Recently, more and more ambitious brain-computer interfaces and nerve repair techniques have become news. Last month, Elon Musk’s company Neuralink demonstrated wireless BCI with more than a thousand flexible electrodes inserted directly into the brain by a specialized robotic surgeon. (So far, the company has only shown short-term use in pigs.) Inserting electrodes is tricky. Although brain surgery is not entirely rocket science, there are risks to whether the surgeon is a robot. Even if the soft and delicate electrodes shown by Neuralink are sufficiently invasive, the brain will try to resist them and coat them with glial cells, thereby reducing their ability to conduct the electrical impulses they seek. Although implanted electrodes (such as the more commonly used “Utah Array”
At the same time, electrodes placed directly on the scalp absorb brain waves (electroencephalogram or EEG), but these electrodes lack the spatial details of implanted electrodes. Neuroscientists roughly know which part of the brain is doing what, but the more you know which neurons are firing, the more clearly you will know what they are firing.
Cortical EEG is a recent innovation that places the electrode net directly on the surface of the brain. Combined with the intelligent spectral processing of the signals received by these electrodes, ECoG is sufficient to convert the movements of the parts of the motor cortex that control the lips, jaws and tongue into text or even speech. There are other ways. Facebook acquired CTRL-labs for as much as $1 billion in 2019 in an attempt to obtain motion signals from wrist neurons. The inner core uses functional near-infrared spectroscopy on the head to sense brain activity.
Oxley and his colleagues’ stent electrodes, if they continue to show good results, will fit the spectrum between the implanted electrode and the EEG. Its inventor hopes that the second thing is closer to the first thing. But this is still in its infancy. Vikash Gilja said: “The core technology and core ideas are very cool, but considering where they access the signal, compared to other human-machine interface strategies, this is a relatively low-fidelity signal.” He is responsible for the transformational neural engineering of the University of San Diego UC laboratory. “We at least know that high-density ECoG recordings from the surface of the brain can convey information beyond the scope of this article.”
One possible problem is that the tissue conducts electrical impulses, but the electrodes in the stent receive signals from the brain through blood vessel cells. This reduces the signal content. “If we were to record the recording of the cortical surface and compare it with the Utah array experiment (a lot of clinical experience with implanted electrodes), I would say that the recording method in ECoG limits the rate,” Gilja said. (Just for transparency reasons, I should point out that Gilja does paid work with BCI companies including Neuralink. In theory, Synchron can one day compete with it.)
Therefore, it may not be good enough for neuroscience, but it may be useful for paralyzed patients who wish to maintain a low BCI and do not need to drill the skull. Andrew Pruszynski, a neuroscientist at Western University in Canada, said: “To become an intrusive technology, there is a trade-off between the degree to which information is collected.” “This is an attempt. Reach the middle ground and insert the catheter close to the nerve activity.” It is obviously invasive, but certainly not as invasive as inserting electrodes into the brain. “