Representations of temporal sleep dynamics: review and synthesis of the literature

Factors associated with poor sleep in children with drug-resistant epilepsy

OBJECTIVE

We aimed to investigate sleep in children with drug-resistant epilepsy (DRE), including developmental and epileptic encephalopathies (DEEs). Next, we examined differences in sleep macrostructure and microstructure and questionnaire outcomes between children with well-controlled epilepsy (WCE) and children with DRE. Furthermore, we wanted to identify factors associated with poor sleep outcome in these children, as some factors might be targets to improve epilepsy and neurodevelopmental outcomes.

METHODS

A cross-sectional study was conducted in children 4 to 18-years-old. Children without epilepsy, with WCE, and with DRE were included. Overnight electroencephalography (EEG), including chin electromyography and electrooculography, to allow sleep staging, was performed. Parents were asked to fill out a sleep questionnaire. Classical five-stage sleep scoring was performed manually, spindles were automatically counted, and slow wave activity (SWA) in the first and last hour of slow wave sleep was calculated.

RESULTS

One hundred eighty-two patients were included: 48 without epilepsy, 75 with WCE, and 59 with DRE. We found that children with DRE have significantly lower sleep efficiency (SE%), less time spent in rapid eye movement (REM) sleep, fewer sleep spindles, and a lower SWA decline over the night compared to children with WCE. Subjectively more severe sleep problems were reported by the caregivers and more daytime sleepiness was present in children with DRE. Least absolute shrinkage and selection operator (LASSO) regression showed that multifocal interictal epileptiform discharges (IEDs), benzodiazepine treatment, and longer duration of epilepsy were associated with lower SE% and lower REM sleep time. The presence of multifocal discharges and cerebral palsy was associated with fewer spindles. Benzodiazepine treatment, drug resistance, seizures during sleep, intellectual disability, and older age were associated with lower SWA decline.

SIGNIFICANCE

Both sleep macrostructure and microstructure are severely impacted in children with DRE, including those with DEEs. Epilepsy parameters play a distinct role in the disruption REM sleep, spindle count, and SWA decline.

Overnight Sleep Staging Using Chest-Worn Accelerometry

Overnight sleep staging is an important part of the diagnosis of various sleep disorders. Polysomnography is the gold standard for sleep staging, but less-obtrusive sensing modalities are of emerging interest. Here, we developed and validated an algorithm to perform "proxy" sleep staging using cardiac and respiratory signals derived from a chest-worn accelerometer. We collected data in two sleep centers, using a chest-worn accelerometer in combination with full PSG. A total of 323 participants were analyzed, aged 13-83 years, with BMI 18-47 kg/m2. We derived cardiac and respiratory features from the accelerometer and then applied a previously developed method for automatic cardio-respiratory sleep staging. We compared the estimated sleep stages against those derived from PSG and determined performance. Epoch-by-epoch agreement with four-class scoring (Wake, REM, N1+N2, N3) reached a Cohen's kappa coefficient of agreement of 0.68 and an accuracy of 80.8%. For Wake vs. Sleep classification, an accuracy of 93.3% was obtained, with a sensitivity of 78.7% and a specificity of 96.6%. We showed that cardiorespiratory signals obtained from a chest-worn accelerometer can be used to estimate sleep stages among a population that is diverse in age, BMI, and prevalence of sleep disorders. This opens up the path towards various clinical applications in sleep medicine.

Customizable automated cleaning of multichannel sleep EEG in SleepTrip

While standard polysomnography has revealed the importance of the sleeping brain in health and disease, more specific insight into the relevant brain circuits requires high-density electroencephalography (EEG). However, identifying and handling sleep EEG artifacts becomes increasingly challenging with higher channel counts and/or volume of recordings. Whereas manual cleaning is time-consuming, subjective, and often yields data loss (e.g., complete removal of channels or epochs), automated approaches suitable and practical for overnight sleep EEG remain limited, especially when control over detection and repair behavior is desired. Here, we introduce a flexible approach for automated cleaning of multichannel sleep recordings, as part of the free Matlab-based toolbox SleepTrip. Key functionality includes 1) channel-wise detection of various artifact types encountered in sleep EEG, 2) channel- and time-resolved marking of data segments for repair through interpolation, and 3) visualization options to review and monitor performance. Functionality for Independent Component Analysis is also included. Extensive customization options allow tailoring cleaning behavior to data properties and analysis goals. By enabling computationally efficient and flexible automated data cleaning, this tool helps to facilitate fundamental and clinical sleep EEG research.

Sleep After Concussion: A Scoping Review of Sensor Technologies

Sleep disturbances following a concussion/mild traumatic brain injury are associated with longer recovery times and more comorbidities. Sensor technologies can directly monitor sleep-related physiology and provide objective sleep metrics. This scoping review determines how sensor technologies are currently used to monitor sleep following a concussion. We searched Ovid (Medline, Embase), Web of Science, CINAHL, Compendex Engineering Village, and PsycINFO from inception to June 20, 2022, following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for scoping reviews. Included studies objectively monitored sleep in participants with concussion. We screened 1081 articles and included 37 in the review. A total of 17 studies implemented polysomnography (PSG) months to years after injury for a median of two nights and provided a wide range of sleep metrics, including sleep-wake times, sleep stages, arousal indices, and periodic limb movements. Twenty-two studies used actigraphy days to weeks after injury for a median of 10 days and nights and provided information limited to sleep-wake times. Sleep stages were most reported in PSG studies, and sleep efficiency was most reported in actigraphy studies. For both technologies there was high variability in reported outcome measures. Sleep sensing technologies may be used to identify how sleep affects concussion recovery. However, high variability in sensor deployment methodologies makes cross-study comparisons difficult and highlights the need for standardization. Consensus on how sleep sensing technologies are used post-concussion may lead to clinical integration with subjective methods for improved sleep monitoring during the recovery period.

Functional changes in sleep-related arousal after ketamine administration in individuals with treatment-resistant depression

The glutamatergic modulator ketamine is associated with changes in sleep, depression, and suicidal ideation (SI). This study sought to evaluate differences in arousal-related sleep metrics between 36 individuals with treatment-resistant major depression (TRD) and 25 healthy volunteers (HVs). It also sought to determine whether ketamine normalizes arousal in individuals with TRD and whether ketamine's effects on arousal mediate its antidepressant and anti-SI effects. This was a secondary analysis of a biomarker-focused, randomized, double-blind, crossover trial of ketamine (0.5 mg/kg) compared to saline placebo. Polysomnography (PSG) studies were conducted one day before and one day after ketamine/placebo infusions. Sleep arousal was measured using spectral power functions over time including alpha (quiet wakefulness), beta (alert wakefulness), and delta (deep sleep) power, as well as macroarchitecture variables, including wakefulness after sleep onset (WASO), total sleep time (TST), rapid eye movement (REM) latency, and Post-Sleep Onset Sleep Efficiency (PSOSE). At baseline, diagnostic differences in sleep macroarchitecture included lower TST (p = 0.006) and shorter REM latency (p = 0.04) in the TRD versus HV group. Ketamine's temporal dynamic effects (relative to placebo) in TRD included increased delta power earlier in the night and increased alpha and delta power later in the night. However, there were no significant diagnostic differences in temporal patterns of alpha, beta, or delta power, no ketamine effects on sleep macroarchitecture arousal metrics, and no mediation effects of sleep variables on ketamine's antidepressant or anti-SI effects. These results highlight the role of sleep-related variables as part of the systemic neurobiological changes initiated after ketamine administration. Clinical Trials Identifier: NCT00088699.

Automated sleep staging on reduced channels in children with epilepsy

Objectives

This study aimed to validate a sleep staging algorithm using in-hospital video-electroencephalogram (EEG) in children without epilepsy, with well-controlled epilepsy (WCE), and with drug-resistant epilepsy (DRE).

Methods

Overnight video-EEG, along with electrooculogram (EOG) and chin electromyogram (EMG), was recorded in children between 4 and 18 years of age. Classical sleep staging was performed manually as a ground truth. An end-to-end hierarchical recurrent neural network for sequence-to-sequence automatic sleep staging (SeqSleepNet) was used to perform automated sleep staging using three channels: C4-A1, EOG, and chin EMG.

Results

In 176 children sleep stages were manually scored: 47 children without epilepsy, 74 with WCE, and 55 with DRE. The 5-class sleep staging accuracy of the automatic sleep staging algorithm was 84.7% for the children without epilepsy, 83.5% for those with WCE, and 80.8% for those with DRE (Kappa of 0.79, 0.77, and 0.73 respectively). Performance per sleep stage was assessed with an F1 score of 0.91 for wake, 0.50 for N1, 0.83 for N2, 0.84 for N3, and 0.86 for rapid eye movement (REM) sleep.

Conclusion

We concluded that the tested algorithm has a high accuracy in children without epilepsy and with WCE. Performance in children with DRE was acceptable, but significantly lower, which could be explained by a tendency of more time spent in N1, and by abundant interictal epileptiform discharges and intellectual disability leading to less recognizable sleep stages. REM sleep time, however, significantly affected in children with DRE, can be detected reliably by the algorithm.Clinical trial registration: ClinicalTrials.gov, identifier NCT04584385.

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.

Certainty about uncertainty in sleep staging: a theoretical framework

Sleep stage classification is an important tool for the diagnosis of sleep disorders. Because sleep staging has such a high impact on clinical outcome, it is important that it is done reliably. However, it is known that uncertainty exists in both expert scorers and automated models. On average, the agreement between human scorers is only 82.6%. In this study, we provide a theoretical framework to facilitate discussion and further analyses of uncertainty in sleep staging. To this end, we introduce two variants of uncertainty, known from statistics and the machine learning community: aleatoric and epistemic uncertainty. We discuss what these types of uncertainties are, why the distinction is useful, where they arise from in sleep staging, and provide recommendations on how this framework can improve sleep staging in the future.