Meta description (around 120–130 characters)
Japan’s Fugaku supercomputer has simulated almost the entire mouse cortex with ~10 million neurons and 26 billion synapses. As a mid-career stroke survivor, I explore what this really means—beyond the hype—for brain science, rehab, BCI, and our future.


—

Table of Contents

1. Introduction | Why This Headline Made My Chest Tighten


2. What Actually Happened? | “Fugaku Fully Recreated a Mouse Cortex”

2-1. The Scale: ~10 Million Neurons, 26 Billion Synapses, 86 Regions

2-2. The Risky Word “Fully”



3. Why It Matters So Much | How This Changes Brain Research Itself

3-1. A “Virtual Brain” You Can Break Over and Over

3-2. Simulating Alzheimer’s and Epilepsy in Silico



4. Looking at the Cortex Through a Stroke Survivor’s Eyes

4-1. The Cortex Owns “Most of What We Call Me”

4-2. What Current Stroke Medicine Still Can’t Do



5. How Fugaku’s Virtual Mouse Cortex Connects to a Stroke Survivor’s Future

5-1. Possibility ①: Understanding the “Dynamics” of Disease

5-2. Possibility ②: Making Brain Stimulation More Precise

5-3. Possibility ③: A Bridge Between Digital Twin Brains and BCI



6. Sober Limits | This Won’t Magically Cure Us Tomorrow

6-1. Still Only Mouse, Still Only Cortex, Plasticity Still to Come

6-2. The Reality: 32 Seconds of Compute for 1 Second of Brain Activity

6-3. A New Question: Who Owns Our Brain Data?



7. Lateral Thinking | How This Could Change the Role of People with Disabilities

7-1. From “Data Provider” to Co-Creator

7-2. Updating the Story of “Living with a Broken Brain”

7-3. Holding Powerlessness and Quiet Hope at the Same Time



8. Conclusion | Standing on the Border Between Sci-Fi and Reality


9. FAQ | Common Questions People Ask About Fugaku and Virtual Brains




—

<a id=”section1″></a>

1. Introduction | Why This Headline Made My Chest Tighten

> “Supercomputer Fugaku fully recreates a mouse cerebral cortex – no longer science fiction.”



The moment I saw that headline, I felt my grip on my smartphone tighten just a little.

As someone who had a brain hemorrhage in mid-life and still lives with its after-effects, this wasn’t just “another cool tech news story” to me.

A few thoughts flashed through my mind at once:

“Wait, we’ve already come that far?”

“But does this really connect to my future?”

“Will this have anything to do with my broken brain one day?”


Excitement, unease, and a kind of nameless restlessness all rose up together.

In this article, I want to start from that restlessness and ask:

What actually happened, factually, with Fugaku’s mouse cortex simulation?

How could this, realistically, connect to a stroke survivor’s future?

Where does current reality end, and where do sci-fi-like expectations begin?


I’ll try to walk through these questions with no hype and no lies, using a medical and scientific lens—but without throwing away hope.


—

<a id=”section2″></a>

2. What Actually Happened? | “Fugaku Fully Recreated a Mouse Cortex”

Let’s calmly sort out the basic facts first.

2-1. The Scale: ~10 Million Neurons, 26 Billion Synapses, 86 Regions

In late 2025, a research team using Japan’s supercomputer Fugaku announced a major result.

They built a virtual model of a mouse cerebral cortex with:

About 10 million neurons

About 26 billion synapses (connections between neurons)

86 cortical regions that are mutually connected in a realistic network


And this wasn’t just a simple network of dots and lines.

They used a biophysically detailed model, which incorporates:

Each neuron’s membrane potential

Branching structures of dendrites as multiple compartments

The time course of electrical signals traveling through synapses


In other words, this is a virtual cortex where neurons don’t just “fire or not,” but actually behave with a fair amount of biological richness.

On top of that, the performance they demonstrated was:

Simulating 1 second of mouse cortical activity

In up to 32 seconds of wall-clock time on Fugaku


It isn’t real-time, but for something this large and detailed, many researchers consider it astonishingly fast.

2-2. The Risky Word “Fully”

However, we have to be careful with one part of the headline:

> The word “fully” in “fully recreated.”



Even the researchers themselves are open about what their model does not yet include.

For example:

The detailed mechanisms of synaptic plasticity, which underlies learning and memory, are still incomplete

Complex roles of neuromodulators like dopamine and serotonin are simplified

Realistic sensory inputs (vision, hearing, touch) are still heavily simplified

Pathological states such as impaired blood flow, inflammation, hemorrhage, stroke are not part of the current model


So at this stage, “fully recreated” more realistically means:

> “We’ve built a detailed, large-scale virtual model of a healthy mouse cortex, including its structure and basic activity patterns.”



It definitely does not mean:

> “We’ve recreated an entire brain, with all memories, feelings, diseases, and recovery processes included.”



If we want to talk about how this connects to stroke survivors, we first have to keep that distinction in mind.


—

<a id=”section3″></a>

3. Why It Matters So Much | How This Changes Brain Research Itself

You might think:

> “If it’s still incomplete, why is everyone so excited?”



Because this result has the potential to change how brain research is done at a fundamental level.

3-1. A “Virtual Brain” You Can Break Over and Over

Until now, neuroscience has mainly relied on:

Animal experiments (mice, rats, monkeys, etc.)

Human data from brain imaging, EEG, or from surgery in very limited situations


But real brains have a few very big constraints:

You can only break a living brain once

You cannot ethically perform large-scale, risky interventions on human brains


That’s where a virtual brain becomes revolutionary.

With a simulation like Fugaku’s virtual cortex, you can:

Run the same experiment hundreds or thousands of times

Gradually weaken a specific region or type of neuron and see what happens

Safely reproduce dangerous states like epileptic seizures or abnormal rhythms


In short, researchers get an “experimental lab” that they can break and rewind as much as they want—without harming a single real brain.

This opens the door to a different style of research:

> Rather than starting by breaking an animal brain,
we can first try to understand the process of breaking inside a virtual one.



3-2. Simulating Alzheimer’s and Epilepsy in Silico

The team has also mentioned long-term goals such as:

Simulating Alzheimer’s disease

Simulating epilepsy and seizure spread

Modeling abnormal cognitive states and changes in consciousness


All inside this kind of large-scale virtual cortex.

That means:

New drugs or interventions might first be tested on virtual brains,
instead of being thrown directly into animal models or clinical trials

The way damage spreads through networks—and which circuits are fragile—
can be studied at a resolution we’ve never had before


Stroke and brain hemorrhage are also about how damage spreads and how the network collapses.

So even if the first targets are Alzheimer’s or epilepsy, the same tools and ideas can eventually be used to think about stroke-related damage too.

From here, let’s shift the perspective more explicitly to that of a stroke survivor.


—

<a id=”section4″></a>

4. Looking at the Cortex Through a Stroke Survivor’s Eyes

4-1. The Cortex Owns “Most of What We Call Me”

Very roughly, the cerebral cortex is responsible for:

Motor areas: commands to move your arms, legs, face, body

Somatosensory areas: touch, pain, temperature, body sense

Visual and auditory areas: seeing and hearing

Language areas: speaking, understanding, reading, writing

Prefrontal areas: planning, decision-making, attention, emotional control


When the cortex or the white matter just beneath it is damaged by stroke or hemorrhage, you can get:

Hemiparesis (paralysis on one side)

Aphasia (difficulty speaking or understanding language)

Visual field defects

Attention and memory problems

Difficulty controlling emotions or behavior


In other words:

> The relationship between “how the cortex breaks” and what kind of disability appears
is almost directly tied to the life of every stroke survivor.



4-2. What Current Stroke Medicine Still Can’t Do

Stroke medicine has improved dramatically in recent decades:

Thrombolytic drugs (like t-PA) for early clot dissolution

Endovascular therapy to physically remove clots

Specialized stroke centers and intensive care

Detailed CT and MRI-based imaging


And yet, in the real world, doctors still struggle with things like:

Accurately predicting how much function a patient can regain

Deciding which rehab approach will work best for which patient, and when

Understanding how the rest of the brain reorganizes itself around the lesion


Experienced clinicians use pattern recognition and their intuition to make predictions, and they do it impressively well—but we’re still far from a complete, network-level understanding.

Which is exactly where a tool like Fugaku’s virtual cortex might begin to matter.


—

<a id=”section5″></a>

5. How Fugaku’s Virtual Mouse Cortex Connects to a Stroke Survivor’s Future

This is the real core of the article for me.

Instead of stopping at “Wow, that’s cool tech,” I want to ask:

> As a person who has survived a brain hemorrhage,
how could this realistically affect my future—or the future of people like me?



5-1. Possibility ①: Understanding the “Dynamics” of Disease

What I personally hope for most is this:

> A better understanding of how damage spreads and how the network responds,
not just which spot on the MRI is dark.



In a virtual mouse cortex, researchers can, in principle:

Gradually remove a certain percentage of neurons in a given region

Weaken inhibitory neurons in a specific circuit

Apply a “stroke-like lesion” to an area and watch the network collapse


Then they can observe:

How do activity patterns break down?

How do distant regions change their rhythms?

Which nearby circuits try to compensate and which simply fail?


We won’t be able to map this directly onto human brains in a simple way.

But these simulations can reveal typical patterns:

“Lesions in this kind of area usually destabilize these specific networks.”

“In this pattern of damage, certain regions are more likely to take over or compensate.”


On top of clinical imaging, that kind of understanding could eventually support:

More individualized rehab strategies:
“Given your lesion pattern, we should focus rehab on this function at this timing, because this network still has reserve capacity.”


That’s not tomorrow, or even next year. But in a 10–20 year perspective, it’s not crazy.

5-2. Possibility ②: Making Brain Stimulation More Precise

Another area where I see realistic potential is neuromodulation—brain stimulation therapies.

We already use:

TMS (transcranial magnetic stimulation)

tDCS (transcranial direct current stimulation)

DBS (deep brain stimulation)


for conditions such as depression, Parkinson’s disease, and epilepsy. TMS is also being tried after stroke to promote motor recovery.

But in practice, it’s still often:

Unclear exactly where to stimulate

Unclear how strong and at what frequency to stimulate

Based heavily on clinical experience and trial-and-error


Large-scale virtual brain models open the door to a different workflow:

1. Reconstruct the post-stroke network state in a virtual brain as best as possible


2. In the simulation, try stimulating different spots with different parameters and observe how the whole network responds


3. Pick a small set of promising stimulation patterns and test them in real patients



Of course, to do that we still need:

High-resolution brain images of each patient

Sophisticated models for how lesions affect connectivity

Long-term plasticity models for how the brain changes over weeks and months


So there is a long road ahead.

Still, it could move us from:

> “Let’s try stimulating the left motor cortex and see what happens…”



towards:

> “Given your specific network configuration, we’ll stimulate here with this pattern,
because the model suggests it will best support your remaining circuits.”



Again, not magic—but a more theory-driven way to choose stimulation.

5-3. Possibility ③: A Bridge Between Digital Twin Brains and BCI

This is where my excitement and fear get strongest:

> The combination of “digital twin brains” and BCI (brain–computer interfaces).



Separate from Fugaku, other researchers are developing low-invasive BCI techniques, such as:

Making a 4–8 mm hole in the skull

Inserting a sheet of electrodes that unfolds over the surface of the brain


For someone who has already had a hole drilled in their skull because of a hemorrhage,
just reading those words makes my body tense up.

But imagine this:

You record real brain activity via electrodes or BCI

You feed that pattern into a virtual cortex like the one running on Fugaku

You use the simulation to ask:
“What happens if we stimulate here, with this timing and intensity?”

You only try in the real brain the few patterns that look promising and safe in the model


This could greatly improve both safety and efficiency when using BCI as:

An interface to control robotic limbs for people with severe paralysis

A way to externalize “inner speech” for people who can’t speak well

A partial assist for certain cognitive functions in higher-brain dysfunction


From the perspective of someone living with a brain injury, this is very much an “offensive” future—a way to actively regain functions, rather than just accommodating loss.

And Fugaku-class simulations might quietly become one of the hidden pillars that make that future less risky and more realistic.


—

<a id=”section6″></a>

6. Sober Limits | This Won’t Magically Cure Us Tomorrow

So far I’ve focused on the hopeful side—but if I want to live “without lying to myself” as a stroke survivor, I have to look just as honestly at the limits.

6-1. Still Only Mouse, Still Only Cortex, Plasticity Still to Come

Important reality check:

1. The current model is still of a mouse, not a human


2. It covers the cortex only, not the thalamus, cerebellum, brainstem, basal ganglia, etc.


3. Long-term synaptic plasticity and life-long network re-wiring are not yet fully modeled



The human brain, by rough estimates, contains:

Around 86 billion neurons

And many times more synapses


Going from ~10 million neurons in a mouse cortex model to ~86 billion neurons in an entire human brain is not just a matter of “scaling up”—we’re talking about orders of magnitude more complexity and computational cost.

So:

> We are not on the verge of simulating an entire human lifetime in a box.



We’re still at the stage of proving the concept and figuring out how far we can push the idea.

6-2. The Reality: 32 Seconds of Compute for 1 Second of Brain Activity

Another easy-to-miss detail:

> Simulating 1 second of brain activity takes up to 32 seconds of computing time.



For researchers, this is an impressive speed.
For clinical use in real time, it’s far from enough.

If you imagine:

A BCI that talks to a simulated brain in real time

Or a treatment that adjusts stimulation based on instant simulation feedback


we’re not there yet—and we won’t be there tomorrow or the day after.

Yes, future improvements are likely:

Models can be simplified where full detail isn’t needed

Algorithms can be optimized

Specialized hardware (neuromorphic chips, etc.) can take over some workloads


But as of 2025, it’s important not to confuse:

> “We’ve reached a remarkable research milestone”



with:

> “Clinical practice will transform overnight.”



6-3. A New Question: Who Owns Our Brain Data?

As the technology progresses, another question becomes unavoidable:

> “Who owns my brain data?”



Imagine a world where:

Your detailed brain activity is recorded by BCI or MRI

A research lab or company builds a digital twin of your brain

That model is used to develop profitable drugs or AI systems

You, the actual person, never hear about it and get nothing back


For people with brain injuries, there is something comforting in the idea that:

> “My broken brain might one day help others.”



But at the same time, we need to think about:

How our data is used

What kind of consent is truly informed and fair

What kind of benefit sharing is right when our brains become “training data”


Fugaku’s mouse cortex is still far from that scenario.
But it’s clearly pointing in the same general direction.

So we, as patients and survivors, will one day have to decide:

> “How far am I willing to let a digital copy of my brain be created and used?”




—

<a id=”section7″></a>

7. Lateral Thinking | How This Could Change the Role of People with Disabilities

Let’s change perspective again.

Fugaku’s virtual mouse cortex doesn’t just have implications for medicine and technology—it also hints at a shift in the role of people with disabilities in research.

7-1. From “Data Provider” to Co-Creator

Traditionally, the role of patients in medical research has often been:

> “Subject who provides data.”



That role is important, and I don’t mean to belittle it.

But as digital twin brains and full-scale simulations become more realistic,
we’ll need something more than “passive data donation.”

Questions like these will arise:

What kind of research do we, as patients, actually want our data used for?

How should the benefits of that research be returned to us and those who come after us?

Where is the line where we say, “This goes too far, I’m not okay with that”?


In that sense, patients and survivors can become:

> Co-creators who help decide the direction of brain research,
rather than just being “subjects” in it.



As a stroke survivor, I feel that the Fugaku news is a good reminder to ask myself:

> “If my brain data were used,
what future would I want it to help build?”



Even just holding that question—without having a final answer yet—already changes how I read these headlines.

7-2. Updating the Story of “Living with a Broken Brain”

Another big shift, at least for me, is this:

> Moving from the story
“My brain is broken and that’s the end of it”
to
“My brain is broken, so how do I design a life together with it?”



When you have a stroke, it’s hard not to think:

“I’ll never go back to who I was.”

“The upward slope of my life is over.”

“From here on out it’s all decline.”


And medically speaking, there is truth in:

Dead neurons don’t come back to life

The longer you wait, the less recovery potential you have


But if tools like Fugaku’s virtual cortex and future human models advance, we may gradually gain:

A clearer picture of which networks are still intact

A sense of which areas are compensating and which can’t

A more objective sense of which functions we can push, and which we’re allowed to let go


In other words, a third path opens up between:

Pure despair (“everything is over”), and

Pure willpower (“never give up no matter what”)


That third path looks like:

> “Understand your brain’s reality as best you can,
then build a life strategy that fits that reality.”



7-3. Holding Powerlessness and Quiet Hope at the Same Time

As a stroke survivor, I often find myself oscillating between two feelings:

A deep sense of powerlessness:
“No matter how much medicine advances, part of my brain is simply gone.”

And a quiet, almost fragile sense of hope:
“Even so, there are people trying seriously to understand what is happening inside my brain.”


The Fugaku news amplifies both.

It does not promise a magic cure

But it does say:
“There are people who are literally simulating your kind of brain damage inside supercomputers, trying to understand it honestly.”


To me, that alone is a small but real comfort.

It doesn’t erase the weight of living with disability.
But it does make the future feel a bit less like a one-way road into darkness.


—

<a id=”section8″></a>

8. Conclusion | Standing on the Border Between Sci-Fi and Reality

Supercomputer Fugaku “fully recreating” a mouse cortex has been widely reported as:

> “No longer science fiction.”



From where I stand—living with a brain that has literally bled inside my skull—it feels more like this:

A surge of technical excitement

A very personal sense of fear

And a cautious, carefully-held hope


In writing this article, I tried to:

1. Look at the facts without exaggeration, but also without cynicism


2. Ask what this technology means from the viewpoint of a stroke survivor’s body, not just from a researcher’s desk


3. See the future not as sudden magic, but as a slow, 10–20-year process of change



Fugaku’s virtual mouse cortex will not change my life tomorrow morning.

But through:

Better understanding of disease dynamics

More precise neuromodulation

Bridges between digital twin brains and BCI

Shifts in how patients participate in research


it might become one of the quiet forces that, over time, make living with a broken brain less cruel and more navigable.

And above all, there is this one simple fact:

> Somewhere, inside a giant machine,
someone is seriously simulating “what it means for a brain to break”
so that they can help real people whose brains have broken.



As someone who has already been on that table, that matters to me.

The line between sci-fi and reality is shaking—not with loud explosions, but silently, steadily—inside Fugaku.

Standing on that trembling line, as a stroke survivor, I want to keep asking:

> “What kind of story do I want to write with my brain and my life now?”



Just asking that question, I think, is one small but important way
to refuse to treat this news as “someone else’s problem.”


—

<a id=”section9″></a>

9. FAQ | Common Questions People Ask About Fugaku and Virtual Brains

To finish, here are a few likely questions from people who might arrive here via search.

Q1. If Fugaku can simulate a mouse cortex, does that mean it can soon simulate a human brain?

Short answer: No—but we’re clearly moving in that direction.

The model is:

Mouse, not human

Cortex, not the whole brain

Missing fully realistic plasticity and lifetime development


Scaling from ~10 million neurons to ~86 billion neurons is not just a matter of “adding more servers.” It’s a jump of several orders of magnitude in complexity.

What the result really shows is:

> “Biophysically detailed, large-scale models can be run at full scale on modern supercomputers.”



That’s a big milestone—but still just one step.


—

Q2. Will this technology directly cure stroke or brain hemorrhage?

Short answer: It won’t “cure” them directly, but it can sharpen our understanding and treatment strategies.

Fugaku-class simulations could support:

A deeper grasp of how damage spreads through brain networks

More personalized rehab strategies

Better tuning of brain stimulation therapies

Safer and more efficient BCI development


So rather than “a miracle cure,” it’s more accurate to think of this as:

> “Improving the map and tools we use to navigate life after stroke.”




—

Q3. As a stroke survivor, what should I actually do about all this?

Speaking only as one person living with a brain injury, I’d say:

1. Learn what you can about your own brain and disability, at a pace that feels okay


2. Form your own view on how you want your brain data and story to be used in research or technology


3. Keep doing today’s rehab and life adjustments, while quietly allowing for the possibility that tomorrow’s tools may be a little better



Fugaku’s mouse cortex simulation may seem like something happening in a distant lab.

But in the longer view, it might be part of the invisible “operating system update” that shapes the future of stroke medicine and rehab.

And if that future is being shaped anyway,
I’d like us—patients and survivors included—to have a say in what kind of future it becomes.