Silicon Valley's Race To Create A Brain-Computer Interface

In effect, how to hack the human brain? JL

Antonio Regalado reports in MIT Technology Review:

Computing keeps achieving new heights, but our ability to interface with silicon is stuck in the keyboard era. Even when speaking to a computer program like Alexa or Siri, you can convey at most about 40 bits per second of information and only for short bursts. Compare that to data transfer records of a trillion bits per second along a fiber-optic cable.
Entrepreneur Bryan Johnson says he wanted to become very rich in order to do something great for humankind.
Last year Johnson, founder of the online payments company Braintree, starting making news when he threw $100 million behind Kernel, a startup he founded to enhance human intelligence by developing brain implants capable of linking people’s thoughts to computers.
Johnson isn’t alone in believing that “neurotechnology” could be the next big thing. To many in Silicon Valley, the brain looks like an unconquered frontier whose importance dwarfs any achievement made in computing or the Web.
According to neuroscientists, several figures from the tech sector are currently scouring labs across the U.S. for technology that might fuse human and artificial intelligence. In addition to Johnson, Elon Musk has been teasing a project called “neural lace,” which he said at a 2016 conference will lead to “symbiosis with machines.” And Mark Zuckerberg declared in a 2015 Q&A that people will one day be able to share “full sensory and emotional experiences,” not just photos. Facebook has been hiring neuroscientists for an undisclosed project at Building 8, its secretive hardware division.
As these people see it, computing keeps achieving new heights, but our ability to interface with silicon is stuck in the keyboard era. Even when speaking to a computer program like Alexa or Siri, you can convey at most about 40 bits per second of information and only for short bursts. Compare that to data transfer records of a trillion bits per second along a fiber-optic cable.
“Ridiculously slow,” Musk complained.
But it turns out that connecting to the brain isn’t so easy. Six months after launching Kernel amid a media blitz, Johnson says he’s dropped his initial plans for a “memory implant,” switched scientific advisors, hired a new team, and decided to instead invest in developing a more general-purpose technology for recording and stimulating the brain using electrodes.
Johnson says the switch-up is part of trying something new. “If you look at the key contributing technologies of society, the ones with the most impact, like rockets, the Internet, biology—there was a transition point from academia to the private sector, and for the most part neuroscience hasn’t made that jump,” says Johnson. “The most critical element is timing, when is the right time to pursue this.”
Memory implants
After making a fortune selling Braintree to eBay for $800 million in 2013, Johnson, now 39, reportedly sought the advice of nearly 200 people on how to invest his new wealth. He settled on neurotechnology and, last August, he announced he’d create Kernel and build the first neural prosthetic for human intelligence enhancement.
But Johnson’s business plan was extremely vague; one scientist called it “metaphysical.” Kernel’s website was plastered with book-jacket-like endorsements from scientific celebrities including J. Craig Venter and Tim O’Reilly, extolling his “great” and “serious” commitment to understanding human intelligence, not to mention the impressive $100 million he later promised to invest in Kernel.The reality is that interfacing with the brain is tough: electronics irritate its tissue and stop working after a while, and no one will get brain surgery just in order to send an e-mail. What’s more, even if you can communicate with the brain, you might not know what it is saying.
“Billionaires entering the broader neurotechnology field are very optimistic and may overlook details of the problem, which is we are far away from meaningfully understanding the brain,” says Konrad Kording, a Northwestern University neuroscientist who has advised Johnson. “But neurotechnology allows you to work on the most interesting questions in the universe while potentially making money, and so that is exciting.”
Johnson’s persona is part buttoned-down Mormon missionary (he once was one), part hard-driving door-to-door credit-processing salesman (he was that too), but now, with his new wealth, he’s also taken on the mantle of a technology prophet. At a 2016 startup conference in Silicon Valley, he showed up with his hair unbrushed, wearing a T-shirt with holes in it, and gave a wide-ranging lecture on human tool use from prehistory into the present, arguing that now “our very existence is programmable” through biology and machine interfaces.
Kernel’s original technology was a memory prosthesis, developed by Theodore Berger of the University of Southern California, who until recently was also the company’s chief scientific officer. Berger’s technology (see “10 Breakthrough Technologies: Memory Implants”) is a way of recording memories of rats and monkeys, storing these patterns on a computer chip, and re-delivering them to the hippocampus. One version of the setup, Berger says, has been tested in a handful of human patients undergoing brain surgery for other reasons.
But a mere six months after starting Kernel, Berger is no longer part of the company, and memory implants are no longer part of Kernel’s near-term plans. Johnson and Berger both confirmed the separation.
Berger’s vision, according to several people, was too complex, too speculative, and too far from becoming a medical reality, while Johnson hoped to see a return on his investment sometime soon. “They have a new direction, but we’re still talking,” says Berger. “The basic reason is it was going to take too long. It’s one thing to think about this and quite another to do it.”
Johnson says he concluded that Berger’s work “is really interesting, but not an entry point” into a commercially viable business.
Brain interface
By last November, Johnson was already exploring a pivot for his company, meeting with Christian Wentz, head of a small Cambridge startup, Kendall Research Systems, that sells equipment for recording in the neurons of mice and other animals. The company spun out of the laboratory of Edward Boyden, a professor at MIT who invents new ways of analyzing brain tissue.
In February, Johnson acquired Wentz’s company (for an undisclosed sum) and with it brought in a new team, including Wentz and Adam Marblestone, a noted theorist of both the limitations and possibilities of brain interfaces, who will become chief strategy officer. Both are former Boyden lab members, as are two other Kernel scientists, Caroline Moore-Kochlacs and Jake Bernstein.Johnson says Kernel will now develop a “generalized human electrophysiology platform”—that is, a flexible way of measuring the electrical impulses from many neurons at once, and stimulating them, too. The eventual objective is to use such electronics to treat major diseases, like depression or Alzheimer’s. “It’s for clinical use,” he says. “We are a for-profit company.”
Wentz says as part of the acquisition he and Johnson agreed that much more R&D on brain interfaces will probably be needed. “We have a very sober view of what can and can’t be done,” Wentz says. “We are not naïve.” He calls Kernel’s effort a “15-year endeavor,” although he adds that “we want to do in that period what has been done in the last 100 years.”
With the pivot, Johnson is effectively jumping on an opportunity created by the Brain Initiative, an Obama-era project which plowed money into new schemes for recording neurons. That influx of cash has spurred the formation of several other startups, including Paradromics and Cortera, also developing novel hardware for collecting brain signals. As part of the government brain project, the defense R&D agency DARPA says it is close to announcing $60 million in contracts under a program to create a “high-fidelity” brain interface able to simultaneously record from one million neurons (the current record is about 200) and stimulate 100,000 at a time.
“It’s time for neuroscience to graduate from academia to a general neuroscience platform,” says Johnson. With such a technology “a whole range of new applications—a lot of white space—would open up.”
Johnson declined to describe the specifics of Kernel’s technological approach to connecting with the brain, as did Boyden and Wentz. However, the team members have been working on well-identified problems. Wentz has been involved with developing electronics for high-speed reading of data emitted by wireless implants. Already, the flow of information that can be collected from a mouse’s brain in real time outruns what a laptop computer can handle. The team also needs a way to interface with the human brain. Boyden’s lab has worked on several concepts to do so, including needle-shaped probes with tiny electrodes etched onto their surface. Another idea is to record neural activity by threading tiny optical fibers through the brain’s capillaries, an idea roughly similar to Musk’s neural lace.More sophisticated means of reading and writing to the brain are seen as potential ways to treat psychiatric disorders. Under a concept that Boyden calls “brain coprocessors,” it may be possible to create closed-loop systems that detect certain brain signals—say, those associated with depression—and shock the brain to reverse them. Some surgeons and doctors funded by another DARPA program are in the early stages of determining whether serious mental conditions can be treated in this way (see “A Shocking Way to Fix the Brain”).
­­Boyden says Johnson’s $100 million makes a big difference to how he and his students view the entrepreneur’s goals. “A lot of neurotechnology has come and gone. But one thing is that it’s very expensive,” he says. “The inventing is expensive, the clinical work is expensive. It’s not easy. And here is someone putting money into the game.”