Mortal Crimes: 7 Novels of Suspense (156 page)

She pulled a syringe from the tray, twisted an 18 gauge needle on it, and buried it in a small vial of lidocaine. She took the needle off and laid it on the tray, then found a smaller needle and injected the skin she had cleaned. Finally she rested a small blue towel with a hole in the middle carefully over Ian’s chest, the hole framing the wound.

She slid the cap off an 11 blade scalpel and made a one-centimeter incision over the wound, dabbed at the wound with some gauze, then held pressure until the bleeding had stopped. Without looking, Julia reached back on the tray for a hemostat.
Third from the left.
Her hands closed around an oblong metal retractor and she almost dropped it. This was not her setup.
What was she doing operating in these conditions?
She closed her eyes and took two long, slow breaths.

She looked at the tray and found a hemostat and a pair of forceps and probed the wound for the shrapnel.
There
. She carefully picked out a small metal fragment and laid it on the table beside Ian. A few minutes of probing later, satisfied nothing else was in the wound, she took the syringe with lidocaine and infiltrated the exposed pectoral muscle. A swift movement with the scalpel and she held the pectoral incision open with a small retractor.

There it was. She could see the bloody outline of the implant. She traced its contour with the forceps. It looked intact.

There has to be something there.
She glanced back at Sarah, who watched her from a distance.

Julia traced the contour of the implant with a probe to search for fractures in the shielded housing, weak spots. She found nothing. There were so many components on the chip. A small integrated circuit? It could be anything.

She remembered what Chang had said about needing to take it out. She looked back to her patient. Ian’s expression was blank. She had to find a way to repair it for Ian’s sake - they were going to take it out one way or another.

She traced the delicate wires emerging from the tip, then felt something give under the slight pressure of her probe. She doubled back, felt it again.

Of course.
She should have thought of that first. The most basic principle of electronics. Components almost always fail at switches or connectors. The wires fed into a microarray of pins that snapped into a housing along the edge of the chip. It was loose. The impact must have severed the connection, and the wire array probably gradually worked itself out of the housing.

Julia rummaged through the tray. She needed a microscope, at least a magnifier. There were thousands of channels in each array. The components had to fit together perfectly. Ian was lucky. If he had walked by something with enough alternating current it could have triggered seizures.

Perfect wasn’t likely to happen. Not here. Especially if the component was bent. Julia meticulously cleaned the contacts with tiny droplets of sterile saline. She blinked her eyes to clear them of sweat. Her hands performed over and over the tedious small movements. She hooked the probe under the implant to stabilize it, then turned to the tray.

She found a small glass tuberculin syringe and using the scalpel as a wedge, she tapped the back of it with a retractor. The syringe severed cleanly.

Julia held a piece of the syringe to the light. Perfect. Then she held the makeshift lens over the incision and secured it in place with a hemostat. She found the tiny black line on the housing, a fraction of a millimeter wide, and lined it up with the matching line on the wire assembly. Meticulously, with movements so small she could only see them through the lens, she slid the assembly back into the housing until it was fully seated. She glanced at her watch. Fifteen minutes had passed.

Her fingers ached. She stretched them out and clenched her fists. Then she removed the probe from under the implant and let it slide back under the muscle. She clasped her bloodied hands together as a reminder to keep them sterile, and walked back to Sarah.

“I think I’ve got it. The wires detached from the chip, but I was able to reconnect them. If that’s not it, there’s nothing more I can do here.”

Sarah’s face brightened. “How can we test it?”

“I’m on it,” Chang said. His laptop was already open. “When the wires are detached, it trips an auto shutoff on the chip. Give me a sec to reset the chip.”

Julia turned back to Ian, put on a fresh pair of sterile gloves, and sutured the wound. She had just finished cleaning the wound when she heard Chang and Markov talking in a low voice.

“It’s working.” Chang grinned, watching diagnostics on the computer screen. “She did it.”

CHAPTER ELEVEN

Julia felt triumphant. She had nailed it, and couldn’t keep herself from smiling. Never mind impressing Sarah, she especially loved showing up Markov and Chang. There was nothing like a well-executed surgery to lift her spirits. That she’d succeeded under substandard conditions made her feel all the better. Sarah didn’t seem to notice when Julia walked up. She was studying Chang’s computer screen over his shoulder. Julia wiped her forehead with her forearm and crowded in next to the others.

“So how do we retrieve the data?” Sarah asked.

Chang didn’t look up from the screen. “Once the diagnostics run, I send a command to the implant to switch to replay mode.”

“So you just rewind a few days and push play?” Sarah asked.

At that question, Chang turned around. “If you’re expecting high definition video feed out of this box, you’re going to be disappointed. There’s a lot of data in there. It’s going to take a few minutes to download, even with a fast wireless connection.” A window opened on the computer and Chang turned his head back to read the diagnostic results.

“And then we can watch it?” Sarah asked.

Julia said, “It’s not quite that simple. The data is stored as tens of thousands of traces. Each one shows the electrical activity in a region of the brain that’s smaller than a millimeter in size.”

Sarah turned to Julia as Chang typed. “So it’s like an EEG? It measures brain waves?” Julia could tell that Markov was listening too. She’d always thought that for someone so involved in the project, Markov had been resistant to spending the time to understand how the implants actually worked.

“Sort of,” Julia answered. “Only at much higher resolution. Even the best EEG recordings measure activity from brain areas the size of a quarter. In an area that size there are tens of millions of brain cells. To figure out what all those cells are doing would be like taking a hot air balloon ride over a football stadium and listening to the crowd. You can hear a roar whenever there’s a touchdown or an interception, but you’ll never pick up the individual conversations.”

“Sports metaphor?” Sarah asked.

“When in Rome,” Julia said with a smile.

“Downloading now,” Chang said.

Julia glanced at the computer to see another window open with a progress bar slowly moving across the screen. She turned her attention back to Sarah and Markov. “Anyway, recording from thousands of tiny electrodes right on the surface of the brain is like having a spy in each section of the stadium. He can pick up individual conversations, get a feel for what’s being discussed. Find out if there was cheering or booing in each area. When you look at the data from all of those spies, you can piece together an awful lot about what the crowd is thinking and experiencing. Put together the whole football game, if you will.”

Sarah paused to think for a minute. “So we just need to find which channels record sound and video and analyze those traces?”

“That’s actually not too far off for video. We have about a thousand electrodes in the primary visual cortex. That’s the part of the brain where vision is first processed. Visual information is stored in at least thirty areas in the brain, each processing a different feature, but that first area behaves like a camera, and each tiny cluster of cells processes information coming from the eye about a specific part of visual space. Think of it as a 100 by 100 grid that works kind of like a low resolution TV monitor.”

“So we can see what he sees?”

Chang put his hands behind his head and leaned back to look at Sarah upside down. “Actually the idea that we can see a picture is a myth. The eyes are constantly moving, analyzing different features of an image. Information about movement, color, texture, context. They’re all processed in different areas of the brain. There is no one spot in the brain where the entire picture is represented.”

“Then what we’ll be able to see is a low resolution feed of where Ian is looking. Like a grainy video?”

“Right,” Chang said. “Think of one of those annoying, jerky art films, where the camera never sits still. Now look at that picture reflected through a mirror at the fun house.”

“Almost all of the visual part of the brain is devoted to a tiny area in the center of the visual field,” Julia added. “Peripheral vision occupies a few pixels at the edge, so to speak. But it’s adequate. We can even read text through the subject’s eyes if he reads slowly enough and focuses on each letter.”

“Not that they ever do,” Chang said with a measure of disgust. “Field agents apparently can’t be bothered. I told them about twenty times and they still don’t hold still enough.”

Julia could see Markov nodding in agreement, even though it was clear this was all new to him. It was petty, but it felt good to finally know something he didn’t. As for Chang, she had to admit she was impressed. His communication skills were plenty good when he got excited about some technical subject.”Get back to the point,” Sarah said. “How long does this take?”

Markov found his opening. “It’s quite impressive,” he said. “I managed to get an entire wing of NSA’s supercomputing center on demand, just for this project. We send the data by satellite, query a specific time frame, and we can get images back almost instantly. There’s some sophisticated programming involved…”

Markov trailed off as Sarah asked Julia, “That’s video. What about audio?”

“Audio is more tricky,” Julia said, glancing over Chang’s shoulder to see how much time was left on the download. She turned back to Sarah. “Seems counterintuitive, given that recording sound is much easier than video in the real world. But not in the brain. It stores audio information as frequencies. And it takes a lot of post processing to coax that sound out. The implant doesn’t have enough resolution to get an audio recording with good fidelity.”

Sarah frowned. “Good enough for speech?”

“If we filter it enough, sure.”

“I’m working on that,” Chang said. “It’s mostly a question of stripping out the noise in the signal. Turns out it’s synchronized in corresponding areas in the left and right side of the brain. So by subtracting the opposite side, you clean out a lot of noise.”

“It matters a lot what the subject is doing,” Julia added. “If they’re moving, there’s interference from electrical activity in the muscles. If they’re holding still, it’s better. Fortunately, we also have recordings from language areas.”

“Language areas?”

“The brain has dedicated language centers. One for hearing what other people are saying, and one for phrasing what we want to say.”

“So there’s a cell that only fires when we hear the word ‘touchdown?’” Markov asked. Sarah’s eyes narrowed and fixed on Julia’s at the reference.

“No, not at all,” Julia answered. “That was a huge breakthrough in this project. There’s an old idea in neuroscience of something called the grandmother cell, meaning a cell that fires only when you recognize your grandmother. It’s been known for a long time that this is mostly wrong, and that information is spread over a lot of cells. It’s large patterns of cells that encode information. Anyway, during training for Ian and Kendall we tried using something called a classifier technique. It worked beautifully.”

Julia looked at Chang. She mostly understood the idea, but the truth was she still felt uncomfortable with the details of how the algorithm worked. She’d tried to get Chang to explain it a couple times during the training period, but he always seemed so impatient and she felt stupid to ask questions. She could only imagine how the non-technical types would find it.

“I should probably know this,” Markov said.

“Kind of complex stuff,” Chang said. “Too complex for lay understanding.”

“Try me,” Sarah said to Chang.