Chapter Three: The Eleven Seconds Before
At 2:16 p.m., I discovered that my new research partner had a talent for making impossible things sound like administrative inconveniences.
"The laboratory is downstairs," Adrian said.
"I know where the laboratory is."
"Then why are you standing here?"
"Because I'm waiting for you to explain why our director looks like he might spontaneously combust whenever you mention the first trial."
Adrian adjusted the strap of his watch.
"Perhaps he has a medical condition."
"Perhaps you have a personality disorder."
"Then it's fortunate that you're a neuroscientist."
I stared at him.
He opened the laboratory door.
"After you, Dr. Saye."
I went inside with the firm intention of disliking him for the rest of the afternoon.
It lasted approximately thirty seconds.
The laboratory was nothing like the rest of the institute. While the upper floors were filled with white walls, glass offices, and the familiar smell of burnt coffee, the room below was almost entirely dark.
A bank of monitors illuminated the central workstation. Behind a transparent partition stood a medical chair surrounded by neural sensors, motion-tracking cameras, and equipment I recognized from the institute's most advanced imaging facility.
At the far end of the room, a single observation window looked into an adjoining chamber.
The chamber was empty.
I approached the workstation.
"Where's the main system?"
Adrian entered a security code. The monitors flickered awake.
"Here."
A series of graphs appeared across the largest screen.
Forty-seven trials.
Forty-seven anomalous responses.
Forty-seven identical patterns.
I leaned closer.
"These are the original records?"
"Yes."
"Unedited?"
"Yes."
"Who verified the timestamps?"
"I did."
"Who verified your verification?"
He glanced at me.
"You really do make collaboration sound delightful."
"I'm trying to establish whether we're studying a scientific anomaly or an elaborate professional humiliation."
"Both remain possible."
I nearly smiled.
Nearly.
I opened the first recording.
A subject sat in the chair, facing a screen. A small white symbol appeared in the center of the display. According to the protocol, the subject was supposed to press a button whenever the symbol changed.
Simple enough.
Then I noticed the timestamps.
The neural response appeared at 14:32:08.
The symbol changed at 14:32:19.
Eleven seconds.
I replayed the recording.
Again.
And again.
The subject's fingers moved before the symbol changed. The machine registered the response, and eleven seconds later, the stimulus appeared.
I leaned back.
"Show me the raw signal."
Adrian opened another file.
"Already did."
"Then show me the unprocessed signal. Before filtering, before artifact correction, before anyone decided which data points looked inconvenient."
For the first time that afternoon, he looked impressed.
"Most people would have started with the statistical model."
"Most people are probably why the statistical model has a reputation for being unreliable."
He turned to the keyboard.
A new graph appeared.
I examined it carefully.
There was no obvious equipment failure. No duplicated waveform. No suspicious discontinuity suggesting that the timestamps had been shifted.
I checked the sampling rate.
Then the synchronization logs.
Then the reference clock.
Nothing.
"That's impossible," I whispered.
Adrian stood beside me.
"That's what I said."
I pointed to a section of the waveform.
"Here. The response begins eleven seconds before the stimulus, but the baseline activity is consistent with the rest of the recording. If this were a synchronization error, we'd expect a discrepancy somewhere else."
"I know."
"And if the subject had received an unintended cue, the response would still follow the cue. It wouldn't precede the event itself."
"I know, Dr. Saye."
"Then stop saying it as if you personally invented the concept of knowing things."
He folded his arms.
"Would you prefer that I pretend to be surprised?"
"No. I'd prefer that you explain why you're so calm."
His expression changed.
Only slightly.
But I had begun to understand that Adrian Vale's smallest expressions were often the most revealing.
"Because I was there when the first subject told us what would happen."
I turned toward him.
"What did she say?"
He looked through the observation window.
"The first time, it was something insignificant. A technician was going to drop a pen. The subject said it eleven seconds before it happened."
"Coincidence."
"That's what we thought."
"And the next time?"
"The subject said the lights would go out."
I glanced at the ceiling.
"They didn't."
"No. They didn't."
"So the subject was wrong."
Adrian's gaze returned to mine.
"At the exact second the lights were supposed to go out, every clock in the building stopped for eleven seconds."
I said nothing.
He continued.
"The backup clocks. The security system. The servers. Even the clock on the technician's phone. Everything stopped."
"That would indicate a synchronization failure across multiple systems."
"Of course."
"Not a prediction."
"Of course."
His agreement was infuriating.
I turned back to the screen.
"You're doing that thing again."
"What thing?"
"Agreeing with me in a tone that suggests you think I'm an idiot."
"I don't think you're an idiot."
"How reassuring."
"I think you're trying very hard to make this fit inside a theory that hasn't been designed to contain it."
I looked at him.
"And what theory have you designed to contain it?"
For a moment, he said nothing.
Then he answered.
"None."
That was the first honest thing he'd said all afternoon.
It changed the room between us.
Not because I suddenly trusted him. I didn't. But uncertainty, when admitted without decoration, was more useful than confidence built on assumptions.
I returned to the recording.
"Show me the subject's file."
Adrian hesitated.
"Dr. Saye."
"Please don't tell me I don't need to concern myself with it."
"You don't."
I slowly turned my head.
He sighed.
"You're going to make this difficult, aren't you?"
"I'm a neuroscientist. Making things difficult is how we distinguish a discovery from an expensive mistake."
He entered another command.
A profile appeared.
SUBJECT 01
Age: 31.
Sex: Female.
Enrollment date: Six months earlier.
Medical history: Unremarkable.
Neurological history: Unremarkable.
Discharge date: Blank.
I read the last line twice.
"Where is she?"
"We don't know."
"Did she leave the facility?"
"There's no record of her leaving."
"Was she reported missing?"
"Yes."
"Then there should be an investigation."
"There was."
"And?"
Adrian's face had gone very still.
"The investigation concluded that she had never been admitted to the facility."
I stared at him.
"That's absurd."
"Yes."
"Her medical records?"
"Missing."
"Consent forms?"
"Missing."
"Security footage?"
"Corrupted."
I looked at the subject's profile again.
The information was there. The experiment had been recorded. The neural activity had been measured.
Yet the person at the center of it had apparently become a statistical ghost.
I scrolled to the bottom of the file.
There was one final entry.
FINAL RECORDED STATEMENT:
You keep asking when it happens.
You're asking the wrong question.
I read it aloud.
Adrian didn't move.
"What does that mean?"
"I don't know."
"Did she explain?"
"No."
"Did you ask?"
"Several times."
"And?"
His voice became quieter.
"She asked me what I remembered about the day before the trial."
I waited.
"What did you tell her?"
"That I remembered arriving at work. I remembered reviewing the equipment. I remembered speaking to the technicians."
"And?"
"And I remembered meeting her."
He looked directly at me.
"But I couldn't remember her face."
A chill travelled down my spine.
"That's not possible."
"That was my conclusion, too."
"Memory is reconstructive. Stress, sleep deprivation, suggestion, neurological interference. There are a dozen possible explanations."
"I know."
"Stop saying that."
"Then stop repeating the same explanations I spent six months testing."
We stared at each other.
I disliked him a little more.
I trusted him a little more, too.
Which was significantly worse.
I returned to the workstation and opened the first trial's video recording.
The subject sat in the chair.
The screen was blank.
The timestamp read 14:32:08.
I watched her face carefully.
Her eyes were open, but her gaze was fixed on something beyond the camera.
At 14:32:08, she spoke.
"He's going to ask me about the lights."
Eleven seconds later, the technician asked her about the lights.
I paused the recording.
"Who was she talking to?"
"She was alone in the chamber."
"That wasn't my question."
Adrian moved closer to the monitor.
"I know."
I pointed to the subject's eyes.
"Look at her pupils. She's tracking something."
"There was nothing in front of her."
"Perhaps she was looking at her own reflection."
"The glass was one-way."
I replayed the segment.
The subject's gaze shifted slightly to the left.
Then she smiled.
Not a nervous smile. Not the strained expression of someone struggling with an unfamiliar test.
A small, knowing smile.
The kind someone gives when they recognize a person they weren't expecting to see.
I stopped the recording.
"Enhance the image."
Adrian did.
The footage sharpened.
The subject's face became clearer.
Her lips moved.
I zoomed in.
The audio was faint, but the system recovered the words.
"Not yet."
I stared at the screen.
"What did she mean?"
Adrian leaned forward.
"I don't know."
For once, I believed him.
I replayed the audio.
The subject's mouth moved again.
Two words.
Not yet.
Then the recording ended.
I sat back, trying to organize the possibilities.
A hidden communication channel. A deliberate manipulation. An unknown stimulus. An error in the recording system.
Every explanation was incomplete.
I hated incomplete explanations.
"There's another file," Adrian said.
He opened a folder marked RESTRICTED.
Inside was a single document.
No title. No author. No date.
Just a short paragraph.
The observed response cannot be adequately explained as anticipation, predictive processing, or ordinary temporal inference. In every confirmed case, the subject reports an impression of recognition rather than expectation.
I read it again.
Recognition.
Not prediction.
The distinction mattered.
Prediction meant the brain was calculating what might happen.
Recognition meant the brain was responding to something it perceived as already familiar.
I turned toward Adrian.
"What if the response isn't anticipating the event?"
He watched me carefully.
"What are you suggesting?"
"What if the subject experiences the event as a memory before it occurs?"
Neither of us spoke for several seconds.
Then Adrian said, "That's the hypothesis I was afraid you'd reach."
My stomach tightened.
"Why?"
"Because if you're right, the question isn't how the brain predicts an event."
He looked at the empty chamber.
"It's how the brain recognizes something that hasn't happened yet."
The lights above us flickered.
Once.
Twice.
Then the monitors went black.
I stood abruptly.
"What happened?"
Adrian was already moving toward the emergency console.
"Stay where you are."
"I'm not a houseplant, Vale."
"Then stop standing in the middle of a room full of equipment that has just lost power."
The emergency lights came on.
A low alarm sounded behind the observation window.
I turned toward the chamber.
The medical chair was still empty.
But a figure stood on the other side of the glass.
A woman.
She wore the same plain grey clothing as the subject in the recording.
Her face was turned away from us.
My mouth went dry.
"Adrian."
"I see her."
The woman lifted one hand and placed it against the glass.
Then she turned.
I recognized her face from the recording.
Subject 01.
The missing woman.
She looked directly at me.
Her lips moved.
I couldn't hear her through the glass.
But I understood the words.
You're late.
The lights returned.
The chamber was empty.
The alarm stopped.
The monitors restarted.
I stumbled backward, my shoulder colliding with Adrian's arm.
Neither of us moved.
Finally, I forced myself to speak.
"Tell me you saw that."
"I saw it."
"Tell me there's a camera recording it."
He looked at the screen.
The footage had resumed.
The chamber was empty in every frame.
Every frame except one.
At 2:47:11 p.m., the camera recorded a single image of the woman standing behind the glass.
The timestamp beneath it read:
RECORDED SIX MONTHS AGO.
And in the corner of the image, someone had written a note.
In my handwriting.
Do not let her see me again.
I had never seen that woman before.
At least, I was certain I hadn't.
Until Adrian quietly asked me a question that made certainty feel like a very fragile thing.
"Dr. Saye, what exactly do you remember about six months ago?"
I opened my mouth.
And discovered that I couldn't remember a single thing.
Not about six months ago.
Not about the first trial.
Not even about the day I had applied for this project.
My last clear memory was of waking up that morning and drinking cold coffee.
Everything before it had vanished.
And on the monitor, beneath my handwriting, a second line appeared.
SUBJECT 02: ACTIVE.