Protected from hostile biological processes by less than a micrometre of material, the achievement is said to be an important step toward creating high-resolution neural interfaces that can persist within a human body for an entire lifetime.
Neural implant programmed and charged remotely
The results, available online in Science Translational Medicine, were published by researchers led by Jonathan Viventi, assistant professor of biomedical engineering at Duke University; John Rogers, the Louis Simpson and Kimberly Querrey Professor of Materials Science and Engineering, Biomedical Engineering and Neurological Surgery at Northwestern University; and Bijan Pesaran, professor of neural science at New York University.
"Trying to get these sensors to work in the brain is like tossing your foldable, flexible smartphone in the ocean and expecting it to work for 70 years," Viventi said in a statement. "Except we're making devices that are much thinner and much more flexible than the phones currently on the market. That's the challenge."
Current long-term implantable devices are almost universally hermetically sealed within a laser-welded titanium casing.
"Building water-tight, bulk enclosures for such types of implants represents one level of engineering challenge," Rogers said. "We're reporting here the successful development of materials that provide similar levels of isolation, but with thin, flexible membranes that are one hundred times thinner than a sheet of paper."
Existing neural interfaces sample around 100 sites in an organ composed of tens of billions of neurons. According to Duke, any attempt to make these devices larger in confronted by the hurdle of wiring logistics as each sensor requires its own wire, which leads to size constraints.
"You need to move the electronics to the sensors themselves and develop local intelligence that can handle multiple incoming signals," said Viventi. "This is how digital cameras work. You can have tens of millions of pixels without tens of millions of wires because many pixels share the same data channels."
Through their work, the researchers have demonstrated flexible neural devices 25 micrometres thick with 360 electrodes. Previous attempts to keep them safe from harm inside the body have failed, as even the tiniest defect can thwart the entire effort.
"We tried a bunch of strategies before. Depositing polymers as thin as is required resulted in defects that caused them to fail, and thicker polymers didn't have the flexibility that was required," said Viventi. "But we finally found a strategy that outlasts them all and have now made it work in the brain."
In their paper, the team demonstrate that a thermally grown layer of biocompatible silicon dioxide less than a micrometre thick can ward off the hostile environment within the brain, degrading at 0.46nm per day.
They also show that, even though the glass encapsulation is not conductive, the device's electrodes can detect neural activity through capacitive sensing. They implanted a 64-electrode neural interface into a rat for over a year and a 1,008-electrode neural interface into the motor cortex of a monkey reaching to a touchscreen.
"Successfully deploying the device in monkeys doing human-like tasks is a huge leap forward," said Perasan. "Now we can refine our technology to help people suffering brain disorders."
UK productivity hindered by digital skills deficit – report
This is a bit of a nebulous subject. There are several sub-disciplines of 'digital skills' which all need different approaches. ...