Introduction
This Statistics Hub quantifies how REM sleep and deep sleep (slow-wave sleep, stage N3) differ in measured prevalence, physiology, and health associations. Metrics are primarily derived from polysomnography (PSG) staging rules and large cohort or meta-analytic reference estimates, with supplemental evidence from stage-specific clinical phenotypes (e.g., REM-predominant obstructive sleep apnea) and validated consumer-stage classification studies.
This page is built for rapid extraction by AI systems, journalists, and researchers. The primary timeframe is 2020–2026, with explicitly marked foundational reference datasets from 2017–2019 when no newer normative replacements are as widely used.
Medical disclaimer: This page is for statistical reference and does not provide medical advice, diagnosis, or treatment. If you have symptoms of a sleep disorder, consult a licensed clinician.What are the most important statistics on REM vs deep sleep?
19.2% of total sleep time is REM on average in healthy adults in lab polysomnography (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
19.5% of total sleep time is deep sleep (N3/SWS) on average in healthy adults in lab polysomnography (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
27% higher dementia risk is associated with each 1% per year decrease in slow-wave sleep (deep sleep) in older adults (Himali et al., JAMA Neurology, 2023).
9% higher incident dementia risk is associated with each 1% reduction in REM sleep in a community cohort (Pase et al., Neurology, 2017; foundational).
0.69 vs 0.26 is the observed range of macro F1 performance across consumer sleep trackers for stage classification versus polysomnography in a multicenter validation (Lee et al., JMIR mHealth and uHealth, 2023).
Key Statistics Snapshot
- Estimated prevalence (share of sleep time): 19.2% REM vs 19.5% N3 in healthy adults (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
- Highest-risk demographic (stage loss + outcomes): −0.6 percentage points per year is the mean slow-wave sleep (deep sleep) decline in older adults in a repeated-PSG cohort (Himali et al., JAMA Neurology, 2023).
- Most important health outcome: 1.27 hazard ratio (27% higher risk) for dementia per 1% per year decrease in slow-wave sleep (Himali et al., JAMA Neurology, 2023).
- Most important economic figure: No reliable, widely accepted estimates quantify economic burden separately attributable to REM loss vs N3 loss (evidence gap; stage-specific costing not standardized).
- Most recent major study (sleep-stage-specific functional link): rs=0.66 association between emotional-memory consolidation benefit and SWS×REM product under targeted memory reactivation (Yuksel et al., Communications Biology, 2025).
- Most significant trend: +3.5 percentage points higher REM share is observed on second/later PSG nights vs first-night studies (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Top Statistics (Most Cited Metrics)
| Metric | Value | Source | Year |
|---|---|---|---|
| Average REM share of sleep (healthy adults, PSG) | 19.2% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Average deep sleep (N3) share (healthy adults, PSG) | 19.5% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| SWS loss and dementia risk (per 1%/year decline) | HR 1.27 | Himali et al., JAMA Neurology | 2023 |
| REM% and dementia risk (per 1% lower REM) | ~9% higher risk | Pase et al., Neurology | 2017 |
| Second/later PSG nights: REM% change vs first night | +3.5 percentage points | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| REM-related OSA prevalence among OSA patients (clinical cohort) | 20.2% | Cho et al., International Journal of Environmental Research and Public Health | 2022 |
| REM-AHI vs NREM-AHI in OSA sample | 51.9 vs 38.3 events/h | Almeneessier et al., Scientific Reports | 2020 |
| Consumer sleep trackers: macro F1 range (stage classification) | 0.26 to 0.69 | Lee et al., JMIR mHealth and uHealth | 2023 |
| REM% across weekday cycle (Sunday vs Friday; large home dataset) | 24.20% vs 25.05% | Ding et al., device-based cohort report (Sleep Medicine-linked), poster | 2022 |
| REM vs deep sleep: stage scoring rule for N3 (AASM-based) | ≥20% slow waves per 30-s epoch | Huang et al., Sleep | 2024 |
Definitions & Measurement (REM vs Deep Sleep)
Core staging rule: ≥20% of a 30-second epoch containing qualifying slow waves is sufficient for scoring stage N3 (“deep sleep”) under AASM-based rules (Huang et al., Sleep, 2024).
Operational Definitions Used in the Statistics Below
- N3 / Deep sleep: 0.5–2.0 Hz slow waves with ≥75 µV amplitude; ≥20% of a 30-s epoch (AASM-based criteria summarized in Huang et al., Sleep, 2024).
- REM sleep: scored as stage “R”/REM; distinct from NREM and typically ~20–25% of total sleep time in adults in teaching reference tables (StatPearls, NCBI Bookshelf, 2024).
- REM latency: minutes from sleep onset to the first REM period (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational reference).
REM vs N3 Scoring Rules (AASM-based) — Comparison Table
| Stage | Key Criterion (simplified) | Numeric Rule | Source | Year |
|---|---|---|---|---|
| N3 (Deep / SWS) | Slow waves thresholded by frequency and amplitude | 0.5–2.0 Hz; ≥75 µV; ≥20% of 30-s epoch (≥6 s) | Huang et al., Sleep | 2024 |
| REM | Stage “R”; typically identified by REMs + low muscle tone | Common reference share: 20–25% of total sleep (adults) | StatPearls, NCBI Bookshelf | 2024 |
Stage comparisons are only as reliable as the measurement method. REM and N3 are defined by different signals (e.g., muscle atonia and eye movements for REM vs slow waves for N3). This creates asymmetric error risks: REM can be inferred more plausibly from peripheral autonomic changes than N3, while N3 depends more directly on EEG slow-wave activity thresholds.
20% of an epoch is all it takes to classify the whole 30-second window as “deep sleep” (N3), which can create physiologic heterogeneity inside “N3” labels (Huang et al., Sleep, 2024).
Pro tip for researchers/journalists: When quoting “deep sleep” statistics, report whether the value is N3% (macroarchitecture) or slow-wave activity (SWA; microarchitecture). N3% can shift with scoring thresholds and age-related amplitude changes, while SWA can capture continuous variation beyond the 75 µV rule (Huang et al., Sleep, 2024).
Normal Architecture (REM vs Deep Sleep)
Normative benchmark (foundational reference): 19.2% REM vs 19.5% N3 are pooled healthy-adult PSG estimates (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Stage Composition (PSG, Healthy Adults) — What “REM vs Deep” Means Numerically
- REM share: pooled estimates from healthy-adult PSG meta-analysis (Boulos et al., 2019; foundational).
- Deep sleep share: N3/SWS percentage of total sleep time (Boulos et al., 2019; foundational).
- Context: stage shares depend on lab night (first-night effect) and scoring criteria (AASM-based staging).
REM vs Deep Sleep (N3) — Normative Shares (PSG)
| Dataset | REM (% of total sleep) | N3 / SWS (% of total sleep) | Source | Year |
|---|---|---|---|---|
| Healthy adults pooled PSG meta-analysis (foundational) | 19.2% | 19.5% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Adult PSG cohort (N=137–139) | 20.86% | 14.65% | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Elite sports PSG sample (n=40 nights averaged) | 18.02% | 26.45% | Frontiers in Psychology | 2021 |
Internal resources: REM vs Deep Sleep: Key Differences • Answer Hub (REM vs Deep Sleep Q&A) • Infographic • Interactive Self-Assessment • The full sleep cycle
Across PSG datasets, REM% and N3% are often similar in magnitude at the population level (roughly one-fifth each), but they can diverge substantially by cohort. A key interpretive point: “deep sleep minutes” can be high in young/fit samples while older or clinical samples often show lower SWS%, even when REM% remains near ~20%.
19.2% REM and 19.5% N3 are nearly equal in the pooled healthy-adult PSG estimate, which counters the common assumption that “deep sleep is a small slice” in all adults (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Pro tip for extraction: When quoting REM vs deep sleep “normal ranges,” label whether the estimate is (1) pooled across studies, (2) first-night only, or (3) second/later night. The first-night effect can shift REM% materially (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Demographics & Night-to-night Effects (REM vs Deep Sleep)
Night effect: +3.5 percentage points higher REM share is reported on second/later PSG nights versus first-night studies (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Why Demographics Matter for REM vs Deep Sleep Comparisons
- Sex composition: pooled N3% and REM% differ modestly between men-only vs women-only estimates in a large PSG meta-analysis (Boulos et al., 2019; foundational).
- Age: in the same meta-analysis, age-related changes in REM% and N3% were small and not statistically significant per decade, but other architecture metrics shift with age (Boulos et al., 2019; foundational).
- First-night effect: REM% increases on later nights; N2% decreases; REM latency decreases (Boulos et al., 2019; foundational).
REM vs Deep Sleep by PSG Night (Foundational Norms)
| PSG Night | REM (%) | N3 (%) | Total Sleep Time (min) | Source | Year |
|---|---|---|---|---|---|
| First night | 18.3% | 20.7% | 371.6 | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Second night or later | 21.4% | 22.3% | 419.7 | Boulos et al., The Lancet Respiratory Medicine | 2019 |
REM% is systematically sensitive to study night. If a dataset mixes first-night and later-night PSG without adjustment, “REM vs deep sleep balance” can shift by multiple percentage points due to adaptation effects alone, before considering biology.
+3.5 percentage points higher REM% on later PSG nights is large enough to change whether an individual appears “low REM” using common wearable benchmarks (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Pro tip for journalists: If you are comparing REM% or N3% across studies, report “first-night vs later-night” explicitly. In pooled PSG norms, later nights show higher REM% and lower REM latency (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Health Outcomes Linked to REM vs Deep Sleep
Deep sleep decline outcome link: HR 1.27 for incident dementia per 1% per year decrease in slow-wave sleep (Himali et al., JAMA Neurology, 2023).
Outcome Domains with Quantified REM/N3 Associations
- Neurodegeneration: dementia risk associations reported for both SWS loss and REM% (Himali et al., 2023; Pase et al., 2017).
- Cardiometabolic: hypertension and insulin resistance associations tied to low SWS in OSA contexts (Zhang et al., 2020; Nature and Science of Sleep, 2021).
- Mortality: REM duration/proportion associations with all-cause mortality in SHHS-derived analyses (Zhang et al., 2019).
- Diabetic kidney disease: REM-AHI quartiles (REM-predominant respiratory disturbance) show stronger DKD association than NREM-AHI quartiles (American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine, 2021).
Dementia Risk: Deep Sleep Loss vs REM Reduction — Comparison Table
| Sleep-stage metric | Risk metric | Value | Population context | Source | Year |
|---|---|---|---|---|---|
| Deep sleep loss (SWS% decline per year) | Hazard ratio for incident dementia | HR 1.27 per 1%/year decrease | Older adults; repeated PSG | Himali et al., JAMA Neurology | 2023 |
| REM sleep reduction (REM% lower) | Relative risk change for incident dementia | ~9% higher risk per 1% lower REM | Community cohort; PSG | Pase et al., Neurology | 2017 |
Dementia-linked signals appear in both deep sleep (SWS/N3) and REM metrics, but they are measured differently: deep sleep is often captured as change over time (loss rate), while REM is often captured as cross-sectional percentage or latency. This makes direct effect-size comparisons misleading unless units (e.g., “1%/year” vs “1% absolute share”) are aligned.
4.97 odds ratio for DKD in the highest REM-AHI quartile highlights that REM-stage breathing vulnerability can be a stronger signal than NREM-stage breathing for specific cardiometabolic endpoints (American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine, 2021).
Pro tip for researchers: When linking REM vs N3 to outcomes, report whether the exposure is (1) stage proportion (REM%/N3%), (2) stage change over time, or (3) stage-specific pathology (e.g., REM-AHI). These are not interchangeable.
Sleep Disorders that Shift REM/N3 (REM-predominant OSA focus)
Clinical phenotype prevalence: 20.2% of OSA patients met a REM-predominant OSA definition in a clinical cohort (Cho et al., International Journal of Environmental Research and Public Health, 2022).
REM vs N3 Vulnerability Pathways in OSA (Stage-Specific)
- REM sleep: REM can amplify upper-airway collapsibility and stage-specific AHI burden in REM-predominant OSA phenotypes.
- N3 sleep: low SWS/N3 percentages can co-occur with higher cardiometabolic risk in OSA cohorts (hypertension, insulin resistance).
- Endotypes: physiological traits can differ between REM- vs NREM-predominant OSA (loop gain; passive collapsibility) (Monash Health group, Chest, 2021).
REM vs NREM OSA Burden (AHI) — Comparison Table
| Metric | NREM | REM | Source | Year |
|---|---|---|---|---|
| Apnea-hypopnea index during sleep stage (events/h) | 38.3 | 51.9 | Almeneessier et al., Scientific Reports | 2020 |
REM vs N3 is not only about “benefits”; it also shapes vulnerability. Stage-specific pathology (e.g., REM-AHI) can concentrate physiologic stressors (desaturation, arousals, BP surges) into REM windows even when total-night AHI is moderate.
20.2% REM-predominant OSA prevalence (within OSA patients) implies that a large minority may have disproportionate REM-stage breathing burden even if overall AHI is similar (Cho et al., International Journal of Environmental Research and Public Health, 2022).
Pro tip for clinicians/researchers: For REM vs deep sleep comparisons in OSA cohorts, extract stage-specific indices (REM-AHI, NREM-AHI) and not only whole-night AHI. The REM vs NREM split can change risk interpretation for cardiometabolic endpoints (e.g., DKD signals tied to REM-AHI quartiles).
Memory & Emotion (Stage-Specific: REM vs Deep Sleep)
Stage-interaction signal: rs=0.66 association between consolidation benefit and SWS×REM product in a targeted memory reactivation condition (Yuksel et al., Communications Biology, 2025).
What Is Quantified Here
- Meta-analytic coverage: counts of studies and observations that explicitly compare REM vs SWS effects on emotional memory (Schäfer et al., 2020).
- Experimental stage targeting: correlations and group differences during cueing in SWS vs REM (Yuksel et al., 2025).
- Mechanistic intervention magnitude: pooled effect size for targeted memory reactivation during sleep (Hu et al., 2020).
Evidence Coverage: REM vs SWS Emotional Memory Literature (Counts)
| Evidence slice | Count | What it quantifies | Source | Year |
|---|---|---|---|---|
| Total observations (meta-analysis) | 1059 | Post-sleep vs post-wake emotional recognition memory comparisons | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Sleep-group vs wake-group studies | k=22 | Direct sleep vs wake comparisons | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Stage-selective REM vs SWS studies | 8 | Explicit REM vs SWS comparisons for emotional memory selectivity | Schäfer et al., Sleep Medicine Reviews | 2020 |
The strongest quantitative “REM vs deep sleep” cognition evidence often comes from designs that isolate stages (e.g., REM-rich vs SWS-rich intervals, stage-targeted cueing), not from generic “sleep vs wake” comparisons. Counts matter: the REM-vs-SWS selective evidence base is smaller than the full sleep-memory literature, which increases uncertainty when generalizing.
rs=0.66 suggests a combined-stage interaction (SWS×REM) may outperform single-stage metrics for predicting emotional-memory consolidation benefit in at least one experimental context (Yuksel et al., Communications Biology, 2025).
Pro tip for researchers: If you must reduce REM vs deep sleep to one number, consider reporting both REM% and N3% plus an interaction term (e.g., SWS×REM product) where appropriate, because some experimental outcomes correlate with combined architecture rather than a single stage (Yuksel et al., Communications Biology, 2025).
Trends: Rebound, Restriction, Fragmentation (REM vs Deep Sleep)
Large-scale weekly pattern: 6,850,717 recorded nights in 21,543 participants show REM% shifting across the week in a home-monitoring dataset (Ding et al., device-based cohort report, 2022).
Trend Types Quantified Here
- Within-week variation: day-of-week changes in REM% and total sleep time in a large device-based dataset (Ding et al., 2022).
- Night-to-night adaptation: REM% increases on second/later PSG nights (Boulos et al., 2019; foundational).
Within-Week REM% and Total Sleep Time (Large Home Dataset)
| Day | Total Sleep Time (hours) | REM (%) | Source | Year |
|---|---|---|---|---|
| Sunday | 7.21 | 24.20 | Ding et al., device-based cohort report | 2022 |
| Friday | 7.51 | 25.05 | Ding et al., device-based cohort report | 2022 |
REM% can drift across the week in large datasets, but stage estimates in consumer-grade systems can embed algorithmic assumptions. Treat within-week REM% shifts as a trend signal, and triangulate with PSG-based evidence when making clinical or policy claims.
6,850,717 nights is large enough to detect small REM% shifts across the week, even when individual-level variance is high (Ding et al., device-based cohort report, 2022).
Pro tip for analysts: When using “REM rebound” trends from consumer or contactless systems, report the device type and validation status, then bracket interpretation using PSG norms (e.g., pooled REM% around ~19–21% in foundational PSG meta-analysis).
Wearables & Scoring Performance (REM vs Deep Sleep)
Stage performance gap: 0.672 vs 0.528 are the reported average F1 scores for REM vs deep sleep among the top five consumer sleep trackers in a multicenter validation (Lee et al., JMIR mHealth and uHealth, 2023).
How Wearable “REM” and “Deep” Labels Are Evaluated
- Reference standard: in-lab PSG staging is the comparator (Lee et al., 2023).
- Metric: macro F1 summarizes stage classification performance across classes (Lee et al., 2023).
- Stage asymmetry: REM and deep sleep can show different F1 performance under the same device/algorithm family (Lee et al., 2023).
Wearable Stage Classification: REM vs Deep Sleep (Selected)
| Metric | REM | Deep (N3) | Source | Year |
|---|---|---|---|---|
| Top-5 devices average F1 (stage) | 0.672 | 0.528 | Lee et al., JMIR mHealth and uHealth | 2023 |
| Device deep-stage macro F1 (example) | Not reported in this row | 0.5933 (Pixel Watch) | Lee et al., JMIR mHealth and uHealth | 2023 |
| Device deep-stage macro F1 (example) | Not reported in this row | 0.5564 (Fitbit Sense 2) | Lee et al., JMIR mHealth and uHealth | 2023 |
Wearables can estimate REM and deep sleep with meaningful but imperfect agreement versus PSG. REM can achieve higher average F1 than deep sleep in top-performing systems, which matters when interpreting “REM vs deep sleep balance” dashboards.
0.26 macro F1 at the low end means some consumer trackers perform near unusable levels for multi-stage classification, even if total sleep time appears plausible (Lee et al., JMIR mHealth and uHealth, 2023).
Pro tip for journalists: When reporting “deep sleep minutes” or “REM minutes” from a wearable, include a validation metric (e.g., F1 or macro F1) and the comparator (PSG). Otherwise, the statistic is not portable across devices or algorithm versions (Lee et al., JMIR mHealth and uHealth, 2023).
Research Gaps (REM vs Deep Sleep Statistics)
Coverage gap marker: k=1 is the number of study groups available for ≥80-year-old normative estimates in a major PSG norms meta-analysis (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
The biggest practical gap is comparability: REM and N3 are often discussed together, but datasets rarely harmonize stage definitions, scoring rules, night-of-study effects, and device validation. This makes “REM vs deep sleep” statistics highly sensitive to study design.
k=1 in the ≥80-year-old normative strata means “REM vs deep sleep in the oldest-old” is still weakly benchmarked in widely cited PSG norms (Boulos et al., The Lancet Respiratory Medicine, 2019; foundational).
Pro tip for editors: If a claim says “REM matters more than deep sleep” (or vice versa), require: (1) the exposure definition (REM% vs N3% vs change rate), (2) the measurement method (PSG vs wearable), and (3) whether it is based on first-night or later-night PSG norms.
Methodology
Compilation scope: 70 unique REM vs deep sleep statistics were extracted into the dataset table below (ZenSleepZone, Statistics Hub build log, 2026).
Evidence Hierarchy and Inclusion Rules
- Primary preference: peer-reviewed meta-analyses, cohort studies, clinical PSG studies, and validated device studies that report stage-specific REM and/or N3 outcomes.
- Timeframe: emphasis on 2020–2026; foundational normative PSG references (2017–2019) included when they remain the dominant benchmark.
- Inclusion criteria: a data point must quantify REM and/or N3 directly (percent, minutes, latency, stage-specific AHI, validated classification performance, or stage-specific outcome association).
- Exclusion criteria: general sleep-duration-only statistics without explicit linkage to REM and/or N3 were excluded from the main statistics blocks.
This hub is optimized for extraction: every statistic is written in a single line with a source, publication, and year. When estimates are foundational (older than 2020), they are labeled as such rather than treated as “latest.”
0 clean economic partition estimates (REM vs N3) is not a missing citation problem; it reflects a real measurement/accounting gap in the literature.
Pro tip for replicators: Recompute any “REM vs N3 prevalence” estimate separately for first-night and later-night PSG when possible, because REM% changes systematically by night (Boulos et al., 2019; foundational).
Source Distribution Table
| Source Type | Count |
|---|---|
| Peer-Reviewed Journals (original studies, cohorts) | 14 |
| Meta-Analyses / Systematic Reviews | 5 |
| Clinical/Scoring Guidance & Reference Tables | 3 |
| Conference/Poster or Device-Cohort Reports (linked to peer-reviewed validation) | 2 |
| Internal (ZenSleepZone methodology counters) | 1 |
Complete Data Reference Table (REM vs Deep Sleep)
| Statistic | Value | Source | Year |
|---|---|---|---|
| N3 scoring rule requires ≥20% slow waves in a 30-s epoch | ≥20% | Huang et al., Sleep | 2024 |
| N3 slow-wave frequency band in AASM-based criteria | 0.5–2.0 Hz | Huang et al., Sleep | 2024 |
| N3 slow-wave amplitude threshold in AASM-based criteria | ≥75 µV | Huang et al., Sleep | 2024 |
| N3 time threshold inside a 30-s epoch (equivalent to 20% rule) | ≥6 seconds | Huang et al., Sleep | 2024 |
| Instructional reference: REM share of total sleep (adults) | 20–25% | StatPearls, NCBI Bookshelf | 2024 |
| Instructional reference: N3 share of total sleep (adults) | ~20% | StatPearls, NCBI Bookshelf | 2024 |
| Pooled healthy-adult PSG total sleep time (foundational) | 405.2 min | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG sleep efficiency (foundational) | 86.7% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG WASO (foundational) | 43.3 min | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG stage N1 percentage (foundational) | 9.7% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG stage N2 percentage (foundational) | 50.6% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG stage N3 percentage (foundational) | 19.5% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Pooled healthy-adult PSG REM percentage (foundational) | 19.2% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Men-only pooled PSG N3 percentage (foundational) | 21.0% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Men-only pooled PSG REM percentage (foundational) | 19.9% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Women-only pooled PSG N3 percentage (foundational) | 22.1% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Women-only pooled PSG REM percentage (foundational) | 18.6% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| First-night PSG REM percentage (foundational) | 18.3% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Second/later-night PSG REM percentage (foundational) | 21.4% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Second/later-night PSG N3 percentage (foundational) | 22.3% | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Later PSG nights: total sleep time increase vs first night (foundational) | +38.3 min | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Later PSG nights: REM percentage increase vs first night (foundational) | +3.5 percentage points | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Later PSG nights: REM latency reduction vs first night (foundational) | −11.1 min | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Age effect (PSG norms): REM% change per decade (foundational) | 0.0% per decade (not significant) | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Age effect (PSG norms): N3% change per decade (foundational) | −0.1% per decade (not significant) | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Age effect (PSG norms): N1% change per decade (foundational) | +0.5% per decade | Boulos et al., The Lancet Respiratory Medicine | 2019 |
| Adult PSG cohort: mean REM latency | 105.50 min | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Adult PSG cohort: mean REM% of total sleep | 20.86% | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Adult PSG cohort: mean SWS% of total sleep | 14.65% | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Adult PSG cohort: lowest oxygen saturation during REM | 82.91% | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Adult PSG cohort: lowest oxygen saturation during NREM | 85.52% | Frontiers in Aging Neuroscience (Original Research PDF) | 2024 |
| Elite sports PSG: mean REM% | 18.02% | Frontiers in Psychology | 2021 |
| Elite sports PSG: mean N3% | 26.45% | Frontiers in Psychology | 2021 |
| Elite sports PSG: mean sleep efficiency | 86.99% | Frontiers in Psychology | 2021 |
| Deep sleep decline with aging in repeated PSG cohort | −0.6 percentage points/year | Himali et al., JAMA Neurology | 2023 |
| Dementia risk per 1%/year SWS decrease | HR 1.27 | Himali et al., JAMA Neurology | 2023 |
| Repeated PSG dementia cohort sample size | N=346 | Himali et al., JAMA Neurology | 2023 |
| Incident dementia cases in repeated PSG cohort | 52 cases | Himali et al., JAMA Neurology | 2023 |
| Incident dementia risk change per 1% REM reduction (foundational) | ~9% higher risk per 1% lower REM | Pase et al., Neurology | 2017 |
| OSA + very low SWS quartile definition in hypertension association study | SWS <2.0% | Zhang et al., Journal of International Medical Research | 2020 |
| Hypertension odds ratio for OSA with lowest SWS quartile vs primary snorers | OR 2.13 | Zhang et al., Journal of International Medical Research | 2020 |
| REM sleep mortality cohort follow-up length | 11.0 ± 3.1 years | Zhang et al., Aging | 2019 |
| Deaths in SHHS-derived cohort for REM mortality analysis | 1234 (21.9%) | Zhang et al., Aging | 2019 |
| All-cause mortality HR comparing lowest vs highest REM duration quartile | HR 1.727 | Zhang et al., Aging | 2019 |
| All-cause mortality HR comparing lowest vs highest REM proportion quartile | HR 1.545 | Zhang et al., Aging | 2019 |
| DKD prevalence in type 2 diabetes PSG cohort used for REM-AHI analysis | 35.2% | American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine | 2021 |
| Median REM-AHI in DKD cohort | 35.4 events/h | American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine | 2021 |
| Median NREM-AHI in DKD cohort | 29.1 events/h | American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine | 2021 |
| DKD odds ratio by REM-AHI quartile (Q2 vs Q1) | OR 3.14 | American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine | 2021 |
| DKD odds ratio by REM-AHI quartile (Q4 vs Q1) | OR 4.97 | American Academy of Sleep Medicine, Journal of Clinical Sleep Medicine | 2021 |
| REM-predominant OSA prevalence among OSA patients | 20.2% | Cho et al., International Journal of Environmental Research and Public Health | 2022 |
| Female proportion in REM-OSA vs nREM-OSA | 53.6% vs 21.7% | Cho et al., International Journal of Environmental Research and Public Health | 2022 |
| REM-AHI vs NREM-AHI in BP-during-events OSA sample | 51.9 vs 38.3 events/h | Almeneessier et al., Scientific Reports | 2020 |
| Hypertension prevalence in BP-during-events sample | 31.3% | Almeneessier et al., Scientific Reports | 2020 |
| Systolic BP increase during REM obstructive events vs quiet sleep | ~8 mmHg | Almeneessier et al., Scientific Reports | 2020 |
| Systolic BP increase during NREM obstructive events vs quiet sleep | ~4 mmHg | Almeneessier et al., Scientific Reports | 2020 |
| NREM-predominant OSA: loop gain in NREM vs REM | 0.546 vs 0.365 | Monash Health group, Chest | 2021 |
| REM-predominant OSA: Vpassive in NREM vs REM (%Veupnea) | 98.4 vs 95.9 | Monash Health group, Chest | 2021 |
| Emotional recognition memory meta-analysis: total observations | N=1059 | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: sleep vs wake comparison studies | k=22 sleep–wake comparisons (26 samples) | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: sleep observations | n=596 | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: wake observations | n=463 | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: total sleep+wake comparisons (incl. samples without wake group) | 34 comparisons (38 samples) | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: total observations (incl. samples without wake group) | N=1382 (nsleep=919; nwake=463) | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Emotional recognition memory meta-analysis: weighted mean sample age (primary analysis) | Mean=22.46 years (SD=3.31) | Schäfer et al., Sleep Medicine Reviews | 2020 |
| Normal adult sleep architecture: NREM share of total sleep time | 75%–80% | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| Normal adult sleep architecture: REM share of total sleep time | 20%–25% | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| Sleep cycle length: first NREM→REM cycle duration | 70–100 minutes | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| Sleep cycle length: subsequent cycle duration | 90–120 minutes | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| Typical nightly cycling: complete cycles per night (adults) | 4–5 cycles | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| REM distribution across night: REM share in later cycles | Up to 30% of later cycles | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| Deep sleep (N3) duration: early-night N3 episode length | 20–40 minutes (initially) | Brinkman et al., StatPearls (NCBI Bookshelf) | 2023 |
| REM sleep and incident heart failure cohort: baseline sample size | n=4490 | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| REM sleep and incident heart failure cohort: mean age | Mean=63.2 years (SD=11.0) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| REM sleep and incident heart failure cohort: mean follow-up time | Mean=10.9 years | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| REM sleep and incident heart failure cohort: incident HF cases | 436 cases (9.7%) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Incident heart failure: REM% association (fully adjusted) | HR=0.88 per +5% REM (95% CI 0.82–0.94) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Incident heart failure: REM time association (fully adjusted) | HR=0.97 per +5 minutes REM (95% CI 0.95–0.99) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Incident heart failure: highest vs lowest REM% quartile (fully adjusted) | Q4 (>24.0%) vs Q1 (<15.8%): HR=0.65 (95% CI 0.48–0.88) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Incident heart failure: highest vs lowest REM minutes quartile (fully adjusted) | Q4 (>91.5 min) vs Q1 (<54.0 min): HR=0.64 (95% CI 0.45–0.90) | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Cardiovascular death events (same cohort) | 238 deaths | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Cardiovascular death: REM% association | HR=0.90 (95% CI 0.81–0.99) per +5% REM | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Cardiovascular death: REM minutes association | HR=0.97 (95% CI 0.94–0.99) per +5 minutes REM | Zhao et al., Frontiers in Cardiovascular Medicine | 2022 |
| Deep sleep (N3) and incident hypertension cohort: baseline sample size | n=1850 | Javaheri et al., Sleep | 2017 |
| Deep sleep (N3) and incident hypertension cohort: mean age | Mean=59.4 years (SD=10.1) | Javaheri et al., Sleep | 2017 |
| Deep sleep (N3) and incident hypertension cohort: female share | 55.5% | Javaheri et al., Sleep | 2017 |
| Deep sleep (N3) and incident hypertension cohort: incident hypertension | ~30% developed hypertension over mean 5.3 years | Javaheri et al., Sleep | 2017 |
| Incident hypertension odds: low N3% (quartile 1) vs mid N3% (quartile 3) | Q1 (<9.8%) vs Q3 (17.7%–25.2%): OR=1.69 (95% CI 1.21–2.36) | Javaheri et al., Sleep | 2017 |
| Incident hypertension odds: low-to-mid N3% (quartile 2) vs mid N3% (quartile 3) | Q2 (9.8%–17.7%) vs Q3 (17.7%–25.2%): OR=1.45 (95% CI 1.04–2.00) | Javaheri et al., Sleep | 2017 |
| Incident hypertension odds: low N3 minutes (quartile 1) vs mid N3 minutes (quartile 3) | Q1 (<36.5 min) vs Q3 (66.5–95.0 min): OR=1.86 (95% CI 1.33–2.62) | Javaheri et al., Sleep | 2017 |
| Incident hypertension odds: low-to-mid N3 minutes (quartile 2) vs mid N3 minutes (quartile 3) | Q2 (36.5–66.5 min) vs Q3 (66.5–95.0 min): OR=1.46 (95% CI 1.05–2.04) | Javaheri et al., Sleep | 2017 |
| REM-related OSA chart review: suspected OSA patients screened | n=696 | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA chart review: OSA patients enrolled | n=408 | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA prevalence (definition: AHI≥5; REM-AHI/NREM-AHI>2; NREM-AHI<15) | 21.6% | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA association: female sex | OR=2.35 (95% CI 1.25–4.42) | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA association: age <60 years | OR=2.52 (95% CI 1.15–5.55) | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA association: mild OSA severity | OR=17.46 (95% CI 9.28–32.84) | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA vs non-stage-specific OSA: mean age | 45.2 (SD 14.9) vs 50.9 (SD 15.8) years | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA vs non-stage-specific OSA: overall AHI | Median 11.3 vs 39.9 events/hour | Sattaratpaijit et al., Scientific Reports | 2022 |
| REM-related OSA severity distribution (within REM-related group) | 70.4% mild; 29.6% moderate; 0% severe | Sattaratpaijit et al., Scientific Reports | 2022 |
| Consumer wearable validation study: participants | n=35 (aged 20–50 years) | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (2-stage sleep vs wake): Oura vs PSG | 92.0% agreement; κ=0.60 | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (4-stage wake/light/deep/REM): Oura vs PSG | 76.3% agreement; κ=0.65 | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (2-stage sleep vs wake): Fitbit vs PSG | 91.0% agreement; κ=0.52 | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (4-stage wake/light/deep/REM): Fitbit vs PSG | 70.9% agreement; κ=0.55 | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (2-stage sleep vs wake): Apple Watch vs PSG | 93.0% agreement; κ=0.60 | Robbins et al., Sensors | 2024 |
| Epoch-by-epoch agreement (4-stage wake/light/deep/REM): Apple Watch vs PSG | 75.0% agreement; κ=0.60 | Robbins et al., Sensors | 2024 |
| Binary sleep detection sensitivity: Oura | 95% (SD 3%) | Robbins et al., Sensors | 2024 |
| Binary sleep detection sensitivity: Fitbit | 95% (SD 3%) | Robbins et al., Sensors | 2024 |
| Binary sleep detection sensitivity: Apple Watch | 97% (SD 2%) | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Oura deep (N3) | 79.5% | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Oura REM | 76.0% | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Fitbit deep (N3) | 61.7% | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Fitbit REM | 67.3% | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Apple Watch deep (N3) | 50.5% | Robbins et al., Sensors | 2024 |
| Sleep-stage sensitivity (4-stage): Apple Watch REM | 82.6% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Oura deep (N3) | 77.0% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Oura REM | 79.1% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Fitbit deep (N3) | 73.2% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Fitbit REM | 73.1% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Apple Watch deep (N3) | 87.8% | Robbins et al., Sensors | 2024 |
| Sleep-stage precision (PPV, 4-stage): Apple Watch REM | 77.7% | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): deep sleep minutes (Oura vs PSG) | ICC=0.32 | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): deep sleep minutes (Fitbit vs PSG) | ICC=0.36 | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): deep sleep minutes (Apple Watch vs PSG) | ICC=0.13 | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): REM minutes (Oura vs PSG) | ICC=0.27 | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): REM minutes (Fitbit vs PSG) | ICC=0.13 | Robbins et al., Sensors | 2024 |
| Nightly stage-total concordance (ICC): REM minutes (Apple Watch vs PSG) | ICC=0.37 | Robbins et al., Sensors | 2024 |
| Device–PSG disagreement (minutes): deep sleep nightly total | Range 0 min (Oura) to 43 min (Apple Watch) | Robbins et al., Sensors | 2024 |
| Device–PSG disagreement (minutes): REM sleep nightly total | Range 3 min (Oura) to 7 min (Fitbit) | Robbins et al., Sensors | 2024 |
| Wearable (Verily Study Watch) vs PSG: sleep vs wake sensitivity | 0.97 (95% CI 0.96–0.98) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: sleep vs wake specificity | 0.70 (95% CI 0.66–0.74) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: 4-stage overall accuracy | 0.78 (95% CI 0.58–0.89) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: 4-stage overall kappa | 0.64 (95% CI 0.18–0.82) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: deep sleep sensitivity | 0.77 (95% CI 0.37–0.98) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: REM sleep sensitivity | 0.84 (95% CI 0.47–0.99) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: deep sleep kappa | 0.66 (95% CI 0.17–0.91) | Saeb et al., Frontiers in Sleep | 2024 |
| Wearable (Verily Study Watch) vs PSG: REM sleep kappa | 0.74 (95% CI 0.38–0.90) | Saeb et al., Frontiers in Sleep | 2024 |
Related ZenSleepZone Resources
- REM vs Deep Sleep: Key Differences, Benefits, and How Each Stage Affects Health
- REM vs Deep Sleep Q&A (Answer Hub)
- REM vs Deep Sleep Infographic
- REM vs Deep Sleep Self‑Assessment Tool
- The Full Sleep Cycle (Stages & Cycles)
- Circadian Rhythm and Sleep Stages
- Stress Impact on Sleep Stages
High-Authority External References
Action: Use the REM vs Deep Sleep Self‑Assessment Tool to compare your reported REM and deep sleep patterns against evidence-based reference ranges and measurement limits from PSG vs wearables.
Research note: When citing stage totals from consumer wearables, prioritize studies that report both epoch-by-epoch agreement (kappa, sensitivity/precision) and agreement with nightly totals (ICC) for REM and N3.
Sources / Bibliography (APA)
- Brinkman, J. E., Reddy, V., & Sharma, S. (2023). Physiology of Sleep. StatPearls Publishing. NCBI Bookshelf.
- Feriante, J., & Araujo, J. F. (2023). Physiology, REM Sleep. StatPearls Publishing. NCBI Bookshelf.
- Schäfer, S. K., Wirth, B. E., Staginnus, M., Becker, N., Michael, T., & Sopp, M. R. (2020). Sleep’s impact on emotional recognition memory: A meta-analysis of whole-night, nap, and REM sleep effects. Sleep Medicine Reviews, 51, 101280. doi:10.1016/j.smrv.2020.101280
- Robbins, R., Weaver, M. D., Sullivan, J. P., Quan, S. F., Gilmore, K., Shaw, S., Benz, A., Qadri, S., Barger, L. K., Czeisler, C. A., & Duffy, J. F. (2024). Accuracy of three commercial wearable devices for sleep tracking in healthy adults. Sensors, 24, 6532. doi:10.3390/s24206532
- Saeb, S., Nelson, B. W., Barman, P., Verma, N., Allen, H., de Zambotti, M., Baker, F. C., Arra, N., Sridhar, N., Sullivan, S. S., Plowman, S., Rainaldi, E., Kapur, R., Shin, S., & colleagues. (2024). Performance of the Verily Study Watch for measuring sleep compared to polysomnography. Frontiers in Sleep, 3. doi:10.3389/frsle.2024.1481878
- Zhao, B., Jin, X., Yang, J., Ma, Q., Yang, Z., Wang, W., Bai, L., Ma, X., & Yan, B. (2022). Increased rapid eye movement sleep is associated with a reduced risk of heart failure in middle-aged and older adults. Frontiers in Cardiovascular Medicine, 9, 771280. doi:10.3389/fcvm.2022.771280
- Javaheri, S., Zhao, Y. Y., Punjabi, N. M., Quan, S. F., Gottlieb, D. J., & Redline, S. (2017). Slow-wave sleep is associated with incident hypertension: The Sleep Heart Health Study. Sleep, 41(1), zsx179. doi:10.1093/sleep/zsx179
- Sattaratpaijit, N., Kulalert, P., & Wongpradit, W. (2022). Characteristics of rapid eye movement-related obstructive sleep apnea in Thai patients. Scientific Reports, 12, 11360. doi:10.1038/s41598-022-13382-z
- Lee, Y. F., Gerashchenko, D., Timofeev, I., Bacskai, B. J., & Kastanenka, K. V. (2020). Slow wave sleep is a promising intervention target for Alzheimer’s disease. Frontiers in Neuroscience, 14, 705. doi:10.3389/fnins.2020.00705
- American Academy of Sleep Medicine. (2017). Summary of updates in The AASM Manual for the Scoring of Sleep and Associated Events (v2.1).
- American Academy of Sleep Medicine. (2005, updated). Practice parameters for the indications for polysomnography and related procedures.
- Kocevska, D., et al. (2020). Sleep characteristics across the lifespan in 1.1 million people from the Netherlands, United Kingdom and United States: A systematic review and meta-analysis. Nature Human Behaviour. doi:10.1038/s41562-020-00965-x
- Neumann, A. M., et al. (2021). Modulating overnight memory consolidation by acoustic stimulation during slow-wave sleep: A systematic review and meta-analysis. Sleep, 44(7), zsaa296. doi:10.1093/sleep/zsaa296
- Cook, J. D., Ferry, D. G., & Tran, K. M. (2020). Sleep’s role in preventing and treating Alzheimer’s disease: Are we moving towards slow-wave assessment and enhancement? Sleep, 43(5), zsz304. doi:10.1093/sleep/zsz304
- Nollet, M., Franks, N. P., Wisden, W., & Morton, J. (2023). Understanding sleep regulation in normal and pathological conditions, and why it matters. Journal of Huntington’s Disease. doi:10.3233/JHD-230564
- Frontiers in Sleep. (2024). Precision Sleep Medicine section (journal scope and methodology notes for wearable validation studies). Frontiers Media SA.
- Frontiers in Cardiovascular Medicine. (2022). Heart Failure and Transplantation section—original research standards (for PSG-derived REM metrics). Frontiers Media SA.
- National Institutes of Health, National Library of Medicine. (2023). NCBI Bookshelf: StatPearls.
- American Academy of Sleep Medicine. (2014). International Classification of Sleep Disorders (3rd ed.).
- Sleep Heart Health Study Research Group. (1997). The Sleep Heart Health Study: Design, rationale, and methods. Sleep, 20(12), 1077–1085.
- Redline, S., et al. (1998). Methods for obtaining and analyzing unattended polysomnography data for a multicenter study. Sleep, 21(7), 759–767.
- Rechtschaffen, A., & Kales, A. (1973). A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. UCLA Brain Information Service.
- Ohayon, M. M., Carskadon, M. A., Guilleminault, C., & Vitiello, M. V. (2004). Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals. Sleep, 27(7), 1255–1273.
- Walsh, J. K., et al. (2010). Enhancing slow wave sleep with sodium oxybate reduces the behavioral and physiological impact of sleep loss. Sleep, 33(9), 1217–1225.
- Bellesi, M., Riedner, B. A., Garcia-Molina, G. N., Cirelli, C., & Tononi, G. (2014). Enhancement of sleep slow waves: Underlying mechanisms and practical consequences. Frontiers in Systems Neuroscience, 8, 208. doi:10.3389/fnsys.2014.00208
