Neuro-engineers at Columbia University in New York say they have created a system that can translate human thoughts into a recognizable speech that will revolutionize not only medicine but also communication.
By monitoring the activity of the brain of subjects, researchers at the Brain Brain Institute of Mortimer B. Zuckerman of Colombia have managed to train artificial intelligence to turn their thoughts into comprehensible sentences, the newspaper published in the journal Scientific Reports writes. The authors see patients with impaired speech due to illness or trauma as the first adopters of the initial technology.
"We have shown that, with the right technology, the thoughts of these people could be decoded and understood by each listener" said Dr Nima Mesgarani, senior author of paper.
Since early attempts by researchers to translate brain activity into a recognizable speech failed, they turned to a computer algorithm that could generate a speech called vocoder. The algorithm improves more "Trained" with footage of human speech.
The researchers translate brain signals directly into Speechhttps: //t.co/qEStGOoPOW
This breakthrough, which strengthens the power of speech synthesizers and artificial intelligence, can lead to new ways for computers to communicate directly with the brain. #ai#BCI#speech# science
– News from Neuroscience (@NewscienceNew) January 29, 2019
"This is the same technology used by Amazon Echo and Apple Siri to give verbal answers to our questions" said Dr. Messagrani, who is also a professor at the Faculty of Engineering and Applied Science at the Fu Foundation.
Vokoder was trained to interpret brain work with the help of Dr. Asheesh Dinesh Mehta, a neurosurgeon at Northwestern Institute of Northwestern Neuroscience in Long Island and the co-author of the paper.
"Working with Dr. Mehta, we asked patients with epilepsy who were already undergoing brain surgery to listen to sentences spoken by different people while we measured brain activity" said Dr. Mesgani. "These nerve models train the vocoder."
After the completion of this training, the next phase started. Patients listened to a person reading numbers from 0 to 9, while the algorithm scanned the brain activity and tried to translate it into sound. The result was robotic numbers for reading voices, which human listeners could understand and repeat with an accuracy of 75 percent.
This may seem rather modest, but Dr. Meggani said that such a result is "Above and beyond any previous attempts." The researchers plan to further improve the system so that it could take them as the input models of the brain to a person thinking about speech, not to listen to it.
"This will be a game changer, which will give everyone who has lost the ability to speak, whether through injury or illness, a renewed chance to connect with the world around them" Dr Mesgrani said.
Technology also needs to work with more complex words and sentences to become more practical. The ultimate goal of the team is to create an implant that synthesises speech directly from thought.
Do you think your friends will be interested? Share this story!