We talk about sleep as though it has a door. You are awake, then you are asleep, and something closes behind you. Research published in Cell Reports in 2026 suggests the boundary between sleep and wakefulness is far more porous than that — people can have bizarre, dream-like experiences while awake, and do rational, organised planning while asleep.
If that is right, the tidy division most of us carry around is a convenience rather than a description.
What the sleep and wakefulness research found
The study examined mental activity across the transition between waking and sleeping, and found traffic running in both directions.
Going one way: dream-like cognition appearing during wakefulness. Not hallucination in a clinical sense, but the loosened, associative, slightly illogical quality normally filed under dreaming — occurring in people who were awake by every external measure.
Going the other: structured, rational thought during sleep. Not the fragmentary impressions sleep is assumed to produce, but planning and reasoning of a kind that looks like ordinary daytime cognition.
Neither of those fits a model where sleep is a switch. They fit a model where the features we associate with each state can come apart and recombine.
Why a porous boundary makes sense
This is less surprising in light of what sleep research has been converging on for years.
Dreaming was once thought to be the business of REM sleep alone. That has not held — dreaming occurs in non-REM sleep too, with different texture and worse recall. The clean mapping of one mental state onto one sleep stage was the first casualty.
Local sleep is the other piece. Studies in animals and humans have found regions of the brain can enter sleep-like activity while the organism is awake and behaving. If parts of a brain can be asleep while the rest is not, then “asleep” is a description of a distribution, not a binary.
Read against that, a porous boundary is close to what you would expect. The puzzle was always the sharp line, not its absence.
The states we already have names for
Everyday language has been describing this boundary for a long time without a scientific framework to hang it on.
Hypnagogia is the drifting state on the way into sleep — the half-formed images, the sentence that made sense a second ago, the sensation of falling that jolts you back. Hypnopompia is its counterpart on the way out, when a dream briefly persists into a room you can already see. Both are common, both are benign in most people, and both sit exactly where a sharp boundary says nothing should exist.
Sleep paralysis is the more alarming version, and the one most often mistaken for something sinister: the body still held in the muscle atonia of REM sleep while awareness has returned. It is frightening precisely because two states that normally travel together have come apart.
Microsleeps run the other way — brief intrusions of sleep into waking, lasting seconds, often unnoticed by the person having them. They are a well-documented hazard in driving research, and they are the clearest everyday evidence that wakefulness is not all-or-nothing.
None of these are exotic. What the newer research adds is a framework in which they stop being anomalies at the edges and start looking like the normal behaviour of a system with soft borders.
The caveats this needs
Studies of conscious experience during sleep rely on waking people and asking what was going through their minds, and that method has limits that no amount of neuroimaging removes.
Reports are retrospective. The person is describing a state they are no longer in, through a mind that has just changed state, using language built for waking experience. Some distortion is guaranteed.
The categories are also fuzzy at the edges. “Dream-like” and “rational” are judgements, usually made by a rater, and the interesting cases are the ambiguous ones — exactly where agreement between raters is weakest.
Laboratory sleep is not home sleep, either. Electrodes, an unfamiliar room, and scheduled awakenings produce lighter and more fragmented sleep than normal, which may make boundary states more common than they would otherwise be.
And findings like this are typically drawn from modest samples. The direction is interesting; the prevalence is not established.
What it means for anyone who sleeps
Mostly it reframes experiences people already have and tend to find unsettling.
Dream-like intrusions while awake are not necessarily a warning sign. The drifting, associative quality that shows up during a dull meeting or a long drive looks, on this account, like a normal feature of a boundary that was never sharp. Persistent or distressing intrusions are worth raising with a doctor; the ordinary kind appears to be ordinary.
Problem-solving in sleep is not folklore. The common experience of waking with an answer fits a brain capable of organised thought while asleep, rather than requiring anything mystical.
“I wasn’t really asleep” may be literally true. People who insist they lay awake all night while a tracker says otherwise are often describing a real experience of a boundary state, not misremembering.
It does not change sleep hygiene. Nothing here alters the practical advice. If you want the better-established lever, our piece on light exposure and sleep timing covers it, and dreams and memory looks at what sleeping minds appear to be doing with the day.
Why the sharp line persisted
If the boundary is this soft, it is worth asking why the binary model lasted so long.
Partly it is the measurement. Sleep staging was built on scoring EEG in fixed epochs, usually thirty seconds, and assigning each one a single label. A method that forces one answer per window will produce clean transitions whether or not the underlying process has them. The tidiness was partly an artefact of the scoring sheet.
Partly it is experience. From the inside, sleep genuinely feels like a gap — you were there, then you were not. But that is a report assembled afterwards by a mind that was absent for the part in question, which makes it a poor witness to its own transitions.
Frequently asked questions
Can you really think rationally while asleep?
This research reports organised, planning-like thought occurring during sleep. It is a finding about what can happen, not a claim that sleep is generally a productive state.
Is having dream-like thoughts while awake normal?
On this evidence it appears to be part of how the waking–sleeping boundary works. Intrusions that are frequent, distressing or accompanied by sudden sleep attacks are different and worth medical attention.
Does this mean dreams are not confined to REM sleep?
Correct, and that was already established. Dreaming occurs in non-REM sleep too. This study extends the point by finding dream-like cognition outside sleep altogether.
Why do I sometimes solve a problem overnight?
Consolidation during sleep is the usual explanation, and a brain capable of structured thought while asleep is consistent with it. Neither establishes that sleeping on a decision reliably improves it.
Can a sleep tracker tell me when I crossed over?
Not precisely. Consumer trackers infer stages from movement and heart rate and are approximate even about basic stages. A boundary this blurred is beyond what they measure.
Where this is heading
The useful shift is conceptual. Sleep looks less like a room you enter and more like a set of processes that arrive and depart on overlapping schedules — usually together, sometimes not.
That makes several ordinary experiences less strange: the hour that passes without you noticing, the answer that was not there last night, the night you would swear you never slept through.
Further reading: TIME on the new science of dreaming and PsyPost on the dream database and consciousness during sleep.
