“Cells To Silicon” Looks At The Future Of Inner Space

Mapping brain regions is starting to allow us to restore lost movement to the paralyzed and grant limited amounts of sight to the blind (lab mice, that is). But will scientists be able to go even further in coming decades, allowing us to upgrade, copy, or even move our brains around to other substrates besides a meat puppet with a shelf life of about 100 years, at best? These and other far-off future concepts took center stage in the program Cells To Silicon: Your Brain in 2050, part of the Big Ideas Series at the 2014 World Science Festival.

“Tonight we can dream a little bit,” moderator and NPR journalist Robert Krulwich said in his introduction. “Science invites crazy dreams.”

Welding a brain to a machine is what allowed Cathy Hutchinson—who was left completely paralyzed by a brain stem stroke—to take her first drink of coffee under her own power in more than a decade. Hutchinson raised the water bottle full of java to her own lips with a brain-linked robotic arm developed by Cells to Silicon panelist and Brown University researcher John Donoghue:

The arm is linked via wires to a hairbrush-shaped brain implant about the size of an M&M.

“It takes information out of her brain as she thinks about moving,” Donoghue said. “That generates a pattern of activity in our computer that is translated into a command for the robot arm.”

How does Donoghue know where to place the implant? Actually, he explained, scientists have known since the 1800s that there’s a strip of the brain going from the top of the head down to the jawbone that controls movement. Patients with brain damage in certain areas of that strip, or conditions like epilepsy, have helped doctors refine their knowledge of what movements correspond to which part of that brain region.

Ordinarily, anywhere from 10s of thousands to perhaps millions of neurons would be involved in raising a drink to your lips. But although Donoghue’s device samples just around 100 neurons, it still manages to get the job done.

“One question is: how can he do as well as he can, with 100 out of a million?” NYU research psychologist and fellow panelist Gary Marcus said.

Donoghue pointed out that one of the finer features of the brain is it’s what scientists call a “robust” system: a single neuron can carry different impulse signals to and from the brain. That way, if you lose a few neurons, you can still move your arm.

But overall, our current understanding of brain activity is still drastically limited.

“It’s kind of like trying to understand politics by looking out an airplane window,” Marcus said. “But people are working on getting us closer to the ground.”

Another panelist, Cornell University neuroscientist Sheila Nirenberg, works on other kinds of brain-interfacing prosthetics (neuroprosthetics), including a chip that can replace a non-functioning retina. In the retina, light entering the eye gets translated into a pattern of “spikes,” also called nerve impulses, where there’s a sudden fluctuation in the cellular membrane of a neuron. This electrical signal travels to the brain, which processes it into what we can recognizably “see,” be it a face, a flower, or a movie.

Without the retina, the information doesn’t get relayed to the brain. So Nirenberg found a way to use math to translate the input of light into series of electrical impulses. Armed with those equations, a microchip can translate the input into the normal output and give some semblance of sight (though it’s not quite a perfect replication yet).

“The brain is just a giant math problem,” Nirenberg said. “But just because we say it’s an equation doesn’t mean there’s not a probabilistic component to it.”

Donoghue echoed the sentiment. “The brain is a statistical equation, not a formal one,” he said.  “It’s always changing; there’s always a kind of noise inside.”

If we had the ability to create backups of our neural state, would it be us? Or another entity?

That inherent noise is one of the main reasons that neuroscientists will probably not be able to predict human behavior, even if we build incredibly faithful simulations of brains. Such a thing would be “possible in principle, but impossible in practice,” Marcus said. “The amount of detail you’d need to know to make the simulations work perfectly faithfully is more than we’re able to collect. We’ll probably never be able to build whole-brain emulations that work as fast as a human brain. And it’s one thing to build an idealized brain, another to build an individualized one.”

One of the problems that concerned panelist and University of California, Berkeley engineer Michel Maharbiz is the difficulty of making a neuroprosthetic that can last for life—the brain isn’t usually hospitable to foreign objects for very long. The key may be in dramatically reducing the size of the device. To that end, Maharbiz and several colleagues have proposed a design for “neural dust,” a group of thousands of tiny, ultrasound-sensitive devices that could record bursts of nerve impulses.

Krulwich raised the idea that in the future, neuroscientists may be asked to use their art to literally change minds, perhaps to boost a person’s brainpower, or tune their moods. “How do you constrain a thing like this—or don’t you?”

Donaghue noted that neuroscientists would hardly be the first to change the brain. The pharmaceutical industry’s already leagues ahead in that department: Aspirin alters your ability to feel pain; college students take Ritalin to study.

Another knotty question the panelists wrestled with was the potential for making copies of the brain. If we had the ability to create backups of our neural state, would it be us? Or another entity? Most of the panelists opined that if such a thing is possible, it would likely be closer to a person’s twin than their “copy.”

“The process would be like making a wax dummy,” Marcus said. “We could build a molecule-for-molecule copy that gets the gist of you, but I still think that’s a copy, not you.”

Nirenberg noted that a brain copy, if made just at one particular point in time and frozen thereafter, would definitely not reflect the “truest” nature of the original for very long.

The copy “would be me at this particular moment, but I’m learning all the time. Even just this conversation has slightly changed my brain in some way.”

A better way to be sure you’re preserving your “real” self, Maharbiz mused, might be to keep hooking up better and better neuroprosthetics to your brain as your fleshy parts failed bit by bit.

But even the brain is made of very spoilable matter. One audience member raised an interesting question: what about when we have the ability to replace the neurons themselves? What if parts of your brain became synthetic over time? It would be akin to the philosophical conundrum known as the Ship of Theseus: if you start replacing a boat bit by bit, and end up with a boat at the end that contains none of the parts it started out with, is it truly the same boat? Would there some crossover point that we stop being the same person, or become more machine than person? Like the philosophical thought experiment, this question is an essentially subjective one.

“We can set up different definitions [for humanity], and live with those choices,” Marcus said.

Krulwich sounded uncomfortable with the brain-centricity of the discussion, though. “Without a body I might be dull to the world,” he said. “Our bodies are very important.”

Nirenberg disagreed. “I am a bag of neurons, working and thinking. Your body does shape a lot of things, but it’s basically three billion neurons entertaining themselves.”