For most of the last century, medicine

told you a hard, flat thing. You are

born with all the brain cells you will

ever have, and from there it is only

downhill. Every year you lose some, and

they never come back. That was the

settled science. Doctors taught it.

Textbooks printed it, and it was wrong.

Here is what I'm going to show you,

because you clicked on proof, and I

intend to deliver proof, not comfort. In

the next few minutes, I'm going to walk

you through the actual studies,

including the two famous laboratories

that looked at the same human brains and

reached flatly opposite conclusions, a

real fight that is still not fully

settled, and then the study from July of

2025 that may have ended the argument by

catching new brain cells in the very act

of dividing inside a 78-year-old brain.

And stay till the end, because I owe you

the honest part, what this proof does

promise you, and what it absolutely does

not. The gap between those two is where

every overblown headline lives, and I am

not going to leave you there. One note

before the evidence, because it is a

brain and memory topic, and honesty is

the whole brand here. This is

educational information from published

research, not a diagnosis, and not a

treatment for any condition. And who is

talking? A researcher who reads this

literature, not a clinician with

patients. This particular corner of

neuroscience is one where good

scientists genuinely disagree, and I'm

going to show you that disagreement

honestly, rather than pretend it is

tidier than it is. That is the

interesting part anyway. Let's get into

it. Now, before the nuclear bomb story,

which is the famous one, and I'm getting

to it, I owe you the very first crack in

the wall, because it came earlier, and

it came in a way almost nobody expected.

The year is 1998. A researcher named

Peter Eriksson, working with a lab run

by Fred Gage, published the first real

evidence that new neurons form in the

adult human hippocampus, the memory

region, the seahorse-shaped structure

buried on each side of your brain that

turns today into a memory you can find

tomorrow. And the way they proved it is

worth pausing on because it tells you

how carefully this had to be done. There

was a group of cancer patients who, as

part of their care, had been given a

compound called BrdU. Think of it as a

dye that only gets stitched into the DNA

of a cell at the moment that cell is

being freshly made. It was given for

medical reasons that had nothing to do

with the brain,

but it left a marker behind. So, after

these patients died, that lab could look

into the memory region, hunt for the

dye, and ask a question no one had been

able to ask cleanly in a living human

before. Are any of these neurons new?

And the answer was yes. Newly born

neurons in the hippocampus of adults,

some of them well into their 60s and

70s. That was the first hole punched in

a 100-year-old dogma, and hold on to

that name, Gage, because his

fingerprints show up again in a minute

in a way I find genuinely moving. But,

here is why 1998

did not settle it. It answered, "Do new

neurons exist?" It could not answer how

many or how fast or whether it kept

going year after year. It was a snapshot

in a small handful of people who

happened to have that dye in them. To

turn a snapshot into a measurement,

science needed a way to read the birth

date of a brain cell across a whole

life.

And that finally is where the bomb tests

come in. Above-ground nuclear testing in

the 1950s and early '60s dusted the

whole planet with a faint, harmless

trace of a carbon isotope, carbon-14.

Every living thing absorbed it. And here

is the clever part. When a new cell is

born, it locks a snapshot of the

atmosphere's carbon into its DNA like a

date stamp, frozen at the moment of

birth. So, a team at the Karolinska

Institute in Sweden, publishing in the

journal Cell in 2013, had a way to read

the birth date of a brain cell. They

could take neurons from the memory

region of adults who had died, read the

carbon stamp, and ask, "When were you

born?" If the old dogma were true, every

neuron in an 80-year-old would carry a

birth stamp from the 1930s, all born

before adulthood, none since. That is

not what they found. They found neurons

stamped with recent dates, cells born in

adulthood, in middle age, in old age.

And now the numbers, because this is the

study that turned belief into

arithmetic. They calculated that this

memory region, the hippocampus, adds on

the order of 700 new neurons a day, and

keeps doing it at a slow and steady rate

into old age. They put a turnover figure

on it, too.

Within the pool of neurons that actually

renews, roughly 1 and 3/4% gets

exchanged every year, old cells

retiring, new cells arriving. And most

strikingly, they estimated that about a

third of the neurons in that region

belong to this renewing pool at all,

meaning a real, sizeable fraction of

your memory hardware is not the same

tissue you were born with. It is being

swapped quietly, decade after decade,

700 a day in a structure most people

were told stopped growing before they

finished school. That was the first

hard, physical, quantified proof the

adult human brain builds new cells. It

is not all downhill. And then, because

science is not a straight line, it got

contentious. In 2018, within about a

month of each other, two respected teams

published studies on exactly this

question and came to opposite

conclusions. Let me give you both in

full, because the detail is the whole

point, and I would rather you leave able

to argue this at your own kitchen table

than take my word for it. The first team

was at Columbia University, publishing

in Cell Stem Cell. They looked at

healthy human hippocampus brains from

people ranging from 14 years old all the

way to 79 and they reported finding

thousands of immature freshly made young

neurons at every single age including in

the very oldest brains. Neurogenesis

persists across the whole lifespan they

said. But here is why I trust this

team's honesty. They did not claim

nothing changes with age. They reported

that two supporting things clearly do

decline as we get older. The growth of

new small blood vessels feeding that

region and the size of the dormant

resting reserve of stem cells waiting to

be called into service. In other words,

the factory keeps producing but the

machine slows down. Fewer fresh supply

lines, a smaller bench of reserves,

persistence but not youth. I find that a

very believable shape for a biological

process. And almost simultaneously a

group at the University of California,

San Francisco published in the journal

Nature and reported close to the

opposite. They examined 59 brain samples

spanning an enormous range from before

birth in fetal tissue up to age 77 and

they traced a steep brutal decline in

young neurons. Something like 1600 of

them per square millimeter at birth

collapsing to a bare handful by around

age 13 and then by their count

undetectable in adults at all. Their

explanation for why other labs kept

seeing new cells was pointed and

specific. They argued the other teams

were miscounting that some of what was

being tallied as brand new neurons were

really just ordinary support cells, the

brain's maintenance and scaffolding

cells being mistaken for something they

were not. I want to sit in this

disagreement honestly with you for a

second because this is exactly the kind

of moment most health channels smooth

over and I think that is a disservice.

Two good labs, the same question,

opposite answers published a month

apart. When I first read them side by

side, my honest reaction was discomfort.

I wanted one of them to be obviously

sloppy, and neither was. One says the

factory runs your whole life. One says

the factory is bricked up and cold

before you finish middle school.

Both use real human brains. Both use

serious methods. So, the real question

became not who is right, but why would

careful people get opposite results from

the same kind of tissue? And the answer,

when it came, turned out to be almost

mundane. And that is where the fight

started to resolve. In 2019, a team in

Spain published what I think is the

hinge of this whole story in the journal

Nature Medicine. They found thousands of

new young neurons in healthy people into

their 80s and 90s. And here the numbers

matter, so let me give them to you. In

the youngest healthy brains they looked

at, they counted something on the order

of 40,000 of these new young neurons per

square millimeter. In the oldest healthy

brains, that number had fallen, but only

to around 25,000.

Fallen, yes. Vanished, absolutely not. A

real decline with age, and a floor that

was still enormous in the very old. That

alone is a striking result. But they

also found the thing that explained the

contradiction between the 2018 labs, and

it was almost embarrassingly ordinary.

It was about how the brain tissue was

handled after death. If the tissue was

left sitting in preservative chemical

too long, more than about 12 hours, the

very signal that marks these new young

neurons faded away and became invisible.

12 hours in the wrong bath, and the

fingerprints were gone. The neurons were

there the whole time.

The preservation had simply erased the

evidence of them. The lab that found

nothing had, in all likelihood, been

looking at tissue where the marks had

already been wiped off before anyone

counted. Same brains, in effect, but one

handling method quietly destroyed the

very thing it was trying to measure.

That is not fraud, and it is not

stupidity. It is how genuinely hard

science actually gets sorted out,

slowly, through method, through the

unglamorous business of arguing about

tissue handling times. And notice what

that does to the fight. It does not

declare one 2018 team the villain and

the other the hero. It reframes the

whole thing. It says the disagreement

was probably never about the biology at

all. It was about a preparation step

nobody had flagged as decisive. The most

important variable in the argument

turned out to be a clock and a chemical,

not a truth about your brain. I keep

coming back to that because it is such a

clean lesson in how easily a real signal

can be scrubbed out by accident, and how

a single careful control can rescue

years of contradictory data. That same

Spanish study added something that

reaches past "Do these cells exist?"

toward "Do they matter?" And this is

where my attention really sharpened. The

number of these new young neurons

dropped sharply, and dropped early, in

people with Alzheimer's disease, and

kept falling further as the disease

advanced. So, it was not just aging that

lowered the count. The disease itself

tracked with fewer new cells, beyond

what age alone would predict. That is a

genuinely provocative finding, and I

will come back in a moment to exactly

how much weight it can bear because it

cannot bear as much as a headline wants

it to. Then, in the same stretch of

years, another study, this one back in

Cell Stem Cell, looked at a group of

people whose average age was around 90.

Around 90, and they were still hunting

for the markers of newborn neurons. And

they found them. Signs of new neuron

birth present even in people in their

90s. But the finding I want you to hold

is this one. The amount varied

enormously from person to person. Not a

tidy line where everyone at 90 has the

same small trickle. Some of these very

old people had a good deal of this young

neuron activity. Others had strikingly

little. And here is the part that leans

toward mattering. The people who carried

more of these young neurons tended to

have better cognitive status, sharper

standing on the memory and thinking

measures than those who carried fewer. I

have to flag the honesty here hard and I

am going to flag it every time it comes

up because it is the exact spot where

this field gets oversold. That is a

correlation. More new neurons went

together with sharper minds. It does not

prove the new neurons cause the sharper

mind. It could run the other way. Or a

third thing could drive both. But it is

the second real hint, alongside the

Alzheimer signal, that the building and

the remembering are somehow linked, not

just a biological curiosity happening

off in a corner where it changes

nothing. So let me draw the line under

where we are before the newest chapter

because the shape matters. 1998

New neurons exist in adults. 2013

Here is the rate, 700 a day, into old

age. 2018 A real and public fight about

whether that is even true. 2019

The fight starts to resolve on a tissue

handling technicality, plus the first

hints that the count tracks with disease

and with cognition. Every one of those

steps, except the last two I owe you a

warning about, they mostly counted

young-looking neurons and then reasoned

backward that they must be newly made.

That inference was strong, but it was an

inference. And the critics in San

Francisco had put their objection in a

way that was, frankly, fair. Show me the

actual dividing parent cells caught in

the act of making a new neuron, or I

will keep believing you are miscounting

mature tissue. For years, nobody could

fully meet that challenge in the adult

human brain. Hold that thought because

that is precisely the door the 2025

study kicks open. Now the freshest

proof, the reason I am making this video

in 2026 and not 2 years ago. In July of

2025, the same Swedish institute that

started the modern chapter with the

carbon dating Karolinska published a

study in the journal Science that did

something none of the earlier work

could. And I will name the lead

scientists because credit belongs on the

record.

The work came out of the group led by

Jonas Frisén

with a researcher named Marta Paterlini

Dumitru among those driving it. Remember

I told you to hold on to the name Gage

from 1998

and Frisén from 2013? This is the same

lineage of researchers coming back a

quarter century later to finish the

argument they helped start. Same

problem, later decade, better tools.

There is something in that I find

quietly stirring a scientific question

handed down and finally closed by the

same hands that opened it. Here is what

they actually did and why it is

different in kind, not just degree.

The earlier studies mostly counted young

neurons and inferred they must be new.

This one went after the parent cells,

the dividing precursor cells, the neural

progenitors that give birth to new

neurons and tried to catch them

mid-division. To do it, they stacked

four different tools on top of each

other. First, they read the full genetic

transcript of individual cell nuclei,

one nucleus at a time, over 200,000 of

them across people spanning ages 13 to

78. Think of that as reading each cell's

complete job description in its own

handwriting instead of judging it by a

single blurry stain. Second, they used a

marker, its name is Ki-67, that only

lights up in a cell that is actively in

the process of dividing right now, not a

cell that looks young. A cell caught in

the act. Third, they brought in machine

learning to sort and classify the cell

types at a scale no human eye could. And

fourth, they used spatial mapping to pin

down exactly where in the tissue these

cells were sitting, and they found them.

Actively dividing neural progenitor

cells, the parent factory cells

themselves in the adult human

hippocampus, specifically in the part

called the dentate gyrus, which is

exactly the subregion where new neurons

are supposed to be born. The oldest

person in whom they caught these

dividing parent cells was 78 years old.

78, and the cell factory was still

running its line. And the same honest

note that keeps recurring, because the

biology keeps insisting on it, they

found huge variation between

individuals. Some adults had a rich

population of these dividing cells. Some

had almost none. Not one uniform gift

handed equally to every human, a real

large person-to-person spread. This is

the study that answers the 2018

objection head-on. The San Francisco

team had essentially said, "Show me the

dividing precursor cells, or I do not

believe it." And this study shows them

not by one ambiguous stain that skeptics

could wave off, but by the cells' entire

molecular fingerprint, its complete

genetic signature caught mid-division,

and mapped to the exact right spot in

the tissue. That is a different order of

evidence than counting shapes under a

microscope and arguing about what they

are. That is why I would say, if you ask

me where I land, the weight of the

evidence now sits clearly on the side of

yes, the adult human brain, including

the old adult human brain, keeps making

new neurons, and it does so from real,

identifiable

dividing parent cells. I hold that with

real confidence. And now I owe you the

other side of my own conclusion. Fairly,

because a brand built on honesty has to

include the honesty that cuts against my

own view. A credible minority of serious

researchers still argues that adult

human neurogenesis is minimal or rare or

improbable enough that it changes little

in practice. And their case is not

foolish. Some of it rests on

evolutionary reasoning that I think is

genuinely worth hearing. The argument

runs like this.

Your long-term memories are stored in

stable circuits, and there may be real

value in leaving those old circuits

alone, rather than constantly threading

brand new cells into them. A memory

system that kept aggressively rebuilding

itself might be a memory system that

keeps overriding the very past it is

supposed to protect. On that view, a

brain that mostly stopped adding new

neurons to its core memory store might

not be a broken brain. It might be a

brain doing something sensible. I do not

think that is where the evidence has

landed, but it is a serious idea held by

serious people, and I would update the

day a better study told me I was wrong.

That is the honest state of it in 2026.

Strongly yes, not unanimously yes, and I

would rather hand you the argument with

its live descent intact than sell you a

false unanimity. But I promised you the

wall, and here it is, because you

deserve the ceiling as clearly as the

hope. This is where the title's promise

your memory will come back has to be

handled with care, and I am going to

handle it carefully. New neurons at 70,

real.

Your memory marching back on command

like a tide returning, not what any of

this shows. 700 new neurons a day is

genuine, and it is also tiny against a

brain of tens of billions. It is a

maintenance and plasticity process, not

a regrowth of lost memories, and not a

reversal of dementia. Let me be blunt,

because it is the single most important

sentence in this video. Neurogenesis is

not a memory cure. The link between more

new neurons and better memory in humans

is, so far, a correlation, not a proven

cause. I said it about the 90-year-olds,

and I will say it again here, because

everything overblown gets built in the

gap between correlation and cause. And

there is one more honesty I will not

skip, and it is the same thread that ran

through both the newest studies. The

amount varies enormously between people.

Some adults had many of these dividing

cells, others had almost none. This is

not a guarantee handed equally to

everyone. So, what is the true

defensible version of your memory will

come back? It is this. Your brain is not

a sealed, only shrinking organ after a

certain birthday. The machinery to build

new neurons, new connections, real

plasticity is still in there at 70, at

80. That is not false hope. It is the

correct reframe against the lie you were

told, that it is all downhill and

nothing can be built. Something can

still be built. What that plasticity

supports is your ability to keep

learning, keep forming, keep adapting,

not the recovery of a specific name you

lost last Tuesday. The honest promise is

capacity, not a cure. And capacity at 70

is worth a great deal. I will give you

one honest thread about what feeds it,

and then leave the fuller how-to for its

own video, because I do not want to

overload this one past the evidence, and

because the science on feeding it is a

different, thinner kind of science than

everything I just showed you. And you

deserve to know that. Here is the gap,

stated plainly. The strongest work on

what actually drives this building comes

from mice, not people. In mice, it is

powerful and clear. When researchers

gave mice a running wheel, the number of

surviving new neurons in the memory

region roughly doubled. That was Fred

Gage's lab again. In 1999, the same Gage

from the 1998

human study. Put mice in a richer, more

stimulating environment, more to

explore, more to do, and the new neuron

count went up as well. That line of work

is tied to a researcher named Gerd

Kempermann. And on the other side of the

ledger, chronic stress and its main

stress hormone, cortisol, appear to

suppress it. Press the animal with

relentless stress, and the building

slows down. That is a clean, consistent

animal story. Move and engage and it

rises. Grind under chronic stress and it

falls. But, and this is the honesty,

those are mice. There is exactly one

genuine human data point I will put real

weight on. In postmortem brain tissue,

depressed people who had been treated

with antidepressants showed more

hippocampal neurogenesis than depressed

people who had not been treated. That is

a real human signal that this process

can be nudged. It is also a single

narrow window. So, the shape of the

advice is honest and unsurprising. The

things that are good for a plastic brain

are movement, engagement, and lower

chronic stress. I just will not oversell

it. Exercise grows new brain cells is

proven in mice, plausible in humans, and

not directly proven by counting neurons

inside a living person because we cannot

yet count neurons inside a living

person. We have measured that the

capacity is there. We have not measured

it growing inside you. Let me hand it

back cleanly. For 100 years you were

told the adult brain only loses cells.

Gage's lab cracked that first in 1998,

new neurons in adult humans caught with

a dye. Carbon dating turned it into a

number. In 2013, 700 new neurons a day

into old age. A third of that region's

cells belonging to a pool that gets

exchanged over a lifetime. Two labs then

fought to a draw in 2018 and the fight

was resolved not by one side winning a

shouting match, but by a mundane

discovery about how long tissue sat in

preservative. And in July 2025,

researchers finally caught the parent

cells dividing in a 78-year-old brain

with the cells' full genetic fingerprint

as the proof. The proof is real. The

promise it earns is real, too, and it is

specifically this. Your brain can still

build. Not a miracle, not a cure, not a

memory conjured back on demand, but a

genuine measured capacity for growth

that the old dogmas were did not exist.

That reframe alone changes how you spend

the next decade. Reading the 2025 paper

was one of the better mornings I have

had with a study in a while, not because

it promises a cure, it does not, but

because it quietly overturns something

an entire century got wrong, and does it

with a 78-year-old's own dividing cells

as the evidence. Same brain, later

decade, still building. Take the honest

version of the hope, not the inflated

one. The honest one is sturdier, and it

is enough. Subscribe if you want the

next paper the day it lands. Look it up

yourself. It is Karolinska 2025,

and I will see you in the next one.