Why You Dream: The Neuroscience of the Strangest Thing Your Brain Does

...and What to Do About It

Every night, your brainstem paralyzes you.

This is not a malfunction. It is one of the most elegant protective mechanisms in vertebrate biology — a dedicated circuit that suppresses your voluntary muscles while your visual cortex, your limbic system, and your motor areas run at near-waking levels of activity, generating experiences so internally consistent that you accept them completely. You are unconscious, immobilized, and fully convinced you are awake in a world that does not exist.

You will do this tonight for approximately two hours. Over the course of a human lifespan, that adds up to roughly six years.

The question science has been circling for decades is not whether dreams are meaningful. The question is what kind of meaning they carry, what biological function justifies the metabolic cost and the evolutionary risk of spending a third of your life in a state of motor paralysis, and whether the modern disruption of this system is contributing to the epidemic of emotional dysregulation we are now measuring across the population.

The answers that have emerged from sleep laboratories over the past forty years are more interesting than either Freud’s interpretive framework or the dismissive “it’s just random noise” counterreaction. Both are wrong, and the truth is stranger and more useful than either.

What the Brain Actually Does During Dreaming

REM sleep — Rapid Eye Movement sleep, named for the visible tracking motion of the eyes under closed lids — was discovered by Eugene Aserinsky and Nathaniel Kleitman at the University of Chicago in 1953.¹ It was the first demonstration that sleep was not a uniform state but a cycling architecture, and it opened the first scientific window into dreaming.

In a typical night of seven to eight hours, you complete four to six sleep cycles, each lasting approximately ninety minutes. Each cycle moves through stages of progressively deeper non-REM sleep and back up to a REM period. The first REM window of the night lasts about ten minutes. The last, in the hours before waking, can extend to sixty minutes or more. Total nightly REM time: approximately two hours, constituting roughly twenty-five percent of sleep.²

What happens during REM is remarkable. Electroencephalographic recordings of the dreaming brain are nearly indistinguishable from those of the waking brain. The visual cortex fires. The amygdala and hippocampus — the emotional and memory processing centers — fire intensely. The motor cortex activates. The acetylcholine-driven arousal system runs at something close to full tilt. If you handed an unlabeled brain scan to a naive observer, they would not easily identify the REM state as sleep.

One major exception: the prefrontal cortex — the region responsible for logical evaluation, working memory, and what cognitive scientists call executive control — shows significantly reduced activity during REM.³ This is the mechanism behind the central paradox of dreaming: that your childhood home has inexplicably become a submarine, and you find this perfectly normal, and then wake up and cannot imagine how. The editorial system that would flag the inconsistency is largely offline. The rest of the brain just runs the story.

A note that a careful reader of the primary literature will want: dreaming is not exclusively a REM phenomenon. A 2017 study by Siclari and colleagues in Nature Neuroscience, using high-density EEG recordings, identified a “posterior hot zone” of cortical activity associated with dreaming that can occur during NREM sleep as well — specifically during windows when slow-wave activity in posterior regions quiets, creating pockets of near-waking neural activity within otherwise deep sleep.⁴ The emotional processing function described in this article is primarily attributed to REM, but the boundaries between dreaming sleep and dreamless sleep are less clean than earlier models assumed. The vivid, recalled, narratively complex dreams are predominantly REM. The architecture is more distributed than the popular account suggests.

The Paralysis Is the Point

While the dreaming brain runs its simulation, the brainstem executes a mechanism worth understanding in detail, because its failures are clinically informative and its purpose is precisely what you would infer from those failures.

The sub-laterodorsal tegmental nucleus and ventral medial medulla project inhibitory signals to motor neurons in the spinal cord via the neurotransmitters GABA and glycine, suppressing voluntary muscle tone throughout the body. The eye muscles and diaphragm are exempt — which is why your eyes move and you keep breathing. Everything else stops. This is called REM atonia, described in detail by Jouvet in 1967 and its neurotransmitter mechanisms clarified by Brooks and Peever in 2012.⁵

The purpose is mechanical and obvious: to prevent you from acting out what you experience in dreams. We know this from the pathology. In REM sleep behavior disorder, the atonia mechanism fails. Affected individuals move during dreaming — sometimes violently. They punch, kick, shout, run. They injure themselves and their partners. They wake to find themselves across the room. The correlation between these behaviors and the concurrent dream content, when it can be recovered, is direct.⁶ They are doing what the dream asks. The atonia, under normal conditions, is what stops them — and stops you.

Sleep paralysis is the reverse condition — when you wake up but the atonia signal is still running. The eyes open, the room is visible, but voluntary movement remains blocked. This affects approximately eight percent of the population at least once in a lifetime.⁷ It is experienced as terror, and the specific hallucinations it produces — a presence in the room, a weight on the chest, a threatening figure — have been reported across culturally isolated populations throughout recorded history, given different names but exhibiting the same perceptual architecture. The temporoparietal junction, misfiring in the liminal state between REM and waking, generates a phantom other. The amygdala simultaneously activates at maximum threat level. The result is subjectively indistinguishable from a supernatural encounter, and neuroscientifically explicable as what happens when the boundary between the dreaming simulation and waking perception collapses in a specific way.⁸

Why Dreams Are Bizarre: Hobson and McCarley’s Enduring Framework

In 1977, Harvard psychiatrists Allan Hobson and Robert McCarley published a paper in the American Journal of Psychiatry that fundamentally reoriented scientific thinking about dreaming.⁹ Their activation-synthesis hypothesis proposed that the brainstem, during REM, fires bursts of essentially random neural activation upward into the forebrain. These signals have no intrinsic narrative. They are the machinery running in the absence of structured sensory input.

The forebrain receives them and does what it always does with input: it synthesizes. It matches the signals to memory fragments, to emotional patterns, to fragments of recent and autobiographical experience, and constructs the most coherent narrative it can from profoundly incoherent raw material — with the prefrontal editor only partially participating.

This is why dreams are vivid but strange. They are not encrypted messages from the unconscious. They are not the return of repressed material in disguise. They are associative, emotionallydriven narrative construction operating without full logical supervision. The strangeness is not censorship, as Freud believed — it is the absence of the censoring system.

This does not mean dreams are devoid of personal meaning. Hobson himself was explicit about this.¹⁰ The associative material the brain draws on is real — your actual memories, your actual emotional preoccupations, your actual fears and desires. What the activation-synthesis model refutes is not the personal relevance of dreams but the interpretive apparatus Freud built on top of it: the symbols, the disguises, the need for a trained analyst to decode what the dream “really means.” The dream means what it appears to mean, roughly. It is bizarre not because its content is hidden but because its construction process is.

Why Dreaming Exists: The Threat Simulation Theory

If REM sleep and dreaming are so metabolically expensive — demanding sustained nearwaking brain activity across two hours every night — evolutionary pressure would have eliminated them if they had no function. They are ancient. They appear across mammals. Whatever dreaming does, it has been doing it for a very long time.

The most compelling functional account is Antti Revonsuo’s Threat Simulation Theory, published in 2000.¹¹ Revonsuo’s argument: the dream-production mechanism preferentially selects threatening scenarios and simulates them, rehearsing the neurocognitive mechanisms required for threat recognition and avoidance. The function is not to process all experience equally but to overrepresent danger — to run the dangerous scenarios repeatedly, in varied combinations, building the neural infrastructure for responding to them.

The evidence Revonsuo and colleagues assembled is substantial. Analysis of hundreds of dream reports found that approximately sixty-six percent of recurrent dreams contain threats, that those threats tend to be physically dangerous and directed at the dreamer, and that the dreamer’s response tends to involve realistic defensive action.¹² Children exposed to genuine threat in waking life — children in war zones, children with documented trauma — have dreams with significantly elevated threatening content compared to children in safe environments.¹³ The simulation system responds to genuine exposure by intensifying its rehearsal.

The uncomfortable implication, which is worth sitting with: the nightmares that feel worst are likely the most functionally active dreams. Posttraumatic nightmares, from this framework, are not a disorder of dreaming. They are the threat simulation system running at maximum activation because it was given an unresolved threat it has not yet succeeded in processing. The nightmare is the mechanism working. The recurrence, the intensity — these signal that the threat has exceeded the system’s normal single-session processing capacity, not that the system is broken.

This connects to a pattern that keeps appearing in this work: the body’s systems are rarely malfunctioning when they produce symptoms. They are usually operating as designed, under conditions they were not designed for. The treatment question is rarely “how do we suppress this response” and more often “what is driving the response, and can we address that.”

The Overnight Therapy: What REM Does to Emotional Memory

The most clinically useful research on dreaming comes from work examining what REM sleep specifically does to emotional memory — not just whether you remember the memory, but whether it still hurts.

Matthew Walker at UC Berkeley has called this the overnight therapy hypothesis, published in Psychological Bulletin in 2009.¹⁴ During REM sleep, norepinephrine — the primary stress arousal neurochemical — drops to near-zero. This is the only period in either sleep or wakefulness when this happens. In this window of chemical quiet, the brain represents emotional memories — replaying the content without the associated autonomic arousal state — allowing the memory to be reencoded with reduced emotional charge.

The mechanism briefly: the memory and its emotional tag are stored together. During REM, in the absence of norepinephrine, the cortex can process the information content of the memory while the limbic system runs with less activation than the original encoding required. The result, over multiple nights, is a gradual decoupling of the memory from its full original distress load. You still know what happened. It hurts less.

The clinical evidence that makes this most concrete comes from Rosalind Cartwright’s work with depressed divorcees. In a study widely cited in sleep medicine, Cartwright recruited women experiencing depression following divorce, brought them into the sleep laboratory, and woke them four times during REM sleep each night to record their dream content. She tracked them clinically for one year.¹⁵ Her finding: the women who dreamed about their ex-partners during REM — who incorporated the emotionally relevant material into their dream content — showed significantly better clinical recovery from depression at the one-year mark than those who had not. The content was not incidental. The processing of the emotional material during REM, while that material was present in the dream, predicted who recovered.

Between seventy-five and ninety-five percent of dream content contains emotional elements, emerging predominantly from REM sleep.¹⁶ If your life contains significant unprocessed emotional material — and this describes most people in modern conditions of chronic stress, abbreviated sleep, and frequent REM disruption — the implications are direct.

The remainder of this article — including the lucid dreaming verification experiments, analysis of what modern life does to REM architecture, and six specific evidence-based interventions with primary citations — is available to paid subscribers.

If you have been reading Popular Rationalism, you already know that the actionable material is where this work is different from what you find elsewhere. The solutions here are not general wellness advice. They are specific, named, and cited — including an intervention for recurrent nightmares with controlled trial evidence, and a technique for lucid dreaming verified in a sleep laboratory that the mainstream sleep hygiene industry has never told you about.

Become a paid subscriber to read the full analysis, including one powerful technique called Image Rehearsal.

For Paid Subscribers

The Signal from Inside the Dream: Lucid Dreaming Verified

The most philosophically striking result in the history of dream research is the laboratory verification of lucid dreaming — and it happened because a researcher was willing to take seriously what most of his colleagues dismissed as impossible.

In April 1975, British researcher Keith Hearne was working in a sleep lab with an experienced lucid dreamer named Alan Worsley. The question was whether a person could be simultaneously in verified REM sleep and consciously aware that they were dreaming — and whether they could, from inside the dream, execute a deliberate act that would leave a verifiable trace on the polygraph recording outside it.