This video masterfully deconstructs the biological machinery of insomnia, turning a midnight crisis into a clear and fascinating lesson in neurochemistry. It is a rare example of high-level science made genuinely accessible without sacrificing intellectual rigor.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
Why YOUR Brain Replays the Same Thought 47 Times Between Midnight and 2AMAdded:
It is 12:47 in the morning.
Your eyes are closed.
You have been in bed for 90 minutes. And the thought is back.
The same one. The one about what you said in that conversation.
The one about the bill.
The one about your child.
Again.
Don't move. Don't open your eyes.
Just notice. The thought arrives. You try to push it away. You start to drift.
And there it is again. The same shape.
Maybe the words are slightly different this time.
But it is the same animal.
This is not a failure of your willpower.
It is not a moral problem. It is a specific event happening inside your skull right now. And it has structure.
And we can describe it. I want to tell you what is going on inside your head between the hours of midnight and 2:00 in the morning. Because if you understand the machinery, you stop fighting yourself. Now then, let me tell you what you have inside your head.
At the back of your brain stem, behind the bone at the base of your skull, there is a small cluster of neurons called the locus coeruleus.
It is a tiny thing.
About 30,000 cells on each side. No bigger than a grain of rice.
It is colored faintly blue.
Which is why somebody named it that.
Locus coeruleus. The blue spot. This blue spot is the principal source of a chemical called norepinephrine for almost your entire brain. When the locus coeruleus fires, your cortex hears it. You feel it as alertness. You feel it as the wired feeling at 1:00 in the morning when sleep won't come.
Behind your forehead, just behind the bone, there is a region called the medial prefrontal cortex. In the middle of your brain, between the two hemispheres, there is a region called the posterior cingulate cortex.
And tucked between them, the precuneus.
These three structures, along with a few neighbors, form a network.
The default mode network. They light up together when your mind has nothing to do. They go quiet when you read a book or solve a problem.
They wake up when you stop.
Above each of your kidneys, sitting on top like a small hat, is the adrenal gland. The outer layer of that gland makes a hormone called cortisol.
Right now, at 12:47 in the morning, that gland is supposed to be putting out almost no cortisol.
The valley of the day.
The lowest point on the curve. Deep in your temporal lobe, on each side there is a curled up structure shaped like a seahorse. The Greeks named it that.
Hippocampus. Just in front of it, an almond-shaped cluster of neurons.
The amygdala. These two have been working all day encoding what happened to you, marking what mattered. They have a lot of homework to do tonight. And running along the blood vessels in your brain, there is a system of channels nobody knew about until 2013. The glymphatic system.
It flushes the waste products of yesterday out of your brain.
But it only runs when you are deeply asleep.
It is waiting for you.
It cannot start until you let go. These are the structures. Now I am going to walk you through what is happening to each one of them at this exact moment in your night. The first thing first, the second thing second.
Because they are connected.
They are not separate problems. They are one chain.
So, the first piece of the puzzle is cortisol.
You have heard about cortisol.
Stress hormone.
People say it like it is the villain of every story.
But cortisol is not a villain. It is a beautifully tuned signal that follows the rotation of the earth. When the sun comes up, cortisol comes up. 30 to 45 minutes after you wake, cortisol peaks.
Through the day it falls.
By the time you go to bed, it is supposed to be down at the bottom.
The valley. A man named Bruce McEwen at Rockefeller University spent his career studying what happens when this curve gets stuck.
He called it allostatic load. The body does not pay the cost of any single stress.
It pays the cost of the curve never coming all the way down.
If your cortisol does not drop to its nadir at midnight, the system that runs on the rhythm cannot work right. And here is the thing.
The cortisol awakening response in healthy people is between 38% and 75%.
That is a real measurement. Saliva, every 15 minutes in laboratory studies.
38% to 75%.
That rise is supposed to come at 4:00 or 5:00 in the morning.
Not midnight. The mechanism behind the curve is worth understanding for a moment. Above the brainstem, in a region called the hypothalamus, there is a small group of cells that produces a peptide called corticotropin-releasing hormone.
The hypothalamus sends that peptide a few millimeters down to the pituitary gland, the gland that hangs from the underside of your brain like a little berry. The pituitary releases a different hormone, adrenocorticotropic hormone, into your bloodstream.
That hormone travels all the way down to the adrenal glands, sitting on top of your kidneys. And the adrenal cortex, in response, releases cortisol.
This whole circuit is called the HPA axis.
Hypothalamus, pituitary, adrenal.
Three stations, three handoffs.
And the entire cascade is coordinated by your master clock, the suprachiasmatic nucleus, sitting just above the place where your two optic nerves cross. When you are healthy and the rhythm is intact, the cascade fires hardest at 5:00 or 6:00 in the morning, building you up for the day.
By midnight, the whole circuit is at low ebb, cortisol below 2 micrograms per deciliter.
The body is supposed to be metabolically dormant.
But in chronic stress, McEwen showed, the negative feedback loops that should turn the cascade off start to misbehave.
The brain stops listening to its own quiet signal.
The midnight nadir does not arrive.
Cortisol stays at 4, 5, sometimes higher.
The body never gets the chemical message that it is night.
But for somebody whose baseline cortisol is already elevated, somebody under chronic stress, somebody with a worry that will not put itself down, the curve never reaches the valley.
It hovers. It stays high enough that the brain reads it as daytime.
The signal does not say sleep. The signal says alert.
Now, here is the part that textbooks don't make obvious.
The cortisol rise is not just a wake-up timer.
It is a permission slip for the rest of the body's stress system to fire.
When cortisol is up, the heart rate variability flattens.
The blood vessels narrow.
The mind feels braced for something.
You lie there, eyes closed, in a perfectly safe bed, and your endocrine system is putting on a coat and looking for the door.
While that hormone is failing to drop, another system in your brain is failing to start.
I want you to picture something.
Imagine your brain at the end of a long day. The neurons have been firing all day, and just like any cellular machine that runs hard, they have produced waste. Proteins that have folded wrong.
Misfolded amyloid beta.
Tau fragments.
Metabolic byproducts. All of it has to be cleared, and the brain does not have the lymphatic system the rest of your body has. Until 2013, nobody knew how the brain cleaned itself.
Then a researcher named Maiken Nedergaard at the University of Rochester published a paper in the journal Science. She found that during deep sleep, the brain cells themselves shrink by about 60%. The space between the cells opens up and cerebrospinal fluid, the fluid that surrounds your brain, washes through the tissue along the channels that run beside your blood vessels. She called this the glymphatic system.
Glia plus lymphatic. And during slow wave sleep, this washing system runs about 10 times more actively than when you are awake.
Tenfold, not a little more.
An order of magnitude.
The fluid flows because of water channels called aquaporin 4 embedded in the membranes of star-shaped cells called astrocytes.
In animals where you knock those channels out, the clearance drops by about 70%.
70%.
The brain stops cleaning.
And what does it not clean?
Among other things, amyloid beta, the protein that over many years accumulates in Alzheimer's disease. I want you to sit with that for a moment.
The protein associated with the most feared neurodegenerative disease of the 20th and 21st centuries is washed out of your brain at full power only when you are asleep. There are decades of epidemiological evidence linking chronic poor sleep to dementia risk and the glymphatic system is the leading mechanistic candidate for the link.
It is not the only candidate, but it is the one that explains the directionality.
Sleep does the work.
Wake does not. The cells shrink.
The fluid flows.
The waste exits along the perivenous channels, eventually reaching the lymphatic vessels in the meninges.
And from there into the rest of the body's drainage system. Now, here is what matters for you lying awake at 1:00 in the morning.
The glymphatic system does not run while you are awake.
It runs when you are in slow-wave sleep, the deep, dreamless kind, the kind you have not entered yet because the cortisol is still up and the locus coeruleus is firing. So, while you lie there, the toxins from yesterday are not being washed away.
The cleaning crew is standing in the hallway waiting for the lights to go out.
They cannot start until you let them. I should tell you that there is some controversy here.
In 2024, a separate group published a paper claiming the glymphatic system actually clears more during wake.
Nedergaard's group fired back showing the opposing experiment was measuring the wrong thing.
As of right now, the weight of the evidence is on her side. But, this is honest science. The story is being refined. The basic picture that the brain clears waste differently in sleep than in waking is solid. The rate, the tenfold figure, is what is being argued about. Keep that in mind. Now, why does the same thought keep coming back? To answer that, I have to tell you about the most counterintuitive discovery in neuroscience in the last quarter century. In 1998, a man named Marcus Raichle at Washington University in St. Louis was looking at brain scans.
He was trying to find which regions of the brain became more active when a person performed a task. Read a word, solve a math problem, recognize a face.
He found something he was not looking for. There was a set of regions that became less active when the person did a task and more active when the person did nothing. He coined the term default mode in 2001.
The default mode network, the network the brain runs when you give it nothing to do.
The medial prefrontal cortex behind your forehead the posterior cingulate in the middle of your brain the angular gyrus on the side. These regions all light up together when you stop.
And what is the network doing? It is thinking about you.
It is constructing your narrative.
It is replaying things you said, rehearsing things you might say, comparing yourself to other people, asking what kind of person you are. The default mode is the I network. It is where the self is built. In 2009, a researcher named Evette Sheline, also at Washington University, published a paper in the Proceedings of the National Academy of Sciences. She showed that in people with depression, the default mode network is hyperconnected. The nodes are talking to each other too much. And this is the important part. The network does not know how to shut off. It keeps running.
It keeps recycling self-referential content. The clinical name for this is rumination.
The folk name is a thought loop.
Sheline's group put depressed patients in scanners and showed them sad faces, neutral faces, words about themselves, words about other people. In a healthy brain when you ask the prefrontal cortex to read a word about somebody else the default mode quiets.
In a depressed brain, it does not. The default mode keeps the floor. Whatever stimulus you give the brain, it bends back to the self. The MPFC stays loud.
The posterior cingulate stays loud. The angular gyrus joins in. The connectivity between them does not break. I want you to think about a particular kind of person now.
The person who is not necessarily depressed, but who is 52, has a long marriage, two adult children, a job with deadlines, an aging parent in another country, the person whose partner falls asleep within 4 minutes of turning out the light, the person who has tried the lavender oil, the chamomile tea, the white noise machine, the weighted blanket, the melatonin gummy, the person who has read three books on sleep hygiene, who knows about screens before bed, who has cut out caffeine after noon, and who still wakes up at 1:00 in the morning with the same loop running.
This is a brain whose default mode network has too many open files. The HPA axis stays elevated because the workload of the day never closes. The amygdala has tagged dozens of small unresolved items. The default mode is doing its job.
It is the only thing trying to integrate all of it.
Sleep is what gets sacrificed. Now, look at what is happening in your bed at 1:00 in the morning. You have given your brain nothing to do.
You are lying still.
Your eyes are closed.
There is no task. The default mode network has the floor.
And in a brain where cortisol is already elevated, where the locus coeruleus is already firing, the default mode does not just hum quietly. It runs hot.
It picks up the thread it dropped this afternoon.
It picks up the bill you have not paid, the conversation you replayed in the shower, the thing you said to your child that you wish you had not said. And it goes around and around and around. The reason it is the same thought is that the default mode network has a finite library of unresolved material to chew on.
The amygdala has tagged certain memories as emotionally hot.
The default mode preferentially returns to hot tags.
There is no off switch.
There is only handing the floor back to the task positive network.
And you cannot hand the floor back because there is no task. You are in bed.
There is one place this is supposed to get fixed. And you are not getting there.
Let me tell you about REM.
In the 1950s, a researcher named Eugene Aserinsky was watching his son sleep.
And he noticed the eyes moving rapidly under the lids in cycles.
Rapid eye movement.
REM.
We have known about REM for 70 years.
We have known it is when most dreaming happens.
But what we did not understand for a long time is what REM does to your memory. A researcher named Matthew Walker at the University of California, Berkeley has spent a lot of his career on this question.
And he has put forward a model, supported by a lot of evidence, though still being refined, that REM sleep is the brain's nightly therapy.
Here is what he means by that.
During REM sleep, the locus coeruleus, that little blue spot at the back of your brainstem, falls silent.
Norepinephrine, the alertness chemical, drops to near zero.
It is the only time in the 24-hour cycle when this happens.
And during this window, this no norepinephrine window, the brain replays the day's emotional memories. The hippocampus and the amygdala fire together.
They walk through what happened.
But they walk through it without the stress chemical.
The story replays.
The fear is not there.
The hot tag gets turned down. Walker's group ran an experiment that gives the model some teeth.
They had people view emotionally charged images.
Then they sent half the participants for a full night's sleep. And they kept the other half awake.
The next day, both groups came back and saw the same images again while the researchers measured amygdala activation.
The group that had slept showed significantly less amygdala response.
The same images, the same eyes, but the emotional pull had decreased. The group that had stayed awake, same images, full amygdala response.
The pull had not decreased. Walker calls REM overnight therapy.
The mind processes the day's emotional content in a chemical environment that lets the charge bleed out of the memory.
You wake up the next day, and the thing that mattered yesterday matters less.
Not because you forgot it, because you metabolized it.
Now I have to tell you the catch. To get to REM, you have to fall asleep first.
You go through stage one, then stage two, then slow wave sleep, and then you cycle into REM.
Most of your REM sleep is in the second half of the night. If you do not fall asleep, if you lie there at 1:00 in the morning ruminating, you do not get to REM. The emotional memory does not get processed.
It stays raw.
Tomorrow morning, the amygdala will hand the same hot tag back to the default mode network.
And tomorrow night, when you lie down, the same loop will start.
And there is a cruel detail here. The longer you go without REM, the more REM you owe. Sleep researchers call this REM rebound.
After several nights of disrupted sleep, the brain pushes harder to enter REM, and the REM you do get is more intense, more emotionally vivid, sometimes more disturbing. The unprocessed material has to come out somewhere.
Sometimes it comes out as the strange, frightening dreams of the third bad night in a row. Your brain is trying to catch up on its homework.
And the reason you cannot fall asleep is the same reason your eyes feel wide open. The blue spot is loud.
The locus coeruleus has a job. The job is to wake you up if something dangerous is happening. A noise, a change in temperature, a thought tagged as threatening. In 1981, two researchers named Aston Jones and Bloom at the National Institute of Mental Health mapped the firing pattern of the locus coeruleus across the sleep cycle.
They found something specific. The locus coeruleus fires most while you are awake.
It fires less in non-REM sleep, and it falls almost completely silent in REM.
Then, near the end of REM, it gradually wakes up again.
Daniel Buysse, a professor of psychiatry at the University of Pittsburgh, has spent his career studying insomnia.
He has put forward what he calls the hyperarousal model.
The idea is simple.
People with chronic insomnia do not have a sleep problem.
They have an arousal problem. Their locus coeruleus does not quiet down at night.
It keeps firing.
The brain stays in the alert mode.
And once the alert mode is established, every thought gets amplified. Every worry gets a bigger neural footprint.
Every rumination has more horsepower behind it. Buysse's group, along with collaborators across multiple sleep labs, has measured this directly.
They have shown that people with chronic insomnia have higher whole body metabolic rates while trying to sleep.
Higher heart rate, higher core body temperature, higher cerebral glucose metabolism on PET scans, particularly in the regions that should be quietest.
The brain is burning fuel as if it were doing something. It is not. It is just running hot.
The locus coeruleus is the engine of that running hot state, and once it is running, it is hard to stop. Here is the trap. The default mode network running in rumination mode produces emotionally charged content. Emotionally charged content keeps the locus coeruleus firing.
The firing locus coeruleus prevents you from dropping into the slow-wave sleep that would shut down the default mode.
So, the rumination keeps going, and the firing keeps going, and the firing prevents the rumination from ending. And around it goes.
This is what engineers call a positive feedback loop.
The output of the system feeds back into the input, and each cycle reinforces the next. Most positive feedback loops in biology have a built-in break.
Hormones get cleared.
Receptors downregulate. Neurons fatigue.
But, the late-night thought loop has weak brakes. The default mode does not exhaust itself easily.
The locus coeruleus can fire for hours without slowing.
Cortisol takes a long time to clear, so the loop runs on. Sometimes for the whole window between midnight and 2:00 in the morning, sometimes longer.
And the same thought keeps showing up because the loop has nowhere else to go.
That is your 1:00 a.m. brain.
That is the loop.
Cortisol failing to drop. Lymphatic system failing to start.
Default mode network running self-referential content. REM not arriving to discharge the emotional charge.
Locus coeruleus keeping the lights on.
Five mechanisms. One chain. Each one feeding the next. Now, let me pause here because this is important. I have told you a story.
I have given you names of researchers and numbers and journals.
But, I want you to know exactly how confident to be in each piece because not all of these are equally settled. The cortisol curve and its early morning rise that is replicated across thousands of studies across decades across many laboratories. You can take that to the bank.
The locus coeruleus arousal mechanism, that is also strongly evidenced in humans and animals going back 40 years.
The default mode network and its role in self-referential thought.
That has been replicated more than I can count.
Marcus Raichle's discovery has reshaped a whole field.
The hyperconnectivity of the default mode and depression, that is well supported with multiple independent groups. The rumination connection, that is a strong correlation and the direction of cause is still being worked out. But the association is there.
The glymphatic system and its enhancement in sleep. That is a powerful model.
And it is what most experts in the field believe. But there is an active debate about how to measure clearance.
And a paper in 2024 challenged some of the original findings.
Nedergaard's group and others have responded.
The story is real.
The exact rate is being refined. Treated as a strong hypothesis with a body of supporting evidence, not a closed case.
Walker's overnight therapy hypothesis that REM strips the emotional charge from memory because of the absence of norepinephrine.
That is a beautiful model supported by some clinical data and some neurochemical work. It is not yet at the level of certainty that the cortisol curve has.
It is the working frame in the field and it is plausible, but if you read the same studies in 5 years, the picture may be more nuanced.
I am telling you this because the worst thing a science writer can do is oversell.
If you understand which pieces are bedrock and which pieces are still being polished, you understand the topic better than most people who write about it.
Now then, let me tell you the part that should make you a little angry.
The pharmaceutical industry has a class of drugs called Z-drugs.
Zolpidem, eszopiclone, zaleplon. They They between $200 and $400 a month in the United States without insurance, more in some places.
These drugs work by binding to a specific subunit of the gamma-aminobutyric acid receptor, the GABA receptor.
GABA is your brain's main inhibitory neurotransmitter.
It tells neurons to be quiet.
When a Z-drug binds the GABA receptor, neurons fire less, including the neurons of the locus coeruleus.
The blue spot quiets.
The cortex stops getting hosed with norepinephrine.
The default mode network slows down.
The cognitive arousal drops. You fall asleep. It is an effective intervention.
I am not going to tell you it does not work.
It does work in the short term for many people.
And I am not going to tell you that pharmaceutical sleep aids have no role.
They have a role, but here is what I want you to see.
Your own brain has a free version of this exact mechanism.
Your own brain produces GABA.
Your own brain has a sleep architecture that when undisturbed, walks the locus coeruleus down from waking arousal into the silent state of REM.
The same molecular endpoint, the same quiet cortex, the same emotional discharge. The body's version takes longer to start. The pharmaceutical version starts in 30 minutes.
That is a real difference. And the cognitive intervention that is first-line treatment for chronic insomnia, cognitive behavioral therapy for insomnia, what the researchers call CBT-I, operates on the same loop. It is not a pill.
It is a behavioral protocol developed by Daniel Buysse and others over the last 30 years.
The protocol does several things. It tells you to get out of bed if you are awake more than 20 minutes, so the bed does not become associated with arousal.
It tells you to write the thought down on paper, so the default mode network has something to hand off to. It tells you to keep the bedroom dark, so the cortisol curve is not pushed earlier. It tells you to avoid clock checking, so the prefrontal cortex does not generate panic about how little time is left. The reason the protocol works is that every step lowers cognitive arousal.
Lower cognitive arousal means lower locus coeruleus firing. Lower locus coeruleus firing allows slow-wave sleep to start.
Slow-wave sleep allows the glymphatic system to wash.
Slow-wave sleep precedes REM. REM strips the emotional charge from the memory.
The next night, the default mode does not have as much hot material to chew on.
The cognitive behavioral therapy for insomnia protocol has efficacy comparable to pharmaceutical sleep aids over the long term.
The cost of the protocol is the cost of a workbook.
Sometimes the cost of one consultation.
The dose response is slower. It takes about 4 to 8 weeks to take hold.
You will not feel rescued tonight.
The drug will. That is the honest dose calibration. But the molecular pathway is the same pathway.
You are not buying access to a mechanism that your body does not have. You are buying speed.
Now, there is one more piece of this I want you to hear because the part that will help you tonight is not a drug, and it is not a 4-week protocol. It is something simpler.
Your grandmother told you when you were a child that you should sleep on it.
That whatever was bothering you would look better in the morning.
She was right.
She was telling you the conclusion of a chapter that the science would not write for another 70 years.
She was saying, and she did not know the molecular basis, but she had observed the effect across a whole life, that the morning version of a problem is less hot than the late night version of it. What she was describing was the emotional discharge of REM sleep.
She was describing the cortisol curve falling, then climbing into the morning rise, then settling into the ordinary daytime band. She was describing the default mode handing the floor back to the task positive network as you got up and brushed your teeth and made breakfast.
She was describing the glymphatic clearing of yesterday's protein waste.
She did not have the words, but she had the pattern.
Sleep on it.
Your grandfather may have told you something else.
He may have said, "Your bed is for sleeping."
He was not being old-fashioned. He was telling you the foundation of CBTI.
If your bed becomes the place where you ruminate, the bed itself becomes a cue for arousal.
Your locus coeruleus learns to fire when you lie down. Keep the bed for sleeping, and the bed becomes a cue for sleep.
Some old family doctor, somewhere in your past, probably told you not to look at the clock when you wake up at night.
That was not folksy advice.
That was a precise neurological instruction.
Looking at the clock activates the prefrontal cortex's anxiety circuitry around lost time, which raises cortisol, which raises locus coeruleus firing, which reinforces the loop you are trying to break. The folk wisdom of an older generation, in the case of sleep, has been almost perfectly validated by modern neuroscience.
The grandmothers were right.
The science was slow.
So, it is 1:00 in the morning.
The thought is back.
Here is what I want you to do.
Name what is happening.
Out loud, in a whisper, to nobody but yourself. The cortisol has not dropped to its valley.
The glymphatic clearing has not begun.
The default mode network is running self-referential content.
The locus coeruleus is firing. REM has not arrived to discharge the memory.
Five mechanisms, one chain.
Then get out of bed.
Sit somewhere dim.
Write the thought down on a piece of paper.
The act of writing it transfers it from the default mode network to the task positive network. The default mode does not need to hold it anymore. It is on the page.
The page can hold it until tomorrow. You do not have to.
Then go back to bed when you feel sleepy, not before. Sleepy. Heavy eyelids.
Slow blink.
The body asking for it. The thought will probably come back tomorrow night.
That is fine.
The mechanism is built into your skull.
It is not going anywhere. But each night you understand it, you fight it less.
And each night you fight it less, the loop is a little shorter. Sleep on it.
Tell me, when the thought comes back tonight, what is the shape of it?
Related Videos
Why can’t Trump take sleep meds?
concussiontalks_slp
14K views•2026-05-29
Recovery pronouns. Neuroplasticity & practical neuroscience tips to help recover from pain & fatigue
Fantasticneuroplastic
907 views•2026-05-31
I Saw the Thing Crash. Then I Lost Hours | Beyond Black Budget
BeyondBlackBudget
148 views•2026-05-30
Neuroanatomy of smell (olfaction)
SamWebster
644 views•2026-05-28
women never forget when you upset them
healsick
745 views•2026-06-01
Your Brain Is Actively Deleting Your Childhood Memories! 🧠🗑️ #Shorts #Anatomy #DidYouKnow
voiceless2345
225 views•2026-06-01
What are you looking at
SuperStaticPro
1K views•2026-05-31
Why Trauma Doesn’t Just 'Go Away'
historyofsimplethings
1K views•2026-05-28











