Chapter 23
Monday, September 14, 2026
97.8%
A few days later, I had an appointment at the Prévenir center in Parc Monceau for a series of additional tests.
I didn't really know what to expect. Eva hadn't told me anything more, and between the drone and the diplomatic incident, I hadn't dared to ask her again. Sometimes, uncertainty is more comfortable than an answer.
Leaving Île Moineau for Paris weighed on me more than expected. Lia was still limping, her paw trapped in the splint, and the Morel had promised to visit her during the day.
Eva also had to return to Paris. At six o'clock, we were on the train. If this was the Miracle Morning¹, the miracle part must have completely eluded me.
Despite the fatigue, it was impossible to sleep on the train. Throughout the journey, the 97.8% lingered somewhere in the back of my mind.
When I arrived at the Prévenir center, the waiting room was almost empty at that hour. I sat in the same place as during my first assessment.
About ten minutes later, I heard voices in the hallway. A patient came out of an office, followed by a man in his forties. They shook hands, then the man turned to me.
— Mr. Mallet?
I stood up. He approached to shake my hand.
— Dr. Huang. Neurologist.
With a nod, he invited me to follow him. I followed him to a glass airlock I hadn't noticed during my first assessment. Beyond the airlock, the decor changed completely. Screens, machines, brain maps: the place reminded me more of the NEX Robot Lab than a medical prevention center.
In his office, Huang began with the only sentence I was waiting for.
— Your first MRI showed no lesions, no masses, no concerning anatomical anomalies.
I felt my shoulders relax.
— Good. At least my brain isn't killing me.
— No, absolutely not. Rest assured. If you're here today, it's because some data collected during your first assessment makes you eligible for a complementary research program. It's only non-invasive measures. No intervention on your brain. And no data will be used outside the protocol described here without your consent.
— Why me?
— Your profile meets several of our criteria. We are particularly studying how some people detect patterns, connect distant information, or make decisions with incomplete information.
— And concretely?
— We will offer you some cognitive exercises and some non-invasive stimulations to establish new physiological measures and functional images. You can stop your participation at any time.
I looked up.
A few months earlier, I would probably have found three other questions. But I had to admit: knowing that everything was fine and that, moreover, science wanted to study my brain had slightly flattered my ego.
And above all, I wanted to know why.
— OK, let's try, I said.
He brought up a form on a tablet. I signed.
A few minutes later, a nurse came to get me for a new MRI session, this time with functional imaging. Back into the awful tunnel. Could I have avoided it by discovering a claustrophobia?
— Can you hear me?
— Yes.
— Good. Exercises will appear on the screen behind you. You will see them in the mirror above your eyes.
A series of shapes appeared in the mirror. I had to choose the one that completed the sequence.
Alone in the machine, I let myself be absorbed by the exercises.
Then, a company had to choose between launching an imperfect product immediately or delaying its launch.
— A or B?
— B.
A competitor was now set to release theirs three months later.
— A.
Then the flaw became a safety issue.
— B. Obviously.
The voice that answered me froze me.
— Yet, you chose A.
I recognized the voice. Anxiety rose immediately.
— Cobot, is that you?
Dr. Huang intervened in my headset.
— Everything okay, Mr. Mallet?
— This voice... seems familiar.
— All protocol voices are artificially generated². They do not reproduce any recorded voice. Try to relax.
A new situation appeared.
Then the same voice:
— Why did you choose answer A?
I jumped. The situations followed one another: vehicle trajectory, contradictory information, data corrected at the last moment. Each time, I had to choose, doubt, sometimes reverse my decision.
After a while, I understood.
— You're not just measuring my answers.
— Not just.
— You're watching how I arrive at them.
— Exactly.
Huang paused for a second.
— We are especially observing what happens when you hesitate, change your mind, or correct a decision.
I didn't find that particularly reassuring.
Upon exiting the MRI, Dr. Huang led me to a large screen.
He scrolled through the first automated analyses of the data we had just obtained. Slices of my brain, curves, then a map crisscrossed with lines.
— I thought we all had more or less the same brain.
— In broad strokes, yes. But the connections between different regions vary from person to person.
— To that extent?
— Why not? We're all different, aren't we?
— It's a kind of brain fingerprint?
— Exactly. We call it your connectome³. It can even be used to identify you.
— Identify me just with my brain?
— You're recognized by your face. Why not by your brain?
— Seen like that...
I looked at the map.
Dr. Huang changed the display.
— Do you want to know its age? Our model estimates it around 44 years from the structural MRI. Another, based on diffusion MRI, gives you 41 years⁴.
— I'll keep 41.
He smiled.
— No need to worry. In both cases, it's less than your age!
A new display appeared.
This time, it wasn't really a brain anymore. More like a constellation of points connected to each other.
— Is that me?
— An approximation. Your virtual brain twin⁵.
— I'm less photogenic than I thought.
I was looking at my own brain with the strange impression that it had just been introduced to me. The model combined the structural and functional data of my brain collected since the first assessment. The following exams were to further refine it.
— You don't do this for everyone who goes through a Prévenir assessment?
— No.
— Then why me?
He looked at the screen.
— Your profile interests us.
— You've already told me that.
He didn't answer.
— What's it for?
— The goal, ultimately, is to simulate certain responses of your brain to an intervention before testing it on you.
— So then you test on it before trying on me?
— That's the idea.
— Then make as many mistakes as you want on that one, and don't touch the original!
The second exam took place in a neighboring room.
This time, no tunnel.
— I'm going to position a net of electrodes on your head. It will record the electrical activity of your brain during a series of exercises.
The nurse began to set up the device.
— High-density EEG⁶, she specified.
She also attached some sensors to monitor my physiological reactions.
Dr. Huang sat in front of a screen.
— Look at your right hand.
A command appeared.
Think about moving your thumb. Do not move it.
The problem with this kind of exercise is that it seems so simple that you immediately start to fear being the only idiot capable of failing it.
I stared at my finger.
I had managed not to move it, yet a curve had just appeared on the screen.
A new command appeared.
Move your thumb.
I did.
Dr. Huang compared the two signals.
— Why does it move in both cases? I asked.
— In the first case, we detect activity associated with the imagination of movement. In the second, we can observe its preparation and then its execution.
— And you see a difference between thinking about moving and moving?
— We mostly see what the signals have in common.
He continued with the index finger, then the left hand.
Imagine.
Wait.
Move.
Each time, something appeared in my brain before becoming a movement.
The machines weren't reading my thoughts.
But the nuance was starting to seem thin.
The exercises then changed nature.
Think of someone you love.
Thomas.
Imagine they are in danger.
My body tensed.
Now imagine you can save them, but another person must suffer the consequences.
I pressed the intercom.
— Is this emotional crash-test approved?
No one answered.
Other dilemmas followed. An autonomous vehicle. A computer flaw. Each time, a detail changed, and with it, sometimes, my decision.
Then four characters appeared.
E.
I thought of Eva.
T.
I thought of Thomas.
L.
I thought of Lia.
Ø.
My heart raced.
GØD.
The image of the drone came back to me, then the message.
— Bring back the previous symbol, Huang requested.
The Ø reappeared.
He observed his screens for a few seconds.
— Does this symbol mean anything to you?
I hesitated.
— A recent event.
— Stressful?
— Rather.
He nodded.
— Your reaction started almost immediately.
— And that's what you're looking for?
— Among other things. We want to know if some of your reactions can be detected even before you can consciously say what triggered them.
He turned to me.
— Has it ever happened to you? To have an intuition without immediately knowing what triggered it?
I remembered a demonstration, back when I was an engineer. All the indicators were normal, but something bothered me. I had asked to postpone it.
An hour later, a component had overheated.
I only understood afterward: two fans didn't have exactly the same rhythm.
I had heard it without knowing I had heard it.
The third exam was different again.
The nurse placed two electrodes on my forearm. Huang returned with a black coil shaped like an eight.
On the screen, I read: TMS⁷.
— This time, we will stimulate a specific area of your brain and observe its response.
He brought the coil close to my head.
A sharp click.
My right thumb moved.
All by itself.
I recoiled.
— I didn't do anything.
— I know.
Huang repeated it several times. At each impulse, my thumb reacted while the nurse monitored the measurements on her screen.
Then he moved the coil to the back of my skull.
A new click.
A brief white flash crossed my field of vision.
— Did you see something?
— A flash.
He repeated it.
Same flash.
I looked down at my hand.
Earlier, their machines had detected my intention to move without my finger moving.
Now, they had made my finger move without my intention.
I didn't like the idea that it worked both ways.
I thought it was over, but Huang asked me to wait a few more minutes.
On his screen, the data from the different exams now scrolled together: my functional fingerprint, the signals recorded by the EEG, my physiological reactions, the responses to the stimulations. As they were integrated, my digital twin seemed to complete itself before my eyes.
I watched the curves evolve without understanding what they meant. Dr. Huang, on the other hand, moved from one display to another with a concentration that began to worry me.
Then a window opened in the center of the screen.
97.8%
My stomach tightened.
— Wait...
Huang didn't respond immediately.
— We will redo some sequences.
— For what reason?
— To verify that some measures remain consistent when reproduced.
Huang subjected me again to several decisions, each time changing information at the last moment to make me hesitate. Then he asked me to repeat the motor intention sequence.
97.9%.
I looked at Huang.
— What exactly are you checking? Did I fail the test?
No answer. He launched a new sequence.
— Why are we starting over?
— Because you succeeded in exactly the same way.
Huang repeated. Other decisions, other hesitations, then the motor intention sequence. I complied, watching his face almost as much as the exercises themselves.
When he finished, a new value appeared.
Match = 98.1%.
He leaned slightly toward the screen.
For several days, I had taken this percentage for a risk.
It wasn't a risk.
It was a match.
But obviously, they were no longer looking for the match. They were testing it.
— It matches with what? I asked.
He kept his eyes on the screen for a few seconds before turning it off.
— We still need to analyze all the results.
— That's not really an answer.
— It's the only one I can give you for now.
I didn't insist.
For the first time since the beginning of the assessment, I was certain of one thing. They hadn't discovered a problem in my brain.
They had found something.
And this discovery, I felt now, was no longer really about preventive medicine.
Eva was waiting for me in the hall.
When she saw me arrive, she stood up.
— So?
— My brain has two ages, a fingerprint, a digital twin, and apparently some security flaws.
She smiled.
— Nothing worrying?
— I don't know. Huang is very good at answering without really answering.
She observed me for a few seconds.
— Did they do the motor intention sequence?
I stopped.
She also realized too late what she had revealed.
— So you know more than the broad lines of the protocol.
Eva looked away.
— I asked them to delve deeper into some results from your first assessment.
— Without telling me?
— I didn't know yet what they meant.
— And now?
— Now, I have more data.
It still wasn't an answer.
— There were also three values. 97.8. Then 97.9. And 98.1.
Her eyes widened.
— They confirmed it three times?
I looked at her.
— How do you know they were confirming something?
Her gaze froze.
— It's a match, Eva. With what?
She took a breath.
— Franck, if my hypothesis is wrong, I'd rather not drag you into this for nothing.
— And if it's true?
She took a few seconds to answer.
— Then we'll talk about it.
For the first time, I had the impression that Eva wasn't trying to protect me from bad news.
She was afraid of being right.
My phone vibrated. I looked down.
They found what they were looking for.
GØD
I froze.
— What's wrong? asked Eva.
— Nothing.
Since the drone, I had never replied to GØD. This time, I typed before I could reconsider.
They found what?
The answer came almost immediately.
Not what. Who.
I reread the sentence. Then everything disappeared: GØD's message, mine, the entire history.
— Franck?
I looked up at Eva. This 97.8%, which I had taken for a risk, had become a match. And now, there was someone on the other end.
— It's not something you were looking for, is it?
Eva didn't answer. Her silence answered for her.
When I returned to the cabin, Thomas was already asleep. Lia came towards me limping before lying down near the sofa. At least her problem was simple: a fracture, a few weeks of rest, and everything would be back to normal.
I let myself fall onto the sofa. My watch vibrated.
High physiological stress.
Then a second notification appeared.
Heart rate: 104.
I didn't even know yet if I was afraid.
The screen lit up again.
Franck, rest. Would you like me to help you?
Cobot
I stared at the question for a few seconds before removing the watch. For the first time in a long time, the sensation of my bare wrist reassured me.
Huang had his measurements. Eva her hypothesis. GØD seemed to have the answer. My watch knew what was happening inside me. And now, the President of the Republic was getting involved.
Yet I had known this brain forever. It was mine. And I had become the one who knew the least about it.
I went upstairs to bed.
For what happened next, I only had one version for a long time.
Source
¹ Le Miracle Morning est une méthode de développement personnel popularisée par Hal Elrod, qui consiste notamment à se lever très tôt pour consacrer le début de sa journée à soi-même.
² Des outils comme Gradium ou ElevenLabs permettent de générer des voix synthétiques ou de cloner une voix à partir d’échantillons audio. https://gradium.ai/
³ Finn E. S. et al., « Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity », Nature Neuroscience, vol. 18, 2015, p. 1664–1671. Les auteurs montrent que les profils de connectivité fonctionnelle du cerveau sont suffisamment propres à chaque individu pour permettre son identification, à la manière d’une « empreinte » cérébrale. https://pubmed.ncbi.nlm.nih.gov/26457551/
⁴ Gao C. et al., « Brain age identification from diffusion MRI synergistically predicts neurodegenerative disease », Imaging Neuroscience, 2025. Les auteurs montrent que l’estimation de l’âge cérébral à partir de différentes mesures issues de l’IRM de diffusion peut contribuer à prédire le risque de maladies neurodégénératives.
⁵ Hashemi M. et al., « Principles and Operation of Virtual Brain Twins », IEEE Reviews in Biomedical Engineering, 2026 : « Cette situation pourrait évoluer grâce à l’utilisation d’un jumeau cérébral virtuel (VBT), une réplique numérique personnalisée du cerveau d’un individu, qui intègre des données cérébrales structurelles et fonctionnelles dans des modèles informatiques avancés et des algorithmes d’inférence. » https://pubmed.ncbi.nlm.nih.gov/40257892/
⁶ EEG (électroencéphalographie) : technique non invasive qui mesure l’activité électrique du cerveau grâce à des électrodes placées sur le cuir chevelu. Elle permet notamment d’en suivre les variations lors d’une perception, d’une décision ou de la préparation d’un mouvement. National Institute of Neurological Disorders and Stroke (NINDS/NIH), « Neurological Diagnostic Tests and Procedures ». https://www.ninds.nih.gov/health-information/disorders/neurological-diagnostic-tests-and-procedures
⁷ TMS (stimulation magnétique transcrânienne) : technique non invasive utilisant des impulsions magnétiques pour stimuler certaines régions du cerveau. Appliquée au cortex moteur, elle peut provoquer un mouvement musculaire involontaire ; au cortex visuel, la perception d’un bref flash lumineux. Transcranial Magnetic Stimulation: A Primer - https://pubmed.ncbi.nlm.nih.gov/17640522/
Sign up to receive the next chapters by email:
©️ 2026 Frédéric Mazzella. All rights reserved.