleep and its interplay with different neurological disorders. Sleep research relies on manual annotation of sleep stages based on rules developed for healthy adults. Automating sleep stage annotation can expedite sleep research and enable us to better understand atypical sleep patterns. Our goal was to create a fully unsupervised approach to label sleep and wake states in human electro-corticography (ECoG) data from epilepsy patients. Here, we demonstrate that with continuous data from a single ECoG electrode, hidden semi-Markov models (HSMM) perform best in classifying sleep/wake states without excessive transitions, with a mean accuracy (n=4) of 85.2% compared to using K-means clustering (72.2%) and hidden Markov models (81.5%). Our results confirm that HSMMs produce meaningful labels for ECoG data and establish the groundwork to apply this model to cluster sleep stages and potentially other behavioral states.

Association of Sleep Electroencephalography-Based Brain Age Index With Dementia

IMPORTANCE

Dementia is an increasing cause of disability and loss of independence in the elderly population yet remains largely underdiagnosed. A biomarker for dementia that can identify individuals with or at risk for developing dementia may help close this diagnostic gap.

OBJECTIVE

To investigate the association between a sleep electroencephalography-based brain age index (BAI), the difference between chronological age and brain age estimated using the sleep electroencephalogram, and dementia.

DESIGN, SETTING, AND PARTICIPANTS

In this retrospective cross-sectional study of 9834 polysomnograms, BAI was computed among individuals with previously determined dementia, mild cognitive impairment (MCI), or cognitive symptoms but no diagnosis of MCI or dementia, and among healthy individuals without dementia from August 22, 2008, to June 4, 2018. Data were analyzed from November 15, 2018, to June 24, 2020.

EXPOSURE

Dementia, MCI, and dementia-related symptoms, such as cognitive change and memory impairment.

MAIN OUTCOMES AND MEASURES

The outcome measures were the trend in BAI when moving from groups ranging from healthy, to symptomatic, to MCI, to dementia and pairwise comparisons of BAI among these groups.

FINDINGS

A total of 5144 sleep studies were included in BAI examinations. Patients in these studies had a median (interquartile range) age of 54 (43-65) years, and 3026 (59%) were men. The patients included 88 with dementia, 44 with MCI, 1075 who were symptomatic, and 2336 without dementia. There was a monotonic increase in mean (SE) BAI from the nondementia group to the dementia group (nondementia: 0.20 [0.42]; symptomatic: 0.58 [0.41]; MCI: 1.65 [1.20]; dementia: 4.18 [1.02]; P < .001).

CONCLUSIONS AND RELEVANCE

These findings suggest that a sleep-state electroencephalography-based BAI shows promise as a biomarker associated with progressive brain processes that ultimately result in dementia.

Improved tracking of sevoflurane anesthetic states with drug-specific machine learning models

OBJECTIVE

The ability to monitor anesthetic states using automated approaches is expected to reduce inaccurate drug dosing and side-effects. Commercially available anesthetic state monitors perform poorly when ketamine is administered as an anesthetic-analgesic adjunct. Poor performance is likely because the models underlying these monitors are not optimized for the electroencephalogram (EEG) oscillations that are unique to the co-administration of ketamine.

APPROACH

In this work, we designed two k-nearest neighbors algorithms for anesthetic state prediction.

MAIN RESULTS

The first algorithm was trained only on sevoflurane EEG data, making it sevoflurane-specific. This algorithm enabled discrimination of the sevoflurane general anesthesia (GA) state from sedated and awake states (true positive rate = 0.87, [95% CI, 0.76, 0.97]). However, it did not enable discrimination of the sevoflurane-plus-ketamine GA state from sedated and awake states (true positive rate = 0.43, [0.19, 0.67]). In our second algorithm, we implemented a cross drug training paradigm by including both sevoflurane and sevoflurane-plus-ketamine EEG data in our training set. This algorithm enabled discrimination of the sevoflurane-plus-ketamine GA state from sedated and awake states (true positive rate = 0.91, [0.84, 0.98]).

SIGNIFICANCE

Instead of a one-algorithm-fits-all-drugs approach to anesthetic state monitoring, our results suggest that drug-specific models are necessary to improve the performance of automated anesthetic state monitors.

Automatic detection of cortical arousals in sleep and their contribution to daytime sleepiness

OBJECTIVE

Significant interscorer variability is found in manual scoring of arousals in polysomnographic recordings (PSGs). We propose a fully automatic method, the Multimodal Arousal Detector (MAD), for detecting arousals.

METHODS

A deep neural network was trained on 2,889 PSGs to detect cortical arousals and wakefulness in 1-second intervals. Furthermore, the relationship between MAD-predicted labels on PSGs and next day mean sleep latency (MSL) on a multiple sleep latency test (MSLT), a reflection of daytime sleepiness, was analyzed in 1447 MSLT instances in 873 subjects.

RESULTS

In a dataset of 1,026 PSGs, the MAD achieved an F1 score of 0.76 for arousal detection, while wakefulness was predicted with an accuracy of 0.95. In 60 PSGs scored by nine expert technicians, the MAD performed comparable to four and significantly outperformed five expert technicians for arousal detection. After controlling for known covariates, a doubling of the arousal index was associated with an average decrease in MSL of 40 seconds (p = 0.0075).

CONCLUSIONS

The MAD performed better or comparable to human expert scorers. The MAD-predicted arousals were shown to be significant predictors of MSL.

SIGNIFICANCE

This study validates a fully automatic method for scoring arousals in PSGs.

Artificial intelligence in sleep medicine: background and implications for clinicians

None

Polysomnography remains the cornerstone of objective testing in sleep medicine and results in massive amounts of electrophysiological data, which is well-suited for analysis with artificial intelligence (AI)-based tools. Combined with other sources of health data, AI is expected to provide new insights to inform the clinical care of sleep disorders and advance our understanding of the integral role sleep plays in human health. Additionally, AI has the potential to streamline day-to-day operations and therefore optimize direct patient care by the sleep disorders team. However, clinicians, scientists, and other stakeholders must develop best practices to integrate this rapidly evolving technology into our daily work while maintaining the highest degree of quality and transparency in health care and research. Ultimately, when harnessed appropriately in conjunction with human expertise, AI will improve the practice of sleep medicine and further sleep science for the health and well-being of our patients.

Excess brain age in the sleep electroencephalogram predicts reduced life expectancy

The brain age index (BAI) measures the difference between an individual's apparent "brain age" (BA; estimated by comparing EEG features during sleep from an individual with age norms), and their chronological age (CA); that is BAI = BA-CA. Here, we evaluate whether BAI predicts life expectancy. Brain age was quantified using a previously published machine learning algorithm for a cohort of participants ≥40 years old who underwent an overnight sleep electroencephalogram (EEG) as part of the Sleep Heart Health Study (n = 4877). Excess brain age (BAI >0) was associated with reduced life expectancy (adjusted hazard ratio: 1.12, [1.03, 1.21], p = 0.002). Life expectancy decreased by -0.81 [-1.44, -0.24] years per standard-deviation increase in BAI. Our findings show that BAI, a sleep EEG-based biomarker of the deviation of sleep microstructure from patterns normal for age, is an independent predictor of life expectancy.

A novel sleep stage scoring system: Combining expert-based features with the generalized linear model

In this study, we aim to automate the sleep stage scoring process of overnight polysomnography (PSG) data while adhering to expert-based rules. We developed a sleep stage scoring algorithm utilizing the generalized linear modelling (GLM) framework and extracted features from electroencephalogram (EEG), electromyography (EMG) and electrooculogram (EOG) signals based on predefined rules of the American Academy of Sleep Medicine (AASM) Manual for Scoring Sleep. Specifically, features were computed in 30-s epochs in the time and frequency domains of the signals and were then used to model the probability of an epoch being in each of five sleep stages: N3, N2, N1, REM or Wake. Finally, each epoch was assigned to a sleep stage based on model predictions. The algorithm was trained and tested on PSG data from 38 healthy individuals with no reported sleep disturbances. The overall scoring accuracy reached on the test set was 81.50 ± 1.14% (Cohen's kappa, κ = 0.73 ± 0.02 ). The test set results were highly comparable to the training set, indicating robustness of the algorithm. Furthermore, our algorithm was compared to three well-known commercialized sleep-staging tools and achieved higher accuracies than all of them. Our results suggest that automatic classification is highly consistent with visual scoring. We conclude that our algorithm can reproduce the judgement of a scoring expert and is also highly interpretable. This tool can assist visual scorers to speed up their process (from hours to minutes) and provides a method for a more robust, quantitative, reproducible and cost-effective PSG evaluation, supporting assessment of sleep and sleep disorders.

Scalable automatic sleep staging in the era of Big Data

Numerous automatic sleep staging approaches have been proposed to provide an eHealth alternative to the current gold-standard - hypnogram scoring by human experts. However, a majority of such studies exploit data of limited scale, which compromises both the validation and the reproducibility and transferability of such automatic sleep staging systems in real clinical settings. In addition, the computational issues and physical meaningfulness of the analysis are typically neglected, yet affordable computation is a key criterion in Big Data analytics. To this end, we establish a comprehensive analysis framework to rigorously evaluate the feasibility of automatic sleep staging from multiple perspectives, including robustness with respect to the number of training subjects, model complexity, and different classifiers. This is achieved for a large collection of publicly accessible polysomnography (PSG) data, recorded over 515 subjects. The trade-off between affordable computation and satisfactory accuracy is shown to be fulfilled by an extreme learning machine (ELM) classifier, which in conjunction with the physically meaningful hidden Markov model (HMM) of the transition between the different sleep stages (smoothing model) is shown to achieve both fast computation and the highest average Cohen's kappa value of κ = 0.73 (Substantial Agreement). Finally, it is shown that for accurate and robust automatic sleep staging, a combination of structural complexity (multi-scale entropy) and frequency-domain (spectral edge frequency) features is both computationally affordable and physically meaningful.