In 2016, Tesla and SpaceX CEO Elon Musk founded the tech startup Neuralink with the aim of developing a high-bandwidth implantable brain-computer interface to enable people to control a computer or smartphone using their minds. On Tuesday, Musk offered the world a first glimpse at what the team is working on.
How the Neuralink technology works
According to a white paper the firm released Tuesday, which hasn't undergone peer review, the technology is a small, thin-film, polymer probe with almost 3,100 electrodes on about 100 flexible wires each 1.6 millimeters long inserted by a surgical robot. According to Tim Harris, from the Howard Hughes Medical Institute's Janelia Research Campus, the wires are long enough to "cover most of the cortex in a human," and they are thinner than a human hair.
The electrodes are packaged in a small, implantable device containing custom-built integrated circuits and detect electrical activity, or action potential "spikes," that make up brain activity. Right now, that implantable device transmits these signals via wires to a USB port outside the brain, which processes and interprets them. (Eventually, the team plans to transmit those signals wirelessly, and has gotten started on a product it calls the "N1 sensor," The Verge reports). According to Neuralink, the electrodes can monitor the activity of more than 1,000 neurons at a time.
To insert the device, Neuralink developed a robot, akin in some ways to a sewing machine, that can target specific areas of the brain using a stiff needle that quickly shoots the probes into the brain. This robot allows surgeons to avoid major blood vessels when installing the probes. In trials in rats, the robot is able to install the electrodes correctly about 87% of the time and can do so in an hour or less.
While current brain-computer interfaces are limited to relatively simple tasks, such as controlling a computer mouse or moving a robotic arm, Musk described far higher aspirations: "a symbiosis with artificial intelligence" that would prevent humans from being "left behind" as AI evolves in the future.
The obstacles Neuralink still must overcome
In the white paper, Neuralink said it had successfully used the device to implant electrodes in the brains of rats, and according to Musk, the device has been tested on monkeys, with one subject being able to control a computer with its brain.
However, experts say the technology has a numerous obstacles to overcome before it can be used on humans. For instance, Harris said the white paper left unanswered how long the electrodes will last after being implanted in the brain. "If you're going to do this for people, you should be aiming for at least five years, minimum," Harris said. "To do an implantation surgery of this level of intricacy, a year or two is not enough."
Further, according to Jacob Robinson, a neuroengineer at Rice University, removing an expired implant and putting in a new one could prove challenging for Neuralink, given that the "threads" of the implant are so small.
Andrew Hires, a neurobiologist at the University of Southern California, said, "The big risk, of course, is if something goes wrong … one more major adverse event could cause a stricter regulatory crackdown."
Loren Frank, a neuroscientist at the University of California, San Francisco, said another concern is potential damage to the brain during insertion. "My guess is that the rapid insertions do substantial long term damage, such that they may not get high quality recordings for very long, but that’s just a guess at this point," he said. Typically, he added, scientists "go into the brain much more slowly, and that's likely an important part of why we were able to get such long lasting recordings" in experiments on rats.
And some of Neuralink's claims struck experts as dubious. For example, in their presentation, Musk and other Neuralink executives said the interface could stimulate the somatosensory cortex in a way that would make paralyzed patients feel like they touched something, but the white paper doesn't elaborate on that ability.
According to Robinson, that's a significant omission, because "given the electrodes they are working with, stimulation is going to be a lot more challenging than recording."
Further, while Harris praises the system's electronics that read and interpret brain activity as being "significantly advanced over prior art," he expressed concern over the system's interface to the brain. According to the white paper, no electrode in the device detected neuronal activity in rat brains as strong as 100 microvolts, which Harris said is the minimum signal "that you can trust."
Still, many experts are excited about the technology's potential. Leigh Hochberg from Brown University, said the technology is "a novel and exciting neurotechnology."
Hochberg added, "Given the great potential that intracortical brain-computer interfaces have to restore neurologic function for people with spinal cord injury, stroke, [amyotrophic lateral sclerosis], traumatic brain injury, or other diseases or injuries of the nervous system, I'm excited to see how [the company will] be translating [its] system toward initial clinical studies."
Andrew Schwartz, a neurobiologist at the University of Pittsburgh Medical Center, said, "Overall, the concept is impressive and so is the progress they've made. But a lot of this still seems to be conceptual. It's hard to tell what's aspirational and what they've actually done."
Regardless, Schwartz said, "to me, they have the right concept, and are doing what needs to be done."
According to Scientific American, Musk talked about implanting the technology into a human brain by as early as next year. However, Neuralink President Max Hodak and the company's head surgeon Matthew MacDougall, who both joined Musk on stage during the presentation, were more restrained in describing the timeline.
"There is a whole FDA process we have to go [through]," Hodak said, "we haven't done that yet."
MacDougall noted that safety is a primary goal. Ultimately, MacDougall said the company wants to refine the technology to be "something more like Lasik" eye surgery, which is a less invasive procedure that does not require general anesthesia.
However, MacDougall said early iterations of the device will require invasive implantation (Begley/Robbins, STAT News, 7/18; Hernandez/Mack, Wall Street Journal, 7/17; Lewis, Scientific American, 7/17; Markoff, New York Times, 7/16; Lopatto, The Verge, 7/16).