Why Some People Sleep Through Noise and Others Don't

Last updated August 8, 2026

Why some people sleep through noise and others don't

Short answer: noise tolerance during sleep is mostly set by your arousal threshold, which varies by sleep stage, by how much sleep debt you're carrying, by age, and by measurable differences in how your thalamus gates incoming signals — the sleep spindle being the best-studied of these. Some of it is stable trait, a lot of it is state, and almost none of it is willpower.

Two people in the same bed, the same street noise, the same night. One never stirs; the other is awake at 3am cataloguing every car. The difference is real and largely involuntary, and the parts we understand well are more interesting than the parts that get repeated.

Arousal threshold is the actual variable

The umbrella concept is arousal threshold: how much stimulus it takes to produce a shift toward wakefulness. It can be measured — sleep researchers do it by delivering graded tones or pressure and finding where a response appears — and it varies enormously between people and within the same person across a night.

The main drivers are unglamorous:

Sleep stage. Threshold is highest in slow-wave sleep, lower in lighter non-REM stages, and variable in REM, where responsiveness depends heavily on what the stimulus is. Slow-wave sleep is front-loaded into the first cycles, which is why the same neighbour's door at 1am and 6am produce different outcomes.

Sleep pressure. Accumulated wake time builds adenosine and other markers of homeostatic drive, and higher sleep pressure raises the threshold across the board. After short nights, everyone becomes a heavy sleeper. This is the largest and most reversible cause of the trait people think they were born with.

Age. Slow-wave sleep declines across adult life and sleep becomes more fragmented, so older adults generally wake more easily to noise. That's a shift in sleep architecture, not in hearing.

Sleep spindles: the best mechanistic story we have

Spindles are short bursts of rhythmic activity, roughly a second long, generated by a loop between the thalamic reticular nucleus and the cortex, and they punctuate lighter non-REM sleep. Their proposed function includes actively gating sensory transmission — the thalamus is the relay point for most incoming sensory traffic, and spindle activity appears to interfere with passing that traffic on to the cortex.

That predicts something testable: people whose sleep produces more spindles should be harder to wake with noise. A small and much-cited experiment did exactly that, playing recorded hospital and traffic sounds to sleeping volunteers and finding that spindle rate tracked how much noise they tolerated. It's an elegant result and it fits the anatomy.

It's also worth keeping in proportion. The samples in this literature are small, spindle rate explains only part of the variation between individuals, and the field is still arguing about how much of a spindle's job is sensory gating versus memory consolidation. Sleep spindles are the most satisfying available explanation for stable individual differences in noise tolerance. They are not a settled one, and the popular coverage is more confident than the evidence.

The related K-complex is more contested still. These large single waveforms often appear in response to sounds, and researchers have argued both that they represent the brain suppressing an awakening and that they represent a miniature arousal. Both cases have support, which usually means the truth is that they do different things in different circumstances.

The sound itself matters more than its loudness

Sleeping brains discriminate. Responses are consistently larger to your own name than to another name matched for length and volume, and parents respond to their own infant's cry far more readily than to comparable sounds. Something is triaging incoming audio for significance while you're unconscious, and that filter cares about meaning.

Several properties of a sound change its odds independently of volume:

Novelty. Unexpected sounds outperform familiar ones. This is the whole mechanism behind alarm tones losing their power.
Rise time. Sudden onsets are more arousing than sounds that fade in. A door slamming beats a steadily louder engine.
Intermittency. Irregular, unpredictable noise disturbs sleep more than continuous noise at the same level, which is why a dripping tap outperforms a motorway.
Meaning. A voice saying your name, a smoke alarm's voice announcement, a child crying. Semantically loaded sounds get through gates that pure tones don't.

Steady, predictable background noise is the least disruptive category, which is the honest basis for white noise machines. They don't mask anything in a physical sense so much as reduce the contrast between the background and the sudden events that punctuate it.

Habituation, and the thing habituation hides

Repeated, consequence-free noise gets downweighted. Studies of populations living near airports and major roads find residents reporting and recording progressively fewer awakenings over years of exposure.

The finding underneath that is the important one. In the same populations, physiological responses to overflights persist — heart rate changes and brief cortical arousals that never reach conscious awakening. The person is not being woken. The person is also not being left alone. Sleeping through noise is a statement about awakening, not about sleep quality, and the environmental noise literature is fairly consistent that fragmented sleep from unnoticed arousals still carries costs during the day.

This is worth internalising if you share a room with someone who insists the noise doesn't bother them. It may well not wake them. That's a weaker claim than it sounds.

Hearing is a real and underrated variable

Age-related hearing loss affects high frequencies first, and a great many alarms and consumer devices sit in exactly that range. Someone with early presbycusis may be sleeping through a sound they genuinely cannot hear well at any volume, which is a different problem from a high arousal threshold and has a different solution. Lower-pitched signals carry better for this group, which is why some emergency alerting research favours them.

The same logic applies more sharply to people with substantial hearing loss, where sound-based waking is not a matter of degree at all. Alarm clocks for deaf and hard of hearing people covers that case properly.

And the mirror image: earplugs reliably raise the threshold for everything, alarm included. That's the point and also the risk.

What about substances and sex differences?

Alcohol raises arousal threshold early in the night and then does the opposite. It suppresses REM in the first half and produces rebound and sympathetic activation in the second, which is why a nightcap makes you unwakeable at midnight and wide awake at four. Sedative-hypnotics raise the threshold more consistently, sometimes to the point of clinical concern for people who need to hear a smoke alarm.

Sex differences get asserted a lot — usually the claim that women wake more readily, particularly to infant sounds. The evidence is genuinely mixed and heavily entangled with who is doing the caregiving and who reports what. Treat confident claims here with suspicion; this is thinner than the popular coverage suggests.

Genetics beyond the spindle work is thinner still. It's reasonable to assume stable individual differences have some heritable component, because most sleep traits do. Anything more specific than that is currently speculation.

What this means practically

The order of operations matters, because most people start with the last item on this list.

Fix sleep duration first. Sleep pressure is the biggest modifiable input to arousal threshold, and it swamps everything else. A person two hours short is a heavy sleeper regardless of their spindles.
Reduce intermittency rather than volume. Steady background sound, better window sealing, moving the bed away from a shared wall — the goal is fewer sudden contrasts, not silence.
Rule out hearing. If sounds that others notice are consistently absent for you rather than ignored, that's an audiology question, not an alarm question.
Rule out apnea if you're unrefreshed regardless. Someone waking exhausted after eight undisturbed hours has a different problem.
Then, and only then, change the wake-up signal. Volume is the weakest lever available, and it costs you a worse sleep inertia on the mornings it does work.

For alarms specifically, the interesting move is often changing sensory channel rather than turning the same channel up. Vibration at the skin bypasses the entire question of whether a sound can be discriminated from background noise, and it's the standard approach for people with hearing loss for exactly that reason. Silent Wake is a wrist device with no speaker at all, which also means it can't wake anyone else in the room — relevant when one person's threshold is low and the other's is high, since the usual compromise is an alarm loud enough for the heavy sleeper and miserable for the light one. Three intensity levels let you find the lowest setting that actually works rather than defaulting to maximum.

Vibration is not exempt from habituation, and it isn't a solution to sleep debt. Nothing is.

Common questions

Is being a heavy sleeper something you're born with?

Partly, but less than most people assume. There are stable individual differences in arousal threshold, and sleep spindle activity is the best-supported physiological correlate of them. The larger share of variation is state rather than trait: sleep debt, alcohol, medication, sleep stage and age all move the threshold substantially, and sleep debt moves it most. People who describe themselves as lifelong heavy sleepers are frequently people who have been chronically short of sleep for years, which is a fixable version of the same symptom.

Why can I sleep through traffic but wake to a whisper?

Because the sleeping brain filters for significance, not decibels. Steady, familiar, predictable noise gets habituated and downweighted; novel, intermittent or meaningful sounds get through. Your name, a child's voice or an unfamiliar footstep carry information that could matter, and responses to those are measurably larger than to matched sounds without meaning. Traffic carries none. That's also why a dripping tap can be more disruptive than a motorway at higher volume — irregularity is the arousing property, not intensity.

Does white noise actually help?

For some people, in a specific way. It doesn't cancel sound; it raises the background floor so that sudden noises stand out less against it, and reduced contrast means fewer arousals. That's a real mechanism and a modest effect. The evidence base is smaller and lower quality than the market implies, and there's a reasonable argument that a constant sound source overnight isn't ideal either. Worth trying if intermittent noise is your problem; not worth expecting much if the problem is sleep debt.

If noise doesn't wake me, is it still harming my sleep?

Probably, at least somewhat. Research on people living near airports and motorways consistently finds physiological responses to noise — heart rate changes, brief cortical arousals — continuing even after reported awakenings have habituated away. Sleep is being fragmented below the level you notice, and that fragmentation has been associated with daytime effects. So "it doesn't bother me" is a claim about awareness rather than about sleep quality. Reducing night-time noise is usually worth doing even for people who sleep through it.

Can I train myself to sleep through noise, or to wake more easily?

Habituation happens on its own, and you can't do much to speed it. The reverse — becoming more responsive — isn't really trainable either, though it does happen situationally: new parents become dramatically more responsive to one specific sound without any deliberate practice, which suggests the filter is tuned by significance rather than effort. What you can change is the inputs. Consistent adequate sleep lowers your threshold more reliably than any technique, and it's the only lever with a large effect size.

Why Some People Sleep Through Noise and Others Don't