Parasomnias are the things that happen during sleep that shouldn't: walking, talking, screaming, eating, thrashing, acting out a dream. Most are harmless and most people who have them never need a clinic.
But they split into two families with completely different implications, and the single most useful piece of information is when in the night the episode happens. Get that right and you know which conversation you're having.
The split: deep-sleep events versus dream-sleep events
Sleep isn't uniform — it cycles between non-REM stages and REM through the night, and what happens in each stage is genuinely different. That matters here because the two parasomnia families arise from the two different states, and the states are distributed unevenly across the night.
Deep non-REM sleep is concentrated in the first third of the night. REM sleep gets longer and more prominent in the second half. So:
| NREM parasomnias (disorders of arousal) | REM sleep behaviour disorder | |
|---|---|---|
| When | Usually first third of the night | Usually second half, early morning |
| Memory of it | None — amnesia is characteristic | Recalls a dream matching the behaviour |
| Behaviour | Complex but clumsy — walking, fumbling, confused | Enacting a dream — punching, kicking, shouting, often defensive |
| On waking | Confused, disoriented, hard to rouse | Alert quickly, can describe the dream |
| Typical age | Commonest in children; can persist | Typically older adults, more often men |
| What it means | Usually benign; often has a treatable trigger | Needs proper assessment — see below |
The NREM family: sleepwalking, sleep terrors, confusional arousals
These are collectively called disorders of arousal, and the name explains the mechanism: the brain partially exits deep sleep without fully waking. Motor systems come online; the parts that make memories and judgements don't.
That explains the presentation. Eyes are often open, and people describe the person as looking straight through them. They can walk, unlock doors, move objects, hold a fragment of conversation — and remember none of it. They're difficult to rouse, and confused and irritable if you manage it.
They're common in childhood, usually outgrown, and run strongly in families — so a parent who sleepwalked makes a sleepwalking child unremarkable.
Sleep terrors are not nightmares
Worth separating clearly, because parents and partners routinely conflate them and the handling differs.
A sleep terror comes out of deep sleep early in the night. There's often a scream, obvious terror, a racing heart, sweating, and the person is inconsolable and not really present. Trying to comfort them frequently makes it worse. There's no dream to report and no memory in the morning. It looks far more frightening to the observer than it is to the sleeper — who isn't experiencing it as fear in any rememberable way.
A nightmare comes out of REM sleep, usually later in the night. The person wakes properly, is oriented, can describe the dream in detail, and can be comforted — and remembers it the next day.
Practically: for a sleep terror, don't try to wake them. Keep them safe, stay nearby, and let it pass. For a nightmare, ordinary reassurance is exactly right.
Two more that people don't mention
Sleep-related eating disorder — eating during partial arousals with no memory of it, sometimes with odd or inedible combinations. It can cause real weight and dental consequences, and it's associated with some sedative-hypnotic medications.
Sexsomnia — sexual behaviour during sleep, with amnesia. It exists, it is recognised, and it carries relationship and occasionally legal consequences serious enough that it warrants a sleep clinic rather than embarrassment. It is also one of the most under-reported things in sleep medicine, for obvious reasons.
The treatable drivers — this is the actionable part
Disorders of arousal need two things: a predisposition, and something that night that primes or fragments deep sleep. You can't change the predisposition. You can usually do something about the rest, and this is where most of the available improvement sits.
- Sleep deprivation. The most reliable provoker. Insufficient sleep produces a rebound with deeper, more consolidated deep sleep — more raw material for a partial arousal. If episodes cluster after short nights or a disrupted schedule, that's the pattern. Start with whether you're spending enough time in bed.
- Alcohol. Fragments the second half of the night and is a common trigger — what a nightcap actually does is worth reading if you've assumed it helps.
- Untreated sleep apnea. The most important item on this list, because it's common, invisible and treatable. Apnea generates hundreds of arousals a night, and arousals are precisely what triggers these events. Treating the apnea often resolves the parasomnia. If there's snoring, witnessed pauses or unrefreshing sleep alongside the episodes, screen for it.
- Restless legs and periodic limb movements — same logic: another source of repeated arousals. How to recognise restless legs, and note that the first step there is iron studies rather than a sleep study.
- Medication. Sedative-hypnotics, the Z-drugs in particular, are associated with complex sleep behaviours — sleepwalking, sleep-driving, sleep-eating — prominently enough to carry regulatory warnings. If episodes started after a new sleep medication, that's the first thing to raise. Don't stop a prescription on the strength of a web page; take the list to the prescriber.
- Fever, stress, and an irregular schedule — including shift work, where the sleep debt and the timing both work against you. The countermeasures are here.
Safety first, and it's mostly boring
For anyone who walks: lock external doors and windows and put the key somewhere not obvious. A gate at the top of the stairs. Nothing sharp or hot reachable. Move car keys. Ground-floor bedroom if it's frequent. If there's a bed partner being hurt, separate beds until it's assessed — that isn't a failure, it's just sensible.
REM sleep behaviour disorder: the one to take seriously
This one is different in kind, not degree.
Normally your muscles are actively paralysed during REM sleep — which is what stops you acting out dreams. In REM sleep behaviour disorder that paralysis fails, and the dream gets physically enacted. Typically in the second half of the night, often violently: punching, kicking, leaping out of bed, shouting. Injuries to the sleeper and to the bed partner are common, and the partner is usually the one who raises it.
The distinguishing feature is that the person wakes and can describe a dream that matches what they were doing — often a dream of being attacked or defending someone.
Why it matters clinically: REM sleep behaviour disorder is strongly associated with later development of Parkinson's disease, dementia with Lewy bodies or a related condition. It can precede any other symptom by years. That is not a reason to avoid finding out — it is the reason to get it properly diagnosed, because it changes what a neurologist monitors, and because this group is exactly who research and early-intervention trials are looking for.
We have written up what the long-term follow-up studies actually show, with the numbers. That piece is with a clinician for review before we publish it, because a prognostic statistic about neurodegenerative disease is not something a supplement company should put out unchecked. We'd rather say that plainly than either publish it unreviewed or pretend the association doesn't exist.
Two important caveats in the meantime:
Thrashing in bed is not the same as having this diagnosis. Several things mimic it — most notably untreated sleep apnea, where the arousals produce movements and confusion that can look very similar. Distinguishing real REM sleep behaviour disorder requires an overnight sleep study demonstrating the loss of REM muscle paralysis. It cannot be diagnosed from a description, and it should not be self-diagnosed from an article.
And it can be drug-induced. Some antidepressants can produce REM-sleep behaviour changes, which is a different situation with a different meaning. Another reason the medication list matters.
When to see someone
Most childhood sleepwalking needs reassurance and a stair gate. Book an appointment if:
- Anyone is getting injured — the sleeper or the partner
- The behaviour is violent, or involves leaving the house or driving
- It started in adulthood, particularly after 50, and especially if dreams are being enacted
- Episodes are frequent, or there's significant daytime sleepiness alongside
- Episodes are brief, stereotyped and repeat several times a night — that pattern raises the question of nocturnal seizures rather than a parasomnia, and the two are genuinely hard to separate without testing
- A new medication coincided with onset
Where we sit
No supplement treats a parasomnia, and nothing in our range does. There's a specific reason we want to be blunt about that here.
The instinct when something alarming happens at night is to reach for a sleep product. In parasomnias that can be exactly wrong: sedating someone whose problem is a partial arousal out of deep sleep does not obviously help, and some sedative-hypnotics are themselves associated with these behaviours. The interventions that work are fixing the sleep debt, treating an underlying apnea or limb-movement disorder, reviewing medication, and making the bedroom safe — none of which we sell.
If what you actually have is insomnia, CBT-I has better evidence than anything you can buy. And if you've noticed unusually vivid dreams since starting a supplement rather than any behaviour during sleep, that's a separate and much less worrying question.
This article is not medical advice and cannot diagnose a parasomnia. REM sleep behaviour disorder in particular requires an overnight sleep study and specialist assessment. Nothing here should be used to start, stop or change a prescribed medication.