The Race For The Most Advanced Brain-Computer Interface

Saturday - April 20 2024, 10:13 UTC - 7 months ago

Neuralink's implant, named Telepathy, featured a paralyzed man playing computer chess with his mind. Companies in the field aim to decipher brain activity for communication, and classify devices based on invasiveness. Greater invasiveness results in greater risks but also higher-resolution data. Companies like Neuralink and Paradromics hope to decode speech. Blackrock Neurotech leads in experience, with their Utah array being used in dozens of people and forming the basis of their MoveAgain device.

In the world of brain-computer interfaces, it can seem as if one company sucks up all the oxygen in the room. Last month, Neuralink posted a video to X showing the first human subject to receive its brain implant, which will be named Telepathy. The recipient, a 29-year-old man who is paralyzed from the shoulders down, played computer chess, moving the cursor around with his mind. Learning to control it was "like using the force," he says in the video.

Neuralink’s announcement of a first-in-human trial made a big splash not because of what the man was able to accomplish—scientists demonstrated using a brain implant to move a cursor in 2006—but because the technology is so advanced. The device is unobtrusive and wireless, and it contains electrodes so thin and fragile they must be stitched into the brain by a specialized robot. It also commanded attention because of the wild promises Neuralink founder Elon Musk has made. It’s no secret that Musk is interested in using his chip to enhance the mind, not just restore function lost to injury or illness.

Most of the companies working in this space have the same goal: capturing enough information from the brain to decipher the user’s intention. The idea is to aid communication for people who can’t easily move or speak, either by helping them navigate a computer cursor or by actually translating their brain activity into speech or text.

There are a variety of ways to classify these devices, but Jacob Robinson, a bioengineer at Rice University, likes to group them by their invasiveness. There’s an inherent trade-off. The deeper the electrodes go, the more invasive the surgery required to implant them, and the greater the risks. But going deeper also puts the electrodes closer to the brain activity these companies hope to record, which means the device can capture higher-resolution information that might, say, allow the device to decode speech. That’s the goal of companies like Neuralink and Paradromics.

Robinson is CEO and cofounder of a company called Motif Neurotech, which is developing a brain-computer interface that only penetrates the skull (more on this later). In contrast, Neuralink’s device has electrodes that go into the cortex, "right in the first couple of millimeters," Robinson says. Two other companies—the Austin-based startup Paradromics and Blackrock Neurotech—have also developed chips designed to penetrate the cortex.

"That allows you to get really close to the neurons and get information about what each brain cell is doing," Robinson says. Proximity to the neurons and a greater number of electrodes that can "listen" to their activity increases the speed of data transfer, or the "bandwidth." And the greater the bandwidth, the more likely it is that the device will be able to translate brain activity into speech or text.

When it comes to the sheer amount of human experience, Blackrock Neurotech is far ahead of the pack. Its Utah array has been implanted in dozens of people since 2004. It’s the array used by academic labs all over the country. And it’s the array that forms the basis of Blackrock’s MoveAgain device, which received an FDA Breakthrough D .