Scientists have created a device that turns brain signals into words

It is so natural and simple for most of us that we do not even think about what a complicated process represents. However, there are people who can not speak for various health issues. However, research shows them a better future. Recent scientists have created a decoder that analyzes brain signals and transforms them into speech.

The most famous case of a man able to communicate despite the fact that he was unable to speak was a brilliant scientist, Stephen Hawking. The voice synthesizer he worked on was based on detecting the movements of his cheeks. Most devices designed to replace human speech rely on muscle or eye movements. Their biggest drawback is that they are slow. While a normal conversation reaches 100-150 words per minute, these devices can "say" less than 10 words per minute.

However, in the latest research, scientists have chosen a totally different approach, more sophisticated, bolder and much more in perspective. They did not focus on the voice as such, but on the cerebral signals that dictate to the authorities to create it. In doing so, they simultaneously coordinate the lips, tongue, throat, jaw, larynx and diaphragm. It is therefore an extremely complex task and very difficult to simulate for medical engineers. However, if we could decode brain signals and transform them into sound with the instruments, we would create an artificial voice organ.

The movements of speech organs are extremely subtle, fast and complex. Therefore, it is very demanding to decode them, even for powerful computer simulations. A team of scientists from the University of California at San Francisco however succeeded. Thanks to five volunteers waiting for a brain operation for epilepsy problems. Scientists have implanted electrodes in the brain to map cerebral signals during surgical preparations. The role of the volunteers was then to read. They read the sentences, the words and the scientists collected data on the brain signals captured by the electrodes.

After collecting the data, the scientists had a difficult task. First, they "translate" the signals into the movements of the voice, before turning into sounds. Finally, the device has learned an algorithm that can compare and match the individual cerebral signals with the corresponding voice movements and give them the necessary sound. The result is a device called a virtual voice path. Driven by the brain, it creates a sound imitation of the human voice.

Although the voice of the device is unnatural and some sounds or words are difficult to understand, it is a significant step forward. An apparatus that a person unable to speak could control with his own mind is a revolutionary solution to simplify the lives of people with disabilities. Patients with Parkinson's disease, amyotrophic lateral sclerosis or throat cancer, in particular, might be helpful.

The Guardian, Science Alert, Nature