With $160 million, Science Corp. is second only to Musk’s startup in total funding.
At Elon Musk’s Neuralink Corp., Max Hodak helped lead development of a computing device designed to be inserted into a person’s brain. Now he has a competing startup with $160 million in funding that seeks to do something even more striking: manipulate the brain without any in-skull implants.
Hodak’s new company Science Corp. is working on a brain-computer interface that’s in some ways similar to Neuralink’s but using a physical science called photonics. The idea is to use light and the eye’s optic nerve as a pathway into the brain, rather than implant chips deep within it.
Scientists and companies have experimented with photonics for years in the hope of restoring patients’ vision, but efforts to make brain-machine interfaces using the optic nerve have fallen short.
Hodak, the former president of Neuralink, has been operating Science in stealth mode since last year. He envisions making a variety of brain-augmenting products beginning with vision. The startup built a prosthetic device, Science Eye, that it says works in rabbits. In the next year or two, Science hopes to try it in humans. The machine is intended to treat retinitis pigmentosa, a disease affecting peripheral vision, and macular degeneration, which affects central vision.
“We’re starting in very disabled patient populations with serious unmet needs,” said Hodak, whose grandfather suffered from retinitis pigmentosa. “But if you refine that technology five or six generations, you get to replace glasses and [virtual reality] googles with just the tiny little implant in the eye.”
The philosophy is similar to one outlined by Musk. If the technology can address disabilities, then eventually healthy people may be willing to try it. The future goal would be to create a new sort of brain operating system that could, for example, project turn-by-turn directions onto the eyeball or conjure more immersive video games. Whether any of this would actually happen is unknown. The risk of opening one’s brain to hackers may be enough to scare away the general population.
For now, Hodak said he’s “focused on pragmatic, near-term things.” Science Eye is based on a 2-millimeter wide, very thin LED film that gets implanted on top of the retina, behind the eyelid. The surgical procedure lasts two hours, but Science said it can eventually be done in half that time. The LED film processes patterns sent wirelessly from glasses embedded with tiny cameras, which convert images into a form that the optical nerve can read.
For the device to function, the cells of the optic nerve must become light sensitive. Using gene therapy, Science delivers an engineered protein via injection that alters the optic nerve cells. By targeting individual cells rather than groups of cells, Science said it can achieve a much higher resolution compared with other cutting-edge approaches to treating blindness or eye disease.
But getting it to work in a rabbit is a far step from doing so in a person. The main way Science knows it is succeeding is by measuring activity in the rabbits’ visual cortexes, the part of the brain that handles sight.
One of the rabbits, Lela, sat patiently in her quarters when a reporter visited the Science headquarters in Alameda, Calif., this month. A small metal plate was attached on her forehead, under fluffy, white ears. As the reporter approached, Lela and some of her bunny colleagues hopped to the front of their hutches to observe.
Science’s technology will eventually be able to read from and write to the brain at the same time but for now is focusing on the writing part. Hodak said he has discussed Science Eye in a preliminary conversation with the US Food and Drug Administration.
Although the company’s technology is unproven in humans, photonics would solve a major flaw of existing brain-computer interfaces. Widespread adoption of such a machine is seen as unlikely if users are required to have holes drilled in their skulls. Finding a way to stimulate targeted sections of the brain in humans without implanting an electrode, though, has long eluded scientists. A practical application would require significant biological breakthroughs, Hodak acknowledged.
Still, he said he became convinced of the viability of photonics after working at Neuralink, which he left abruptly last year. He declined to comment on his time there or his departure from the company. He is still a shareholder in Neuralink, he said, as well as another competing startup, Synchron. Both are pursuing electrode-based systems.
Hodak runs Science with his co-founders, the biologist Alan Mardinly, microfabrication director Yifan Kong, engineer Corey Wolin and software lead Emma Zhou. The startup’s funding haul of $160 million is second in the field only to Neuralink. Among Science’s backers are founders of the crypto companies Ripple Labs, Protocol Labs and Paradigm as well as the former GV partner Blake Byers. Artis Ventures is also an investor.