Sleep Stages Research Articles

Multi-night acoustic stimulation is associated with better sleep, amyloid dynamics, and memory in older adults with cognitive impairment.

Sleep is a potential early, modifiable risk factor for cognitive decline and dementia. Impaired slow wave sleep (SWS) is pronounced in individuals with cognitive impairment (CI). Cognitive decline and impairments of SWS are bi-directionally linked in a vicious cycle. SWS can be enhanced non-invasively using phase-locked acoustic stimulation (PLAS), potentially breaking this vicious cycle. Eighteen healthy older adults (HC, agemean±sd, 68.3 ± 5.1) and 16 older adults (agemean±sd, 71.9 ± 3.9) with CI (Montreal Cognitive Assessment ≤ 25) underwent one baseline (sham-PLAS) night and three consecutive stimulation nights (real-PLAS). EEG responses and blood-plasma amyloid beta Aβ42/Aβ40 ratio were measured pre- and post-intervention, as was episodic memory. The latter was again evaluated 1week and 3months after the intervention. In both groups, PLAS induced a significant electrophysiological response in both voltage- and time-frequency analyses, and memory performance improved in association with the magnitude of this response. In the CI group, both electrophysiological and associated memory effects were delayed compared to the healthy group. After 3 intervention nights, electrophysiological response to PLAS was no longer different between CI and HC groups. Only in the CI sample, stronger electrophysiological responses were significantly associated with improving post-intervention Aβ42/Aβ40 ratios. PLAS seems to improve SWS electrophysiology, memory, and amyloid dynamics in older adults with CI. However, effects on memory require more time to unfold compared to healthy older adults. This indicates that PLAS may become a potential tool to ameliorate cognitive decline, but longer interventions are necessary to compensate for declining brain integrity. This study was pre-registered (clinicaltrials.gov: NCT04277104).

Selective Enhancement of REM Sleep in Male Rats through Activation of MT1 Receptors Located in the Locus Coeruleus Norepinephrine neurons.

Sleep disorders affect millions of people around the world and have a high comorbidity with psychiatric disorders. While current hypnotics mostly increase non-rapid eye movement sleep (NREMS), drugs acting selectively on enhancing rapid eye movement sleep (REMS) are lacking. This polysomnographic study in male rats showed that the first-in-class selective melatonin MT1 receptor partial agonist UCM871 increases the duration of REMs without affecting that of NREMS. The REMS-promoting effects of UCM871 occurred by inhibiting, in a dose-response manner, the firing activity of the locus coeruleus (LC) norepinephrine (NE) neurons, which express MT1 receptors. The increase of REMS duration and the inhibition of LC-NE neuronal activity by UCM871 were abolished by MT1 pharmacological antagonism and by an adeno-associated viral (AAV) vector which selectively knocked down MT1 receptors in the LC-NE neurons. In conclusion, MT1 receptor agonism inhibits LC-NE neurons and triggers REMS, thus representing a novel mechanism and target for REMS disorders and/or psychiatric disorders associated with REMS impairments.Significance Statement Rapid eye movement sleep (REMS) is involved in the processes of memory consolidation and emotional regulation, but drugs selectively enhancing REMS are scant. Herein, we show that the first-in-class selective melatonin MT1 receptor agonist UCM871, by inhibiting the activity of norepinephrine neurons in the locus coeruleus, an important nucleus regulating the sleep/wake cycle, selectively increases the duration of REMS. These findings enhance our current understanding of the neurobiology and pharmacology of REMS and provide a possible novel mechanism and target for disorders associated with REMS dysfunctions.

Sleep, Sleepiness, and Memory

Abstract Objective To evaluate the relationship between sleep and sleepiness with memory complaints. Materials and Methods Patients who were submitted to polysomnography between May and September of 2022 and answered the prospective and retrospective memory questionnaire and the Epworth sleepiness scale were included, respectively. Data were entered into an Excel spreadsheet and converted to a file compatible with the SPSS software. Results The sample consisted of 98 subjects, 62.2% male, mean age of 45.9 years, 73.4% overweight, 54.1% with comorbidities, and 51% with excessive sleepiness. There was a significant difference in sleep efficiency, respiratory disturbance index (RDI), slow wave sleep (SWS), and rapid eye movement (REM) sleep for the group with comorbidities; in latency to sleep and SWS between genders; and in RDI for the body mass index group. No correlation between RDI and memory could be identified, but there were statistically significant correlations between REM and sleep efficiency; RDI and REM sleep; RDI and SWS; SWS and sleep efficiency; and sleep efficiency and latency to sleep onset. Older adults performed better on memory tests when total sleep time (TST) is longer than 5 hours and excessive daytime sleepiness is related to complaints of prospective, retrospective, and total memory. Conclusion Elderly people with TST longer than 5 hours have a better memory. Although a correlation between RDI and memory was not observed, a correlation between excessive daytime sleepiness—one of the main symptoms of patients with sleep disorders—and memory was.

Respiration-triggered olfactory stimulation reduces obstructive sleep apnea severity: A prospective pilot study.

Obstructive sleep apnea is a prevalent sleep-disordered breathing condition characterized by repetitive reduction in breathing during sleep. The current care standard for obstructive sleep apnea is continuous positive air pressure devices, often suffering from low tolerance due to limited adherence. Capitalizing on the unique neurocircuitry of olfactory perception and its retained function during sleep, we conducted a pilot study to test transient, respiration-based olfactory stimulation as a treatment for obstructive sleep apnea markers. Thirty-two patients with obstructive sleep apnea (apnea-hypopnea index ≥ 15 events per hr) underwent two polysomnography sessions, "Odour" and "Control", in random order. In "Odour" nights, patients were presented with transient respiratory-based olfactory stimulation delivered via a computer-controlled commercial olfactometer (Scentific). The olfactometer, equipped with a wireless monitoring, analysed respiratory patterns and presented odour upon detection of respiratory events. No odours were presented in "Control" nights. Following exclusions, 17 patients entered the analysis (four women, 47.4 (10.5) years, body mass index: 29.4 (6.3) kg m-2). We observed that olfactory stimulation during sleep reduced the apnea-hypopnea index ("Odour": 17.2 (20.9), "Control": 28.2 (18.6), z = -3.337, p = 0.000846, BF10 [Bayesian Factor 10]= 57.9), reflecting an average decrease of 31.3% in the number of events. Relatedly, stimulation reduced the oxygen desaturation index by 26.9% ("Odour": 12.5 (15.8), "Control": 25.7 (25.9), z = -3.337, p = 0.000846, BF10 = 9.522). This effect was not linked to the severity of baseline obstructive sleep apnea markers (ρ = -0.042, p = 0.87). Olfactory stimulation did not arouse from sleep or affect sleep structure, measured as time per sleep stage (F1,16 = 0.088, p = 0.77). In conclusion, olfactory stimulation during sleep was effective in reducing the severity of obstructive sleep apnea markers without inducing arousals, and may provide a novel treatment for obstructive sleep apnea, prompting continued research.

Serum levels of neurotensin, pannexin-1, and sestrin-2 and the correlations with sleep quality or/and cognitive function in the patients with chronic insomnia disorder

ObjectivesTo examine serum concentrations of neurotensin, pannexin-1 and sestrin-2, and their correlations with subjective and objective sleep quality and cognitive function in the patients with chronic insomnia disorder (CID).MethodsSixty-five CID patients were enrolled continuously and fifty-six good sleepers in the same period were served as healthy controls (HCs). Serum levels of neurotensin, pannexin-1 and sestrin-2 were measured by enzyme-linked immunosorbent assays. Sleep quality was assessed with the Pittsburgh Sleep Quality Index (PSQI) and polysomnography, and mood was evaluated by 17-item Hamilton Depression Rating Scale. General cognitive function was assessed with the Chinese-Beijing Version of Montreal Cognitive Assessment and spatial memory was evaluated by Blue Velvet Arena Test (BVAT).ResultsRelative to the HCs, the CID sufferers had higher levels of neurotensin (t=5.210, p<0.001) and pannexin-1 (Z=−4.169, p<0.001), and lower level of sestrin-2 (Z=−2.438, p=0.015). In terms of objective sleep measures, pannexin-1 was positively associated with total sleep time (r=0.562, p=0.002) and sleep efficiency (r=0.588, p=0.001), and negatively with wake time after sleep onset (r=−0.590, p=0.001) and wake time (r=−0.590, p=0.001); sestrin-2 was positively associated with percentage of rapid eye movement sleep (r=0.442, p=0.016) and negatively with non-rapid eye movement sleep stage 2 in the percentage (r=−0.394, p=0.034). Adjusted for sex, age and HAMD, pannexin-1 was still associated with the above objective sleep measures, but sestrin-2 was only negatively with wake time (r=−0.446, p=0.022). However, these biomarkers showed no significant correlations with subjective sleep quality (PSQI score). Serum concentrations of neurotensin and pannexin-1 were positively associated with the mean erroneous distance in the BVAT. Adjusted for sex, age and depression, neurotensin was negatively associated with MoCA score (r=−0.257, p=0.044), pannexin-1 was positively associated with the mean erroneous distance in the BVAT (r=0.270, p=0.033).ConclusionsThe CID patients had increased neurotensin and pannexin-1 and decreased sestrin-2 in the serum levels, indicating neuron dysfunction, which could be related to poor sleep quality and cognitive dysfunction measured objectively.

Movement Termination of Slow-Wave Sleep—A Potential Biomarker?

Movement Termination of Slow-Wave Sleep—A Potential Biomarker?

FDA-cleared home sleep apnea testing devices

The demand for home sleep apnea testing (HSAT) devices is escalating, particularly in the context of the coronavirus 2019 (COVID-19) pandemic. The absence of standardized development and verification procedures poses a significant challenge. This study meticulously analyzed the approval process characteristics of HSAT devices by the U.S. Food and Drug Administration (FDA) from September 1, 2003, to September 1, 2023, with a primary focus on ensuring safety and clinical effectiveness. We examined 58 reports out of 1046 that underwent FDA clearance via the 510(k) and de novo pathways. A substantial surge in certifications after the 2022 pandemic was observed. Type-3 devices dominated, signifying a growing trend for both home and clinical use. Key measurement items included respiration and sleep analysis, with the apnea–hypopnea index (AHI) and sleep stage emerging as pivotal indicators. The majority of FDA-cleared HSAT devices adhered to electrical safety and biocompatibility standards. Critical considerations encompass performance and function testing, usability, and cybersecurity. This study emphasized the nearly indispensable role of clinical trials in ensuring the clinical effectiveness of HSAT devices. Future studies should propose guidances that specify stringent requirements, robust clinical trial designs, and comprehensive performance criteria to guarantee the minimum safety and clinical effectiveness of HSATs.

Automated sleep staging on reduced channels in children with epilepsy

ObjectivesThis 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).MethodsOvernight 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.ResultsIn 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.ConclusionWe 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.

Insights from the 2nd China intelligent sleep staging competition.

Artificial intelligence (AI) is quickly advancing in the field of sleep medicine, which bodes well for the potential of actual clinical use. In this study, an analysis of the 2nd China Intelligent Sleep Staging Competition was conducted to gain insights into the general level and constraints of AI-assisted sleep staging in China. The outcomes of 10 teams from the children's track and 13 teams from the adult track were investigated in this study. The analysis included overall performance, differences between five different sleep stages, variations across subjects, and performance during stage transitions. The adult track's accuracy peaked at 80.46%, while the children's track's accuracy peaked at 88.96%. On average, accuracy rates stood at 71.43% for children and 68.40% for adults. All results were produced within a mere 5-min timeframe. The N1 stage was prone to misclassification as W, N2, and R stages. In the adult track, significant differences were apparent among subjects (p < 0.05), whereas in the children's track, such differences were not observed. Nonetheless, both tracks experienced a performance decline during stage transitions. The computational speed of AI is remarkably fast, simultaneously holding the potential to surpass the accuracy of physicians. Improving the machine learning model's classification of the N1 stage and transitional periods between stages, along with bolstering its robustness to individual subject variations, is imperative for maximizing its ability in assisting clinical scoring.

Validation of spectral sleep scoring with polysomnography using forehead EEG device

IntroductionVisual scoring of sleep electroencephalography (EEG) has long been considered the gold standard for sleep staging. However, it has several drawbacks, including high cost, time-intensiveness, vulnerability to human variability, discomfort to patients, lack of visualization to validate the hypnogram, and no acknowledgment of differences between delta and slow oscillation deep sleep. This report highlights a spectral scoring approach that addresses all these shortcomings of visual scoring. Past algorithms have used spectral information to help classify traditional visual stages. The current method used the clearly visible spectral patterns to develop new spectral stages, which are similar to but not the same as visual stages. Importantly, spectral scoring delivers both a hypnogram and a whole-night spectrogram, which can be visually inspected to ensure accurate scoring.MethodsThis study compared traditional visual scoring of 32-channel polysomnography with forehead-only spectral scoring from an EEG patch worn concurrently. The PSG was visually scored by trained technicians and the forehead patch was scored spectrally. Because non-rapid eye movement (NREM) stage divisions in spectral scoring are not based on visual NREM stages, the agreements are not expected to be as high as other automated sleep scoring algorithms. Rather, they are a guide to understanding spectral stages as they relate to the more widely understood visual stages and to emphasize reasons for the differences.ResultsThe results showed that visual REM was highly recognized as spectral REM (89%). Visual wake was only scored as spectral Wake 47% of the time, partly because of excessive visual scoring of wake during Light and REM sleep. The majority of spectral Light (predominance of spindle power) was scored as N2 (74%), while less N2 was scored as Light (65%), mostly because of incorrect visual staging of Lo Deep sleep due to high-pass filtering. N3 was scored as both Hi Deep (13 Hz power, 42%) and Lo Deep (0–1 Hz power, 39%), constituting a total of 81% of N3.DiscussionThe results show that spectral scoring better identifies clinically relevant physiology at a substantially lower cost and in a more reproducible fashion than visual scoring, supporting further work exploring its use in clinical and research settings.

Two Cases of Sleep-Related Dissociative Disorder with Episodes of Nocturnal Eating

AbstractSleep-related dissociative disorder (SRDD) is a female-predominant psychiatric parasomnia that was first identified as a condition that mimics sleepwalking in 1989, and was included in the International Classification of Sleep Disorders, 2nd edition, in 2005, with a subsequent expanding literature of case series and case reports. The objective hallmark of SRDD, found in about half of the reported cases, is sustained electroencephalogram (EEG) wakefulness during dissociative episodes emerging during wake-sleep transitions or after awakenings from light non-rapid eye movement (NREM) sleep or rapid eye movement (REM) sleep. Herein are reported two additional cases of SRDD in two female patients aged 53 and 40 years with prominent histories of multimodal abuse (typical of SRDD), with childhood emotional and food deprivation abuse in Case 1, and childhood emotional, sexual, and physical abuse in Case 2. Both patients were affected by “sleep phobia” and had recurrent nocturnal eating episodes. Major findings from the cumulative literature on SRDD are reinforced by these cases, with additional findings being described, particularly nocturnal eating behaviors and priming/triggering factors.

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.

Neurophysiological features of STN LFP underlying sleep fragmentation in Parkinson’s disease

BackgroundSleep fragmentation is a persistent problem throughout the course of Parkinson’s disease (PD). However, the related neurophysiological patterns and the underlying mechanisms remained unclear.MethodWe recorded subthalamic nucleus (STN) local field...

Older adults at greater risk for Alzheimer’s disease show stronger associations between sleep apnea severity in REM sleep and verbal memory

BackgroundObstructive sleep apnea (OSA) increases risk for cognitive decline and Alzheimer’s disease (AD). While the underlying mechanisms remain unclear, hypoxemia during OSA has been implicated in cognitive impairment. OSA during rapid eye movement (REM) sleep is usually more severe than in non-rapid eye movement (NREM) sleep, but the relative effect of oxyhemoglobin desaturation during REM versus NREM sleep on memory is not completely characterized. Here, we examined the impact of OSA, as well as the moderating effects of AD risk factors, on verbal memory in a sample of middle-aged and older adults with heightened AD risk.MethodsEighty-one adults (mean age:61.7 ± 6.0 years, 62% females, 32% apolipoprotein E ε4 allele (APOE4) carriers, and 70% with parental history of AD) underwent clinical polysomnography including assessment of OSA. OSA features were derived in total, NREM, and REM sleep. REM-NREM ratios of OSA features were also calculated. Verbal memory was assessed with the Rey Auditory Verbal Learning Test (RAVLT). Multiple regression models evaluated the relationships between OSA features and RAVLT scores while adjusting for sex, age, time between assessments, education years, body mass index (BMI), and APOE4 status or parental history of AD. The significant main effects of OSA features on RAVLT performance and the moderating effects of AD risk factors (i.e., sex, age, APOE4 status, and parental history of AD) were examined.ResultsApnea–hypopnea index (AHI), respiratory disturbance index (RDI), and oxyhemoglobin desaturation index (ODI) during REM sleep were negatively associated with RAVLT total learning and long-delay recall. Further, greater REM-NREM ratios of AHI, RDI, and ODI (i.e., more events in REM than NREM) were related to worse total learning and recall. We found specifically that the negative association between REM ODI and total learning was driven by adults 60 + years old. In addition, the negative relationships between REM-NREM ODI ratio and total learning, and REM-NREM RDI ratio and long-delay recall were driven by APOE4 carriers.ConclusionGreater OSA severity, particularly during REM sleep, negatively affects verbal memory, especially for people with greater AD risk. These findings underscore the potential importance of proactive screening and treatment of REM OSA even if overall AHI appears low.

Sleep stage continuity is associated with objective daytime sleepiness in patients with suspected obstructive sleep apnea.

Excessive daytime sleepiness (EDS) in patients with obstructive sleep apnea (OSA) is poorly explained by standard clinical sleep architecture metrics. We hypothesized that reduced sleep stage continuity mediates this connection independently from standard sleep architecture metrics. 1,907 patients with suspected OSA with daytime sleepiness complaints underwent in-lab diagnostic polysomnography and next-day Multiple Sleep Latency Test (MSLT). Sleep architecture was evaluated with novel sleep-stage continuity quantifications (mean sleep stage duration and probability of remaining in each sleep stage), and conventional metrics (total N1, N2, N3 and REM times; and sleep onset latency). Multivariate analyses were utilized to identify variables associated with moderate EDS (5 ≤ mean daytime sleep latency (MSL) ≤ 10 minutes) and severe EDS (MSL < 5 minutes). Compared to those without EDS, participants with severe EDS had lower N3 sleep continuity (mean N3 period duration 10.4 vs 13.7 minutes, p<0.05), less N3 time (53.8 vs 76.5 minutes, p<0.05); greater total sleep time (374.0 vs 352.5 minutes, p<0.05) and greater N2 time (227.5 vs 186.8 minutes, p<0.05). After adjusting for standard sleep architecture metrics using multivariate logistic regression, decreased mean wake and N3 period duration, and the decreased probability of remaining in N2 and N3 sleep remained significantly associated with severe EDS, while the decreased probability of remaining in wake and N2 sleep were associated with moderate EDS. Patients with OSA with EDS experience lower sleep continuity, noticeable especially during N3 sleep and wake. Sleep-stage continuity quantifications assist in characterizing the sleep architecture and are associated with objective daytime sleepiness highlighting the need for more detailed evaluations of sleep quality.

Shared EEG correlates between non-REM parasomnia experiences and dreams

Sleepwalking and related parasomnias result from incomplete awakenings out of non-rapid eye movement sleep. Behavioral episodes can occur without consciousness or recollection, or in relation to dream-like experiences. To understand what accounts for these differences in consciousness and recall, here we recorded parasomnia episodes with high-density electroencephalography (EEG) and interviewed participants immediately afterward about their experiences. Compared to reports of no experience (19%), reports of conscious experience (56%) were preceded by high-amplitude EEG slow waves in anterior cortical regions and activation of posterior cortical regions, similar to previously described EEG correlates of dreaming. Recall of the content of the experience (56%), compared to no recall (25%), was associated with higher EEG activation in the right medial temporal region before movement onset. Our work suggests that the EEG correlates of parasomnia experiences are similar to those reported for dreams and may thus reflect core physiological processes involved in sleep consciousness.

Retuning of hippocampal representations during sleep.

Hippocampal representations that underlie spatial memory undergo continuous refinement following formation1. Here, to track the spatial tuning of neurons dynamically during offline states, we used a new Bayesian learning approach based on the spike-triggered average decoded position in ensemble recordings from freely moving rats. Measuring these tunings, we found spatial representations within hippocampal sharp-wave ripples that were stable for hours during sleep and were strongly aligned with place fields initially observed during maze exploration. These representations were explained by a combination of factors that included preconfigured structure before maze exposure and representations that emerged during θ-oscillations and awake sharp-wave ripples whileon the maze, revealing the contribution of these events in forming ensembles. Strikingly, the ripple representations during sleep predicted the future place fields of neurons during re-exposure to the maze, even when those fields deviated from previous place preferences. By contrast, we observed tunings with poor alignment to maze place fields during sleep and rest before maze exposure and in the later stages of sleep. In sum, the new decoding approach allowed us to infer and characterize the stability and retuning of place fields during offline periods, revealing the rapid emergence of representations following new exploration and the role of sleep in the representational dynamics of the hippocampus.

Age-Related Changes in Sleep and Its Implications for Cognitive Decline in Aging Persons With Schizophrenia: A Critical Review.

Cognitive impairment is a core feature of schizophrenia that worsens with aging and interferes with quality of life. Recent work identifies sleep as an actionable target to alleviate cognitive deficits. Cardinal non-rapid eye movement (NREM) sleep oscillations such as sleep spindles and slow oscillations are critical for cognition. People living with schizophrenia (PLWS) and their first-degree relatives have a specific reduction in sleep spindles and an abnormality in their temporal coordination with slow oscillations that predict impaired memory consolidation. While NREM oscillatory activity is reduced in typical aging, it is not known how further disruption in these oscillations contributes to cognitive decline in older PLWS. Another understudied risk factor for cognitive deficits among older PLWS is obstructive sleep apnea (OSA) which may contribute to cognitive decline. We conducted a narrative review to examine the published literature on aging, OSA, and NREM sleep oscillations in PLWS. Spindles are propagated via thalamocortical feedback loops, and this circuitry shows abnormal hyperconnectivity in schizophrenia as revealed by structural and functional MRI studies. While the risk and severity of OSA increase with age, older PLWS are particularly vulnerable to OSA-related cognitive deficits because OSA is often underdiagnosed and undertreated, and OSA adds further damage to the circuitry that generates NREM sleep oscillations. We highlight the critical need to study NREM sleep in older PWLS and propose that identifying and treating OSA in older PLWS will provide an avenue to potentially mitigate and prevent cognitive decline.

Machine learning-empowered sleep staging classification using multi-modality signals

The goal is to enhance an automated sleep staging system's performance by leveraging the diverse signals captured through multi-modal polysomnography recordings. Three modalities of PSG signals, namely electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG), were considered to obtain the optimal fusions of the PSG signals, where 63 features were extracted. These include frequency-based, time-based, statistical-based, entropy-based, and non-linear-based features. We adopted the ReliefF (ReF) feature selection algorithms to find the suitable parts for each signal and superposition of PSG signals. Twelve top features were selected while correlated with the extracted feature sets' sleep stages. The selected features were fed into the AdaBoost with Random Forest (ADB + RF) classifier to validate the chosen segments and classify the sleep stages. This study's experiments were investigated by obtaining two testing schemes: epoch-wise testing and subject-wise testing. The suggested research was conducted using three publicly available datasets: ISRUC-Sleep subgroup1 (ISRUC-SG1), sleep-EDF(S-EDF), Physio bank CAP sleep database (PB-CAPSDB), and S-EDF-78 respectively. This work demonstrated that the proposed fusion strategy overestimates the common individual usage of PSG signals.

A multi-task learning model using RR intervals and respiratory effort to assess sleep disordered breathing

BackgroundSleep-disordered breathing (SDB) affects a significant portion of the population. As such, there is a need for accessible and affordable assessment methods for diagnosis but also case-finding and long-term follow-up. Research has focused on exploiting cardiac and respiratory signals to extract proxy measures for sleep combined with SDB event detection. We introduce a novel multi-task model combining cardiac activity and respiratory effort to perform sleep–wake classification and SDB event detection in order to automatically estimate the apnea–hypopnea index (AHI) as severity indicator.MethodsThe proposed multi-task model utilized both convolutional and recurrent neural networks and was formed by a shared part for common feature extraction, a task-specific part for sleep–wake classification, and a task-specific part for SDB event detection. The model was trained with RR intervals derived from electrocardiogram and respiratory effort signals. To assess performance, overnight polysomnography (PSG) recordings from 198 patients with varying degree of SDB were included, with manually annotated sleep stages and SDB events.ResultsWe achieved a Cohen’s kappa of 0.70 in the sleep–wake classification task, corresponding to a Spearman’s correlation coefficient (R) of 0.830 between the estimated total sleep time (TST) and the TST obtained from PSG-based sleep scoring. Combining the sleep–wake classification and SDB detection results of the multi-task model, we obtained an R of 0.891 between the estimated and the reference AHI. For severity classification of SBD groups based on AHI, a Cohen’s kappa of 0.58 was achieved. The multi-task model performed better than a single-task model proposed in a previous study for AHI estimation, in particular for patients with a lower sleep efficiency (R of 0.861 with the multi-task model and R of 0.746 with single-task model with subjects having sleep efficiency < 60%).ConclusionAssisted with automatic sleep–wake classification, our multi-task model demonstrated proficiency in estimating AHI and assessing SDB severity based on AHI in a fully automatic manner using RR intervals and respiratory effort. This shows the potential for improving SDB screening with unobtrusive sensors also for subjects with low sleep efficiency without adding additional sensors for sleep–wake detection.