Last week, we learned that Elon Musk will start a mind-computer interface company called Neuralink. The name added a brainy new entry to Musk’s growing scroll of big ideas—Tesla, SolarCity, SpaceX, the Hyperloop.
But as the news of Musk’s nascent venture to merge man and machine spread across social media, an electrical engineer in Ohio named Pedram Mohseni must have been slapping his forehead.
That's because in January he'd agreed to sell the name Neuralink to Musk without realizing it.
Mohseni, a professor at Case Western Reserve University, and his scientific partner, Randolph Nudo of Kansas University Medical Center, had owned the trademark on "NeuraLink" since 2015 after creating their own startup company.
The pair of longtime neurotech researchers had developed a device that might help people with brain injuries. But their initial contacts with investors hadn’t advanced very far when a stranger approached them offering tens of thousands of dollars for their company’s name. They accepted. No one mentioned that Musk, whose net worth is $14.7 billion according to Forbes, was behind it.
“They approached us, we negotiated, and now Elon Musk will be the rightful owner of Neuralink,” says Mohseni.
Instead of hard feelings, Mohseni says he’s excited. Finally, tech titans are throwing money behind some far-out ideas that a small number of neuroscientists have long championed and doggedly sought to advance.
In addition to Musk, the online payments entrepreneur Bryan Johnson is putting $100 million in a company called Kernel, which is also developing brain implants.
In uncovering details of Musk’s venture, the Wall Street Journal last week reported the company will develop new ways to treat disease but ultimately also a means of fusing human and machine intelligence. That’s something Musk seems to think is necessary to counter the risk of runaway artificial intelligence.
It’s “difficult to dedicate the time” to yet another high-tech venture, in addition to electric cars and space rockets, Musk tweeted, “but existential risk is too high not to.”
Just how brain technology will let humanity keep up with AI is anyone’s guess—and Musk’s company hasn’t said what it intends. But Rikky Muller, a professor at the University of California, Berkeley, says treating medical conditions and the aim of connecting consciousness to computers “are not unrelated, because anything implanted in the human body has to meet all the standards of a medical device.”
No one knows that better than Nudo and Mohseni. Their story—the story of the original Neuralink—shows the kind of challenges Musk will face trying to fill the brain with electronics.
Starting in 2011, Mohseni, a bioengineer, and Nudo, a brain specialist, began exploring an idea for an electronic brain chip to treat traumatic brain injury. Their idea: reëstablish damaged connections by recording neurons in one part of the brain, then transmitting the chatter to another. By 2013, they’d even demonstrated that their prototype could help brain-damaged rats.
That’s when the duo formed NeuraLink (which they spelled with a capital "L"). But raising money proved difficult. Any device that’s going to end up in the human brain needs to be as reliable as a Swiss clock and could easily take $200 million to develop and test. What’s more, while Nudo and Mohseni had some provocative data, they couldn’t say for sure the system would help anyone. Even if they did, there might not be enough eligible patients to justify the big expense. That’s also been a problem for researchers developing devices that read the brains of paralyzed people and allow them to move robotic arms. “Even though it’s a terrible condition, it’s not that many people,” says Nudo. “The thing in neurotech is that even if it works, it’s hard to see profitability.”
Nudo adds: “The feeling among investors was reluctance to invest in invasive brain technology, unless there is a very strong proof of principle. The place where our startup was is that we had a name without a product.”
Now, Musk is in the same position. But Mohseni thinks the billionaire might be able to blast through obstacles. “The whole idea of uploading or downloading thoughts to a healthy person, well, that is pie in the sky, but he has the credibility and vision to talk about those things,” Mohseni says. “We still have to advance our work a bit more, obtain some preliminary human data, before we can go to the investment community. But Mr. Musk doesn’t have that problem.”
A spokesperson for Musk declined to say why the entrepreneur wanted the Neuralink name badly enough to pay for it, but Mohseni believes it was worth every penny. “The name Neuralink really nicely captures what is happening in the field of neuromodulation,” he says.
Only a very few types of electronic brain implants have ever reached the market. The most widely employed, and sold by medical device giant Medtronic, is a “deep brain stimulator” able to stop the tremors of people with Parkinson’s disease. More than 140,000 patients have received versions of Medtronic’s stimulator, and the company’s brain modulation division has about $500 million in annual sales.
The Medtronic stimulator is in some ways low tech—it’s based on 1980s technology—and uses just one or two electrodes to continually send zaps of electricity into the brain. In fact, no one is exactly sure why it works. Very roughly, it’s the neuroscience equivalent of banging on a TV to adjust the picture.
Lothar Krinke, who manages that business for Medtronic, says the company continues to invest in making the system smaller and adding features for surgeons who implant it. “These things take much longer than you think to bring to market,” says Krinke. “When you talk about these systems, you have to talk about reliability. A brain implant has to perform for [decades].”
More recently, a company called NeuroPace began selling the first “closed-loop” brain implant for epilepsy patients. That’s a leap forward because the device can both detect a seizure coming on and then zap the brain to stop it, creating an automatic control loop. It's a neuralink, if you will.
But other ventures have not gone so well. The list of failed brain-interface companies includes BrainGate and Northstar, a company that liquidated itself in 2009 after spending $132 million in an attempt to help stroke patients recover with a brain implant.
Nudo and Mohseni, who have funding from the U.S. Army and the Paralyzed Veterans of America, say they would still like to raise money from investors to advance their idea towards commercialization.
Now that they sold off the name Neuralink to Musk, Nudo says he has been thinking up new names for their company. “But I don’t want to tell you what they are. Someone else would buy the trademark before we do,” he says.