Before a dawn run, the promise is tempting: simply count backwards from your alarm in ninety-minute units and you will wake precisely at the end of a sleep cycle. The formula is tidy; the night is not such a regular timetable. Sleep does consist of stages and recurring cycles, but an average duration does not imply a personal boundary known in advance to the minute.
Understanding sleep cycles can still help. It explains why the composition of the night changes, what a sleep investigation can show and why a sports watch’s colourful graph is not a direct measurement of brain activity. For runners, that can support workable time planning: do not shorten the opportunity for sleep for an attractive multiplication, or judge all your recovery from one awakening.
What is a sleep cycle?
The two broad types of sleep are REM and non-REM sleep. REM stands for rapid eye movement; the three non-REM stages are distinguished using brain activity measured during sleep investigation. NHLBI describes a recurring cycle of approximately 80–100 minutes and notes that brief awakenings may occur between cycles. That range alone shows why treating every cycle as exactly ninety minutes is misleading. NHLBI: sleep stages.
A cycle describes a physiological pattern. It is not a package that starts at the press of a bedtime button and repeats with identical contents. Non-REM N1 is the transition from wakefulness into sleep; N2 is a distinct sleep stage; N3 is also called deep or slow-wave sleep. REM features characteristic rapid eye movements and different brain activity. These concepts are not performance levels invented by a sports watch. The scientific basis of sleep stages.
In everyday conversation, light and deep sleep sometimes sound like value judgements: one is good, the other less useful. Stage names are not competition grades. If an app displays several colours, your task is not to maximise the same one every night. First establish which signal the device estimates from, how it names its categories and what evidence supports the display’s meaning.
The beginning and end of the night differ in composition
NICHD explains that deep sleep is generally longer and REM periods shorter early in the night; later, REM duration increases and deep sleep decreases. The page mentions approximately ninety minutes until the first REM period. That is not the same as claiming that every subsequent cycle in every person lasts ninety minutes and can be counted from bedtime. “Approximately” is an essential part of the physiological description. NICHD: changing sleep patterns across the night.
The night therefore cannot be assembled from identical, interchangeable bricks. An earlier awakening does not simply remove the same package as a later bedtime. You do not need to calculate personal REM or deep-sleep quotas from this. The useful practical point is that the whole opportunity for sleep has value; an attractive cycle count alone does not justify shortening the night.
NHLBI also describes sleep patterns changing with age. This is another reason not to use a differently aged running friend’s watch graph as your personal model. A fellow runner may organise the day differently, experience waking differently and use another device. Shared experiences can be interesting, but you need not adjust your sleep to the colours on someone else’s screen. NHLBI: changes in sleep patterns.
Enlarge imageCounting backwards becomes uncertain at the starting point
An online calculator often asks when you want to get up, then lists several bedtimes. Its calculation must assume when sleep begins and how long recurring cycles last. If you fall asleep later than assumed, the first stage already shifts relative to the plan. If later cycle lengths are not constant either, further uncertainty joins that initial difference.
A purely mathematical example helps: five ninety-minute units make seven and a half hours. Imagine one actual, variable unit being shorter and another longer, and the fifth endpoint shifts. This is not a model of your cycles; it illustrates that uncertain components do not become certain because the result is displayed precisely. A minute on a screen does not remove the input assumptions.
A similar distinction exists between departure and arrival when travelling to a race. An average journey time can support planning without promising traffic lights and connections known to the second. With sleep, you cannot voluntarily start the next stage either. The example concerns interpreting an estimate, rather than replacing the night with yet another calculation method.
How you feel on waking depends on more than a cycle endpoint
Temporarily reduced alertness and performance after awakening is called sleep inertia. Scheer and colleagues’ 2008 laboratory study examined repeated awakenings and arithmetic performance in twelve adults at different phases of the internal daily rhythm. They found greater immediate post-awakening impairment during the biological night. The rhythm’s effect could not simply be explained by the proportions of sleep stages before awakening. Scheer and colleagues: circadian rhythm and performance on waking.
This is original human research with a small sample in a specific laboratory setting, rather than a comparison of dawn running programmes. It gives no universal waiting period for when driving is safe or when training should begin. It does illustrate why the post-awakening state cannot be reduced to one cycle boundary. Internal timing and the circumstances of awakening may also contribute.
NHLBI explains that sleep is regulated both by the internal daily rhythm and by sleep pressure increasing during wakefulness, while environmental light cues influence timing too. The time shown on an alarm clock is therefore not a full description of every biological process. Feeling groggy despite an alarm ringing at “exactly the right minute” need not indicate an arithmetic mistake: the simple formula may ask too much of one number. NHLBI: the sleep–wake cycle.
A laboratory stage and a coloured watch band are different data
Polysomnography obtains information from brain activity and eye movements, among other signals; these are important foundations for classifying sleep stages. A wrist-worn consumer device does not become a direct brain measurement by displaying the same stage names. Specific comparative research can support the usefulness of a signal-based estimate. Display detail and classification accuracy are separate properties. NHLBI: what a sleep study can measure.
Chinoy and colleagues’ original study, published in 2021, compared seven consumer devices with polysomnography across three laboratory nights in 34 healthy young adults, including disrupted sleep. Several devices showed promise in detecting sleep, but individual-stage classification was inconsistent, and performance generally worsened on more disrupted nights. Estimating total sleep and reliably identifying each stage are therefore different tasks. Chinoy and colleagues: comparison of seven sleep trackers.
The study examined particular earlier devices and settings; it is not a test of a current watch running different firmware, or a result for everyone living with a sleep disorder. It implies neither that every new device is wrong nor that any current smart alarm guarantees refreshed waking. If a particular product makes such a promise, seek evidence for that specific claim: who was studied, compared with what and under which conditions?
What should you ask about a smart alarm’s promise?
| Claim or display | Useful clarifying question | What does not follow from it |
|---|---|---|
| Cycle calculator | How does it estimate sleep onset and cycle length? | The calculated minute is the end of your actual cycle. |
| Sleep-stage graph | What reference investigation was used to check the classification? | The coloured band is a direct brain measurement. |
| Flexible alarm window | What determines the choice within the window, and when will it ring at the latest? | The chosen moment makes everyone feel better. |
| Feeling good in the morning | What happens during the rest of the day, and how much sleep do you regularly get? | One refreshed morning settles every sleep question. |
The table is a set of questions for interpreting products, rather than a rating of one device. An alarm operating within a time window and a fixed ninety-minute calculation can be technically different solutions. Evidence for one cannot automatically be transferred to the other. If a feature makes the alarm ring earlier than you wanted, notice the actual sleep time available too, rather than only the feature’s name.
The AASM’s 2018 position statement emphasises that consumer sleep technology does not replace medical assessment, although the data it collects may support discussion. This concerns the limits of healthcare interpretation; it does not claim an app can offer no everyday convenience. A convenience feature and a diagnosis remain separate even on the same screen. AASM: consumer sleep technology.
Enlarge imageAs a runner, leave room for the whole night
If a group run requires an early start, choosing the nearest neat cycle multiple instead of a longer sleep opportunity may be tempting. The AASM and SRS adult recommendation concerns regularly obtaining at least seven hours of sleep to support health, rather than completing a cycle count. It is not a personalised need for everyone or individual medical clearance to run. A calculator’s tidy numbers alone do not make a shorter night adequate. The adult sleep-duration recommendation.
The useful organisational question is simpler: which departure time fits your actual routine? Driving before the start, packing and family tasks take time just as running does. If the portion left for sleep regularly shrinks on several days of the week, there is planning to discuss. You do not need to predict the next onset of REM or purchase another alarm app to do that.
One good or bad awakening is not personal experimental proof of a method. If you are interested in how you feel in the morning, briefly note the actual context: usual schedule, early travel, noisy accommodation or several awakenings. Such notes support memory, but need not validate a calculator. The aim is to fit the morning programme to your life, rather than force your night into the shape of an online table.
Do not treat persistent symptoms by optimising cycles
Regular difficulty falling asleep, frequent wakefulness, persistent daytime fatigue or symptoms disrupting daily life warrant talking to a doctor. The NHS also advises not driving when sleepy. If you would need to drive to dawn training, a favourable app display does not override actual sleepiness. Personal safety is not a test of whether a calculation method succeeds. NHS: insomnia and seeking help.
Knowledge of sleep cycles becomes a useful reference when it reveals the night’s complexity. You can understand that different stages exist, their proportions change from early to late in the night, and morning alertness has several determinants. These concepts help you ask better questions about a device or your own symptoms. Instead of a promise written to the exact minute, you gain clearer sleep knowledge.
Sources
- NHLBI: sleep stages
- NICHD: what happens during sleep?
- NHLBI: regulation of sleep and wakefulness
- Scheer and colleagues: sleep inertia and circadian phase, 2008
- Chinoy and colleagues: seven sleep trackers compared with polysomnography, 2021
- AASM: consumer sleep technology, 2018 position statement
- AASM and SRS: adult sleep-duration recommendation, 2015
- NHS: insomnia
- NHLBI: sleep-study measurements
How Futószív articles are prepared: our editorial guidelines.


