Why you hear your alarm in a dream and sleep through it
Short answer: your ears keep working while you sleep, but the sleeping brain runs a crude significance check on every sound before it interrupts anything — and one of the things it does with sounds that fail the check is fold them into the dream already in progress. Incorporation means the alarm arrived and lost the argument.
It's a particular kind of infuriating, because you heard it. There was a phone ringing somewhere in the dream, or a reversing lorry, or a smoke alarm in a building you've never been to. The signal got in. What didn't happen is the step where the signal gets classified as your problem.
Sound doesn't stop at the ear when you fall asleep
Hearing has no off switch. The cochlea keeps transducing air pressure into nerve impulses all night, and the auditory brainstem keeps passing them upward. What changes is the thalamus, the relay station that sits between the senses and the cortex. During sleep it becomes a far less cooperative gate, damping how much of the incoming traffic reaches cortical areas that could do something about it.
Recordings taken from the scalp of sleeping volunteers show auditory responses persisting through every sleep stage — smaller than in waking, and reshaped, but present. The brain is still receiving. Sleeping through a noise is not a failure of the ears. It's an editorial decision taken several layers below anything you'd call a decision.
Incorporation is what happens to a signal that isn't acted on
Sleep laboratories have been sticking stimuli into people's dreams for a long time: sprayed water, spoken words, pressure cuffs, tones. The general finding is that external stimuli do sometimes turn up in dream reports, and that the rate is modest. Most stimuli are simply ignored. When something is incorporated, it almost never appears as itself.
That last part is the interesting bit. Your alarm doesn't show up in the dream as an alarm you are failing to turn off. It shows up as an ambulance, a kettle, a videogame, a bird. The dreaming brain is a relentless explainer — its whole business is generating a scene that accounts for whatever activity is currently running. Handed an unexplained rhythmic beep, it does what it does with everything else, which is invent a source. Ambiguous, repetitive sounds are the easiest raw material for that, and most alarm tones are exactly that.
Why an explanation is cheaper than an awakening
The gate isn't indiscriminate. Sleeping brains reliably respond more strongly to some sounds than others — your own name over a stranger's, a baby's cry over a comparable noise, an unfamiliar tone in a stream of familiar ones. Something is triaging for significance while you're unconscious.
An alarm tone you have heard several hundred times, which has never once preceded danger, is a poor candidate for that triage. It is known, predictable and — in the brain's accounting — already accounted for. Waking is metabolically and functionally expensive, particularly out of deep sleep. Generating a narrative patch is not. When a sound is loud enough to register but not meaningful enough to alarm, the patch is the path of least resistance.
This is also why the volume-first approach disappoints so many people. Loudness raises the odds of a response but does nothing about the significance problem, and past a point it mostly buys you a worse awakening rather than a more reliable one. Why loud alarms don't work for heavy sleepers goes further into that trade.
Habituation is why last month's alarm stopped working
Nervous systems systematically downweight stimuli that repeat without consequence. It's one of the oldest findings in the field and it applies to sleep as clearly as anywhere else. Research on people living under flight paths and beside motorways finds something telling: over time, residents report and record fewer full awakenings, while measurable physiological responses to the same noise — heart rate changes, brief cortical stirrings — continue. The response to the sound survives. The awakening is what gets pruned.
Your alarm is a stimulus that repeats daily and, from your sleeping brain's point of view, never predicts anything worse than a commute. The classic pattern — a new alarm works beautifully for two weeks, then quietly stops — is habituation doing exactly what it does. Changing the sound restores some novelty, which is why that trick helps, and why it also stops helping after a while.
Stage matters, but not the way folklore says
The arousal threshold — how much stimulus it takes to rouse you — is genuinely stage-dependent. It's highest in slow-wave sleep, which is concentrated in the first part of the night, and lower in lighter stages. But incorporation into dreams isn't exclusive to REM, and it isn't the case that dreaming makes you unwakeable.
Part of why alarms specifically end up in dreams is just scheduling. Alarms go off in the last hours of sleep, and the last hours are proportionally rich in REM, when vivid, narrative, well-recalled dreaming is most likely. The alarm didn't seek out your dream. It arrived during the shift when the dream department is busiest and the reports are most memorable.
The dream may be assembled at the moment you wake
Dream reports are reconstructions. What you retrieve on waking is a memory formed under unusual conditions, and it can be stitched together retrospectively — which means "I heard it ringing in the dream for ten minutes" is not reliable evidence about duration. A sound that was present for a few seconds can be given a backstory that feels like it ran for a scene.
There's a related trap. False awakenings — dreaming that you got up, showered, dressed, and left — are common around alarms, and they can be convincing enough that you're genuinely surprised to find yourself in bed. The dream is doing what your intention asked of it, which does nothing whatsoever for your actual morning.