The short answer
Sleep moves between three non-REM stages, called N1, N2 and N3, and REM sleep. Their balance changes across the night: deep sleep is more prominent earlier, while REM becomes more prominent later. Cycles are variable, so multiplying 90 minutes cannot reveal your exact stage or guarantee an easy awakening. A stage chart needs a measurement context; it is not a timetable your night must follow.
![Uneven folds of pale-blue, grey and indigo bedding cross a bed in a softly lit room.](data:image/svg+xml,%3Csvg xmlns="http://www.w3.org/2000/svg" width="1536" height="1024"/%3E)
What a sleep stage actually measures
Sleep stages are categories assigned to physiological recordings. In laboratory scoring, specialists examine short intervals of signals and label them wake, N1, N2, N3 or REM. An electroencephalogram, or EEG, records electrical activity from the scalp; eye-movement recordings also help distinguish states. You may see REM written as R and wake as W on a chart.
These labels make a complex recording easier to describe. They do not mean the brain turns off between stages. A journal commentary on automated scoring highlights disagreement between human scorers and the continuous nature of some changes. A sharp boundary drawn on a graph should not be mistaken for a perfectly abrupt biological switch.
The four sleep labels in plain English
N1 describes the transition from waking into sleep. N2 is another non-REM stage: being outside deep sleep does not mean you are awake. N3 is called deep or slow-wave sleep because of the slow-wave pattern in the brain recording. REM stands for rapid eye movement; its brain activity has features resembling waking, while muscle tone is normally reduced.
The numbers name non-REM categories. REM is not a fourth level of increasing depth, and the labels are not a ranking of which minutes count as worthwhile sleep. If an app groups N1 and N2 under a different label, check its explanation before comparing that total with a laboratory report.
The balance changes as the night continues
An ordinary night contains repeated passages through non-REM and REM sleep, with more N3 earlier and more REM later. Brief waking can occur between cycles. This is the broad pattern described by NHLBI, not a requirement to enter every stage in the same order on every return.
This distinction helps when reading a hypnogram, the name for a graph of sleep stages over time. Look first at the whole sequence and its time axis. A total for one stage tells you how much was classified that way; it does not show where those minutes fell or how often sleep was interrupted.
An illustrative night, without a countdown
This text illustration shows the broad pattern a hypnogram can display. It is deliberately schematic, with no measured minutes or individual data. Early in the sleep period: wake, then N1, N2, N3, a return towards N2, and REM. Later: further passages through non-REM and REM, with shorter or absent N3 portions and more space occupied by REM. Brief waking may interrupt the sequence.
Read the example from earlier to later. Its purpose is to show a changing mixture, not identical blocks copied along a timeline. It cannot tell you what was happening at a particular hour last night. Adding an exact bedtime to this illustration would not turn it into a measurement of your sleep.
Why a 90-minute calculator adds false precision
A cycle is a recurring pattern, not a fixed unit like a minute. In their original 1979 study, Feinberg and Floyd compared successive cycles using definitions based on non-REM or REM onsets. Their abstract describes different cycle lengths across the night in the healthy comparison groups and warns that the overall trends may not be apparent on a single night. Only that original abstract was inspected here; it does not supply a personalised timing rule.
Suppose a calculator offers 06:00 or 07:30 because each is a whole number of 90-minute blocks after bedtime. Its arithmetic can be correct while its biological prediction is unsupported. It also needs to know when sleep actually began. An estimated cycle boundary is not a reason to cut a sleep opportunity short.
Age differences are context, not a stage target
A meta-analysis by Ohayon and colleagues brought together 65 studies of objectively recorded sleep in healthy participants across the lifespan. Its abstract reports less slow-wave sleep with increasing adult age, but also shows that screening for health conditions and other factors changed the size of age associations. The complete paper was not reviewed here.
That finding helps explain why a single illustration cannot represent every adult. It does not set a compulsory deep-sleep percentage for your age or identify the cause of a change in one person's graph. Before comparing two records, check whether they used the same measurement method and whether they describe one night or a longer period.
Keep dream reports and tracker estimates separate
Dreaming is reported after both REM and non-REM sleep. The original abstract of Siclari and colleagues' EEG study supports that distinction. A vivid story therefore cannot identify the stage behind it. Dreams in REM and non-REM sleep explores the report evidence; this guide concerns the arrangement of the sleep stages themselves.
A consumer graph needs its own evidence too. In a 2022 study of 21 healthy young adults sleeping at home, Chinoy and colleagues compared four wearables with a mobile EEG headband. Stage estimates varied substantially. The inspected abstract concerns those devices and that reference method, not every current product. Ask what validates your device's stage labels instead of treating their precision on screen as proof of precision in measurement.
Choose the question the chart can help answer
A useful first question is whether you want to understand sleep physiology, a remembered dream, or a difficulty in daily life. Those require different information. For example, “I had little time available for sleep” is a scheduling observation. “My watch displayed little deep sleep” is a device estimate. “I keep struggling to stay awake at work” is a waking concern. Preserve the actual observation when describing each one.
For duration, continue with How much sleep do adults need? For interpreting device outputs, use Sleep trackers and their limits. Body clock and sleep pressure explains why timing and previous waking both matter. The examples here are educational; they do not turn stage totals into a personal diagnosis or a stage-optimisation programme.
Questions to take with you
Use these optional prompts to consider the evidence or describe your own experience. They are not a test or a task you need to complete.
- Does this graph contain measured stages, estimated stages or an illustration?
- Am I comparing stage totals, their order or an actual waking difficulty?
- What evidence supports an exact timing claim?
Common questions
Are all sleep cycles exactly 90 minutes?
No. Cycle length changes, and the way a cycle is defined matters. A fixed 90-minute schedule cannot identify the stage you will be in at a particular alarm time.
Is N2 sleep the same as being awake?
No. N2 is one of the three non-REM sleep stages. A category described as lighter sleep still describes sleep.
Does remembering a dream prove I woke from REM?
No. Reports of dreaming also occur after non-REM sleep. The remembered story and the measured sleep stage are separate records.
Can I use the illustrative sequence to judge last night's sleep?
It shows an idea, not your recording. It has no individual measurements, exact durations or pass/fail stage targets. To understand an actual graph, start with how its labels were obtained.
Sources & their limits
These sources support the statements indicated above. Their authors do not endorse this website or provide an assessment of your experience.
- NHLBI — How Sleep Works: Sleep Phases and Stages
Official public overview inspected: stage definitions and the changing distribution of N3 and REM. Its approximate cycle description is not used as an individual prediction.
Cited in this guide · 3 places
- American Academy of Sleep Medicine — Scoring for Sleep Stage
Official scoring-interface instructions inspected for the W, N1, N2, N3 and R labels. This is not a full review of the proprietary scoring manual.
Cited in this guide · 3 places
- Nasiri et al. (2023) — Exploiting labels from multiple experts in automated sleep scoring
Journal commentary opening paragraphs inspected for epoch scoring, expert disagreement and the continuous biological changes represented by discrete categories. No model-performance result is used here.
Cited in this guide · 1 place
- Feinberg and Floyd (1979) — Systematic Trends Across the Night in Human Sleep Cycles
Original abstract only: definitions and changing cycle durations across the night in the comparison groups. Full methods and individual cycle ranges were not inspected.
Cited in this guide · 2 places
- Ohayon et al. (2004) — Meta-analysis of quantitative sleep parameters across the lifespan
Original meta-analysis abstract only: age associations and the influence of participant screening. Descriptive sleep measurements do not establish individual sleep requirements.
Cited in this guide · 1 place
- Siclari et al. (2017) — The neural correlates of dreaming
Original abstract inspected for dream reports in REM and non-REM. This new source use does not claim a fresh full-text review of the paper.
Cited in this guide · 2 places
- Chinoy et al. (2022) — Performance of Four Commercial Wearable Sleep-Tracking Devices Tested at Home
Original abstract only: 21 healthy young adults, one week at home, mobile EEG headband reference and variable stage tracking. Not a current product ranking or a clinical validation of every wearable.
Cited in this guide · 1 place