The news is encouraging and brings a very new perspective to the approach of some neurological deficits. The entire scientific community dealing with human-machine interfaces has echoed the news.

On July 15, 2021, a study was published in the New England Journal of Medicine by a team of neurosurgeons from the University of California at San Francisco, in which the researchers implanted a prosthesis - or rather a "neuroprosthesis" - capable of decoding neuronal activities. The feat is unprecedented.

But, what is a neuroprothesis? Basically, they are devices that could decode sensory signals (from our vision, motor activities, sensory inputs, or even cognitive signals for language tasks in this case) - of single or a bunch of neurons at the same time - and then, based on computer signals processing (you can call this a decoding stuff) and then return that information for an individual to perform a functional they lost or would like do "enhance".

What caught my attention in this case was the fact that this device could be an interesting solution for some specific cognitive problems, where other areas of the brain are relatively intact, but there is a potential benefit for a function that is assigned to an anatomically delimited area. Watch the following video on the scientific journal's Instagram.

In the illustration, it is possible to see a representation of the patient with the neuroprosthesis.

Unfortunately, from a medical point of view, individuals who have lesions in extensive areas or that have more severe repercussions, tend to be poor candidates for these therapies, since as in the possible case of this patient, although the motor area of speech could be compromised, apparently there is a preservation in the brain area responsible for language, a function located in areas relatively well studied and mapped.

The implantation, therefore, took place in areas related to human speech and language, and through advanced computer algorithms, captured by a sensor and specially designed decoders, it was possible to decipher the brain activity from computational decoding, using technologies related to machine learning. The patient in question had suffered a stroke" in an area of the spinal cord - a region that is not usually related to language problems - and had suffered the injury 15 years earlier.

Post from the official account of the Department of Neurosurgery at the University of California San Francisco (UCSF), about the research results.

This type of technique shows that people who have had lesions in certain parts of the brain that are still preserved can gain advantages when this decoding is possible to be translated into clear signals, understood by other people who do not have the disease, bringing an improvement in the functional capacity of individuals with these conditions.

Other types of devices, as we mentioned, had already been developed using prostheses related to motor activities, but this is one of the few times - possibly the first - in which cognitive activities could be decoded using more advanced algorithms.

For this, as I said, the encoded neuronal region was most likely still fully functional (i.e., with neuronal activity intact or very close to normal), which would be very difficult to occur in people who have lesions in these areas and who no longer produce these neuronal stimuli.

Video produced by the University of California at San Francisco.

There is undoubtedly a great recent advance in computer machine interfaces capable of decoding neuronal activity. These technologies usually called human machine or brain machine interfaces are still in very restricted use in research, but there have been very consistent advances and this is certainly encouraging news. In this video, you can get an idea of some of the things that have been emerging in this area of research.

Originally published on the blog on July 18, 2021, 10:45 AM (UTC-3 BRT). Last updated on July 18, 2021, at 11:05 AM (UTC-3 BRT). Cover photo credit to Daniel Fazio.