Sleep Masks, White Noise and Blackout: What Changes Your Dreams

01The short version, before you buy a fourth thing

You have probably done this in order. The fan first, because it was already there. Then a rain app, because the fan was too cold in winter. Then blackout curtains, or a mask, because the streetlight was the real problem all along. And somewhere in that sequence your dreams changed, or you think they did, and now you cannot tell whether you fixed something or broke something.

That confusion is built into how this is sold. A sound machine is marketed on sleep, reviewed on sleep, and then quietly credited with anything that happens after dark, dreams included. The two claims sit in the same paragraph on almost every product page you will read, and they carry wildly different amounts of evidence.

So, here is the split. There is a modest, poor quality, genuinely mixed body of evidence that continuous broadband sound affects how you sleep. There is essentially no evidence that it affects what you dream about. And because better or worse sleep changes how many dreams survive to morning, a change in your dream life is exactly what you would expect from a change in your sleep, even if the dreams themselves are untouched.

02What the noise and dreams literature actually contains

The best single place to look is a 2024 systematic review in Sleep Medicine Reviews by Leila Salvesen, Elena Capriglia, Martin Dresler and Giulio Bernardi, Influencing dreams through sensory stimulation. It gathered every study it could find in which researchers deliberately fed a sensation to a sleeping person and then asked what they had dreamed. Fifty one publications, across sound, touch, smell, sight, movement and combinations of those. Twenty one were auditory, and white noise appears among the stimuli used.

Two things about that review matter to you, and they pull in opposite directions.

The first is that stimulation during sleep can genuinely turn up in a dream. The review gives the example of white noise being incorporated not as static but as the sound of waves, or even as a visual scene of the ocean, which is a fair description of something a lot of people have experienced with a fan. The brain takes the input and dresses it.

The second is how unreliably that happens. Across the studies, the reported frequency of stimulus dependent dream changes ranged from roughly zero percent to about eighty percent, which the authors put down to heterogeneity in definitions and methods. Plain pure tones had the lowest incorporation rate of all, around nine percent, against roughly twenty three percent for visual stimuli and forty two percent for touch. A sound with personal relevance got in; a featureless hiss mostly did not. Average sample size was about seventeen people, and the authors rate the field's methodological quality at thirteen out of twenty five, concluding that a paradigm shift is required for meaningful progress.

That is the entire scientific foundation under "sound changes your dreams." A small, inconsistent literature in which the least meaningful sounds, the category your machine produces, performed worst.

The sleep side has more studies and a clearer verdict. Noise as a sleep aid, a 2021 systematic review in Sleep Medicine Reviews by Samantha Riedy, Michael Smith, Sarah Rocha and Mathias Basner, pulled together 38 studies of continuous white noise or similar broadband sound. Results varied so much that the extremes were continuous noise improving sleep and continuous noise disrupting it, and following the GRADE criteria, the authors rated the quality of evidence for continuous noise improving sleep as very low, which they note contradicts its widespread use. They also flagged that it may negatively affect sleep and hearing.

A second systematic review, Capezuti and colleagues in the Journal of Clinical Sleep Medicine in 2022, sorted 34 studies covering 1,103 people into white noise, pink noise and mixed audio. Nineteen of the 34 reported positive sleep outcomes, but only 6 of the 18 white noise studies did. Their conclusion is worth keeping in both hands: no strong evidence supports the use of auditory stimulation, and none of the studies reported any adverse effects from short term use.

Neither review measured dreams. That is simply what the field has done, and it is why the honest answer to the question in the title has to begin with an admission rather than a number.

Two people worth listening to have said as much in public. In Harvard Health Publishing, Dr. Sogol Javaheri, a physician in the division of sleep and circadian rhythm disorders at Brigham and Women's Hospital, warns that white noise "may interrupt important stages of sleep, such as REM sleep or deep sleep, and affect hearing if it is too loud." And Basner, asked about sound machines after that review, gave one of the more honest answers a researcher can give: it is possible they could be beneficial for sleep, and also plausible that they could be detrimental. That was the state of the question. Then his laboratory measured it.

03The 2026 study that finally measured it

In February 2026 the picture got sharper, and the sharpening came from the same laboratory that had co-authored the 2021 review. Mathias Basner and colleagues at the University of Pennsylvania published Efficacy of pink noise and earplugs for mitigating the effects of intermittent environmental noise exposure on sleep in the journal Sleep.

Twenty five healthy adults, average age twenty eight, slept seven nights in a laboratory with full polysomnography. Each night was a different condition: a quiet control night, intermittent aircraft noise, pink noise alone at 50 decibels, aircraft noise combined with pink noise at 40 and at 50 decibels, and aircraft noise with foam earplugs. The results are unusually clean for this subject.

Read the bottom row twice. Earplugs, on that night, produced sleep that was statistically indistinguishable from a quiet room. The pink noise rows did not.

A companion paper from the same experiment, published in Sleep Health in August 2026, gives pink noise its due: it did mitigate noise induced arousals and awakenings, in a dose response way, meaning more of it worked better. It was simply outperformed by earplugs. That paper also found the two hardest sounds to sleep through were a crying baby and a fire alarm, worth knowing if you run a machine in a house with either.

Now the caveats, because they may be the part that applies to you. This was a short exposure study in a fairly small sample, and the authors say plainly that the findings may not generalize to longer exposure. Every participant was required to be new to broadband sound at night, so nobody in that room had slept beside a fan for fifteen years. If you are that person, this study did not test you. It tested what happens when the sound is new.

What it does establish is that the trade is real and not free. Steady sound covers an intermittent disturbance and takes something back in REM. Nobody selling one of these will tell you the second half.

04Why remembering fewer dreams does not mean dreaming less

This is the idea that resolves most of what people notice, and once you have it you will stop misreading your own nights.

You dream whether or not you remember it. What decides whether a dream reaches your morning is largely whether you woke up near it. The clearest evidence for that comes from van Wyk, Solms and Lipinska in 2019, who took nineteen people who remember dreams often and seventeen who rarely do, and recorded them overnight on two nights each with polysomnography.

The frequent recallers spent significantly more time awake after falling asleep and had significantly more awakenings across the night. What they did not have was more REM sleep, or more of the eye movement activity taken as a proxy for dream production. And the awakenings that separated the groups were not from REM at all. They came from stage two non-REM sleep. The authors conclude that differences in dream recall between healthy people cannot be put down to differences in REM sleep, and that awakenings from non-REM sleep are what matters.

Sit that beside a sound machine and the arithmetic goes the opposite way from the one everybody assumes. If your setup genuinely quiets the street and you stop surfacing four times a night, you will remember fewer dreams. You have not lost your dreams. The same confound is why supplements sold for dreaming are so hard to judge, and whether any dream supplement actually works takes it apart in full. You have lost the awakenings that were letting you catch them, which is the thing you were trying to buy. Our explainer on why you forget your dreams goes through that mechanism, and if you want the recall back without giving up the sleep, how to remember your dreams is the page, because the fix is in your first two waking minutes rather than in your hardware.

It runs the other way too. If a machine is fragmenting your sleep, and the 2026 data says that is possible, more awakenings mean more dreams caught, and the ones you catch skew vivid and strange because those survive a jolt. Somebody who starts a noise machine and reports wilder dreams may be describing worse sleep. Why your dreams feel so vivid covers what else pushes that way.

05If your nightmares got worse after you started

If this is why you are here, everything above is background.

Nobody has run a study on whether brown noise, or white noise, or a rain app causes nightmares. There is no trial establishing that it does, and none establishing that it does not. Anybody telling you either with confidence is telling you something the literature does not contain.

What can honestly be said is this. Broadband noise at ordinary bedroom volume has now been shown, in a laboratory, to reduce REM sleep and, when layered on top of an existing disturbance, to increase time spent awake. Nightmares happen mostly in REM, and fragmented sleep is associated with remembering more of what happens in it. So, a route by which a machine could leave you with a worse night and more bad dreams to recall from it is not far fetched. It is just not tested. And "not tested" is a different statement from "proven harmless," which is how the category prefers to be read.

The practical move costs nothing. Turn it off for two weeks and change only that. If the nightmares ease, you have your answer and you did not need a study. If they do not, the machine was not the problem and you have ruled something out.

If bad dreams are frequent, distressing, or making you reluctant to go to bed, that deserves better than gear. The American Academy of Sleep Medicine's position paper on the treatment of nightmare disorder in adults recommends image rehearsal therapy, a structured daytime practice, for nightmare disorder and for nightmares linked to post traumatic stress. It is the only therapy in that document carrying a plain recommendation, and it involves no purchase at all. Our pages on what nightmares mean and what to do after a bad dream cover the reflective side.

06The one real finding about sound and nightmares, and why you cannot buy it

Here is the study that gets quoted in every article about sound machines and dreams, usually stripped of the detail that makes it mean anything.

In October 2022, Sophie Schwartz, Alice Clerget and Lampros Perogamvros published Enhancing imagery rehearsal therapy for nightmares with targeted memory reactivation01477-4) in Current Biology. Thirty six patients with nightmare disorder, all having at least one nightmare a week. Everyone got image rehearsal therapy, the treatment the AASM recommends, in which you take your recurring nightmare and deliberately rewrite its ending into something better, then rehearse the new version.

Here is the part that is always cut. While one half of the group sat in that therapy session imagining their rewritten dream, a one second neutral piano chord, at about 42 decibels, played every ten seconds. That is all that separated the two groups at the start. Then, for the next two weeks, every patient in the study, in both groups, wore an EEG headband to bed that scored their sleep stages in real time and played that same chord every ten seconds during REM sleep specifically. Both groups. Same sound, same nights, same delivery.

After two weeks, the group whose chord had been paired with the rewritten dream was down to 0.19 nightmares a week. The group who had the identical sound played into their REM sleep, without the pairing, was at 1.02. Three months later the gap had widened, 0.33 against 1.48. The paired group also reported markedly more joy in their dreams.

So, the sound alone did nothing that the sound plus a learned association did. That comparison sits inside the study, which is what makes it so useful. Playing a tone into somebody's REM sleep all night is what the control group got, and the control group did worse.

Four things had to be true for that result, and a consumer machine delivers none of them:

• The sound had been bonded to a specific rewritten dream in a therapy session with a clinician.
• An EEG headband detected REM sleep as it happened and fired only then.
• The sound was a one second cue every ten seconds, not a continuous wash.
• The patients did image rehearsal therapy every evening for a fortnight.

A sound machine on a nightstand does not know what stage you are in, has never been paired with anything, plays continuously, and comes with no therapy. It is not a weaker version of this protocol. It is a different thing that happens to also make noise.

The authors are careful about their own limits. There was no group that received no stimulation at all, since the point was to test whether the cue added to a therapy already known to work. They did not record sleep before stimulation began, so they cannot say what other effects the sound had on sleep. And 22 percent of participants dropped out before the three month follow up.

07White, pink, brown: does the color change anything

Brown noise sounds deeper because more of its energy sits at low frequencies. Pink sits between brown and white. That is a genuine acoustic difference, and if one is more pleasant to your ear, that preference is real and worth respecting.

What it is not is evidence. In the Capezuti review, pink noise studies reported positive sleep outcomes more often than white noise studies did, but those were different studies of different quality measuring different things, not a head to head test. And in the 2026 Penn laboratory study it was pink noise, the better performing color in that review, that cost nineteen minutes of REM.

The place people go wrong is confusing a playlist with a laboratory technique that shares a word. Closed loop acoustic stimulation is a research method in which an EEG rig watches for the upswing of a slow brain wave and fires a very brief, very quiet click precisely on it. It is timed to your brain and it is nothing like a continuous track. Even that has produced modest results. A meta analysis by Wunderlin and colleagues in Sleep in 2021 pooled 10 studies and 177 healthy participants and found an overall effect on overnight memory of Hedges' g equal to 0.25, which did not reach significance at p equals 0.07. In the best subgroup, young adults with phase locked stimulation, it rose to 0.44. Their conclusion states directly that the evidence is not currently sufficient to recommend the use of commercially available devices.

What happens when the timing is removed is instructive. Roüast, Kumral, Gais and Schönauer, publishing in iScience in 2026, played random rather than phase locked sounds during an afternoon nap. Slow wave sleep went down, slow wave count went down, and declarative recall was impaired in proportion to the lost sleep depth. Untimed sound during sleep did not help memory. It got in the way.

The 2021 Riedy review explicitly excluded studies using intermittent white noise to enhance slow wave activity, precisely because that is a separate intervention. When a product page cites "pink noise boosts your brain overnight," it is borrowing from a literature about a clicking EEG rig, not about the thing in the box.

08The light half of the question

Sleep masks and blackout curtains are a different subject and deserve their own honest treatment rather than a share of the sound verdict. Light has a real literature on circadian timing and sleep continuity, and a dark room is not folklore.

The best specific evidence for a mask is Greco and colleagues in Sleep in 2023. In their first experiment, ninety four people aged 18 to 35 wore an eye mask every night for a week and went a second week without one, then did a cognitive battery. Episodic encoding and alertness were both better in the mask week. In a second experiment, thirty five people repeated it with a wearable sleep monitor, the encoding benefit replicated, and its size was predicted by time spent in slow wave sleep.

There is also Hu and colleagues in Critical Care in 2010, who put fourteen healthy volunteers through polysomnography under recorded intensive care unit noise and light. That condition produced lighter sleep, longer REM latency and less REM. Adding earplugs and eye masks brought REM time back up, shortened REM latency, reduced arousals and raised overnight melatonin. Real result, real boundary: a simulated ICU is an extreme environment, not a bedroom with a streetlight.

So, put plainly. Blocking light has a small, replicated benefit for next day learning and alertness, and a clear benefit in genuinely bright conditions. Not one of those studies measured what anybody dreamed about. A mask is a cheap, low risk, evidence supported thing to try if light is your problem. It is not a dream tool, and the sound half of this page does not get to borrow its credibility.

09Volume, children, and the part that is an actual risk

The documented hazard in this category is not your dreams. It is decibels.

A 2024 scoping review in Sleep Medicine by De Jong, Davis, Chelf, Marinelli, Erbele and Bowe looked at continuous white noise and childhood development across 20 studies covering 83 animal and 9,428 human subjects. The headline finding is blunt: white noise machines can exceed 91 decibels at maximum volume, above the National Institute for Occupational Safety and Health exposure guideline for a two hour work shift in an adult. Their conclusion is that existing data support limiting the maximum intensity and duration of these devices.

They are fair about the other side. Animal evidence suggests harm from continuous moderate intensity exposure during early development, human data broadly agrees, and yet some studies suggest low intensity exposure during sleep may be beneficial. Their call is for research on optimal intensity and duration, not a ban. Basner made a related point after the 2026 study, cautioning specifically against broadband noise for newborns and toddlers, because children spend far more of the night in REM than adults do.

For scale, the World Health Organization's guidance on noise puts the bedroom target under 30 A weighted decibels for sleep of good quality. The pink noise that cost nineteen minutes of REM in the Penn laboratory was played at 50.

So, three practical things. Keep it quieter than a background conversation. Put it across the room rather than beside the head, and never in or on a crib. And know that a machine loud enough to fully cover a genuine disturbance is louder than the level anybody has tested as safe for nightly use.

10A two week test that costs nothing

You do not need another purchase to answer this, and a fortnight of honest notes will tell you more than any article can, this one included. Change exactly one variable. Not the machine and the mask and your bedtime. One.

Each morning, before anything else, write down five things: what time you got in and out of bed, how many times you remember waking and why if you know, how rested you feel out of ten, whether you recalled a dream at all, and if it was unpleasant, how distressing it was out of ten. Five lines. Ninety seconds.

Then read the whole fortnight at once rather than judging it night by night, because one memorable dream is noise and a trend is signal. What you are looking for is fewer awakenings caused by the specific outside sound you were covering, and better function during the day. Those are the outcomes the research measured, so they are the ones you can meaningfully track.

Stop if you are waking more, if the sound irritates you rather than fading, if you feel worse in the afternoons, or if bad dreams are getting more frequent. And if two weeks changes nothing at all, that is a real result: the thing is not doing anything for you, and the next purchase will not either. Any notebook works for this, and so does the notes app on your phone. If you want the dream side of the log to still make sense in a month, our guide to starting a dream journal covers how to record an entry.

11What I would actually do

Nobody wants a page that ends in "more research is needed," so here is the decision, by situation.

The uncomfortable version of that table is that the two things with the best evidence in it, foam earplugs and a piece of fabric over your eyes, are the two cheapest things in the room.

None of this makes you foolish for buying a sound machine. Most people bought one because a real problem was keeping them awake, and covering that up is a reasonable thing to try. It is only worth knowing that you were sold two promises, and only one of them has ever been to a laboratory.

12Common questions about white noise, sleep masks and dreams

Does white noise affect dreams?

Not in any way that has been demonstrated. The systematic review of sensory stimulation and dreaming found that featureless sounds like white noise had the lowest rate of turning up in dream content of any stimulus tested, around nine percent, and the two large reviews of white noise and sleep did not measure dreams at all. What sound can change is your sleep, and that changes how many dreams you remember, which is the effect most people are actually noticing.

Can white noise or brown noise cause nightmares?

No study has tested it, in either direction. What is known is that broadband noise at 50 decibels reduced REM sleep in a 2026 laboratory study, that fragmented sleep is associated with recalling more of what happens in it, and that nightmares occur mostly in REM. That is a plausible route, not a demonstrated one. If yours got worse after you started, turn it off for two weeks and change nothing else. If they are frequent or distressing, that is a clinician's question rather than an equipment one.

Does white noise increase REM sleep or make dreams more vivid?

There is no good evidence that it increases REM, and the best recent measurement points the other way: pink noise at 50 decibels was associated with about nineteen minutes less REM per night than a quiet control night. No controlled study has established that continuous sound changes vividness. If your dreams feel more vivid, disturbed sleep producing more awakenings is the likelier explanation.

Should I run a sound machine all night?

Only if it is masking a genuine, irregular disturbance, and only at the lowest volume that does the job. The 2026 laboratory data showed a cost in REM sleep, and the same experiment found foam earplugs did the masking better with no measurable penalty. If your room is already quiet, continuous sound is more likely to be a new disturbance than a solution.

Is white noise safe for a baby?

Volume and placement are the issue. A 2024 scoping review found machines can exceed 91 decibels at maximum, above the occupational guideline for a two hour adult shift, and recommended limiting device intensity and duration. Researchers have specifically cautioned against broadband noise for newborns and toddlers, partly because children spend far more of the night in REM sleep than adults. Keep it low, keep it across the room, never in or on the crib, and raise it with your pediatrician if you are unsure.

Do sleep masks or blackout curtains change your dreams?

There is no evidence that they change dream content. There is decent evidence a mask helps other things: in a 2023 study, ninety four people who wore one for a week showed better next day episodic encoding and alertness than in a week without, and the benefit tracked with time in slow wave sleep. If light is your problem, a mask is cheap and reasonable. It is not a dream tool.

Has anything been shown to reduce nightmares using sound?

Yes, and the detail is the point. A 2022 trial paired a piano chord with a rewritten version of a patient's nightmare during image rehearsal therapy, then replayed that chord during REM sleep for two weeks using an EEG headband that detected the stage in real time. That group ended with far fewer nightmares than a control group who got the identical sound during REM sleep without the pairing. The sound was a cue for a memory built in therapy. No consumer device reproduces the therapy, the pairing, or the stage detection.

If you came here with a cart open, the honest close is that you can probably shut it. The two things on this page with real evidence behind them are foam earplugs and something over your eyes, neither of which changes what you dream about, and the one genuine finding about sound and nightmares is a course of therapy rather than a device. Whatever you decide, keep the log for a fortnight, because your own nights are the only dataset that is actually about you. DreamTold treats dreams as a window into your own mind rather than something to be engineered from a nightstand, so once the bedroom question is settled, take whatever keeps showing up to the dream dictionary. This guide is educational and meant for reflection, not medical advice. If nightmares are frequent or distressing, if you cannot sleep, or if you are worried about hearing or a child's exposure, speak with a qualified clinician.

13Sources

Every claim on this page that comes from somewhere else, with the somewhere else.

• PubMed Central: Influencing dreams through sensory stimulation, Salvesen, Capriglia, Dresler and Bernardi, Sleep Medicine Reviews, 2024
• PubMed: Noise as a sleep aid: A systematic review, Riedy, Smith, Rocha and Basner, Sleep Medicine Reviews, 2021
• PubMed Central: Efficacy of pink noise and earplugs for mitigating the effects of intermittent environmental noise exposure on sleep, Basner and colleagues, Sleep, 2026
• PubMed: The efficacy of pink noise and earplugs for mitigating sleep disruption induced by different environmental noise sources, Basner and colleagues, Sleep Health, 2026
• Current Biology: Enhancing imagery rehearsal therapy for nightmares with targeted memory reactivation01477-4), Schwartz, Clerget and Perogamvros, 2022
• PubMed: Systematic review: auditory stimulation and sleep, Capezuti and colleagues, Journal of Clinical Sleep Medicine, 2022
• PubMed: Modulating overnight memory consolidation by acoustic stimulation during slow-wave sleep, Wunderlin and colleagues, Sleep, 2021
• PubMed: Random auditory stimulation during sleep disturbs traveling slow waves and declarative memory, Roüast, Kumral, Gais and Schönauer, iScience, 2026
• PubMed: Increased Awakenings From Non-rapid Eye Movement Sleep Explain Differences in Dream Recall Frequency in Healthy Individuals, van Wyk, Solms and Lipinska, Frontiers in Human Neuroscience, 2019
• PubMed: Wearing an eye mask during overnight sleep improves episodic learning and alertness, Greco and colleagues, Sleep, 2023
• PubMed: Effects of earplugs and eye masks on nocturnal sleep, melatonin and cortisol in a simulated intensive care unit environment, Hu and colleagues, Critical Care, 2010
• PubMed: Continuous white noise exposure during sleep and childhood development: A scoping review, De Jong and colleagues, Sleep Medicine, 2024
• PubMed Central: Position Paper for the Treatment of Nightmare Disorder in Adults, Morgenthaler and colleagues, American Academy of Sleep Medicine, 2018
• Harvard Health Publishing: Can white noise really help you sleep better?
• World Health Organization: Noise fact sheet