No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications
... The gold standard for SDB diagnosis is interpreting the multichannel signals recorded through polysomnography (PSG) [8]; However, this method has several limitations. For example, patients feel uncomfortable because numerous sensors might interfere with sleep; PSG measurement is highly labor-intensive and must be performed in a special environment, thus limiting its application to the whole population. ...
... Quantitatively, an apnea event (OSA or CSA) is identified when the airflow breathing amplitude decreases more than 90% for a duration ranging from 10 to 120 seconds, whereas an HYP event is identified when either of the following two conditions holds: (1) the airflow breathing amplitude decreases more than 30% of the pre-event baseline with ≥ 4% oxygen desaturation; (2) the airflow breathing amplitude decrease more than 50% of the pre-event baseline with ≥ 3% oxygen desaturation or with an arousal for a duration ranging from 10 to 120 seconds, but does not fulfill the criteria for apnea. In this study, we followed the AASM 2007 [8] and did not classify an HYP event as a central or obstructive in nature, although the AASM updated the scoring criteria for sleep disordered breathing events in 2012 [31]. The 2012 scoring rule [31] for sleep apnea is identical to that in 2007, whereas the HYP rule is modified as "a HYP event is identified when the airflow breathing amplitude decreases over 30% of the pre-event baseline with ≥ 3% oxygen desaturation or with an arousal". ...
... Ground truth by sleep experts. We considered the respiratory activity scored by sleep experts according the AASM 2007 guideline [8] as the ground truth. We referred to the expert's score as the "PSG state." ...
Physiologically, the thoracic (THO) and abdominal (ABD) movement signals, captured using wearable piezo-electric bands, provide information about various types of apnea, including central sleep apnea (CSA) and obstructive sleep apnea (OSA). However, the use of piezo-electric wearables in detecting sleep apnea events has been seldom explored in the literature. This study explored the possibility of identifying sleep apnea events, including OSA and CSA, by solely analyzing {one or both the THO and ABD signals. An adaptive non-harmonic model was introduced to model the THO and ABD signals, which allows us to design features for sleep apnea events. To confirm the suitability of the extracted features, a support vector machine was applied to classify three categories -- normal and hypopnea, OSA, and CSA. According to a database of} 34 subjects, the overall classification accuracies were on average and , respectively, based on the cross validation. When the features determined from the THO and ABD signals were combined, the overall classification accuracy became . These features were applied for designing a state machine for online apnea event detection. Two event-by-event accuracy indices, S and I, were proposed for evaluating the performance {of the state machine. For the same database, the} S index was , and the I index was . The results indicate the considerable potential of applying the proposed algorithm to clinical examinations for both screening and homecare purposes.
... Various biomarkers across the temporal and frequency domains play a pivotal role in sleep stage classification. These include characteristics such as sleep spindles, K-complexes, slow-wave activity, power spectral density, and frequency bands (e.g., delta, theta, alpha, beta, and gamma) [30][31][32][33] . Figure 1 visualises typical examples of these wave patterns found during the various stages of normal sleep 34 , Figure 10.2. ...
... For both EEG-SESM and EEGNet, the highest accuracy was observed for the N3 class. Conversely, the lowest accuracy was recorded for N1 sleep, which often only involves subtle changes in EEG signals compared to restful wakefulness 31 . Additionally, N1 sleep is the most under-represented class in the dataset. ...
... Features that contain more frequency-rich information were selected with less accuracy, such as slow waves in N2 and N3 sleep. The inability of the model to identify spectral-based bio-markers may explain the lower accuracy for the N1 and N3 classes, as the EEG features for these classes are strongly related to phenomenon in the frequency domain 31 . To the author's knowledge, this prototype generation and analysis represents the first examination of prototypes sampled from data in real-time aligning to EEG biomarkers. ...
Prototype-based methods in deep learning offer interpretable explanations for decisions by comparing inputs to typical representatives in the data. This study explores the adaptation of SESM, a self-attention-based prototype method successful in electrocardiogram (ECG) tasks, for electroencephalogram (EEG) signals. The architecture is evaluated on sleep stage classification, exploring its efficacy in predicting stages with single-channel EEG. The model achieves comparable test accuracy compared to EEGNet, a state-of-the-art black-box architecture for EEG classification. The generated prototypical components are exaimed qualitatively and using the area over the perterbation curve (AOPC) indicate some alignment with expected bio-markers for different sleep stages such as alpha spindles and slow waves in non-REM sleep, but the results are severely limited by the model’s ability to only extract and present information in the time-domain. Ablation studies are used to explore the impact of kernel size, number of heads, and diversity threshold on model performance and explainability. This study represents the first application of a self-attention based prototype method to EEG data and provides a step forward in explainable AI for EEG data analysis.
... PSG equipment was used to record sleep EEG along with electromyography, ECG and electro-oculography (collectively referred to as PSG). PSG scoring was based on the guidelines of the American Academy of Sleep Medicine and was assessed at a blinded laboratory [22]. ...
... A meta-analysis of randomized controlled trials involving FDAapproved antidepressant agents [24] reported that compared to placebo, the mean weighted effect-size for these agents was 0.37 (95% CI, 0.33-0.41) in published studies and 0.15 (95% CI, 0.08-0. 22) in unpublished studies. In contrast, in the current study, which involved an MDD sample enriched by removing early placebo responders, the effect size for the 20 mg seltorexant dose group was 0.58. ...
The antidepressant efficacy and safety of seltorexant monotherapy in major depressive disorder (MDD) was investigated in a placebo-controlled, placebo lead-in, randomized, double-blind, phase 1b study. Participants were randomized to receive seltorexant (20 mg or 40 mg) or placebo. The treatment effect was assessed by changes in the Hamilton Rating Scale for Depression-17 item (HDRS17) from treatment-period baseline to week 5 in lead-in placebo non-responders (“enriched” intent-to-treat analysis set). As a secondary outcome, the effect of seltorexant on HDRS17 was assessed in patients with and without subjective insomnia. Seltorexant’s effects on polysomnography, serum cortisol, and cortisol waking response were also measured. In total, 128 participants were enrolled, including 86 in the enriched sample (lead-in placebo non-responders). The mean changes from baseline (SD) in HDRS17 score at week 5 differed significantly across arms: −7.0 (5.04) for seltorexant 20 mg, −5.5 (4.34) for seltorexant 40 mg, and −4.4 (3.67) for placebo (p = 0.0456), which was attributable to the difference between the 20 mg and placebo arms (p = 0.0049). Improvement in depression severity at week 5 for seltorexant 20 mg was greater in patients with higher baseline insomnia severity (nominal p = 0.0059). The treatment benefit in the 20 mg arm remained significant when HDRS scores were adjusted by removing the sleep items (nominal p = 0.0289). The mean HDRS17 change versus placebo was numerically larger in the 20 mg than the 40 mg arm, consistent with data from a previous study in which seltorexant was administered adjunctively to conventional antidepressants. In secondary analyses, the waking cortisol response decreased in the 20 mg arm but not the 40 mg or placebo arms, and while total sleep increased more in the 40 mg arm, this arm also showed reduced REM onset latency and increased stage N1 sleep, which were not evident in the 20 mg arm. These biomarker data suggest mechanistic hypotheses that may account for the apparent curvilinear dose-response relationship of seltorexant. Trial Registration: ClinicalTrials.gov, NCT03374475.
... The classification process ensures a systematic and accurate assessment of sleep architecture. NREM can be further divided into four stages-S1, S2, S3, and S4-according to the R&K manual [7] or into three stages-N1, N2, and N3-according to the American Medical Sleep Association (AASM) manual [8]. In clinical practice, sleep staging is usually performed manually by an experienced physician. ...
... ), θ (4-8 Hz), α(8)(9)(10)(11)(12)(13), β(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), and γ (30-35 Hz) bands are extracted, as well as the ratio of fast to slow waves.Secondly, the extracted features are subjected to principal component analysis (PCA) to reduce the redundant information of the features. ...
The pivotal role of sleep has led to extensive research endeavors aimed at automatic sleep stage classification. However, existing methods perform poorly when classifying small groups or individuals, and these results are often considered outliers in terms of overall performance. These outliers may introduce bias during model training, adversely affecting feature selection and diminishing model performance. To address the above issues, this paper proposes an ensemble-based sequential convolutional neural network (E-SCNN) that incorporates a clustering module and neural networks. E-SCNN effectively ensembles machine learning and deep learning techniques to minimize outliers, thereby enhancing model robustness at the individual level. Specifically, the clustering module categorizes individuals based on similarities in feature distribution and assigns personalized weights accordingly. Subsequently, by combining these tailored weights with the robust feature extraction capabilities of convolutional neural networks, the model generates more accurate sleep stage classifications. The proposed model was verified on two public datasets, and experimental results demonstrate that the proposed method obtains overall accuracies of 84.8% on the Sleep-EDF Expanded dataset and 85.5% on the MASS dataset. E-SCNN can alleviate the outlier problem, which is important for improving sleep quality monitoring for individuals.
... In 1968, Rechtschaffen and Kales proposed in the old standard (R&K) that sleep is divided into six stages: Wakefulness (W); Non-Rapid Eye Movement (NREM), which is divided into stage 1 (S1), stage 2 (S2), stage 3 (S3), stage 4 (S4), and Rapid Eye Movement (REM) [4]. Subsequently, the American Academy of Sleep Medicine (AASM) revised these standards in 2007 [5], amalgamating stages 3 (S3) and 4 (S4) into N3. Thus, the ASM-defined sleep stages were W, N1, N2, N3, and R. Polysomnography (PSG) is commonly employed in clinical practice for sleep assessment, enabling the recording of various signals, such as electroencephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), electrocardiogram (ECG), and other physiological indicators. ...
... Previous studies have demonstrated transitional relationships between various sleep stages [5], for instance, the R stage seldom precedes the W and N2 stages. Recurrent Neural Networks (RNNs) have found extensive application in research domains involving sequential data. ...
Polysomnography (PSG) is commonly used to diagnose sleep disorders. However, manual sleep staging is a time-consuming task due to high human effort and technical thresholds, and it involves certain subjective factors. To improve the efficiency of sleep staging, this paper proposes a deep learning automatic sleep staging method based on multidimensional sleep data, the MCSSN model. This model uses continuous wavelet transform (CWT) to first extract sleep features, followed by data enhancement of time-frequency features using the SMOTE algorithm. Finally, a two-step training algorithm is used to learn the sleep features and sleep state transition rules. The evaluation results show that the model achieved an excellent overall accuracy of 85.87% on a sample of 104 subjects, with a macro average score (
F
1 M
) of 0.82 and a Kappa of 0.80. The model achieved excellent classification performance for N1 stage detection, with an F1 of 0.64, outperforming the other models. The experimental results show that the SMOTE oversampling algorithm plays an active role in detecting sleep stages, especially the N1 stage, which is difficult to identify. In addition, learning sleep transition rules through a two-step training algorithm helps to improve the performance, and the MCSSN algorithm provides a reference for the automatic design of subsequent sleep classification networks.
... • Obstructive and central Apnea: the definition of apnea is given in the previous paragraph on apnea detection; obstructive apnea is characterized by the presence of inspiratory effort throughout the period of reduced airflow, whereas in central apnea inspiratory effort does not appear at all. A mixed event is characterised by the absence of inspiratory effort in the initial part, followed by a period of inspiratory effort [253]. • Obstructive and central Hypopnea: the hypopnea definition is given in the previous paragraph on apnea detection; Obstructive hypopnea is characterized by snoring during the event, increased flattening of the inspiratory portion of the nasal pressure signal, or thoracoabdominal paradox, which is asynchronous movement of the chest and abdomen during breathing. ...
... The RADAR alternative is attractive for sleep medicine because it is completely noninvasive and is certainly much more convenient for the subject than wearable sensors. For the next summary we will use the following nomenclature of sleep stages [253]: ...
In the field of physiological signals monitoring and its applications, non-contact technology is often proposed as a possible alternative to traditional contact devices. The ability to extract information about a patient’s health status in an unobtrusive way, without stressing the subject and without the need of qualified personnel, fuels research in this growing field. Among the various methodologies, RADAR-based non-contact technology is gaining great interest. This scoping review aims to summarize the main research lines concerning RADAR-based physiological sensing and machine learning applications reporting recent trends, issues and gaps with the scientific literature, best methodological practices, employed standards to be followed, challenges, and future directions. After a systematic search and screening, two hundred and seven papers were collected following the guidelines of PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses). The included records covered two macro-areas being regression of physiological signals or physiological features (
n=77
papers) and the other a cluster of papers regarding the processing of RADAR-based physiological signals and features; the latter cluster concerns four fields of interest, being RADAR-based diagnosis (
n=77
), RADAR-based human behaviour monitoring (
n=25
), RADAR-based biometric authentication (
n=19
) and RADAR-based affective computing (
n=9
). Papers collected under the diagnosis category were further divided, on the basis of their aims: in breath pattern classification (
n=41
), infection detection (
n=10
), sleep stage classification (
n=9
), heart disease detection (
n=9
) and quality detection (
n=8
). Papers collected under the human behaviour monitoring were further divided based on their aims: fatigue detection (
n=9
), human detection (
n=7
), human localisation (
n=4
), human orientation (
n=2
), and activities classification (
n=3
).
... SWS was identified if there were at least two continuous, synchronized, high-amplitude slow waves occurred within a 5-second epoch, and accordingly, the remaining NREM periods were classified as NS ( Figure 1B,I). Our offline analysis showed that, consistent with human studies, [28,29] SWS in mice exhibited significantly higher power spectrum density (PSD) in delta band (0.8-4 Hz) and longer duration of SO than the NS ( Figure 1D,G); while NS showed significantly longer duration of spindles than SWS ( Figure 1H). Sleep cycles in rodents are shorter and more fragmented than in humans. ...
... The spectrum power density (PSD) was calculated by Welch's power spectrum density estimate with Hanning window tapering. The individual frequency band was defined as the delta band (0.8-4 Hz), theta band (4-8 Hz), alpha band (8)(9)(10)(11)(12)(13)(14) and beta band (14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30). The spindle band was defined as 10-14 Hz. ...
Sleep stabilizes memories for their consolidation, but how to modify specific fear memory during sleep remains unclear. Here, it is reported that using targeted memory reactivation (TMR) to reactivate prior fear learning experience in non‐slow wave sleep (NS) inhibits fear memory consolidation, while TMR during slow wave sleep (SWS) enhances fear memory in mice. Replaying conditioned stimulus (CS) during sleep affects sleep spindle occurrence, leading to the reduction or enhancement of slow oscillation‐spindle (SO‐spindle) coupling in NS and SWS, respectively. Optogenetic inhibition of pyramidal neurons in the frontal association cortex (FrA) during TMR abolishes the behavioral effects of NS‐TMR and SWS‐TMR by modulating SO‐spindle coupling. Notably, calcium imaging of the L2/3 pyramidal neurons in the FrA shows that CS during SWS selectively enhances the activity of neurons previously activated during fear conditioning (FC+ neurons), which significantly correlates with CS‐elicited spindle power spectrum density. Intriguingly, these TMR‐induced calcium activity changes of FC+ neurons further correlate with mice freezing behavior, suggesting their contributions to the consolidation of fear memories. The findings indicate that TMR can selectively weaken or strengthen fear memory, in correlation with modulating SO‐spindle coupling and the reactivation of FC+ neurons during substages of non‐rapid eye movement (NREM) sleep.
... According to the American Academy of Sleep Medicine manual (AASM manual), 26 we classified modifications of the EEG frequency (with increased chin tone if in REM stage) as "arousal" if lasting 3-15 s, and as "awakening" if lasting more than 15 s. The representation of NREM3 sleep and REM sleep was considered reduced when accounting for less than 20% and 25% of TST, respectively. ...
... Conversely, the representation of NREM1 sleep was considered increased when accounting for more than 5% of TST. 26 ...
Objectives
People with neurodevelopmental disorders frequently experience sleep disturbances, negatively impairing their quality of life. We aimed to determine the prevalence and nature of sleep disturbances in patients with SCN8A ‐related disorders.
Methods
Through a collaborative network of caregivers and clinicians, we collected data about epilepsy, cognitive/motor abilities, medications, and relevant comorbidities of patients harboring a pathogenic SCN8A variant. The Sleep Disturbance Scale for Children (SDSC), the Children's Sleep Habits Questionnaire (CSHQ‐22‐items), and the Pediatric Daytime Sleepiness Scale (PDSS) were distributed and evaluated by factor scores and Composite Sleep Index. Video‐EEG‐polysomnographic recordings were performed.
Results
We enrolled 47 patients (age range: 2–39 years), whose phenotypes ranged from SCN8A ‐DEE to intellectual disability without epilepsy. In the majority of them (82%), sleep disturbances were reported and/or observed at the SDSC. The most frequent were difficulty in initiating and maintaining sleep (64%), followed by sleep breathing disorder (43%), sleep–wake transition disorder (34%), and daytime sleepiness (34%). Sleep disturbances were more frequent in patients with severe DEE (96%) and ongoing seizures (93%) and were more severe in patients with sleep‐related seizures. The CSHQ and PDSS confirmed difficulty in initiating and maintaining sleep. Polysomnographic recordings (9 patients/20 nights) showed an altered sleep structure in 95%, with frequent arousals, mainly not seizure related.
Significance
More than 80% of patients with SCN8A ‐related disorders presented with sleep disturbances, primarily consisting of sleep instability with difficulty of initiating and maintaining sleep. Animal studies showed sleep disturbances in SCN8A‐ and SCN1A ‐Dravet Syndrome mice models, suggesting a role for voltage‐gated sodium channels in the regulation of sleep. Understanding the effects of SCN8A dysfunction on sleep stability may guide future therapeutic efforts to alleviate this often distressing symptom also in seizure‐free SCN8A patients.
Plain Language Summary
In this study, we analyzed sleep disturbances in patients with disorders related to a genetic mutation in the gene SCN8A . We found that the majority of the patients experienced sleep disturbances, mainly consisting in difficulty of initiating and maintaining sleep. Sleep disturbances were more frequent in patients with severe cognitive impairment and active epilepsy and more severe in patients with seizures during sleep.
... The PSG was divided into 30-sec epochs and was analyzed independently by two experts employing standard criteria [11] and taking account of amendments introduced by the American Academy of Sleep Medicine [12]. The latent period of falling asleep, total sleep duration, total duration of nocturnal wakings, duration of REM sleep, and durations of stages 1, 2, and 3 of slow-wave sleep were calculated. ...
... The latent period of falling asleep, total sleep duration, total duration of nocturnal wakings, duration of REM sleep, and durations of stages 1, 2, and 3 of slow-wave sleep were calculated. In each PSG, numbers of activations were also counted separately for the slow-wave and REM sleep phases in accordance with the criteria of the American Academy of Sleep Medicine [12]. The numbers of activations in the slow-wave and REM phases were related to total sleep duration. ...
... These included arousal index (AI; arousals per hour of sleep), oxygen desaturation index (ODI, 3% drops per hour of sleep, logtransformed in regression analysis), percentage of time in stage N3, sleep efficiency (% sleep time during the recording period), total sleep time, oAHI, and minimal oxygen saturation. The oAHI was defined using AASM pediatric criteria and was log-transformed to account for skewed distribution [10]. ...
Purpose
Although asthma is common in children with sleep-disordered breathing (SDB), it is unclear whether and to what extent asthma is associated with SDB-related outcomes. Our objectives are to describe risk factors for asthma among children with mild SDB (mSDB) and assess the association between asthma and the severity of sleep-related outcomes.
Methods
Cross-sectional analyses were conducted for children aged 3–12.9 years with mSDB enrolled in Pediatric Adenotonsillectomy for Snoring Children Study. Sleep-related outcomes included SDB symptoms (Pediatric Sleep Questionnaire-Sleep-Related Breathing Disorder scale (PSQ-SRBD)), SDB-specific quality of life (OSA-18), sleepiness (modified Epworth Sleepiness Score) and polysomnographic and actigraphic measures. Asthma was defined by caregiver-reported diagnosis with current asthma symptoms and medication use, or a Composite Asthma Severity Index (CASI) score ≥ 4. Asthma was further categorized into mild (CASI < 4) and moderate-to-severe (CASI ≥ 4). Regression analyses were conducted to identify asthma risk factors and estimate the associations between mild and moderate-to-severe asthma with sleep-related outcomes.
Results
The sample included 425 children (20.3%-Black, 17.4%-Hispanic; 51.7%-female). The prevalence of asthma was 19.1% (7.1% moderate-to-severe, 12.0% mild). Environmental tobacco smoke exposure and markers of atopy were associated with asthma in multivariable-adjusted analyses. Moderate-to-severe asthma was associated with increased OSA symptoms measured by PSQ-SRBD (adjusted effect estimate for moderate-to-severe vs. no asthma (adj; 95%CI): 0.08; 0.01, 0.15)) and decreased quality of life measured by OSA-18 (adj; 95%CI: 7.5; 1.20, 13.82)), and a small increase in the arousal index (adj; 95%CI: 0.80; 0.09, 1.51)).
Conclusion
Moderate-to-severe asthma was associated with worse QoL and greater SDB symptoms among children with mSDB. The co-occurrence of common risk factors for mSDB and asthma and worse symptoms and quality of life in children with both conditions support coordinated strategies for prevention and co-management of both disorders.
Clinical trial
Pediatric Adenotonsillectomy for Snoring (PATS), NCT02562040, https://clinicaltrials.gov/study/NCT02562040
... We extracted TST, and percentage of N1, N2, N3 and REM sleep, REML and SL. Following the American Academy Sleep Medicine (AASM) scoring rules, N3 represents slow wave sleep (SWS) and replaces stage 3 and stage 4 from the Rechtschaffen and Kales (R&K) scoring rules [23]. Additional PSG variables included REM density, REM intensity, awakening time, and number of awakenings. ...
Background
To explore the polysomnographic differences between patients with obsessive-compulsive disorder (OCD) and healthy controls.
Methods
An electronic literature search was conducted in MEDLINE, EMBASE, All EBM databases, Web of Science, and CNKI from inception to Sep 2023.
Results
Pooled analyses revealed significant reductions in sleep efficiency, total sleep time, and rapid eye movement (REM) sleep latency, and increases in awakening time, REM density and REM intensity in patients with OCD compared with healthy controls. Meta-regression analyses reveal no significant influence of OCD severity and disease duration on polysomnographic changes of patients with OCD.
Conclusions
Significant polysomnographic changes are present in OCD. Our findings underscore the need for a comprehensive PSG assessment of sleep changes in patients with OCD.
... CSR was diagnosed with an AHI ≥ 5/h of central apneas and/or central hypopneas associated with at least 3 consecutive cycles of crescendo/decrescendo breathing patterns or as episodes of at least 10 consecutive minutes of crescendo and decrescendo changes in breathing amplitude. OSAS was diagnosed with an obstructive apnea-hypopnea index (AHI) ≥ 15/h or ≥5/h plus symptoms suggestive of SDB [26]. All sleep studies were reviewed manually by trained personnel and checked by clinical sleep experts. ...
Background: Impairment in autonomic activity is a prognostic marker in patients with heart failure (HF), and its involvement has been suggested in cardiovascular complications of obstructive sleep apnea syndrome (OSAS) and Cheyne–Stokes respiration (CSR). This prospective observational study aims to investigate the implications of sleep-disordered breathing (SDB) on hemodynamic regulation and autonomic activity in chronic HF patients. Methods: Chronic HF patients, providing confirmation of reduced ejection fraction (≤35%), underwent polysomnography, real-time hemodynamic, heart rate variability (HRV), and baroreceptor reflex sensitivity (BRS) assessments using the Task Force Monitor. BRS was assessed using the sequencing method during resting conditions and stress testing. Results: Our study population (n = 58) was predominantly male (41 vs. 17), with a median age of 61 (±11) yrs and a median BMI of 30 (±5) kg/m2. Patients diagnosed with CSR were 13.8% (8/58) and 50.0% (29/58) with OSAS. No differences in the real-time assessment of hemodynamic regulation, heart rate variability, or baroreceptor reflex function were found between patients with OSAS, CSR, and patients without SDB. A subgroup analysis of BRS and HRV in patients with severe SDB (AHI > 30/h) and without SDB (AHI < 5) revealed numerically reduced BRS and increased LF/HF-RRI values under resting conditions, as well as during mental testing in patients with severe SDB. Patients with moderate-to-severe SDB had a shorter overall survival, which was, however, dependent upon age. Conclusions: Chronic HF patients with severe SDB may exhibit lower baroreceptor function and impaired cardiovascular autonomic function in comparison with HF patients without SDB.
... All PSGs were scored by a registered sleep technologist and subsequently reviewed by experienced pediatric sleep medicine physicians. Sleep stages and respiratory events were scored according to the standard pediatrics criteria of the American Association of Sleep Medicine (AASM) guidelines [17,18] using the AASM scoring manual. Obstructive apneas were defined as the absence of airflow with continued increased inspiratory effort for duration of at least two breaths. ...
Purpose
To assess OSA prevalence, comorbidities, and the influence of sleep stages and body positions on respiratory events distribution in toddlers aged 12–24 months.
Methods
A single center retrospective study that included toddlers aged 12–24 months old who underwent overnight PSG. OSA severity was categorized by obstructive apnea-hypopnea index (OAHI) as mild (1–4.9 events/h), moderate (5–9.9 events/h), and severe (≥ 10 events/h).
Results
283 PSG data were included with a median age of 18 months (IQR 16–20.25) for the OSA group (168/283) and 19 months (IQR 16–22) for the non-OSA group (115/283) (p = 0.047). OSA prevalence was 68.5% (42.3% mild, 18.5% moderate, and 39% severe). 38.1% of children had no comorbidities, 24.4% had a history of prematurity and 11.3% had Down syndrome. Multivariate binominal regression analysis showed that children with history of prematurity (p = 0.017) and Down syndrome (p = 0.043) had higher odds of having OSA. The mean SaO2 in REM sleep was lower, and the mean time spent with oxygen saturation below 90% was higher in children with neuromuscular disease compared to those with other comorbidities. In toddlers without comorbidities, the median REM OAHI was 29.8 events/h (IQR: 58.48), whereas the median non-REM OAHI was 4.1 events/h (IQR: 10.4 p < 0.001). Supine OAHI was 7.9 (IQR: 24.9), and off supine OAHI was 10.5 (IQR: 18.1, p = 0.407).
Conclusion
In toddlers aged 12–24 months, history of prematurity and Down syndrome were significantly associated with OSA. Obstructive respiratory events occurred predominantly in REM sleep, and no significant positional relations were noted.
... PG records were subsequently evaluated independently using established scoring guidelines. 19 Total sleep time (TST) was standardized for age according to the National Sleep Foundation's sleep time duration recommendation [20]. Sleep efficiency was calculated as follows: TST/time in bed*100%. ...
Background
Adenotonsillar hypertrophy is the most frequent cause for obstructive sleep apnea (OSAS) in children. In patients with small tonsils and where adenoid size cannot be assessed, the indication for adenoidectomy often relies on clinical symptoms. However, data on the association of clinical parameters and adenoid hypertrophy with OSAS severity in children undergoing an adenoidectomy is sparse.
Aim
To investigate the correlation of patient characteristics, adenoid hypertrophy, and clinical symptoms with OSAS severity in pediatric patients indicated for an adenoidectomy.
Methods
We performed a retrospective chart review of all pediatric patients at our tertiary referral center between 2018 and 2023 who underwent polygraphy (PG) for OSAS diagnostics. Adenoid hypertrophy was assessed as adenoid-choanal ratio (AC-ratio) via nasal endoscopy and clinical symptom score (CS) via physical examination and parental survey. We included all symptomatic children with mild to severe OSAS (apnea–hypopnea index (AHI) ≥ 1). Exclusion criteria were obesity according to BMI and/or the presence of systemic diseases. The patients were divided according to age in a preschool and school cohort. Patient characteristics and PG data were compared between both groups. Linear regression analysis was used to investigate the association of AC-ratio, CS and BMI with the AHI.
Results
A total of 121 patients were identified of which 81 were included in our study, resulting in 42 and 39 patients from 3–5 and 6–14 years of age, respectively. We observed a significant correlation between CS and BMI (p = 0.026) and the CS and AC-ratio (p < 0.001). Univariable regression analysis showed significant association of the AC-ratio and CS with AHI-score for the total (p < 0.001), the preschool (p < 0.001), and the school cohort (p < 0.001). In multivariable regression analysis, the significant association of AC-ratio and CS remained in the total (p = 0.014; p < 0.001), and the preschool cohort (p = 0.029; p = 0.002). However, only the CS remained as positive predictor in the school cohort.
Conclusion
AC-ratio and clinical symptoms seem to be reliable predictors for OSAS severity in patients between 3–14 years of age. Moreover, only clinical symptoms were associated with OSAS severity in schoolchildren. Future investigation should contribute to the validation of our results
... Wake after sleep onset was defined as the duration of time spent awake after sleep onset and before final awakening. Outcome data reported as N3 and NREM stage four (N4) sleep [28][29][30][31][32], slow wave sleep [23,24,26,27,[33][34][35][36][37], or deep sleep [38] were classified as N3 sleep in line with the current American Academy of Sleep Medicine guidelines [39], except for three studies [40][41][42] where N3 and N4 sleep were reported independently. ...
... This dataset, a resource of polysomnographic (PSG) recordings designed for sleep research, was originally annotated with six distinct sleep stages: wake, REM, S1, S2, S3, and S4. However, based on the standards introduced by the American Academy of Sleep Medicine (AASM) in 2007 [65], current studies typically consider only five stages: wake, S1, S2, S3, and REM. To align with this established framework, stages S3 and S4 are merged to represent the deep sleep stage. ...
The HeartBert model is introduced with three primary objectives: reducing the need for labeled data, minimizing computational resources, and simultaneously improving performance in machine learning systems that analyze Electrocardiogram (ECG) signals. Inspired by Bidirectional Encoder Representations from Transformers (BERT) in natural language processing and enhanced with a self-supervised learning approach, the HeartBert model-built on the RoBERTa architecture-generates sophisticated embeddings tailored for ECG-based projects in the medical domain. To demonstrate the versatility, generalizability, and efficiency of the proposed model, two key downstream tasks have been selected: sleep stage detection and heartbeat classification. HeartBERT-based systems, utilizing bidirectional LSTM heads, are designed to address complex challenges. A series of practical experiments have been conducted to demonstrate the superiority and advancements of HeartBERT, particularly in terms of its ability to perform well with smaller training datasets, reduced learning parameters, and effective performance compared to rival models. The code and data are publicly available at https://github.com/ecgResearch/HeartBert.
... (2) documented occurrence of sleep-related injurious, potentially injurious, or disruptive behaviors through patient history or PSG monitoring; (3) lack of epileptiform activity on EEG during REM sleep; and (4) absence of any alternative explanation for the behavior. Given the concerns surrounding the perceived arbitrariness and leniency of the RSWA criteria in the ICSD-3, we decided to adopt a stricter definition of RSWA by referring to the guidelines outlined in the American Academy of Sleep Medicine (AASM) manual [33]. Once a patient was diagnosed as having RBD, we arranged the patient a neuropsychological test and referred the patient to a neurologist for neurological examination. ...
Background: This study was aimed at analyzing cognitive function and quantitative electroencephalogram (qEEG) in patients with isolated REM sleep behavior disorder (iRBD) based on the presence of depression and at evaluating the impact of depression on phenoconversion to neurodegenerative diseases.
Methods: Individuals diagnosed with iRBD via polysomnography were included. Based on the presence of depression, patients were categorized into two groups. Neuropsychological tests and qEEG were conducted following the diagnosis of iRBD, and outcomes were compared between the two groups. Patients were regularly followed to monitor their phenoconversion status. Cox regression analysis was performed to assess the hazard ratio associated with depression.
Results: Ninety iRBD patients (70% males) were included, with a median age of 66.3 years. Depression was identified in 26 (28.9%) of these patients. The depressed group showed significantly poorer performance only in color reading subtest of Stroop (p=0.029) compared to the nondepressed group, showing reduced processing speed. In qEEG, relative gamma power (p=0.034) and high gamma power (p=0.020) in the parietal region were significantly higher in the depressed group than in the nondepressed group. Depression was associated with a hazard ratio of 3.32 for the risk of phenoconversion to neurodegenerative diseases in iRBD patients (p=0.011).
Conclusion: Depressive symptoms in iRBD patients should be closely monitored as they could aggravate cognitive dysfunction and increase the risk of phenoconversion to neurodegenerative diseases.
... Lastly, the EEG was re-referenced to the averaged mastoids and a lowpass filter of 60 Hz was applied. Following preprocessing, sleep stages were scored by an expert polysomnographic technologist following established guidelines 73 . The classification was performed using the EEGLABcompatible 74 Counting Sheep PSG toolbox (https://github.com/ ...
Sleep is essential for the optimal consolidation of newly acquired memories. This study examines the neurophysiological processes underlying memory consolidation during sleep, via reactivation. Here, we investigated the impact of slow wave - spindle (SW-SP) coupling on regionally-task-specific brain reactivations following motor sequence learning. Utilizing simultaneous EEG-fMRI during sleep, our findings revealed that memory reactivation occured time-locked to coupled SW-SP complexes, and specifically in areas critical for motor sequence learning. Notably, these reactivations were confined to the hemisphere actively involved in learning the task. This regional specificity highlights a precise and targeted neural mechanism, underscoring the crucial role of SW-SP coupling. In addition, we observed double-dissociation whereby primary sensory areas were recruited time-locked to uncoupled spindles; suggesting a role for uncoupled spindles in sleep maintenance. These findings advance our understanding the functional significance of SW-SP coupling for enhancing memory in a regionally-specific manner, that is functionally dissociable from uncoupled spindles.
... The sleep stages of NREM 1-3 (N1 to N3), wake, and REM sleep were scored offline and manually according to the criteria of the American Academy of Sleep Medicine (AASM) by visual inspection of the signals of the frontal, central, and occipital electrodes over 30 s epochs (Iber et al., 2007). Based on offline scoring, we confirmed TMR exposure during N2 and N3 and no significant differences (p-values >0.05) of sleep parameters between the cueing groups (see Supplementary file 5). ...
Sleep associated memory consolidation and reactivation play an important role in language acquisition and learning of new words. However, it is unclear to what extent properties of word learning difficulty impact sleep associated memory reactivation. To address this gap, we investigated in 22 young healthy adults the effectiveness of auditory targeted memory reactivation (TMR) during non-rapid eye movement sleep of artificial words with easy and difficult to learn phonotactical properties. Here, we found that TMR of the easy words improved their overnight memory performance, whereas TMR of the difficult words had no effect. By comparing EEG activities after TMR presentations, we found an increase in slow wave density independent of word difficulty, whereas the spindle-band power nested during the slow wave up-states – as an assumed underlying activity of memory reactivation – was significantly higher in the easy/effective compared to the difficult/ineffective condition. Our findings indicate that word learning difficulty by phonotactics impacts the effectiveness of TMR and further emphasize the critical role of prior encoding depth in sleep associated memory reactivation.
... Notably, slow oscillations (< 1 Hz) and delta waves differ in many aspects, such as their underlying mechanisms 20,60 and roles in cognitive functions including memory 36 .Entrainment of these two bands of neural oscillations may result in different effects. Moreover, 1 Hz is around the boundary between slow oscillations (< 1 Hz) and delta waves (1)(2)(3)(4) 36,61 . Therefore, BB stimulation at 1 Hz may have induced more complex processes in our study and resulted in the discrepancy between the effects of 0.25-Hz and 1-Hz BBs on sleep. ...
Binaural beats can entrain neural oscillations and modulate behavioral states. However, the effect of binaural beats, particularly those with slow frequencies (< 1 Hz), on sleep remains poorly understood. We hypothesized that 0.25-Hz beats can entrain neural oscillations and enhance slow-wave sleep by shortening its latency or increasing its duration. To investigate this, we included 12 healthy participants (six women; mean age, 25.4 ± 2.6 years) who underwent four 90-min afternoon nap sessions, comprising a sham condition (without acoustic stimulation) and three binaural-beat conditions (0, 0.25, or 1 Hz) with a 250-Hz carrier tone. The acoustic stimuli, delivered through earphones, were sustained throughout the 90-min nap period. Both N2- and N3- latencies were shorter in the 0.25-Hz binaural beats condition than in the sham condition. We observed no significant results regarding neural entrainment at slow frequencies, such as 0.25 and 1 Hz, and the modulation of sleep oscillations, including delta and sigma activity, by binaural beats. In conclusion, this study demonstrated the potential of binaural beats at slow frequencies, specifically 0.25 Hz, for inducing slow-wave sleep in generally healthy populations.
... Sleep was recorded by polysomnography (Nihon Kohden, Rosbach, Germany) each night. Recordings were scored offline in 30 s-epochs by two independent researchers using standard criteria (34). In the morning of day 2 (07:00 a.m.), participants were woken up and received a standardized meal at 08:30 a.m., 1:30 p.m., and 08:00 p.m., the macronutrient ratio adapted to the respective C+, F+ or CON diet. ...
Background
Daily dietary intake of macronutrients and energy is closely associated with long-term metabolic health outcomes, but whether 24-h nutritional intervention under isocaloric conditions leads to changes in metabolism remains unclear. Moreover, the short-term effect of diets with different macronutrient composition on hedonic appetite regulation is less clear.
Methods
This study examined the impact of an acute high-fat (F+) and high-carbohydrate (C+) diet on glucose metabolism and hedonic regulation of food intake in young healthy men under controlled conditions. Using a cross-over design, 19 male participants received a one-day isocaloric diet with different macronutrient composition (F+ = 11% carbohydrates, 74% fat; C+ = 79% carbohydrates, 6% fat) compared to a control diet (CON = 55% carbohydrates, 30% fat). Protein content was set at 15% of energy in all diets. The feeling of hunger, as well as “liking” and “wanting” for foods, was assessed through visual analog scales, and blood samples for glucose, insulin, and cortisol levels were assessed repeatedly during the experimental day. An intravenous glucose tolerance test was conducted the next morning.
Results
Postprandial glucose and insulin levels were lowest in F+ over the 24 h. Except for dinner, the CON diet showed the highest mean values in glucose. F+ diet improved insulin resistance, lowering Homeostatis Model Assessment Insulin Resistance (HOMA-IR) values. Changes in hedonic regulation of food intake were not observed during the intervention between the diets, except for higher feelings of satiety under the CON diet.
Conclusion
An acute, isocaloric, high-fat diet improved insulin resistance even in healthy individuals but did not affect hedonic food intake regulation. Macronutrient composition modulate glucose metabolism even under short-term (24-h) and isocaloric diets, which should be considered for personalized preventive dietary treatments.
... These parameters typically consist of electroencephalogram (EEG), eye movements (EOG), electromyogram (EMG), electrocardiogram (ECG), respiratory airflow, oxygen saturation, and other physiological data. PSG is regarded as the gold standard for the diagnosis of sleep disorders and the conduct of sleep research (Iber et al., 2007). In recent years, as technology has evolved, an increasing number of wearable devices have been utilized to measure sleep patterns and quality. ...
Background
Healthcare workers often encounter inadequate sleep conditions. However, limited research has examined the underlying sleep patterns among healthcare workers. This study aimed to identify sleep patterns in healthcare workers, explore predictors associated with various sleep patterns, and investigate the relationship between sleep patterns and psychiatric symptoms.
Methods
This cross-sectional study was conducted in Shenzhen, China, from April 2023 to June 2023. In total, data from 1,292 participants were included using a convenience sampling method. A latent profile analysis was conducted to identify sleep patterns based on the seven dimensions of the Pittsburgh Sleep Quality Index. Multinomial logistic regression analysis was conducted to investigate the influence of socio-demographic variables on each profile. A one-way ANOVA test was employed to examine the relationships between sleep patterns and psychiatric symptoms.
Results
Three distinct profiles were identified: good sleepers (63.9%), inefficient sleepers (30.3%), and poor sleepers (5.7%). Multinomial logistic regression analysis indicated that gender and marital status were predictors of various sleep patterns. The ANOVA revealed significant differences in psychiatric symptoms scores among the three sleep patterns; poor sleepers exhibited the highest levels of mental distress.
Conclusion
This study identified three distinct sleep patterns in healthcare workers and their significant associations with psychiatric symptoms. These findings contribute to the development of targeted intervention strategies aimed at improving sleep and reducing psychiatric symptoms among healthcare workers.
... Sleep fragmentation and high-frequency oscillatory activity (HFO) analysis Sleep fragmentation (SF), which is hypothesized to be a clinical biomarker of recovery after TBI, is measured by the sleep fragmentation index, representing the total number of awakenings and sleep stage shifts divided by total sleep time. 23,24 24-hour EEG recordings from 1, 7, and 28 dpi were manually scored for wakefulness, high delta, and theta activity using the AASM Sleep-Scoring Guidelines 25 (Fig. 2B). SF was compared across groups and time using a linear mixed model analysis (SPSS Statistical Software, version 29.0). ...
Functional connectivity (FC) after traumatic brain injury (TBI) is affected by an altered excitatory-inhibitory balance due to neuronal dysfunction, and the mechanistic changes observed could be reflected differently by contrasting methods. Local gamma event coupling FC (GEC-FC) is believed to represent multiunit fluctuations due to inhibitory dysfunction, and we hypothesized that FC derived from widespread, broadband amplitude signal (BBA-FC) would be different, reflecting broader mechanisms of functional disconnection. We tested this during sleep and active periods defined by high delta and theta electroencephalographic activity, respectively, at 1, 7, and 28 d after rat fluid-percussion-injury or sham injury (n = 6/group) using 10 indwelling, bilateral cortical, and hippocampal electrodes. We also measured seizure and high-frequency oscillatory activity (HFOs) as markers of electrophysiological burden. BBA-FC analysis showed early hyperconnectivity constrained to ipsilateral sensory-cortex-to-CA1-hippocampus that transformed to mainly ipsilateral FC deficits by 28 d compared with shams. These changes were conserved over active epochs, except at 28 d when there were no differences to shams. In comparison, GEC-FC analysis showed large regions of hyperconnectivity early after injury within similar ipsilateral and intra-hemispheric networks. GEC-FC weakened with time, but hyperconnectivity persisted at 28 d compared with sham. Edge and global connectivity measures revealed injury-related differences across time in GEC-FC as compared with BBA-FC, demonstrating greater sensitivity to FC changes post-injury. There was no significant association between sleep fragmentation, HFOs, or seizures with FC changes. The within-animal, spatial-temporal differences in BBA-FC and GEC-FC after injury may represent different mechanisms driving FC changes as a result of primary disconnection and interneuron loss.
... Sleep scoring was performed following the current criteria 40 . PSG derivations were placed according to recommended rules 40 to evaluate sleep features, respiratory, cardiac, and limb events. ...
The phenoconversion trajectory from idiopathic/isolated Rapid eye movement (REM) sleep behavior disorder (iRBD) towards either Parkinson's Disease (PD) or Dementia with Lewy Bodies (DLB) is currently uncertain. We investigated the capability of baseline brain [18F]FDG-PET in differentiating between iRBD patients eventually phenoconverting to PD or DLB, by deriving the denovoPDRBD-related pattern (denovoPDRBD-RP) from 32 de novo PD patients; and the denovoDLBRBD-RP from 30 de novo DLB patients, both with evidence of RBD at diagnosis. To explore [18F]FDG-PET phenoconversion trajectories prediction power, we applied these two patterns on a group of 115 iRBD patients followed longitudinally. At follow-up (25.6 ± 17.2 months), 42 iRBD patients progressed through overt alpha-synucleinopathy (21 iRBD-PD and 21 iRBD-DLB converters), while 73 patients remained stable at the last follow-up visit (43.2 ± 27.6 months). At survival analysis, both patterns were significantly associated with the phenoconversion trajectories. Brain [18F]FDG-PET is a promising biomarker to study progression trajectories in the alpha-synucleinopathy continuum.
... Delta waves are the slowest and highest amplitude brainwaves, with frequencies of < 4 Hz [32], and are believed to originate predominantly from the thalamus and cortex [33]. They are usually most prominent frontally in adults and posteriorly in children, with slow-wave delta rhythms being commonly associated with sleep [34], especially deep sleep beyond Stage 3, or N3 slow-wave sleep [35,36]. ...
During the last four decades there has been a significant growth of interest in mindfulness-based practices and their potential to foster improvements in health, wellbeing and human functioning in a variety of clinical and nonclinical populations. With this growth has come a renewed interest in understanding the psychological processes involved as well as the neuropsychological mechanisms by which such practices operate and effect transformative personal experiences and positive change. The current perspective paper (i) presents a basic taxonomy of meditation types and the structure and function of the processes believed to be involved, (ii) describes these components in terms of key neuroanatomical regions of interest, and (iii) critically appraises current findings regarding EEG measures as they relate to different aspects of meditation, functional activity and connectivity across regions of interest. The correlates between mindfulness and EEG are well described in terms of attentional and interoceptive processes and neuroanatomical regions of interest. To a lesser extent, there is also a growing understanding of such correlates for meditation techniques centred on compassion and loving-kindness meditation. However, the same does not apply to wisdom-based and null-state meditation practices where consistent associations between neuropsychological processes and EEG characteristics have proven elusive. These latter practices are viewed by many as key to fostering the deeper transformative experiences underlying psychological and spiritual development, and although studies of null-state meditation have yielded promising theoretical developments, more research is required. Future research could also benefit from better standardisation of EEG measures and analytic techniques to allow more robust metanalyses, and greater consistency of terminology regarding the fundamental components of meditation practice.
... Standardized in-home PSG was utilized on participants using a portable monitoring system (Philips Alice 6 lDe; Philips Healthcare). Sleep stages were recorded in 30-s epochs following the criteria of Iber (2007) [48]. Respiratory events were scored according to the American Academy of Sleep Medicine manual (version 2.6). ...
Understanding how sleep affects the glymphatic system and human brain networks is crucial for elucidating the neurophysiological mechanism underpinning aging-related memory declines. We analyzed a multimodal dataset collected through magnetic resonance imaging (MRI) and polysomnographic recording from 72 older adults. A proxy of the glymphatic functioning was obtained from the Diffusion Tensor Image Analysis along the Perivascular Space (DTI-ALPS) index. Structural and functional brain networks were constructed based on MRI data, and coupling between the two networks (SC-FC coupling) was also calculated. Correlation analyses revealed that DTI-ALPS was negatively correlated with sleep quality measures [e.g., Pittsburgh Sleep Quality Index (PSQI) and apnea-hypopnea index]. Regarding human brain networks, DTI-ALPS was associated with the strength of both functional connectivity (FC) and structural connectivity (SC) involving regions such as the middle temporal gyrus and parahippocampal gyrus, as well as with the SC-FC coupling of rich-club connections. Furthermore, we found that DTI-ALPS positively mediated the association between sleep quality and rich-club SC-FC coupling. The rich-club SC-FC coupling further mediated the association between DTI-ALPS and memory function in good sleepers but not in poor sleepers. The results suggest a disrupted glymphatic-brain relationship in poor sleepers, which underlies memory decline. Our findings add important evidence that sleep quality affects cognitive health through the underlying neural relationships and the interplay between the glymphatic system and multimodal brain networks.
... Sleep recordings were manually scored by a neurologist blinded to the patient's status (XD). Duration of REM sleep and non-REM sleep stages including light sleep (sleep stages N1 and N2) and deep sleep (sleep stage N3) was assessed using the standard 2007 criteria of the American Academy of Sleep Medicine (17). The presence of atypical sleep was detected according to a modified classification (9), and the duration of atypical sleep was counted in addition to other usual sleep stages. ...
Background:
Sleep deprivation is common in intensive care units (ICUs) and may alter respiratory performance. Few studies have assessed the role of sleep disturbances on outcomes in critically ill patients.
Objectives:
We hypothesized that sleep disturbances may be associated with poor outcomes in ICUs.
Methods:
Post-hoc analysis pooling three observational studies assessing sleep by complete polysomnography in 131 conscious and non-sedated patients included at different times of their ICU stay. Sleep was assessed early in a group of patients admitted for acute respiratory failure while breathing spontaneously (n = 34), or under mechanical ventilation in patients with weaning difficulties (n = 45), or immediately after extubation (n = 52). Patients admitted for acute respiratory failure who required intubation, those under mechanical ventilation who had prolonged weaning, and those who required reintubation after extubation were considered as having poor clinical outcomes. Durations of deep sleep, rapid eye movement (REM) sleep, and atypical sleep were compared according to the timing of polysomnography and the clinical outcomes.
Results:
Whereas deep sleep remained preserved in patients admitted for acute respiratory failure, it was markedly reduced under mechanical ventilation and after extubation (p < 0.01). Atypical sleep was significantly more frequent in patients under mechanical ventilation than in those breathing spontaneously (p < 0.01). REM sleep was uncommon at any time of their ICU stay. Patients with complete disappearance of REM sleep (50% of patients) were more likely to have poor clinical outcomes than those with persistent REM sleep (24% vs. 9%, p = 0.03).
Conclusion:
Complete disappearance of REM sleep was significantly associated with poor clinical outcomes in critically ill patients.
... PSG is a comprehensive recording of biophysiological signals during sleep which includes electroencephalogram (EEG), electrooculogram (EOG), electromyogram (EMG), electrocardiogram (ECG), airflow, respiratory efforts, and pulse oximetry [8]. It is utilized as diagnostic tool upon sleep stage scoring, respiratory events scoring, and movement scoring during sleep [9], [10]. These scores are involved in diagnosis of sleep-related disorders. ...
Obstructive Sleep Apnea (OSA) is one of sleep-disordered breathing characterized by repetitive episodes of partial or complete obstruction of upper airway during sleep, leading to significant health risks. Polysomnography (PSG) is gold standard for diagnosing OSA, but the process is labor-intensive and time-consuming which is often inaccessible. This study proposes a novel deep learning-based framework for severity classification of OSA using physiological signals from PSG type III devices. The proposed method comprises two key models: ApneaDetectNet for detecting apnea/hypopnea events called respiratory events and SleepDetectNet for classifying sleep stages. ApneaDetectNet utilizes abdominal respiratory effort (Abdo), nasal pressure (AIRFLOW), and oxygen saturation (SpO2) signals, while SleepDetectNet uses only Abdo signal. Both models are built using convolutional neural networks (CNNs) and bidirectional long short-term memory (Bi-LSTM) networks. An AHI Estimator consequently calibrates the predicted Apnea-Hypopnea Index (AHI) to resemble clinical observations. Three public datasets were employed: Multi-Ethnic Study of Atherosclerosis (MESA), Wisconsin Sleep Cohort (WSC), and University College Dublin Sleep Apnea (UCDDB). These datasets were split into training, validation, and test sets. Results indicated that the proposed model achieved superior or on par performance compared with state-of-the-art models. The MESA model achieved an accuracy of 71.84% and a Kappa of 0.6057, while the WSC model reached an accuracy of 62.16% and a Kappa of 0.4843. The combined MESA&WSC models showed improved overall metrices. External validation with UCDDB dataset demonstrated 84.00% accuracy with 0.7743 Kappa showing model's robustness. The accuracy and generalizability of the model suggests potential for real-world application.
... Sleep stages and breathing events were scored visually following the American Academy of Sleep Medicine scoring guidelines. 53 Breathing event scoring followed the initially recommended (Rule 4A), 54 then acceptable (Rule 1B), 53,55 hypopnoea scoring rule, which accounts for events with a reduction in flow amplitude of at least 30% for a minimum of 10 s, terminating in a ≥4% drop in oxygen saturation or arousal. Our split-night protocol consisted of a diagnostic portion followed by a PAP titration portion. ...
Obstructive sleep apnea (OSA) is associated with an increased risk for cognitive impairment and dementia, which likely involves Alzheimer’s disease (AD) pathology. Non-rapid eye movement (NREM) slow wave activity (SWA) has been implicated in amyloid clearance, but it has not been studied in the context of longitudinal amyloid accumulation in OSA. This longitudinal retrospective study aims to investigate the relationship between polysomnographic (PSG) and electrophysiological SWA features and amyloid accumulation.
From the Mayo Clinic Study of Aging cohort, we identified 71 participants >=60 years old with OSA (mean baseline age=72.9 ± 7.5 years, 60.6% male, 93% cognitively unimpaired) who had at least two consecutive Amyloid Pittsburgh compound B (PiB)-PET scans and a PSG study within 5 years of the baseline scan and before the second scan. Annualized PiB-PET accumulation (global ΔPiB[log]/year) was estimated by the difference between the second and first log-transformed global PiB-PET uptake estimations divided by the interval between scans (years). Sixty-four participants were included for SWA analysis. SWA was characterized by the mean relative spectral power density (%) in slow oscillation (SO) (0.5-0.9 Hz) and delta (1-3.9 Hz) frequency bands, and by their downslopes (SO-slopes and delta-slope, respectively) during the diagnostic portion of PSG. We fit linear regression models to test for associations between global ΔPiB(log)/year, SWA features (mean SO% and delta% or mean SO-slope and delta-slope), and OSA severity markers, after adjusting for age at baseline PiB-PET, APOE ε4, and baseline amyloid positivity.
For 1 s.d. increase in SO% and SO-slope, global ΔPiB(log)/year increased by 0.0033 (95% CI 0.0001; 0.0064, p=0.042) and 0.0069 (95% CI: 0.0009; 0.0129, P=0.026), which were comparable to 32% and 59% of the effect size associated with baseline amyloid positivity, respectively. Delta-slope was associated with a reduction in global ΔPiB(log)/year by -0.0082 (95% CI: -0.0143; -0.0021, P=0.009). Sleep apnea severity was not associated with amyloid accumulation. Regional associations were stronger in the prefrontal region. Both slow wave slopes had more significant and widespread regional associations.
Annualized PiB-PET accumulation was positively associated with SO and SO-slopes, which may reflect altered sleep homeostasis due to increased homeostatic pressure in the setting of unmet sleep needs, increased synaptic strength and/or hyperexcitability in OSA. Delta-slope was inversely associated with PiB-PET accumulation, suggesting it may represent residual physiological activity. Further investigation of SWA dynamics in the presence of sleep disorders before and after treatment is necessary for understanding the relationship between amyloid accumulation and SWA physiology.
... This goes back to the year 1968, published by Rechtschaffen and Kales [19]. In 2007, the American Academy of Sleep Medicine introduced a new classification by merging stage 3 and stage 4, which is internationally recognised and used, see [20]. ...
Sleep stage classification is a widely discussed topic, due to its importance in the diagnosis of sleep disorders, e.g. insomnia. Analysis of the brain activity during sleep is necessary to gain further insight into the processing that occurs in our brains. We want to use permutation entropy as a model for this analysis. Therefore, the signal processing in terms of electroencephalography is described. This results in a time discrete signal, that can be further processed by applying the method of permutation entropy, which is a modification of the Shannon entropy as a measure of information processing. The method is applied to 18 data sets, nine electroencephalography measurements of patients suffering from insomnia and nine of people without a sleep disorder. A strong correlation between the permutation entropy value and the sleep stages was found during the simulation runs. The results are analysed and presented using boxplot diagrams of the permutation entropy over the sleep stages. Furthermore, it is investigated that there is a steady decrease in the value when the patient is in a deeper sleep. This suggests that the method is a good parameter for sleep stage classification. Finally, we propose an extension of the conceptual model to other pathological conditions and also to the analysis of brain activity during surgery.
... Using the original 30 s epochs as a reference for the presence of lighter NREM sleep, we marked 3 types of epochs: those in which a spindle occurred but not a K-complex ('N2 Spindle'), those in which a K-complex occurred ('N2 K-complex'), and those in which neither was evident ('N2 Plain'). Note that we adopt the American Academy of Sleep Medicine 'N2' abbreviation in this paper to denote epochs in lighter NREM sleep (Iber, 2007), as it is more recognizable to many researchers and is practically equivalent to Rechtschaffen and Kales' stage 2 sleep ('S2'). In 30 s epochs classified as deep NREM sleep, 5 s windows in which clear slow wave activity (i.e. ...
Theta oscillations (4-8 Hz) in frontal cortical regions are present to different degrees across states of consciousness. In sleep, theta is prominent in periods of rapid eye-movement (REM) sleep. Theta has been linked to processes of memory consolidation; however, its mechanistic contribution specifically during REM sleep is not well understood. Interestingly, in the wake state, frontal theta activity increases during effortful cognitive tasks involving executive functions such as working memory, hinting at similarities in circuitry, and potentially, function. The aim of the present work is to create a spatially resolved, whole-brain characterisation of REM oscillatory activity in healthy human subjects, distinguishing theta from neighbouring frequency bands, differentiating substages of REM sleep (phasic and tonic REM), and comparing REM theta to that which is evoked during a working memory task. To that end, we analysed magneto- and electroencephalography (M/EEG) data recorded during overnight sleep in 10 healthy subjects, and similar data from 17 healthy subjects who performed a working memory task, using a novel whole-brain, source-localised MEG approach. Our results show that (i) theta activity has a frontal midline topography that is distinct from those of other prominent frequency bands in REM (delta, alpha, beta), (ii) theta activity in frontal midline regions is best observed within a focused 5-7 Hz range, separating it from occipital alpha activity, (iii) REM theta is dominant over the frontal midline but is also observed in several sub-cortical areas, (iv) theta is more widespread in tonic than phasic REM sleep, and (v) the focused frontal midline theta pattern observed in REM phasic sleep is the most similar of all observed sleep substages to theta evoked by a working memory task. These results enhance our understanding of theta physiology in REM sleep and suggest future targets for research into REM's role in learning and memory.
... The PSG recordings included electroencephalography with frontal, central, and occipital electrodes; 1-lead electrocardiography; electromyography of extraocular eye movement, chin, and bilateral anterior tibialis muscles; nasal airflow and thermistor; peripheral oxygen saturation; sleep position; and chest and abdominal plethysmography. Sleep staging and scoring for respiratory events and movements followed the guidelines of the American Academy of Sleep Medicine and were conducted by two sleep technicians with >10 years of experience [18,19]. Three neurologists (WSH, SC, and KMK), experienced in PSG interpretation and responsible for training residents in this field, meticulously reviewed all PSG data. ...
Background: Delirium affects up to 50% of patients following high-risk surgeries and is associated with poor long-term prognosis. This study employed machine learning to predict delirium using polysomnography (PSG) and sleep-disorder questionnaire data, and aimed to identify key sleep-related factors for improved interventions and patient outcomes. Methods: We studied 912 adults who underwent surgery under general anesthesia at a tertiary hospital (2013–2024) and had PSG within 5 years of surgery. Delirium was assessed via clinical diagnoses, antipsychotic prescriptions, and psychiatric consultations within 14 days postoperatively. Sleep-related data were collected using PSG and questionnaires. Machine learning predictions were performed to identify postoperative delirium, focusing on model accuracy and feature importance. Results: This study divided the 912 patients into an internal training set (700) and an external test set (212). Univariate analysis identified significant delirium risk factors: midazolam use, prolonged surgery duration, and hypoalbuminemia. Sleep-related variables such as fewer rapid eye movement (REM) episodes and higher daytime sleepiness were also linked to delirium. An extreme gradient-boosting-based classification task achieved an AUC of 0.81 with clinical variables, 0.60 with PSG data alone, and 0.84 with both, demonstrating the added value of PSG data. Analysis of Shapley additive explanations values highlighted important predictors: surgery duration, age, midazolam use, PSG-derived oxygen saturation nadir, periodic limb movement index, and REM episodes, demonstrating the relationship between sleep patterns and the risk of delirium. Conclusions: The artificial intelligence model integrates clinical and sleep variables and reliably identifies postoperative delirium, demonstrating that sleep-related factors contribute to its identification. Predicting patients at high risk of developing postoperative delirium and closely monitoring them could reduce the costs and complications associated with delirium.
EEG spectral analysis provides a more sensitive measure of sleep disruption than conventional sleep macro‐architecture. We aimed to examine the use of this technique applied to overnight polysomnography in distinguishing children with narcolepsy and idiopathic hypersomnia (IH) from subjectively sleepy children with a non‐diagnostic multiple sleep latency test. The relative power was calculated for delta (0.5–3.9 Hz), theta (4–7.9 Hz), alpha (8–11.9 Hz), sigma (12–13.9 Hz), and beta power (14–30 Hz). A mean value for each frequency was calculated for each 30 s epoch then averaged for each sleep stage within each child. Data are presented as median and interquartile range. Twenty‐eight children with narcolepsy, 11 with IH, and 26 with subjective sleepiness were included and individually matched for age and sex with a control child. In N2, the F4 beta power was lower in the narcolepsy compared with the IH group ( p < 0.05). The F4 theta power was higher in the narcolepsy compared with the subjectively sleepy group during wake ( p < 0.001), N2 ( p < 0.01), N3 ( p < 0.05), and total sleep ( p < 0.01). During total sleep the F4 delta power was lower in both the narcolepsy and IH groups compared with the subjectively sleepy group ( p < 0.05 for both). Our study identified specific EEG frequencies which differed between groups of children referred for assessment of EDS. In particular, differences in theta and delta power in children with narcolepsy and IH compared with others with subjective sleepiness may provide insights into the pathophysiology associated these conditions.
Accurate sleep stage detection is essential for diagnosing sleep disorders and understanding sleep patterns. Traditional methods, such as manual scoring of polysomnographic data and the use of multi-channel EEG, while effective, are resource-intensive and impractical for home-based monitoring. This study explores the application of machine learning models, specifically deep learning techniques like Convolutional Neural Networks (CNNs), to improve the accuracy of sleep stage detection using single-channel EEG data. By comparing the performance of CNNs, Random Forest, and Support Vector Machines (SVM), the study demonstrates that CNNs significantly outperform other models in terms of accuracy, sensitivity, and specificity, achieving an accuracy of 91.4%. The results suggest that machine learning, particularly CNNs, offers a practical solution for reducing the complexity of sleep monitoring while maintaining high accuracy, making it viable for both clinical and non-clinical settings, including wearable and home-based devices. These findings highlight the potential of machine learning in transforming sleep stage detection into a more accessible and user-friendly process.
Clinical sleep diagnosis traditionally relies on polysomnography (PSG) and expert manual classification of sleep stages. Recent advancements in deep learning have shown promise in automating sleep stage classification using a single PSG channel. However, variations in PSG acquisition devices and environments mean that the number of PSG channels can differ across sleep centers. To integrate a sleep staging method into clinical practice effectively, it must accommodate a flexible number of PSG channels. In this paper, we proposed FlexSleepTransformer, a transformer-based model designed to handle varying number of input channels, making it adaptable to diverse sleep staging datasets. We evaluated FlexSleepTransformer using two distinct datasets: the public SleepEDF-78 dataset and the local SleepUHS dataset. Notably, FlexSleepTransformer is the first model capable of simultaneously training on datasets with differing number of PSG channels. Our experiments showed that FlexSleepTransformer trained on both datasets together achieved 98% of the accuracy compared to models trained on each dataset individually. Furthermore, it outperformed models trained exclusively on one dataset when tested on the other dataset. Additionally, FlexSleepTransformer surpassed state-of-the-art CNN and RNN-based models on both datasets. Due to its adaptability with varying channels numbers, FlexSleepTransformer holds significant potential for clinical adoption, especially when trained with data from a wide range of sleep centers.
Menstruation is an inflammatory process that involves changes in women's physiology leading to mental and physical complaints. Sleep is essential for optimal hormonal release, immune response, and wellbeing, becoming an important factor to be evaluated. We compared sleep, inflammatory mediators, fatigue, anxiety and depression symptoms, and quality of life in menstruating and non‐menstruating women. We used the polysomnographic data of 232 women from EPISONO 2007, an epidemiological study from São Paulo city, Brazil. Women were distributed into menstruating ( N = 61) and non‐menstruating groups ( N = 171). We applied questionnaires related to sleep quality, sleepiness, insomnia, fatigue, anxiety and depression symptoms, and quality of life. The serum levels of interleukin 6, tumour necrosis factor‐alpha, and C‐reactive protein were analysed. For statistical analysis the significance level adopted was p < 0.05. Sleep efficiency was statistically lower in menstruating women (81% ± 13) compared with the non‐menstruating group (84.2% ± 13.3, p < 0.023). No statistical differences between the two groups were found in respect to the other parameters analysed. Both groups scored for fatigue symptoms, but no statistical significance was observed between the groups. Our findings indicate that menstruation was associated with lower objective sleep efficiency, suggesting that menstruation may be a physiological factor impairing sleep. Further studies evaluating menstrual variables, and each phase of the menstrual cycle, should be undertaken to detect the main factors associated with sleep complaints, fatigue, and objective parameters of sleep.
Background/Objectives: This study evaluated the impact of continuous positive airway pressure (C-PAP) therapy combined with a rigorous diet regimen on obese patients with obstructive sleep apnea syndrome (OSAS). Methods: Sixty obese patients (BMI ≥ 30) diagnosed with severe OSAS were recruited in order to establish the evaluation of CPAP therapy with different extents of adherence to a rigorous diet regimen. After one year, significant improvements were observed. Results: BMI reduced by 12.32%, apnea–hypopnea index (AHI) by 22.04%, oxygen desaturation index (ODI) by 15.87%, total sleep time with oxygen saturation below 90% (TST90%) by 25.2%, and Epworth Sleepiness Scale (ESS) scores by 21.74%. Patients were, then, divided into three groups, based on adherence to the restricted diet, as well as to the correct use of the nocturnal C-PAP, showing different reductions in BMI, AHI, ODI, TST90%, and ESS, according to their adherence, based on the sum of % reduction in BMI + AHI into three groups. Conclusions: These findings underscore the effectiveness of combining C-PAP therapy with a strict diet in improving OSAS symptoms and overall health in obese patients. Future studies with larger cohorts and longer follow-up periods are needed to confirm these results and explore the long-term benefits of this integrated approach.
Background/Objectives: Given the existence of higher suicidality in apneic individuals, this study aimed to determine the potential role played by subtypes of the comorbid insomnia disorder (CID) in the occurrence of suicidal ideation for this specific subpopulation. Methods: To perform our analyses, 1488 apneic individuals were retrospectively extracted from the Sleep Laboratory hospitalization register. Only apneic individuals with suicidal ideation highlighted during the psychiatric interview and/or with a score ≥1 on item G of the Beck Depression Inventory confirmed during the clinical interview were included in the group with suicidal ideation. The likelihood of suicidal ideation associated with CID subtypes was investigated using logistic regression analyses. Results: The prevalence of suicidal ideation was 9.3% in our sample of apneic individuals. After hierarchically introducing the significant confounders for adjustment, multivariate logistic regression analyses demonstrated that unlike short sleep duration alone and CID without short sleep duration, the likelihood of suicidal ideation was only higher for CID with short sleep duration in apneic individuals. Conclusions: Thus, we highlighted in this study that CID with short sleep duration could play a major role in higher suicidality for apneic individuals, which seems to require systematic screening and appropriate treatment of this comorbid sleep disorder to enable better management of suicidal risk in this specific subpopulation.
Citation: Dragonieri, S.; Portacci, A.; Quaranta, V.N.; Carratu, P.; Lazar, Z.; Carpagnano, G.E.; Bikov, A.
In clinical practice, sleep stage classification (SSC) is a crucial step for physicians in sleep assessment and sleep disorder diagnosis. However, traditional sleep stage classification relies on manual work by sleep experts, which is time-consuming and labor-intensive. Faced with this obstacle, computer-aided diagnosis (CAD) has the potential to become an intelligent assistant tool for sleep experts, aiding doctors in the assessment and decision-making process. In fact, in recent years, CAD supported by artificial intelligence, especially deep learning (DL) techniques, has been widely applied in SSC. DL offers higher accuracy and lower costs, making a significant impact. In this paper, we will systematically review SSC research based on DL methods (DL-SSC). We explores DL-SSC from several important perspectives, including signal and data representation, data preprocessing, deep learning models, and performance evaluation. Specifically, this paper addresses three main questions: (1) What signals can DL-SSC use? (2) What are the various methods to represent these signals? (3) What are the effective DL models? Through addressing on these questions, this paper will provide a comprehensive overview of DL-SSC.
ResearchGate has not been able to resolve any references for this publication.