Home / Sleep Science Hub / Dreams

The Hive Mind · Issue 015 · 15 min read

A 1978 sleep lab at Rush-Presbyterian-St. Luke's, a recently divorced woman in her mid-forties, and what Rosalind Cartwright found in eight hours of paper-recorder tracings.

Your brain is editing your pain while you sleep — and forgetting the dream is the proof it worked

🌙 15-minute read — written for the first quiet moment you find today, or the hour before you try tonight.

In the spring of 1978, a clinical psychologist named Rosalind Cartwright was sitting in a small, instrument-crowded sleep lab at Rush-Presbyterian-St. Luke's Medical Center in Chicago, watching a paper recorder trace the brain activity of a recently divorced woman.

The woman was in her mid-forties. She had been referred to Cartwright's lab not because of a sleep disorder in the traditional sense, but because she was drowning. The divorce had not been clean. There had been years of difficulty before the end, and now the end itself felt like another kind of difficulty — a sustained, grinding grief that followed her into bed every night and seemed to be waiting for her when she woke.

She told Cartwright she dreamed constantly. She remembered the dreams in terrible detail. Her husband appeared in them, sometimes cruel, sometimes indifferent, sometimes simply present in the way he used to be before everything went wrong. She woke from these dreams not rested but wrung out, as though she had spent the night not sleeping but arguing with a version of the past that wouldn't stay still.

Cartwright had been studying sleep and dreams for more than a decade by then. She was not a romantic about it. She was a clinician. She wanted to understand what dreams actually did, functionally, in the architecture of a night — not what they symbolized, not what Freud or Jung had said, but what job they were performing in the brain while the body lay still.

The paper recorder kept moving. The woman's REM periods — those bursts of rapid-eye-movement sleep where the most vivid dreaming happens — were long. Longer than average. And the emotional content Cartwright was documenting from the dream reports was intense, specific, and repetitive. The same characters. The same emotional register. The same unresolved quality.

But here was the thing that didn't make sense yet: some of the other women in her study — women who had been through equally painful separations — were not like this. They dreamed, but differently. Their REM dreams incorporated the painful material in a looser, stranger way. The ex-husband appeared, but so did unrelated figures, disconnected scenes, memories from decades earlier woven in alongside the recent wound. And when these women woke, they didn't feel wrung out. They felt, most mornings, a little lighter than the day before.

Cartwright started pulling threads. What was the difference between the women who woke exhausted by their dreams and the women who woke lighter? And why did the brain seem to be doing something so different with essentially the same raw material?

The answer she arrived at over the next several years would change the way sleep researchers think about what REM sleep is actually for. It would also explain something that had puzzled philosophers, poets, and insomniacs for centuries: why we forget most of our dreams. And why that forgetting — that clean erasure — might be the whole point.

The night shift your brain runs without telling you

To understand what Cartwright found, it helps to understand what REM sleep looks like from the inside of a brain scan.

REM sleep is not quiet. The brainstem is firing. The eyes are moving. The motor cortex is active in ways that would produce thrashing movement if the body weren't essentially paralyzed — a feature Michel Jouvet described in 1959 in his work with cats at the University of Lyon, which we covered in Issue 013. The brainstem actively disconnects the muscles during REM so the body doesn't act out what the dreaming brain is experiencing. This is not a glitch. It is a design.

But here is what makes REM sleep unusual beyond the paralysis and the eye movement: the neurochemistry is radically different from any other state of consciousness, waking or sleeping.

During REM, levels of norepinephrine — the brain's primary stress-signaling molecule — fall to nearly zero. This is one of the only times in a human's entire 24-hour cycle when norepinephrine essentially disappears from the brain. Even during the deepest stages of non-REM sleep, some norepinephrine persists. During REM, it doesn't.

A neuroscientist at the University of California, Berkeley named Matthew Walker has described this as the brain running a kind of overnight therapy session — but without the therapist in the room, and without the stress hormone that makes painful experiences feel painful. You are replaying emotional memory, Walker argues, but you are doing it in a brain chemically stripped of the capacity to feel the full weight of that emotion.

The implication is almost too elegant to believe. The brain is reviewing what hurt, but in a condition where the hurt cannot fully register. It is re-filing the memory, but the emotional charge — the sharp, hot, present-tense feeling of the experience — is being gradually removed each time the memory is replayed under those conditions.

This is what Cartwright was watching on that paper recorder in 1978, before the language existed to fully describe it.

The women who were getting lighter each morning were the women whose REM dreams were doing this successfully — weaving painful recent memories into older, contextually richer material, replaying them in the flat affect of a norepinephrine-depleted brain, and filing them away with the charge reduced. The women who woke exhausted were the women whose REM dreams were looping without resolution — replaying the pain but not completing the process. Their brains were trying to do the work and failing to finish it.

Cartwright called this the mood-regulatory function of REM sleep. The term was careful, clinical. What it described was remarkable.

What the brain is actually editing

The mechanism becomes clearer when you look at what memory consolidation actually involves.

Every emotional experience you have is encoded twice. The first encoding happens fast, in the heat of the moment, and it carries the full emotional freight — the cortisol, the adrenaline, the physical sensation of fear or grief or anger layered into the neural trace. This encoding happens in the amygdala, the brain's threat-detection and emotional-tagging center. The amygdala is not subtle. It stamps experiences with emotional intensity the way a customs officer stamps a passport — quickly, emphatically, without nuance.

The second encoding is slower. It happens during sleep — particularly during REM — when the hippocampus, which handles memory storage and retrieval, begins moving the day's experiences into longer-term storage. This transfer is not passive. It is not a simple copy-paste. It is more like a re-editing, in which the brain strips certain metadata from the original file before archiving it.

The metadata it strips, in successful emotional processing, is the acute distress signal. What remains is the information — what happened, who was there, what it meant in context — but the raw, visceral, present-tense pain of the experience is reduced. This is why a memory that was agonizing at three weeks can feel bearable at three months. Sleep — specifically REM sleep — has been working on it every night in between.

A 2011 study by Els van der Helm and Matthew Walker at UC Berkeley made this concrete. They showed participants emotionally disturbing images and measured their amygdala response. Then they split the group: one half slept normally, one half were kept awake. The next day, they showed the same images again. In the group that had slept, the amygdala response to the images was significantly reduced — the brain had processed the emotional charge overnight. In the group that had stayed awake, the amygdala fired just as intensely as it had the first time. Without sleep, the brain had made no progress on the emotional load.

The images hadn't changed. The memories hadn't changed. What had changed — or failed to change — was the emotional tag attached to the memory. REM sleep was the agent of that change.

This is where Cartwright's earlier clinical work and the neuroscience of a later era start to read as chapters in the same story. She saw the process from the outside — in dream reports, in morning mood assessments, in the difference between women who were slowly recovering from a devastating loss and women who were not. Walker and van der Helm were watching the same process from the inside, in neural imaging data, watching the amygdala cool as REM sleep did its work.

The paradox of forgetting

So why do we forget our dreams?

This is the question the previous teaser was pointing toward, and the answer is less mysterious than it sounds once you understand the mechanism.

When a dream is functioning properly — when the REM session of emotional re-editing completes — the brain has done what it needed to do. The memory has been re-filed. The emotional charge has been reduced. The experience has been integrated into the broader landscape of autobiographical memory, given context, given proportion, connected to older related experiences that help the brain categorize it.

The dream itself — the particular sequence of images and characters and impossible logic that served as the processing session — is no longer needed. It was the workspace, not the product. Keeping the workspace would be like a surgeon keeping the gauze. The surgery is done. What matters is what healed.

Forgetting the dream is, in this frame, the signature of success. The brain discards the working copy because the permanent copy — the memory with its emotional charge metabolized — has been filed correctly.

This is a radical reframe for anyone who has been told that dream recall is a measure of sleep quality, or that failing to remember dreams means something was lost. The evidence points in almost the opposite direction. Persistent, vivid dream recall — especially of emotionally intense dreams that feel unresolved — is often a sign that the processing is incomplete. The brain is still trying to close the file. It hasn't finished yet.

Cartwright found this in her data from the Chicago lab. The women who were recovering well from their divorces did remember some dreams early in the process — particularly the strange, collaged dreams that blended the painful recent material with older memories. But over weeks and months, as the recovery progressed, dream recall faded. Not because they were sleeping less deeply, but because the processing was completing. The dreams had done their work. They were being forgotten because they had succeeded.

The women who were not recovering were remembering their dreams vividly. Every morning, a clear report of the same loop, the same cast of characters, the same unresolved emotional register. The memory was not being edited. It was being replayed without progress.

The dream that tells you it's working, and the one that tells you it isn't

There is a clinical observation embedded in Cartwright's work that has an almost uncanny accuracy when you hold it against your own experience.

A dream that incorporates emotionally difficult material alongside older, unrelated, contextually rich memories — the strange dream where your ex-husband somehow appears at your childhood home alongside your third-grade teacher and an unfamiliar coastline — is a dream where the processing is likely happening. The brain is doing exactly what it's supposed to do: pulling the new painful material into contact with older emotional memories, giving it context, giving it proportion, reducing its singularity. It feels weird. It is weird. It is also the mechanism working.

A dream that replays the difficult material straight — same faces, same argument, same feeling of helplessness or grief, same resolution-less ending — is a dream where the processing is struggling. The brain is attempting the work but not completing it. It may be that the emotional load is too heavy, or that something in the REM architecture is disrupted — fragmented sleep, alcohol metabolism, elevated cortisol — that is preventing the full norepinephrine withdrawal that makes emotional reprocessing possible.

This distinction has immediate practical weight. It means that the quality of your dreams is not a fixed trait of your neurology. It is a variable that responds to the conditions of your sleep. A night of fragmented REM — broken by waking, compressed by alcohol, cut short by an early alarm — is a night where the editing session was interrupted. The file stays open. The emotional charge remains.

As we covered in Issue 014 on alcohol's effect on sleep architecture, even moderate evening drinking suppresses REM in the first half of the night, then produces a rebound of shallow, disrupted REM in the second half. The rebound REM is not the same as the uninterrupted version. It is more like a photocopied copy than the original — lower resolution, less effective at the fine work of emotional re-encoding.

Why this matters more after 50

There is a cruel irony in the way REM sleep changes as we age, because it changes in exactly the direction that makes the emotional-processing work harder to complete.

REM sleep declines with age. This is well-documented and well-studied. A paper by Aging and Sleep researcher Sebastiaan Overeem and colleagues at Eindhoven University of Technology, reviewing decades of polysomnographic data, confirms the pattern: total REM sleep duration decreases steadily from young adulthood through the fifties and sixties. The architecture also changes — REM periods that used to be long and sustained become shorter, more frequently interrupted, more likely to be cut short by waking.

For a woman in her fifties lying awake at 3am, this is not abstract. The REM periods that do the deepest emotional processing happen predominantly in the second half of the night — the hours between roughly 3am and 6am. This is when REM cycles lengthen, when the dreaming is most vivid, when the brain does the majority of its emotional-memory work for the night.

Waking at 3am and staying awake — or drifting in and out of light, anxious half-sleep — is not merely losing rest time. It is specifically disrupting the window when REM does its most important work. The files that needed editing don't get edited. The emotional charge that was supposed to be metabolized overnight carries forward into the morning, into the day, into the next night.

This is why grief, anxiety, and emotional overwhelm tend to feel so much more persistent and crushing during periods of poor sleep. It is not that bad sleep makes you more emotionally fragile in some vague, general sense. It is that the specific overnight process by which difficult emotional memories are stripped of their acute charge is being interrupted — and the charge is accumulating.

Cartwright saw this in her clinical patients. She saw that the women who were sleeping through the night, getting full REM cycles, were recovering from devastating losses at a meaningfully faster rate than the women whose sleep was fragmented. The difference was not therapy. It was not attitude. It was sleep architecture. It was whether the brain was getting the hours it needed to complete the editing process.

For women in midlife and beyond, the hormonal shifts of perimenopause and menopause add another layer of complication. Declining estrogen affects the depth and stability of sleep, increases the frequency of nighttime waking, and — crucially — increases the likelihood of elevated nighttime cortisol, which competes directly with the norepinephrine withdrawal that makes REM's emotional processing possible. A brain that can't fully lower its norepinephrine can't fully complete the editing session.

The 3am wake is not random. It is not weakness. It is a confluence of age-related REM compression, hormonal disruption, and the particular cruelty of nighttime cortisol elevation — a cascade that tends to peak in the midlife years and that directly targets the window of sleep most responsible for emotional renewal.

Knowing this does not fix it. But it changes what it means. The 3am wake is not a sign that something is broken beyond repair. It is a sign that a specific, identifiable process is being disrupted — and that the disruption, understood correctly, points toward specific, identifiable interventions.

· · ·

Until next issue

Next: The hidden reason why your body temperature drops before you fall asleep — and what happens to your sleep when it can't.

---

About Beezy Beez. Beezy Beez crafts botanical extract honey for people navigating sleep changes after 50. The Hive Mind is the brand's editorial letter on the science and history of rest.

← Back to the Sleep Science Hub