People who have lost their ability to speak due to severe paralysis may hope to find a better way to communicate, thanks to technology that essentially reads their thoughts and translates them into whole words and sentences in real time. By clicking on the part of the brain that controls the voice box, the researchers figured out how to read these signals and convert them into words with sophisticated artificial intelligence.
Apart from the high-precision algorithms, the new method hits the part of the brain that sends signals to voicemail. Previous techniques used areas of the brain that control the hand or arm. It seems almost obvious that scientists should have used this part of the brain from the start, but this is neurological research.
While developing the system, UCSF researchers recorded volunteers' brain signals with uninterrupted speech. The scientists fed the patterns to neural networks, which they learned to decode in real time. They also used a statistical language model to improve the accuracy of the algorithm.
The researchers' main obstacle was whether the region that controls speech has the same function in people with the disorder as it does in people with normal speech. So they hired an anonymous volunteer (Bravo1) to create a simple 50-word vocabulary to feed the algorithms — practical words he could use in everyday life like water, well, me, yes, no, etc. use.
During the experiment, the team asked Bravo1 simple questions such as "How are you?" or "Do you need anything?" When Bravo1 attempted to respond verbally, the computer translated his motives into phrases such as "I'm fine" or "No, I don't need anything."
"Edward Chang," UCSF neurosurgeon and Professor Jane Robertson, the study's lead author, said, "In our view, this is the first demonstration of success." Direct coding is the full expression of the brain activity of a person who is paralyzed and unable to speak.” “It shows powerful promise for communication by tapping into the brain’s natural speech machine.”
The system is still relatively slow. It can translate up to to 18 words per minute - for comparison, ideal people can speak about 200 words per minute. Still faster than any previous neuroprosthetic device. Translation accuracy 93% and average 75%.
However, this is just the beginning of a study Expanded.The next steps for this team are to increase the number of subjects, expand the vocabulary of the system and improve the rate of speech translation.
Researchers have developed a deep learning method to translate audio signals from the brain into text
It is sometimes difficult to write objectively about the progress of SpaceX. The i...
Last Thursday, the Russian space station's new large unit, Navka, was finally conn...
We are now - often horrifyingly - watching what happens to the virus and ...