It’s easy to be dazzled by the latest news from Neuralink, the Elon Musk-financed company that, last week, announced its brain-machine interface (BMI) would be ready for human testing as early as 2020. In an event on July 16, Musk and the Neuralink team explained how they built on decades of BMI research to develop thin, flexible electrodes that could be “sewn” into the brain by a robot, allowing it to safely record brain activity.
A big question that remains is the same one that’s loomed over BMI research for decades: Will the implants last?
The human brain, ever defensive, generally reacts badly to things being inserted into it — even very sophisticated ones.
Cary Kuliasha, Ph.D., a University of Florida materials science researcher who studies how the body environment affects biomedical devices, wonders about durability.
“In my opinion, the primary hurdles researchers face is creating neural implants that can maintain functional integrity during the full duration of the patient’s lifetime,” Kuliasha tells Inverse.
Implantable devices, he says, must be able to withstand the body’s “harsh” environment for years or even decades. MIT Technology Review called longevity the “bugbear” for Neuralink.
Christopher Bettinger, Ph.D., a professor of materials science and biomedical engineering at Carnegie Mellon University, tells Inverse that invasive implants should “‘work’ (e.g. bring benefit to the user) for at least 5-10 years to justify the surgery/implantation in the first place.”
Whether Neuralink can withstand the harsh conditions of the brain remains to be seen, but Kuliasha and Bettinger suggest there’s no ruling it out just yet.
For a brain and a machine to interface, they must be able to communicate with one another. Researchers in the field have settled on “spikes” in neural activity — peaks in the electrical field created by an active neuron — as a common language. Electrodes, traditionally made of metal or silicon, inserted near these neurons can pick up electrical activity and send that information to a computer.
Neuralink’s flexible polymer electrodes, which are less than the width of a human hair, show promise in their ability to be inserted safely into the brain with the help of the company’s new neurosurgery robot, which can reportedly thread the electrodes between blood vessels safely. The team has tested this procedure out on rats (and allegedly monkeys, though there’s no evidence for that yet) recording from around 1,000 neurons, but it doesn’t have much data on the longevity of the electrodes (and their associated chips) in the brain. The company, after all, has only been around since 2016.
"I think it’s very impressive."
“I think it’s very impressive,” says Bettinger of Neuralink’s flexible polymer electrodes. “I’d be interested to see a detailed accelerated lifetime study and see how the device performs over longer (simulated) periods of time.”
The Human Body Doesn’t Like Invaders
Longevity is an issue for neural implants — and implants more generally — because the body generally doesn’t like invaders. “The human body typically does not respond well to foreign objects and does everything it can to either digest/degrade the device or encapsulate it from the rest of the body,” explains Kuliasha. “Both of these interactions can cause neural implants to fail over time.”
In an interview with Technology Review, a “brain-computer interface pioneer” named Nathan Copeland, who had a neural implant inserted following an accident that left him paralyzed from the chest down, explained what must be done to maintain his device, known as a Utah array: “I do have pedestals on top of my head, and I have to worry about keeping it clean, and the skin receding.” And that’s just the external part.
Internally, says Kuliasha, the body will attempt to isolate a “foreign object” from everything else.
"Furthermore, the body’s own immune system typically tries to degrade the device itself."
“Since a neural implant is used to provide direct electrical feedback with the brain, any sort of encapsulating material/tissue that forms around an implant can drastically reduce the efficacy of the device to perform its function,” he says. “Think of it as the body wrapping a blanket of insulating material around an electrical wire.” Doing so would hamper communication between the device and brain tissue, rendering the neural implant useless.
“Furthermore, the body’s own immune system typically tries to degrade the device itself,” he continues.
This range of bodily attacks makes it hard for neural implants to last over time.
“Classically, the failure modes are associated with shorting of the device and degradation of the insulation layers,” says Bettinger. “Reliable connectors that can withstand implant micromotion are also very challenging to develop.”
The Issue of “Biocompatibility”
Attempts to isolate and degrade neural devices aside, Bettinger explains that another problem that influences longevity is “biocompatibility,” which has less to do with how long the device can last in the brain and more to do with how well (and how safely) it can interact with its surroundings.
“It has been shown that long-term function of the device correlates with the ability to insert devices while avoiding the rupture of blood vessels,” he says. To Bettinger, the biggest safety-related issue for invasive implants is the surgery itself, together with complications related to infection and rejection of the device.
“In the case of the Neuralink device,” he says, “the technology is very exciting because (as they claim) the device is small and can be inserted in a targeted way to avoid the rupture of blood vessels, thereby increasing the likelihood of long-term biocompatibility.
Why It’s Hard to Achieve Longevity
The goal of materials scientists and engineers is to design materials that can work with the harsh environment of the body or at least camouflage devices from the immune system, says Kuliasha.
Part of his work on BMIs — “Our end goal is basically to allow for prosthetic limbs to have the same functionality as Luke Skywalker’s in The Empire Strikes Back,” he says — is making flexible polymeric microelectrode arrays and assessing their long-term reliability.
Bettinger’s team chooses to forgo the perils of invasive implants altogether, instead working with flexible adhesive electrodes that use hydrogels, which makes them less rigid and thus more compatible with brain tissue. The devices “have their own challenges,” he says, “but we are using these devices to interface with tissues in the peripheral nervous system.”
Neuralink’s electronics, as the newly published white paper describes, are packaged “for long-term implantation”. The tiny chips that gather data from the electrodes are packaged in “titanium cases which are coated with parylene-c, which serves as a moisture barrier to prevent fluid ingress and prolong functional lifetime.” For how long, we don’t know yet.
Neuralink’s use of flexible polymeric electrodes in place of stiff metallic or silicon versions, Kuliasha points out, isn’t that new — he uses them in his own research — and have been used in neural interface devices for the past decade or so. To him, it’s the robotic “sewing machine” that Musk’s team uses to implant the electrodes, that’s the novel part of the announcement.
“It is especially innovative with the number of individual devices and electrode sites they are planning to implant,” he says.