Musk has moved the field into the limelight with his talk of using brain implants to help humans keep up with advances in artificial intelligence, even suggesting he might get one himself. Yet in plowing his extraordinary resources and star power into Neuralink, Musk has raised the profile of rivals who in some ways are ahead of him, legitimizing them in the eyes of investors and accelerating the race to connect human brains with electronic devices.
Paradromics and three other leading companies have raised more than $240 million since Musk launched Neuralink in 2017, according to a review of financial disclosures, and his fellow billionaires Bill Gates and Jeff Bezos have entered the picture to back a rival, Synchron Inc. At least 42 people globally have used brain-computer implants in clinical trials, including a paralyzed man who used a robotic hand to fist-bump Barack Obama in 2016.
In an interview by text message with The Washington Post, Rodney Gorham, an Australian in his sixties with amyotrophic lateral sclerosis (ALS), used a Synchron implant to type his responses. Asked about his hopes, he responded one minute later. “I will try to improve the system for other people,” he wrote.
To successfully bring such a device to market, companies will have to prove to the Food and Drug Administration that their technology is sufficiently safe and reliable to be implanted in people. They also face profound ethical and security questions raised by a device that could one day give a cognitive advantage to healthy people who get an implant.
Though their approaches vary, the major players are trying to hack the brain’s instructions to the body and transmit them directly to an electronic device — enabling, for example, a person to move a cursor by thinking about it. While the initial applications of the technology would be for people with severe disabilities, Musk has also talked openly of enhancement.
“We’re confident that someone who has basically no other interface to the outside world would be able to control their phone better than someone who has working hands,” Musk told an audience in November.
Neuralink, by far the biggest operation, has staked out perhaps the most ambitious goal: building a high-speed interface between brain and device for the public, and finding ways of treating spine and brain injuries along the way.
Neuralink has designed a computer chip to be stitched into the surface of the brain, and a robot to perform the surgery. Musk envisions people regularly upgrading their brain implants, saying at the November event, “I’m pretty sure you would not want the iPhone 1 stuck in your head if the iPhone 14 is available.”
Nearly all executives and investors in this niche of neurotechnology acknowledge Musk’s impact on the field, though some say it is double-edged.
Marcus Gerhardt, chief executive of Blackrock Neurotech, credits Musk with “the rising tide that kick-started broader interest” of consumers and investors. Nonetheless, he worries that Neuralink “may try things that the FDA may disapprove of,” and “if anyone behaves in an irresponsible manner it can put the field back decades.”
Matt Angle, Paradromics’s CEO and in some ways Neuralink’s closest rival, said that all of his competitors “have responsible people” who "want to make sure the devices are safe.”
Reuters reported Thursday that the FDA had rejected Neuralink’s application last year to conduct trials in people, citing anonymous sources. Musk said in November that the company had submitted most of its paperwork to the FDA and expected to begin human trials in six months.
Musk and Neuralink didn’t respond to interview requests. A reporter who approached Neuralink’s Austin campus was asked to leave.
Neuralink’s competitors emphasize they are focused on helping people stricken by paralysis recover control of the body. Their prowess is increasingly impressive — from that presidential fist bump to converting the garbled vocalizations of a woman with ALS to text at 62 words per minute, according to a January study by Stanford researchers that hasn’t yet been peer-reviewed.
Scientists have explored the brain’s electrical signals over the past century, but the modern era of connecting brain activity to computers accelerated in the early 2000s with a pioneering company called Cyberkinetics, which ran out of money. The field has separated into companies aiming to read brain activity with external devices that are worn and those implanted inside the body.
People in the field, known as brain-computer interface technology, often offer an analogy to a sporting event. For devices that go on top of the head, it’s like hearing the crowd roar from outside the stadium. For those that penetrate the brain, it’s like lowering microphones into the stands and picking up conversations of individual people.
Several companies designing implants want to record from as many of these metaphorical microphones as possible, betting this will provide the clearest signal of brain activity and the fastest way to transmit it to a computer. Others say they can get a decent signal without piercing the brain, and can do so with less risk.
A study published in January found relatively few adverse events among 14 adults who’ve received a brain-computer implant going back to 2004. The outcomes involved no deaths or disabling complications, leading the authors to conclude its safety record is “comparable to other chronically implanted medical devices.”
The brain-computer implant in that study is now manufactured by Blackrock and has been used by 35 patients. The Utah Array, as it is known, resembles a tiny hairbrush with about 100 spikelike electrodes and is part of a system Blackrock hopes to bring to market this year. The device also has been found to produce inflammation where it penetrates the brain that can eventually damage tissue.
Such side effects can be an acceptable risk for people with severe disabilities, like Ian Burkhart, who had a diving accident at age 19 that left him paralyzed. His Blackrock-built brain implant allowed him — while hooked up to a computer in a lab — to move individual fingers and grasp objects with his right hand for the first time since his accident. Surgeons removed the device, which protruded from his head like a bottle cap, after about seven and a half years due in part to an infection at the site.
With the implant gone, Burkhart felt a sense of loss. “I still feel myself being careful when I’m brushing my hair,” he said in an interview. “I miss the possibility of going back into the lab and moving my hand again.”
Burkhart has become an advocate for people with spinal-cord injuries and hopes he may be able to get a new implant — in the hemisphere of his brain without scar tissue. “There’s some virgin territory that could be useful,” he said.
Some companies promise an implant that is less invasive. In a TED Talk last year, Tom Oxley, Synchron’s chief executive, called up an enlarged image of a Utah Array on a giant screen behind him. As he paced the stage, he drew a breath over his teeth and said, “The brain doesn’t really like having needles put into it.” Synchron, he said, has found a “secret backdoor” to the brain — threading a stent-like device through the jugular vein to rest atop the brain’s motor cortex, allowing it to eavesdrop on neural chatter without being inside the brain.
Synchron is among the furthest along of companies seeking to commercialize a brain implant, launching a clinical trial with its device in seven patients so far. One of them is Gorham, an Australian former software salesman who has ALS.
In an interview by text message, Gorham displayed the device’s promise and, rivals would say, its shortcomings. Answering how the implant has changed his life, he took five minutes to write “‘t has made my life easier.”
Gorham was assisted by eye-tracking software, which helps him move the cursor faster. To click, he must think about pressing a switch with his foot. Synchron’s device then decodes the neural signal, recognizing his intent and executing a computer command. A peer-reviewed, Synchron-funded study earlier this year found that four patients averaged 16.6 correct characters per minute, roughly equivalent to about three or four words per minute, when using its implant and eye-tracking software.
“We’re able to go safely where nobody’s gone before,” said Kurt Haggstrom, Synchron’s commercial chief. “We don’t know yet what the limitations are.”
At a recent cocktail party in a D.C. hotel, top executives of five rival brain-tech companies mingled and talked shop while dining on garlic-studded leg of lamb and parmesan truffle potatoes. They had come to present at a Commerce Department conference on Feb. 16 seeking information on whether brain-computer technology could give the United States or adversaries a military or intelligence advantage, as regulators weigh restricting exports.
The after-party — sponsored by Paradromics — was a friendly affair. But in separate interviews, executives are harsh judges of their rivals. Though they say there’s room for multiple types of brain implants to serve different patient needs in the future, they are keenly aware of their present — a battle royale for the investor capital they need to develop their devices.
Gerhardt, Blackrock’s CEO, said Synchron’s stent technique might have useful applications, but the bandwidth is too limited to meaningfully restore people’s function. “Without getting data from the brain, that’s not going to be possible now,” he said.
Ben Rapoport, a neurosurgeon and electrical engineer, indirectly criticizes devices that penetrate the brain. After a stint at Neuralink, Rapoport co-founded Precision Neuroscience Corp. to build an ultrathin, electrode-laden implant that can be inserted through narrow slits in the skull and rest atop the brain’s surface. His goal is to cover the brain in electrodes to transfer more data “in a way that doesn’t damage the brain,” he said.
Angle, Paradromics’ CEO, argues that implanting a device on the surface of the brain is too far from neurons to read their individual signals. He is betting on a variation of the Utah Array, configuring a device for higher bandwidth and less risk of damaging brain tissue.
In the company’s North Austin lab, two machines use an electrified brass wire to chisel 400 spikelike electrodes into a square of platinum-iridium smaller than a fingertip. The electrodes measure 1.5 millimeters long and half the width of a human hair, designed to penetrate the brain to where neurons fire and be thin enough to avoid provoking inflammation, according to company executives.
The components have to be hermetically sealed to withstand the body’s moisture, and a transceiver — implanted in the chest to ferry data from brain to computer — is designed not to overheat. The system is powered by a device that fits over the chest; on a translucent dummy, it is held in place with a gun holster.
So far, Paradromics has only tested its devices in sheep, but it is aiming to start a clinical trial in humans within a year. Of all the risks the company faces, from surgical to regulatory, the one Angle worries about most is money.
Paradromics received $18 million in government grants and has raised $47 million from venture capital, but Angle knows it will take more to get to market.
While Paradromics lacks the deep pockets behind Neuralink, the company presses a different advantage: speed.
That’s what Amy Kruse, a neuroscientist and venture capitalist who sits on Paradromics’ board, highlighted in her firm’s decision to invest in the company. “I think they’re going to get to market first,” she said.