The song of birds can be reconstructed by an AI based on their brain activity

0
33
destacada 16 730x411.jpg
destacada 16 730x411.jpg

The brain activity of birds, when processed by an algorithm specially dedicated for this purpose, can provide information to reconstruct their characteristic whistles.

An investigation that experiments with these elements, was presented as proof of the possibilities that in the future could be counted on to develop vocal prostheses for humans.

Vocal prostheses for humans could base their operation on the song of birds

This test, which just passed its first phase of studies, was presented by a team from the University of California in San Diego, United States. The researchers were able to demonstrate that it is possible to reproduce the complex vocalizations of a songbird, considering aspects such as the original tone, volume and timbre of the animal in question.

To conduct the study, the researchers implanted silicon electrodes in zebra finches adult males and analyzed their neural activity while singing. These recordings were based mainly on the electrical activity of multiple populations of neurons, in the sensorimotor part of the brain that, among other functions, activates the muscles responsible for singing.

The recorded recordings were used as training material for an algorithm that, once prepared, had the mission of being able to build songs based only on their brain impulses, considering that separately this machine learning system already knows the correlation between this activity. and the performance of the song.

The first shared results show really significant progress. The first part of the accompanying recording shows a natural song, recorded in a non-intervened environment. The final part shows a recording reconstructed by the presented algorithm.

SEE ALSO  Nothing confirms the Phone (2a): this will be its expected cheap smartphone to sweep sales in Spain

Achieving this was not easy, according to the team. To achieve this, it was necessary to obtain simple representations of the vocalization patterns of the birds. In simple words, by this they mean mathematical equations that model the changes in pressure and tension that occur in the vocal organ of finches when they sing. The researchers then trained their AI to map neural activity directly to these representations.

“By having this simple representation of the complex vocal behavior of songbirds, our system can learn maps that are more robust and more generalizable to a wider range of conditions and behaviors.”, he pointed Vikash Gilja, professor at the university that carried out this study, in which he collaborated as an adjunct researcher.

Regarding the value of this research, its lead author, Timothy Gentner, professor of psychology and neurobiology at the University of California, San Diego, commented that “The current state of the art in communication prostheses is implantable devices that allow generating textual results, writing up to 20 words per minute”. «Now imagine a vocal prosthesis that allows you to communicate naturally with speech, saying out loud what you are thinking almost as you are thinking it. That is our ultimate goal, and it is the next frontier in functional recovery »added.

Why do the bird study?

In the presentation of the investigation report, it is stated at the beginning that “Language is exclusively human, but its acquisition may involve cognitive capacities shared with other species”.

It is still curious, since the communication mechanisms of birds seem to be quite different from ours. However, according to the same research, the vocalizations of a songbird are similar to human speech in several ways, being complex procedures, which depend on learned behaviors.

SEE ALSO  OPPO's most recent mid-range comes with a surprise: the company gives away headphones with your purchase

“With our collaboration, we are leveraging 40 years of bird research to build a speech prosthesis for humans, a device that would not simply convert a person’s brain signals into a rudimentary set of whole words, but would give them the ability to do any sound and therefore any word they can imagine, freeing them to communicate whatever they want », Gentner indicated in this regard.

The report that compiles the exercises and findings of this study was published in Current Biology.