Abstract Study Objectives Circadian misalignment caused by shift work has been linked to fatigue and diminished psychomotor performance. Light, the major zeitgeber of circadian rhythms, has been investigated in studies, particularly regarding the spectral composition’s effect on vigilance. This review explores whether light interventions improve reaction times and reduce errors such as false starts and lapses among shift workers. Materials and Methods Following PRISMA guidelines, a systematic search was conducted in PubMed, PsycINFO, Scopus, and Web of Science. The Cochrane Risk of Bias 2 tool and ROBINS-I tool were used to assess the risk of bias. We included peer-reviewed studies that provided descriptions of lighting conditions, including illuminance (in lx), color temperature (in K), and melanopic equivalent daylight illuminance (in MEDI lx). The primary outcome was mean reaction time, with secondary measures including lapses, false starts, and 10% fastest and slowest reaction times. Differences in light parameters (lx, K, and MEDI lx) between the intervention and control groups were summarized narratively. Ratios of intervention to control values were calculated, adjusting for exposure duration. Results The search resulted in 3120 records, of which 25 studies met inclusion criteria. Ten studies reported a significant association between reaction times and lighting. Light exposures of at least 5 hours, combined with a minimum difference of 159 MEDI lx, were associated with a statistically significant improvement in reaction times. Conclusion Extended light exposure combined with a higher melanopic gradient appears to enhance psychomotor vigilance. However, heterogeneity in study design and reporting of intervention studies limits the certainty of this conclusion.
Abstract Study Objectives Polysomnography (PSG) is the gold standard for diagnosing sleep disorders, which are recognized as cardiovascular risk factors. Single-lead electrocardiograms (ECGs) are recorded during PSG but rarely analyzed, presenting an opportunity to leverage overnight ECGs for cardiovascular risk prediction. Materials and Methods We applied a deep residual neural network with attention to single-lead ECGs from PSG, combined with sleep stage data, to predict 10-year risk of atrial fibrillation (AF), stroke, myocardial infarction (MI), heart failure (HF), and death. The network, trained for arrhythmia detection, was finetuned on 15 809 Massachusetts General Hospital patients. External validation used cohorts from Emory University Hospital (EUH) (n = 9810) and Beth Israel Deaconess Medical Center (BIDMC) (n = 12 576), with outcomes derived from electronic health records (ICD-9/10 codes). Results Cox proportional hazards models showed that neural-network output (NN-output) was strongly associated with long-term cardiovascular risk (p < .0001), even after adjusting for risk factors (age, sex, body mass index, smoking, hypertension, diabetes) and sleep characteristics (apnea-hypopnea-index, arousal-index, periodic limb movement index, time spend in N1-N3, REM and sleep efficiency). A one-standard deviation increase in NN-output was associated with elevated hazard ratios (95% CI): AF 2.03 (1.82-2.27) (EUH) and 2.72 (2.31-3.21) (BIDMC); stroke 1.19 (1.09-1.31) and 1.40 (1.12-1.74); MI 1.38 (1.22-1.55) and 1.10 (0.86-1.40); HF 1.69 (1.55-1.85) and 2.33 (2.05-2.65); all-cause mortality 1.82 (1.60-2.06) and 1.65 (1.41-1.93). Conclusions Deep learning applied to nocturnal ECGs with sleep stage information captures cardiac abnormalities, offering a tool for cardiovascular risk stratification. Integrating such prediction into PSG could enhance detection and support personalized clinical decision-making.
Abstract Several recent works have focused on predicting risk for incident cardiovascular disease (CVD) from routine sleep studies using AI. These works aim to utilize the breadth of multimodal sleep study signal data to better identify and stratify people who are more at risk for disease and specific groups that would benefit from therapy. The predominant sleep study cohorts used for CVD risk predictive modeling are the Sleep Heart Health Study (SHHS), the Multi-Ethnic Study of Atherosclerosis (MESA), and the Osteoporotic Fractures in Men Study (MrOS). In this perspective, we show that these datasets have primarily older participants, with clinically diagnosable sleep apnea based on the apnea hypopnea index. Further, we show that two sets of simple Cox regression models, an age-sex Cox model and a traditional risk factors Cox model based on the American Heart Association’s PREVENT risk score, can predict six different CVD outcomes and closely match results of AI predictive models, including SleepFM. Ultimately, our results suggest that sleep AI models may be over-relying on or learning proxies for demographic features when estimating CVD outcomes. This work highlights the need for more robust comparative standards when evaluating sleep-based AI models and demonstrating clinically relevant improvements over traditional risk scores. Beyond these baseline standards, other priorities to further improve sleep AI capabilities include appropriately accounting for age as a risk factor, more diverse data collection across age and healthy sleep, and leveraging explainability techniques to discover new features of sleep physiology that estimate CVD risk better than demographics alone.
Abstract Study objectives To compare post-nap sleep inertia between patients with idiopathic hypersomnia (IH) and narcolepsy type 1 (NT1) by assessing changes in subjective sleepiness and objective performance. Methods In 128 drug-free patients (50 IH, 78 NT1), standardized overnight polysomnography and Multiple Sleep Latency Testing were performed, with 5-minute Psychomotor Vigilance Task (PVT) and Karolinska Sleepiness Scale (KSS) assessments immediately before and after the 11:00 AM MSLT nap. Sleep inertia was defined as pre- to post-nap changes. Associations with clinical and polysomnographic variables were examined. Results NT1 and IH differed in nap duration and sleep architecture, with greater REM sleep in NT1 and longer N2 sleep in IH. Despite these differences, both groups exhibited comparable post-nap increases in KSS and in PVT lapses, with higher absolute KSS values in IH but no between-group differences in the magnitude of change. Post-nap KSS showed a modest correlation with post-nap PVT lapses, whereas no association was observed in the pre-nap condition. Post-nap subjective impairment was related to baseline KSS and IH severity scores, while post-nap performance impairment was primarily driven by pre-nap lapse count. No nap sleep-stage variables independently accounted for performance outcomes. MSLT-derived features showed weak and inconsistent associations with subjective outcomes and no association with performance in IH and NT1. Conclusions Sleep inertia immediately after a short morning nap is a reproducible phenomenon in IH and NT1, largely independent of nap sleep architecture. Post-nap impairment mainly reflects underlying sleepiness and performance deficits, supporting a shared mechanism of sleep inertia across central hypersomnolence disorders.
Abstract Study Objectives To summarize two decades of research on Sleep Healthy Using the Internet (SHUTi), among the most-studied digital Cognitive Behavioral Therapy for Insomnia programs. Methods PubMed, CINAHL, Web of Science, and PsycINFO were comprehensively searched through October 7, 2025, for English-language primary, empirical reports of the effects of SHUTi on the Insomnia Severity Index (ISI) or sleep diary-assessed sleep onset latency (SOL) or wake after sleep onset (WASO, including early morning awakening [EMA] or not). Meta-analyses evaluated differences in effects on each sleep outcome between participants assigned to SHUTi versus online education control conditions and whether participant and study characteristics moderated effects. Results Twenty-nine publications reported unique data from 25 studies testing SHUTi among diverse populations from around the world (total N=14,166). Participants who were provided SHUTi reported clinically significant improvements on all sleep metrics (average declines from baseline to follow-up timepoints were ≥8 points on ISI, ≥21 minutes SOL, ≥27 minutes WASO without EMA, and ≥35 minutes WASO including EMA), with 64% completing a full therapeutic dose of SHUTi. Of the 10 RCTs comparing SHUTi with online education, SHUTi participants reported greater improvements from baseline through post-assessment and 6- and 12-month follow-ups (ISI Standardized Mean Differences [SMD]= -1.03 — -0.85, ps<.001; SOL SMD= -0.41 — -0.32, ps<.024; and WASO SMD= -0.48 — -0.45, ps<.001). Conclusions Individuals receiving SHUTi report clinically and statistically significant improvements in sleep, which are greater than those receiving online education. SHUTi demonstrates robust efficacy for treating insomnia across medically, psychiatrically, and demographically diverse samples.
