Night Terrors: What to Do While It Is Happening

01What to do in the next sixty seconds

Stay in the room. You are not there to fix anything. You are there so that nothing in the room becomes a problem and so that somebody is present when the episode finishes. That is enough, and it is more than most people manage on the first night.

Do not try to wake them. This is the instruction every clinical source agrees on, and the reason is practical rather than mysterious. Their brain is partly in deep sleep and partly not, so shaking them does not deliver a person who suddenly understands what is going on. It delivers a more confused one. Cleveland Clinic's guidance is blunt about the consequence: "attempting to wake up someone may prolong a sleep terror episode." The clinical review by Alexander Leung and colleagues in *Current Pediatric Reviews* puts the pediatric version of the same point more sharply, warning that intervention "may confuse and frighten the child even more and can lead to paradoxical increase in aggression."

Move the room, not the person. This is the single most useful thing your hands can do. Push the nightstand back. Slide the lamp cord out of the way. Close the bedroom door if there are stairs beyond it. Clear a path, then stand in it.

Use the least contact that keeps them safe. Children's Hospital Colorado's home-care advice sets the bar exactly where it should be: "Hold your child only if it seems to help them feel better." If holding calms them, hold them. If they push you off and get more agitated, let go. Restraining somebody who is thrashing is where injuries happen, to them and to you. Cleveland Clinic names the same risk directly, noting that trying to hold down or stop somebody mid-episode "can be dangerous for both you and the person experiencing a night terror."

If they leave the bed, guide rather than intercept. Walking alongside somebody and steering them with a hand on the shoulder works. Standing in a doorway to stop them does not, because a body blocking the way is one more thing for a frightened, half-asleep person to push against. Children's Health's advice is to "gently direct them back to bed, if they sleepwalk or try to leave the bed," and gently is doing real work in that sentence.

Say very little, and say it the same way each time. Children's Hospital Colorado suggests "soothing comments such as, 'You are all right. You are home in your own bed. You can rest now,'" delivered calmly and repetitively. That is a good script precisely because it is boring. It asks nothing. It just puts a familiar voice in the room.

Do not ask them anything. Not what is wrong, not what they can see, not whether they know who you are. More on why below, because this is the one instruction that feels most unnatural and matters most.

Then let them go back to sleep. When it finishes, most people lie down and are asleep again within a minute. Do not wake them properly to check they are fine. Tuck them in and go.

02Why they look awake and cannot hear you

The reason none of this behaves the way you expect is that a night terror, or sleep terror, is not a bad dream that somebody is stuck inside. It is an incomplete arousal out of the deepest stage of non-REM sleep, which is why it lands in the first part of the night, usually one to three hours after they fell asleep, and why it looks so specifically wrong from a doorway.

Think of it as parts of the brain waking up at different speeds. The systems that produce movement and fear come online. The systems that make memories, weigh a situation and recognize a face stay asleep. The 2022 review of disorders of arousal by Yannis Idir, Delphine Oudiette and Isabelle Arnulf in the Journal of Sleep Research describes this as a dissociated pattern, "with a wake-like activation in motor and limbic regions and a preserved (or even increased) sleep intensity over a frontoparietal network." In a hallway at midnight that translates as open eyes, a racing heart, a face full of terror, and nobody home behind it.

So, the sitting up and the screaming are not evidence of what they are going through. They are evidence of which parts of the brain arrived first.

One short passage on what this is not, because you have probably already opened a page about it. A nightmare is a REM event, later in the night, and the sleeper wakes up and remembers the dream and wants you; our guide to what nightmares are and where they come from covers that comparison properly, and what to do after a nightmare in the first ten minutes is the page for the version where they are awake and talking to you. This is also not the same thing as physically acting out dreams while asleep, which happens in REM sleep in the last third of the night and is a different event with different implications. If what you are watching started early and they will not surface, you are almost certainly looking at the non-REM kind.

03Lights on, quiet voice, or say nothing at all?

If you have three tabs open, you have already noticed that the advice is not identical, and this is the point where a frightened parent starts to feel that nobody actually knows. The sources do agree on the big thing. They diverge on how much to do.

Set them side by side and the disagreement is smaller than it looks.

Here is what I would do, and why.

Light: enough to see the room, not enough to be an event. Children's Hospital Colorado's stated reason for turning the lights on is shadows, and that reason is sound. A half-lit room full of shapes gives a frightened, half-awake brain more raw material to work with, and it also makes it harder for you to see the corner of the bedframe you are about to walk into. But the full ceiling light in somebody's face is itself a jolt. A bedside lamp or a landing light does the whole job. This one is a judgment rather than a finding, and it is the one I hold most loosely.

Voice: talk, but say almost nothing. The NHS says not to talk to them at all, and Cleveland Clinic and Children's Hospital Colorado both say to speak calmly and quietly. Those are less contradictory than they read, because they are answering different questions. The NHS is warning against engaging, against reasoning, against trying to bring somebody round with conversation. Cleveland Clinic and Colorado are describing a low repeated murmur that carries no information at all. A parent saying "you're all right, you're in your bed" on a loop is not talking to them in the sense the NHS means. If in doubt, fewer words is the safe direction, and silence is never wrong.

Questions: none. This is where I would be firmest, and there is now real evidence for why, from a 2025 study in the journal *Sleep* by Yannis Idir, Régis Lopez and colleagues at the Paris sleep unit, which set out to test something everybody had assumed rather than measured: whether you can actually talk to somebody during one of these episodes. In their first study, 81 percent of 61 adult patients reported having occasional conversations during episodes. So, the person is not simply unreachable. But what came back is the important part. The researchers found that "patients' ongoing mental content influenced both their responses to questions during episodes and their perception of the outside world (including their surroundings and the identity of their interlocutor)."

Read that twice, because it explains the whole instruction. What you say does not arrive as information. It arrives as material, and it gets folded into whatever the person is already experiencing, including who they think you are. A soothing repeated phrase folds in harmlessly. A question does not: it demands an answer from somebody who cannot produce one, and it hands the scene a new character. Children's Health's version, "resist the urge to question them about what's going on," is the right instruction, and the urge really is strong. Resist it anyway.

04How long it lasts, and what the morning looks like

The published figures for duration are wider than you would expect, and knowing that is worth something at two in the morning when you are eleven minutes in and starting to panic.

The NHS puts episodes at up to 15 minutes. Cleveland Clinic says "between one and 30 minutes." Children's Hospital Colorado says 10 to 30. Leung and colleagues say a typical episode "usually lasts no more than a few minutes, but may be protracted, lasting up to an hour." Most episodes are short. Some are not, and a long one is still an ordinary one. If you are watching the clock and it has been twenty minutes, you are inside the range that experienced clinicians describe, not outside it.

Then comes the part that nobody warns you about, and it is the most reassuring fact available on this subject.

They will be fine in the morning and you will not be.

Most children have no memory of the episode whatsoever. Leung and colleagues describe straightforward retrograde amnesia for the event the following morning. That is not a small mercy, it is the whole shape of the thing: the person who screamed is untouched, and the person who watched is the one carrying it. You will come downstairs having barely slept, and they will want breakfast.

Which leads to the instruction that follows from it. Do not tell them about it. Children's Health's guidance is explicit that a child "will not remember having the night terror when they wake up the next day," and the sensible response is not to hand them a frightening story about themselves that they have no memory of and no way to check. If they ask, because they heard you talking about it, keep it short and unalarming. Nothing is gained by a debrief, and a child who learns that something scary happens to them in the night can start dreading bedtime, which makes tired sleep, which makes more episodes.

05If the person having them is an adult

Everything above was written for a parent, and almost every page you will find tonight was too. If you are here because your partner sat bolt upright and shouted at three in the morning, or because somebody told you at breakfast what you did, you have been reading a pediatric page for an hour and quietly concluding that you are unusual.

You are not. The Idir, Oudiette and Arnulf review states it plainly: "growing evidence has shown that DOA are not restricted to children but are also prevalent in adults (2%-4% of the adult population)." Leung and colleagues put the lifetime prevalence of sleep terrors at roughly 10 percent. These are common events that simply get written about as though only children have them.

Three things change for adults, and all three matter.

The room matters more. Cleveland Clinic notes that adults "may be more injury-prone than children, as many people suddenly get out of bed" during an episode. A child in a bed with a rail is contained. An adult is six feet tall, moving fast, and in a house with stairs. The practical version of this is dull and it works: clear the floor on both sides of the bed, keep the bedroom door shut, put anything glass out of reach, and if there is a habit of getting as far as the landing, deal with the stairs before you deal with anything else.

They probably will remember something, and that changes the conversation. This is the finding the pediatric pages leave out, and it comes from a 2021 study in *Nature and Science of Sleep* by Anna Castelnovo, Giuseppe Loddo, Federica Provini and colleagues that interviewed 25 adults and 20 children referred for these disorders and asked the same questions of both. The result inverts cleanly. "More than two-third of the children (n = 14) could not recall any mental activity associated with their episodes, whereas more than two-third (n = 16) of the adults recalled at least one mental experience." The authors' conclusion is that the diagnostic criteria themselves "are tailored around the typical presentation of DOA in children, and do not always fit to adult patients."

So, "they won't remember a thing" is a children's fact. Telling an adult that, when they are sitting there with a vivid fragment they cannot shake, is not reassurance. It is an invitation to think something is wrong with them.

What adults remember is a single frightening image, and it explains the behavior. A 2009 study in *Sleep* by Delphine Oudiette, Isabelle Arnulf and colleagues interviewed 43 adults referred for severe sleepwalking and sleep terrors, average age 26. Seventy-one percent reported at least one dreamlike mentation attached to an episode. Almost all of them were a single visual scene rather than a story, and 84 percent were negatively perceived, "with apprehension, fear, or terror." The examples in the paper are ordinary and awful in the way real ones are: the ceiling was collapsing, she was going to get run over by a truck, his girlfriend was in danger. And the detail that matters most to the person who woke up ashamed: "the dreamlike mentation action corresponded with the observed behavior."

Put that beside the Idir finding that ongoing mental content shapes the perceived identity of the person in front of you, and the two most distressing adult experiences stop being mysterious. You shoved somebody because something was falling. You did not know who they were because the part of you that recognizes faces was still asleep and the part that was awake was busy with a ceiling.

One person described the aftermath better than any clinical summary I found, in a post about her own night terror and her husband: "it's so funny to me that I did such a mean horrible thing and was like, proud of it? But also obviously still feel horrible. Night terror me is awful, and I'm not sure who I thought he was... my husband feels safe (I hope!!!)". The parenthesis is the register of this whole subject for adults. It is not fear of illness. It is shame about what you did, and worry about whether the person beside you is all right.

Both of those deserve a straight answer. The behavior was not a choice and it was not a window into what you secretly think of anybody, and where the episodes are new rather than lifelong, what changed to start them is the question worth asking next, because the machinery that makes choices and holds opinions was offline. And the person beside you is allowed to have been frightened. Those two facts sit together fine. The conversation that goes badly is the one where the person who was asleep needs to be told it was nothing, and the one who was awake needs to say that it was not. Let it be both.

06What makes another one likelier, and the one thing that reliably helps

Nothing on this page can promise you a quiet night. What it can do is name the things that stack the odds, because most of them are unglamorous and several of them are fixable.

Sleep debt is the big one. Deep non-REM sleep rebounds after a short night, which means a tired week produces a heavier, deeper first third of the night and more of exactly the sleep these episodes arise out of. Leung and colleagues make the practical point directly: because sleep deprivation predisposes to sleep terrors, good sleep hygiene and an appropriate sleeping environment matter. If your child has been going to bed an hour late all week, an earlier bedtime is not a fussy suggestion, it is the highest-yield thing available to you.

The rest of the list from the same review is worth reading once and then mostly forgetting: a fever or intercurrent illness, a full bladder, a noisy room, unusual physical exhaustion, emotional stress, and untreated breathing problems in sleep such as obstructive sleep apnea. Cleveland Clinic's list overlaps almost entirely and adds alcohol and caffeine for adults. Notice the shape of it. Almost everything on there either deepens early sleep or fragments it, which is the same mechanism from two directions.

Then there is the one technique with real standing, and it surfaces surprisingly rarely. If the episodes happen at roughly the same time each night, you can pre-empt them by waking the person briefly before that time, for a couple of weeks. Leung and colleagues call it anticipatory awakening, "performed approximately half an hour before the child is most likely to experience a sleep terror episode," and describe it as often effective for frequently occurring sleep terrors. The NHS gives a shorter interval for the same technique: "if they're having a night terror at the same time every night, try waking them up 15 minutes before the night terror every night for 2 weeks. This may stop the night terrors happening."

Two things are worth saying about that. The published intervals differ, 15 minutes against 30, which tells you the precision here is not high; pick one and be consistent. And it only works if there is a pattern to pre-empt, which means the boring part comes first. For a week, write down the time they fell asleep and the time the episode started. That is the whole preparation, and it is also the single most useful thing you can hand a doctor.

07What a doctor would want to know

Most of these never need an appointment. A handful do, and the difference is a pattern rather than a bad night.

The published thresholds are not identical, which is useful to know before you decide you are overreacting or underreacting. The NHS suggests seeing a GP if a child is still having night terrors after the age of 12, or starts having them after the age of 5. Cleveland Clinic sets it on frequency instead, suggesting a conversation if episodes happen a few times a week, if sleep quality is suffering, or if somebody is getting hurt. Between them they cover the honest answer, which is that the trigger for a phone call is disruption or injury, not fear.

A few specific things are worth mentioning to a clinician if they apply, and they are worth framing as things a doctor finds informative rather than as symptoms of anything. Cleveland Clinic lists several: episodes happening repeatedly in a single night, episodes in the second half of the night rather than the first, episodes that are unusually brief, and loud snoring, mouth breathing or daytime sleepiness alongside them. None of those means something is wrong. They are simply the details that help somebody qualified tell one nighttime event from another, and a page cannot do that job. Daytime sleepiness in particular is worth raising in adults: in a case-control study of 100 adult sleepwalkers by Régis Lopez, Yves Dauvilliers and colleagues in *Sleep*, 42.2 percent reported excessive daytime sleepiness against 11 percent of matched controls, and 63.6 percent of that group reported night terrors too. That is a specialist clinic population rather than the general public, so it describes people whose episodes were bad enough to be referred, but it does say something real: in adults, these events are more often part of a wider picture of unrefreshing sleep, and that picture is treatable.

Go straight to urgent care for the things you would go for anyway: a serious injury, difficulty breathing, or an episode that does not look like the pattern you know. An ordinary night terror is frightening to watch and is not a medical emergency, and one of the quiet costs of a frightening subject is that people either overreact for years or dismiss something that genuinely warranted a look.

One more thing an adult should raise, plainly and without drama. Cleveland Clinic's page states that night terrors in adults "usually point to an underlying mental health condition, like post-traumatic stress disorder or anxiety disorder." You are likely to meet that sentence, so it is better to meet it here with some context. Take it as the reason a clinician will ask about stress, about sleep, and about anything difficult you have been through, rather than as a verdict delivered by a website. What is established is narrower and more useful: episodes cluster around things that deepen or fragment the first part of the night, most adult episodes have a findable precipitant, and being asked about your life is not being diagnosed with anything.

08Common questions about night terrors

Should I wake them up?

No, and the reason is practical rather than mystical. Their brain is partly in deep sleep, so waking is difficult and what you get is a more confused, more agitated person. Cleveland Clinic notes that attempting it may prolong the episode. Stay close, keep the space safe, and let it finish on its own.

Can I hold them?

Only if it helps. Children's Hospital Colorado's guidance is to hold your child only if it seems to make them feel better. If they push you away or get more agitated, stop. Restraining somebody who is thrashing risks injury to both of you, and it is the most common way a frightening but harmless event turns into an actual accident.

How long does a night terror last?

Usually a few minutes. Published ranges vary from up to 15 minutes to up to 30, and a clinical review notes that episodes can occasionally run as long as an hour. A long episode is still an ordinary one. If it is dragging on, keep doing the same thing: stay, keep the area clear, say very little.

Will they remember it in the morning?

A child almost certainly will not. Adults usually will remember something, often a single frightening image rather than a story, which is a genuine difference between the two and not a sign that anything is wrong. If a child does not remember, there is nothing to be gained by telling them.

Should I talk about it the next day?

With a child, no. They have no memory of it and no way to make sense of the account, and giving them one can make bedtime something to dread. With an adult partner, yes, but keep it short and factual, and let both people say their piece: the one who was asleep did not choose any of it, and the one who was awake is allowed to have found it frightening.

Can adults have night terrors?

Yes. A review of disorders of arousal puts the adult figure at around 2 to 4 percent of the population, and a clinical review puts the lifetime prevalence near 10 percent. Adults are more likely to leave the bed and more likely to be injured doing it, which makes the state of the bedroom floor the practical priority.

Is it a night terror or a nightmare?

The timing usually settles it. Night terrors come out of deep non-REM sleep in the first part of the night, and the person cannot be reached and typically remembers nothing. A nightmare happens in REM sleep later on, and the person wakes up properly, remembers the dream, and wants comfort. Our guide to nightmares and where they come from covers the comparison in full.

Could it be a seizure?

It is not something this page or any page can settle, and it is a reasonable question rather than a catastrophic one. The details that help a clinician tell them apart are the ones worth writing down: what time it happened relative to falling asleep, how long it lasted, whether the movements were repetitive and similar each time, and whether it happened more than once in a night. Bring the notes, not a theory.

Does a night terror mean my child is stressed or traumatized?

Not by itself. These are events of sleep architecture rather than events of the emotional life, they run strongly in families, and the most reliable trigger is simply not enough sleep. Stress is on the list of things that can make them likelier, along with fever, noise and a full bladder, which gives you a sense of the company it keeps.

If you came here mid-episode and you are reading this afterward with the house quiet again, the thing to take away is that you almost certainly did it right, because the correct response is mostly restraint. DreamTold exists to help you think about what happens in the night without turning every strange one into a symptom, and if an image from your own rough night has stayed with you, you can look it up in the dream dictionary and let it be a dream again. This guide is educational and meant for reflection, not medical advice, and it cannot diagnose anything. If episodes are frequent, causing injury, or new in an adult, speak with a qualified clinician.

09Sources

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

• Children's Health: Resist the urge to question them about what's going on
• Children's Hospital Colorado: Turn on the lights so that your child is less confused by shadows
• Cleveland Clinic: attempting to wake up someone may prolong a sleep terror episode
• NHS: try waking them up 15 minutes before the night terror every night for 2 weeks
• PubMed: Yannis Idir, Delphine Oudiette and Isabelle Arnulf in the Journal of Sleep Research
• PubMed: Yannis Idir, Régis Lopez and colleagues at the Paris sleep unit
• PubMed: Anna Castelnovo, Giuseppe Loddo, Federica Provini and colleagues
• PubMed Central: Alexander Leung and colleagues in Current Pediatric Reviews
• PubMed Central: Delphine Oudiette, Isabelle Arnulf and colleagues
• PubMed Central: Régis Lopez, Yves Dauvilliers and colleagues in Sleep
• tiktok.com: a post about her own night terror and her husband