Hypnagogic States: The Images Just Before Sleep

01First, the question you actually came with

Almost nobody arrives at this subject wanting a mechanism. They want to know whether this happens to other people.

It does, in numbers large enough that the surveys are the reassuring part. In 1996, Ohayon and colleagues interviewed a community sample of 4,972 people across the UK, aged fifteen to a hundred, and found that 37 percent reported hypnagogic experiences and 12.5 percent reported them on waking instead. Their own conclusion was that these were "much more common than expected," far more common than the sleep disorders they had been assumed to belong to.

Four years later, Ohayon ran a larger survey across the UK, Germany and Italy with 13,057 participants and found 24.8 percent reporting sleep-onset experiences. That paper also carries the line most worth reading twice: in more than half of the cases, these experiences had no relationship to any specific pathology at all.

The numbers move around a lot, which is worth knowing so a low figure somewhere else does not frighten you. Romain Ghibellini and Beat Meier's 2023 review of hypnagogia research lays the surveys side by side and the spread runs from 6 percent to 85 percent. What changed between those studies was the question, not human beings. Ask whether someone has ever had a distressing or threatening illusion at the edge of sleep and you get a small number. Read them a list that includes hearing their own name called, or shapes and colors drifting past, and the number climbs steeply, because that list describes something most people recognize.

That last row is the sentence to hold on to. It is not a blogger's reassurance; it is how a 2024 review written for sleep clinicians summarizes the evidence.

02About the word hallucination

There is no gentle way around the vocabulary, so it is better met head on. The clinical term is hypnagogic hallucinations, and for most people that word arrives attached to serious mental illness. The association does real harm here, because it is not what the word means in sleep medicine.

In this context a hallucination is a perception with no outside source. That is the whole definition. It describes where the image came from, not the person having it. By that definition everything you see in an ordinary dream is a hallucination too, and nobody thinks twice about dreaming. Sleep-onset imagery is the same process catching you a few seconds early.

What sleep medicine treats as clinically interesting is much narrower, and it is not the image. It is the surrounding picture: daytime sleepiness, experiences while fully awake, distress bad enough to interfere with sleeping at all. Those are covered further down, and they are specific.

03What people see, and what people hear

The catalog is more consistent than you would expect from something so private. Visual experiences lead, and within them faces are among the most common single things people report: a stranger's face, a face made mostly of eyes, several faces at once, a face that changes as you look at it. After that come colors, geometric patterns, shapes drifting, and short scenes, often of somewhere you cannot place. In a survey of 492 people, summarized in the Ghibellini and Meier review, visual experiences, the sensation of falling, and the feeling that somebody is present in the room came out as the three most frequently reported, with sounds, touch and movement behind them, and smell and taste rarest of all.

The falling sensation deserves a note, because it is so ordinary that people rarely connect it to the rest of this. The lurch that jerks you awake as you drop off belongs to the same borderland, which sits right beside the spiritual meanings people attach to falling in dreams without being the same thing.

The auditory side is where the reports get eerily specific. Hearing your own name called is common enough that it appears as a checklist item in a validated research questionnaire, which is a strange kind of comfort: whole studies have been built on the assumption that enough people will say yes. The same review summarizes work comparing hypnagogic speech with dream speech, and the differences are exactly the ones that make it unnerving. What people hear at sleep onset is more often a single clear word than a sentence, more often a familiar voice than a stranger's, and more often addressed directly to them. A voice that says your name, in a voice you know, is close to the standard form. If yours arrived as a sound rather than a voice, a bell or a knock or a bang, hearing a doorbell as you fall asleep covers that version.

Here is the part worth sitting with. Some of the most exact descriptions of this experience come from people who read it spiritually. A medium posting under the name wendy mccallum medium describes it as "faces, eyes, people you don't recognise, colours, shapes or flashes of images appearing just as you're drifting off to sleep". The host of a paranormal podcast, describing the more overwhelming version, says that sometimes "it's more like 50 faces coming towards you, or a black and white scene of somewhere you've never been".

Read those descriptions against the research literature and they match it closely. Faces, eyes, colors, unfamiliar people, a scene from nowhere. These are not distorted accounts of the experience. They are careful ones. What differs is not the description but the attribution, which is a genuinely separate question and gets its own section below.

04Why it happens right there

Sleep does not switch on. It arrives in stages, and the first of them, called N1, is a short window in which parts of the brain have started behaving as though you are asleep while others have not caught up. Imagery is what that mismatch looks like from the inside.

Kjaer and colleagues put eight people through PET scans while they drifted, comparing awake rest with stage-1 sleep, and found relative blood flow rising in the occipital lobes, where visual processing happens, while it fell in the thalamus, the posterior parietal cortex, the premotor cortex and the cerebellum. In plain terms, the visual system was busy while the systems for orienting yourself and preparing to act had begun to stand down. Their interview data matched: thoughts in stage 1 became markedly more unrealistic and more leaping. The authors described stage-1 sleep as "the dreaming state of wakefulness" as against REM sleep, which they called the dreaming state of the sleeping brain. Two caveats belong with that study. Eight people is a small sample, and the researchers were careful to say that the hypnagogic experiences themselves did not line up with any flow change they could measure. What they showed is the state, not the picture inside it.

Somebody explaining this on Instagram put the same idea more briefly than any of the papers manage, writing that hypnagogic hallucinations "occur during the transition between wakefulness and sleep, when parts of the brain responsible for dreaming activate before the body fully shuts down". That is a fair summary of what the imaging suggests.

The second half of the answer is about content, and it is one of the more charming results in sleep science. In 2000, Stickgold and colleagues had people play Tetris and watched what turned up at sleep onset. Falling blocks, over and over, reported as players drifted off. The striking part was the three amnesic participants with extensive damage to the medial temporal lobes: they reported the same falling-block imagery despite being unable to remember playing at all. Some players also saw versions of Tetris they had played years earlier. So, the imagery draws on recent experience and on old memory, without needing the part of the brain that would let you consciously recall any of it.

That explains a good deal about why the content feels both familiar and unplaceable. A face at the edge of sleep is assembled from a face-processing system that has spent all day looking at people, running without the usual supervision. A street you have never walked down may be several streets you have.

05Is this sleep paralysis? Is it a dream?

These experiences sit in a crowded neighborhood, and telling them apart takes most of the fear out of the question.

The clearest dividing line is movement. If images were arriving and you could have rolled over, that is the hypnagogic version. If you were awake, saw your actual room, and could not move or speak, that is sleep paralysis, and our full explainer on why sleep paralysis happens covers the mechanism and the sensed presence that so often comes with it. The two are relatives rather than the same event, and plenty of people have one for years and never the other.

If what has changed for you lately is not the borderland but the dreams themselves, running louder and stranger and staying with you into the morning, that is a different question with its own answer in our guide to why dreams get more vivid.

06When it is not charming

Plenty of writing on this subject treats hypnagogia as a delight, a private art gallery you get free every night. For some people it is exactly that. For others it is not, and the charming version is no use to them.

Someone posting about living with it wrote about the mental rollercoaster of it, that "you could be in such a good mood before you go to sleep and when you wake up you couldn't feel more different", and described the particular frustration of it: you can do everything within your power while conscious to steady your mood, and then be undone by something you have no control over. That is a fair account of the hard version, and it is not solved by being told the experience is common.

The research points the same way about which experiences matter. In Ohayon's three-country survey, the frightening sleep-onset experiences were the ones more often found alongside something else, such as narcolepsy, obstructive sleep apnea, or an anxiety disorder. Neutral or interesting imagery sat in the majority with no link to a specific condition. The useful question, then, is not what the image was but how it landed and what else is going on around it.

One more thing needs scoping properly, because it gets misused. There is a body of research on perception during very long stretches without sleep, and Waters and colleagues reviewed 21 such studies in 2018. Perceptual changes turned up almost universally, progressing over days into something resembling acute psychosis. Those studies involved an average of 72 to 92 hours awake, up to eleven nights at the extreme. A short night or a stressful week is not that, and nothing in that literature suggests otherwise. It is worth knowing mainly because it shows how far away that outer edge sits.

07When to talk to a doctor

None of this is a reason to panic, and all of it is easy to mention at an appointment you were having anyway.

The signals that matter are not about the images. They are about sleep and about wakefulness. Raise it with a clinician if you have overwhelming daytime sleepiness that a full night does not fix. The National Institute of Neurological Disorders and Stroke describes that pattern as sleepiness that arrives like a sleep attack, an overwhelming wave that comes on quickly, and which does not improve even after enough sleep at night. Raise it if you get sudden episodes of muscle weakness triggered by strong emotion such as laughter or surprise, which is called cataplexy and is worth assessing properly. Raise it if the experiences began or changed sharply after starting or changing a medication.

There is one more distinction that carries genuine clinical weight, and it is the awake one. Ohayon's survey separated experiences at the edge of sleep from those reported during the day, and the daytime visual and auditory ones were the ones strongly associated with other conditions. So, if you are seeing or hearing things in the middle of an ordinary afternoon, fully awake, that is a different conversation from a face at the edge of sleep and it deserves a proper one.

And distress counts on its own. If this has you afraid to go to bed, that is a legitimate reason to ask for help, whatever the images turn out to be. Nobody needs to earn a clinical conversation by having the right symptom.

08What tends to help

There is no reliable way to stop hypnagogic imagery, and any page promising one is overreaching. What there is, is a short list of things that make the borderland less ragged and the episodes easier to sit through.

Protect the schedule first. Sleep loss and irregular hours make the boundary between waking and sleeping less clean, and that boundary is where all of this lives. Roughly the same bedtime and wake time, with enough hours inside them, is dull advice that moves the actual lever.

Inside an episode, knowing what it is takes most of the power out of it. These are seconds long. You can move. Opening your eyes usually ends it.

The person quoted above, who lives with the difficult version, listed what works for her, and it is more practical than most clinical advice on the subject. Turn a light on, which will usually stop the imagery. Ground yourself in something unmistakably real, music or the news or your phone, as confirmation of where you are. Leave the room for five minutes if you need to settle. Then do whatever it takes to be present for the rest of the day, including writing the thing down if that helps.

And if the image stayed with you and you would rather think about it than shake it off, write it down before the coffee, in three words if that is all you have.

09How people read it

Everything in this section is belief and interpretation, described as such. It is here because a great many people who have these experiences reach for a meaning, and pretending otherwise would be dishonest about the subject.

The most visible reading right now is the clairvoyant one, offered by mediums and psychic-development teachers, and their answer is that this is a sense opening rather than a symptom. One creator puts it flatly: seeing faces as you fall asleep might mean "you might be psychic," and could be "clairvoyance trying to connect". The medium quoted earlier is more careful about her own framing, writing that "from a spiritual perspective, some people interpret this as their clairvoyance" becoming more noticeable when the mind is quiet, which is a claim about interpretation rather than about anatomy.

Older traditions did not have a category for sleep-onset imagery, but several treated the borderland around sleep as a place where meaning could arrive. In the Hebrew Bible, Numbers 12:6 has God say of prophets, "I the Lord make myself known to them in visions; I speak to them in dreams." A widely transmitted hadith recorded in Sahih Muslim 2263a sorts dreams into three kinds: good tidings from Allah, distressing ones attributed to Satan, and a third category described as "a suggestion of one's own mind." That third category is a striking piece of restraint, and it is worth noting that the tradition itself declines to treat every night image as a message.

There is a third reading, older than the clinical vocabulary and quieter than either of the others, which treats the state as a workshop. The word hypnagogic was coined by Alfred Maury in 1848, from the Greek for sleep and for a conductor or leader, and Frederic Myers named the waking counterpart hypnopompic in 1904, a history the Ghibellini and Meier review sets out. Long before either name, artists and inventors were deliberately fishing in that window. Thomas Edison is said to have napped holding metal spheres so that the clatter of them dropping would wake him at the moment the ideas started.

That story is usually told as an anecdote, and in 2021 somebody tested it. Lacaux and colleagues gave 103 participants a set of math problems containing a hidden shortcut, then gave them a twenty minute break holding an object, monitored with full polysomnography. Spending at least fifteen seconds in N1, the same sleep stage the imagery belongs to, tripled the chance of discovering the hidden rule, 83 percent against 30 percent for those who stayed awake. The effect disappeared in those who fell into deeper sleep. Their conclusion was that there is a creative sweet spot inside sleep onset, and that catching it means falling asleep easily without falling asleep too far.

So, the sleep-science account and the spiritual account are describing the same few seconds and reaching for different words. Sleep research can say with real confidence where in the night this happens, how common it is, and what the brain is doing while it does. What it does not have is an instrument that measures meaning. There is no study that can tell you whether a particular face carried something, and nothing in this literature predicts anything about your future or anyone else's. Whether the image mattered to you is a question the research was never designed to answer, and it stays yours. Neither of these accounts requires the other to be wrong.

10Common questions about hypnagogic images

Is it normal to see faces as I fall asleep?

Yes, in the sense that matters: it is common, well documented, and happens to a large share of people who have nothing wrong with them. Population surveys have put the figure anywhere from a quarter to over a third, and one clinical review states that these experiences, occurring with no other symptoms and within normal sleep, are most likely non-pathological.

Why do I hear my name being called?

Hearing your own name at sleep onset is one of the standard forms this takes, common enough that it appears as an item on a research questionnaire. Survey work comparing sleep-onset speech with dream speech found it is more often a single clear word than a sentence, more often a familiar voice than a stranger's, and more often addressed directly to the listener. That combination is exactly what makes it feel like somebody really called you.

Is this sleep paralysis?

Not necessarily, and usually not. Sleep paralysis means being awake and aware in your real room and briefly unable to move or speak. Hypnagogic imagery can happen without any paralysis at all, and most of the time it does. The two are related, they share a borderland, and they can occur together, but having one does not mean you have the other.

Does this mean I have narcolepsy?

Not on its own, and the research is clear on that point. The 1996 UK survey found these experiences far too common to be explained by narcolepsy, which affects a tiny fraction of the population. What makes narcolepsy worth investigating is the surrounding pattern, especially overwhelming daytime sleepiness and sudden muscle weakness triggered by strong emotion. Imagery without those is not a diagnosis of anything.

Can stress or a bad week make it worse?

It can. Disrupted and insufficient sleep make the boundary between waking and sleeping less clean, and that boundary is where all of this takes place. Researchers also found the frightening versions specifically were more often reported alongside anxiety and sleep disorders, while neutral imagery mostly sat on its own. If yours has ramped up lately, the sleep you have been getting is a more useful place to look than the images.

Is it dangerous?

An isolated image at the edge of sleep is not dangerous, and it does not damage anything. The genuine harm in this territory is distress, which is real and worth addressing. If it has you dreading bedtime, that alone is a good reason to raise it with a clinician, regardless of what the images turn out to be.

Does it mean something spiritually?

Many people hold that it does, and mediums and psychic-development teachers commonly read it as a form of clear seeing opening up. That is a belief, held sincerely, and this page is not in the business of adjudicating it. What can be said is what sleep research has measured: when it happens, how common it is, and what the brain is doing at the time. Research has no instrument for meaning and it predicts nothing about anyone's future. Holding one account does not oblige you to drop the other.

If you came here worried, the short version is that you found something with a name, a location in the night, and decades of survey data behind it, and that an occasional face or voice at the edge of sleep is not evidence that anything is wrong with you. If you came here because it fascinates you, nothing above takes that away either. DreamTold treats what your nights hand you as a window into your own mind rather than a code to be cracked, so if a particular face or place has stayed with you, look up what it brings to mind in the dream dictionary. This page is educational and meant for reflection, not medical advice, and it cannot diagnose anything. If daytime sleepiness, sudden muscle weakness, waking hallucinations or fear of sleep is affecting your life, speak with a qualified clinician.

11Sources

Every claim on this page that came from somewhere else, with the somewhere else, so you can read it in context rather than trusting a summary.

• BibleGateway: Numbers 12:6
• Instagram: hypnagogic hallucinations occur during the transition between wakefulness and sleep, and separately the mental rollercoaster of hypnagogic hallucinations
• National Institute of Neurological Disorders and Stroke: narcolepsy, excessive daytime sleepiness and cataplexy
• PubMed Central: Romain Ghibellini and Beat Meier's review of the hypnagogic state
• PubMed: Ohayon and colleagues on hypnagogic and hypnopompic hallucinations in 4,972 people
• PubMed: Ohayon on the prevalence of hallucinations in the general population
• PubMed: Waters and colleagues on sleep-related hallucinations
• PubMed: Waters and colleagues on severe sleep deprivation and perception
• PubMed: Kjaer and colleagues on regional cerebral blood flow during light sleep
• PubMed: Stickgold and colleagues on hypnagogic images in normals and amnesics
• PubMed: Lacaux and colleagues on sleep onset as a creative sweet spot
• Sunnah.com: Sahih Muslim 2263a, the three kinds of dreams
• tiktok.com: faces, eyes, people you don't recognise, colours, sometimes it's more like 50 faces coming towards you and you might be psychic, it could be clairvoyance trying to connect