Brain implant translates thoughts into language

The transplant could help patients with neurological diseases such as amyotrophic lateral sclerosis.

Neuroscientists Duke University in North Carolina, USAThey develop a brain implant that can convert thoughts into language. It aims primarily to help people who cannot speak due to neurological diseases communicate again.

These include debilitating movement disorders such as: ALS (amyotrophic lateral sclerosis) Or that Engaged in symptoms, where those affected cannot move or speak but remain fully conscious. The famous physicist Stephen Hawking suffered from amyotrophic lateral sclerosis.

➤ Read more: No more headaches? Researchers develop a brain implant to combat migraines

“The tools currently available to help them communicate are generally very slow and cumbersome,” says one of the project’s lead researchers. Gregory Kogana neurologist at Duke University School of Medicine, Towards the platform Interesting geometry.

The implant consists of a flexible piece of medical plastic – about the size of a postage stamp – and is equipped with 256 small sensors. It has been tested so far on 4 patients.

Patients had to solve a task using the implant during brain surgery

These are patients who had to undergo brain surgery anyway due to an existing disease, such as treatment for Parkinson’s disease or a tumor. They received the implant during the operation and had to solve a task with it. All this happened in a short period of time during the operation.

“I like to compare it to a NASCAR crew,” Coogan said, referring to American motorsports. “We didn’t want to allow extra time for the process, so we had to do it in the meantime 15 minutes To be terminated. Once the surgeon and medical team say “Go!” “We got down to business and the patient did the job,” he said.

➤ Read more: Feel it again with synthetic leather

The task asked patients to repeat certain sounds while the device coordinated the movements of about 100 muscles that move the lips, tongue, jaw and larynx. The device also recorded the activity of each patient’s speech motor cortex, which the researchers want to use for the AI ​​system. This would enable patients to pronounce sounds.

The implant was 84 percent accurate

To test how well the machine learning algorithm could identify sounds produced by recordings of brain activity alone, the researchers fed it neurological and linguistic data collected in the operating room. It turns out that sounds can be decoded with an average accuracy of up to 40 percent. Some sounds even achieved accuracy Up to 84 percent.

This innovation will have the potential to expand our knowledge of the brain and significantly improve the quality of life for people with a range of neurological disorders that prevent them from speaking.

➤ Read more: Salzburg: The patient received a hood from the 3D printer

However, brain transplants require invasive procedures and can impair brain function, which is why they raise questions regarding ethics, safety and data protection. To ensure that these devices are safe and effective, and that ethical considerations are taken into account when using them, researchers and medical experts work hard to constantly adapt and improve their devices.

The study was published in the specialized journal Nature Communications Published and can here Can be read.