Extraordinary technology is transforming the science and medical fields, with the potential to radically change peoples’ lives in ways never thought possible or envisaged just a few decades ago, reports TIME Magazine.
It's been a while since Alice Charton (87) got a good look at a human face. There are plenty of people moving through her world, of course – her husband, friends, doctors, neighbours – but judging just by what she can see, she’d have to take it as on faith that any one person was there at all.
Five years ago, the retired schoolteacher, living in Paris, first noticed her eyesight failing, with a point in the middle of her field of vision going hazy, muddy and dim. That point grew into a spot, and the spot into a blotch, until it became impossible for her to recognise people, read a book, or navigate unfamiliar places on the streets.
The cause of the problem was age-related macular degeneration (AMD), which afflicts some 200m people worldwide and involves a breakdown of the cells in the retina, particularly in the area known as the macula – responsible for central vision. AMD does not typically cause blindness, but vision can be severely impaired. As for a cure for AMD? Non-existent.
“I always worked with children, teaching them how to read,” says Charton. “So it was especially devastating for me not to be able to read.”
But three years ago, everything changed. After battling two years of slowly deteriorating vision, Charton was able to claw back a small portion of her lost world. Today, while she still can’t see faces or walk the streets unassisted, she does read: not very much; just an hour in the morning and an hour in the afternoon. But it was transformative. “It literally changed my life,” she says.
The breakthrough was thanks to two decades of work now being led by Science Corp, a four-year-old neuroscience company in San Francisco and led by biomedical engineer Max Hodak. In an experimental procedure dubbed Prima, which has now been performed on a few dozen people, surgeons implant a 2 mm-by-2 mm computer chip with 400 hexagonal electrodes directly on the spot in the retina that the AMD has destroyed.
Patients like Charton then don a pair of bulky, black plastic glasses equipped with a tiny camera that looks out on the world and beams what it sees in an infrared impulse directly to the chip. The system uses the infrared wavelength – invisible to the naked eye – as opposed to visible light to prevent the signals from interfering with the residual peripheral vision the subjects still have.
From the chip, the signal is transmitted to the optic nerve and then to the brain, restoring something resembling normal vision.
The chip that works this optical magic is not much to see. Under a powerful microscope it resembles an oversize circuit board. To the naked eye, it is a tiny flake of nothing, but brings sight – imperfect, maybe, but sight all the same – to the nearly blind.
“There is an eye chart that (healthy) people are supposed to be able to read at a distance of 4m; even at 1m, untreated patients can barely read the biggest letters on the top line, using their peripheral vision,” says Hodak. “In a clinical trial of Prima, patients were able to read down to the fifth line on the eye chart.”
That trial, just published in the New England Journal of Medicine, involved 38 patients, including Charton, recruited from across Europe, all of whom underwent the Prima procedure. Post-surgery, nearly 80% of them improved their performance on the eye chart by 20 letters, and 84% of them could read letters, numbers, and words at home.
“AMD patients in our clinical trial were able to read and write again, not just letter by letter but word by word,” said Daniel Palanker, Professor of Ophthalmology and Electrical Engineering at Stanford University.
Palanker conceived of the Prima system in 2004, and has been working closely with Hodak and the Science team, serving as a part-time consultant on the Prima project. “The next-generation implant should have pixels that are five times smaller and more of them, going from about 400 in the current implant to 10 000. This should allow for visual acuity of 20/80, and with the help of the camera’s zoom function could even reach the equivalent of 20/20 resolution.”
Science Corp is not stopping there. The company’s researchers are also developing technology that involves implanting a chip directly on the brain, which could allow people who are paralysed by a stroke, an accident, or amyotrophic lateral sclerosis (ALS) to operate a computer, a smartphone, a wheelchair, or even lights and appliances, with their thoughts alone.
For those whose condition has robbed them of speech, the chip could one day make it possible to translate thoughts into words and sentences and paragraphs on a screen. The technology could even translate those thoughts into spoken, computer-generated words – in the person’s own voice, if video or other recordings of them speaking before their illness were available, which the AI loaded into the computer could copy.
In this system, the implanted computer chip would not just sit on the brain, but become part of the brain. Using a technology Hodak calls the biohybrid model, the chip would be seeded with stem cells which would grow into the brain tissue, forging useful connections with neurons that govern thought, speech, creativity, and more.
“You can imagine making a chip with 100 000 electrodes that, when this grows into the brain, you could get a billion synapses,” Hodak says. “Right now you can get information into the brain very easily. Getting information out of the brain is limited. Imagine if you could get imagery or audio or imagination or memories out of the brain.”
Science Corp is not remotely alone in pursuing this union of the computer and the brain – this wedding of cold silicon and warm carbon. The World Economic Forum says up to 680 companies worldwide are least dabbling in brain-computer interface (BCI) technology, making for a sector valued at $1.74bn in 2022, and expected to grow to $6.2bn by 2030.
BCI today is what the personal computer was in the early 1980s, an infant technology that could grow in globe-shaking ways, with some of the companies talking about not just treating patients with ALS or other forms of paralysis, but also using the technology with firefighters, the military, and other first-responders, speeding reaction times and communication.
BCI could even be used by the general public – or at least those who want to have mind-to-mind access to AI systems.
“People with brain implants will be able to interact with AI in ways that people without brain implants do not,” says Matt Angle, CEO and founder of Texas-based BCI company Paradromics. “That is in some sense a superpower.”
The new science is causing not just a technological sensation, but also a cultural one, twanging a live wire in the popular mind. No sooner were the Covid vaccines released in 2020 than unfounded rumours swirled that they contained microchips that would be injected in the body, giving the government access to your thoughts.
“Over the past 20 years, every time there was an advance in this technology, the principal investigators would get calls saying that someone – the government, their wives – had put a chip in them,” says Florian Solzbacher, co-founder and chief science officer of Utah-based Blackrock Neurotech, a BCI company.
“There’s a lack of training in critical thinking.”
BCI, for better or worse, is here. The job now is for scientists to figure out how to use it, and for lay-people to figure out what to make of it.
The best known of the BCI companies, thanks to the ubiquitous presence and deep pockets of its founder, Elon Musk, is Neuralink, in California. Founded in 2016, the company has so far placed its implants in the brains of 12 people, hoping to allow them to operate a computer or smartphone with their thoughts.
Neuralink is currently running a clinical programme dubbed Prime, seeking to enrol patients aged 22 and older who have quadriplegia and are willing to have a 1 024-electrode chip, about the size of a coin, implanted for a study expected to last six years.
In January 2024, the company implanted its first chip, into the brain of Noland Arbaugh, a 29-year-old quadriplegia patient who’d lost movement below the shoulders in a diving accident.
The implant allows him to control a cursor on a screen with only his thoughts – playing video games, surfing the web, and communicating with friends. Hodak was part of this ground-breaking work, as one of Neuralink’s founders and its president before leaving to launch Science.
There is, too, San Francisco–based Echo Technologies, led by University of California-San Francisco neurosurgeon Dr Edward Chang. In 2021, Chang and his colleagues published a paper in the New England Journal of Medicine reporting that they had developed a so-called neuroprosthesis allowing a paralysed man who could not speak to generate words on a computer screen with nothing but his thoughts.
In 2023, as reported in Nature, they improved the system to include computer voice synthesis along with the text, as well as a facial avatar that can display emotions and expressions as it speaks, reflecting the subject’s words. In 2024, as reported in Nature Biomedical Engineering, Echo upgraded the hardware to allow another patient, who was bilingual, to toggle between English and Spanish.
“Our system is fully wireless,” says Chang. “The onscreen avatar is designed to resemble the person who’s doing the speaking. But in reality it could be anything. It could even be an emoji if that’s what the person wanted.”
With the rise of AI in information processing it’s no surprise that it’s at play here too. BCI speech systems rely on so-called large language models that interpret speech and predict the next word or words – much as word-processing programs will suggest the word juice if you type out orange, or States if you type United.
“The things we’re decoding are not just single words but the probability of any single word,” says Chang. “We’ve been working to learn how the brain processes words, how the electrical activity of the brain gives rise to consonants and vowels, how they give rise to the planning of words.”
The system also recognises the parts of the brain that control the lips, jaw, tongue, and larynx. By thinking about speaking, even people who have lost the ability activate these brain centres in ways that would form any given word.
Elsewhere, Blackrock Neurotech has implanted more than 50 people with brain chips, and boasts of amassing thousands of patient-days without adverse events. The most common of these events are infection of brain tissue at the site of the implant; malfunction of the implant, causing it to send spurious signals that would damage the brain; or fibrous encapsulation of the chip, as tissue grows around it, causing it to fail.
With these risks avoided, Blackrock focuses on using its system to allow patients to operate computers and, as with Echo, speak via an on-screen avatar. Solzbacher describes one ALS patient whose disease had progressed to what is known as locked-in syndrome, in which the mind remains alert but with no way to communicate with the outside world. That patient underwent surgery to have a chip implanted and a computer voice created.
“He was able to talk to his three-year-old daughter,” says Solzbacher. “That was the first time that happened in his daughter’s life and it’s quite powerful, actually.”
The matter of how a locked-in patient gives consent to the surgery is a tricky one. Typically, says Solzbacher, consent is given earlier in the course of the disease, before the subject slips into a completely locked-in state. Relatives may also be in possession of advance declarations the patient made while still able to communicate.
Blackrock has been at this work for a while. In 2014 it ran a clinical trial in which Ian Burkhart – paralysed from the elbows down at 19 when he dived into a wave that pushed him into a sandbar – was implanted with a brain chip and then outfitted with electrodes on the skin of his forearm, hand, and elsewhere.
Merely by thinking of moving his extremities he could activate the electrodes, which would cause the arm or hand to move as commanded, allowing him to grasp and hold objects and even play Guitar Hero.
Brain-computer interface technology sometimes doesn’t even require scientists to bother the brain at all. Even minimal, lightly invasive brain surgery is still, well, brain surgery, and at New York City–based Synchron Inc they’re able to avoid it. Instead, they thread a probe carrying a chip through the radial artery in the forearm or the femoral artery in the thigh up to the brain and deposit the chip in the main vein between the brain’s two motor cortexes.
From there, says Kurt Haggstrom, Synchron’s chief commercial officer, “you can actually listen to the brain and understand it, without ever having to touch the brain itself”.
The first BCI surgery was in 1998, when neurologist Philip Kennedy implanted a chip in the brain of a man who suffered from locked-in syndrome caused by a brain-stem stroke. After intensive practice, the patient was able to move a cursor on a screen, a significant achievement, but a painstaking one.
The limitations in the results were partly the result of the limitations of the chip, which was a four-channel model that was able to carry only minimal information.
“It was a very primitive device,” says Jamie Brannigan, a resident neurologist at Mount Sinai in New York and a BCI expert. “But it was the first example of an in-human brain-computer interface.”
Since then, a more powerful chip, the Utah array, has become the default device for the BCI field. The chip measures 4 mm by 4 mm and includes 100 needle-like probes, each measuring 1.5 mm, which penetrate brain tissue. It was first implanted in a human being in 2004, and has been the go-to chip for most BCI work since.
“The Utah array has a proven track record of safety, reliability, and longevity,” says Solzbacher.
Blackrock’s 50 implant surgeries certainly suggest there’s evidence behind what it claims it can do with the chip, but the company’s competitors aren’t so certain. For starters, even at 4 mm by 4 mm, the Utah array would be too big and clumsy a hunk of hardware for Science to implant in the eye or Synchron to thread through a vein. And the 100 probes, while a not inconsiderable number, put a ceiling on how much data the system can carry.
“The thing about the Utah array is that it’s a 1990s device,” says Brannigan, “and if you were using a 1990s chip in your smartphone, you’d very quickly know about it.”
There is also the sort of controlled trauma inflicted on the brain when the 100-wire chip punches 100 tiny holes in its surface. At Paradromics, researchers have developed a chip with thinner wires than the Utah array, something that reduces, but does not eliminate, the damage done to brain tissue.
“There’s a fairly considerable amount of brain injury and cell loss with those technologies,” says Chang. “There is a term in the field called butcher ratio, and it refers to the number of cells killed for every one neuron you can record from. The more electrodes you put into the brain, the cumulative injury rises. The worst-case scenario is that a patient has some residual function, and this is lost as a result of the implantation.”
His company, Echo, sidesteps the problem, using not a chip with probes, but a thin film that sits on the brain without penetrating it. “The film is just lying safely on the brain surface monitoring signals from there,” says Chang.
It’s Hodak’s biohybrid model that would represent the real revolution in chip design, but the technology is not yet ready for human experimentation.
Hodak concedes the biohybrid model poses perils. There is always a chance the stem cells could grow uncontrollably, crowding out native cells and damaging the brain. To prevent that, the system has a “kill switch” in the form of an antiviral drug called ganciclovir that could be used off-label to attack the new cells and stop the growth process.
The hardware the BCI players are building might be impressive, but it’s very much in the beta stage – not remotely ready for release.
The Prima system has given Alice Charton the ability to read the newspaper, but the glasses are wired to an almost 1kg plastic brick that houses the processing computer and the battery.
Buckhart, whose Blackrock implant enabled him to move despite his paralysis, had the chip removed seven years after it was
implanted. The system included a wire that ran from the implant to a small hole in his skull to which a cable was screwed when he was using the chip.
Over time, the skin of his scalp would try to grow over the port, leading to repeated infections and causing him enough discomfort that he gave up on the system. He would have it re-implanted, he says, but only if it were wireless and required no port.
See more from MedicalBrief archives:
First wireless brain chip implant ‘a success’ – Elon Musk
FDA green-lights Musk’s brain implant for human study
Step forward in human studies for Neuralink brain implants