Scientists convert brain signals into speech using Artificial Intelligence system

Scientists have created a system that translates thoughts into intelligible, recognisable speech, an vance that may help people who cant speak regain ir ability to communicate with outside world.

By monitoring someone's brain activity, techlogy developed by researchers from Columbia University in US can reconstruct words a person hears with unprecedented clarity. 

breakthrough, which harnesses power of speech synsisers and artificial intelligence, could le to new ways for computers to communicate directly with brain. 

It also lays groundwork for helping people who cant speak, such as those living with as amyotrophic lateral sclerosis (ALS) or recovering from stroke, regain ir ability to communicate with outside world, researchers said.

ALSO RE | Artificial skin could give humans 'superpowers'

"Our voices help connect us to our friends, family and world around us, which is why losing power of one's voice due to injury or disease is so devastating," said Nima Mesgarani, of Columbia University in US.

"With today's study, we have a potential way to restore that power. We've shown that, with right techlogy, se people's thoughts could be decoded and understood by any listener," said Mesgarani, a principal investigator of study published in journal Scientific Reports.

Deces of research has shown that when people speak -- or even imagine speaking -- telltale patterns of activity appear in ir brain. 

Distinct pattern of signals also emerge when we listen to someone speak, or imagine listening.

ALSO RE | Scientists convert Wi-Fi signals into electricity using a flexible device: Here's how

Experts, trying to record and decode se patterns, see a future in which thoughts need t remain hidden inside brain -- but inste could be translated into verbal speech at will.

However, accomplishing this feat has proven challenging. Early efforts to decode brain signals by researchers focused on simple computer models that analysed spectrograms, which are visual representations of sound frequencies.

However, because this approach has failed to produce anything resembling intelligible speech, team turned inste to a vocoder, a computer algorithm that can synsise speech after being trained on recordings of people talking.

ALSO RE | Scientists develop flexible robots that can swim through blood vessels to deliver drugs

"This is same techlogy used by Amazon Echo and Apple Siri to give verbal responses to our questions," said Mesgarani.

Researchers plan to test more complicated words and sentences next, and y want to run same tests on brain signals emitted when a person speaks or imagines speaking. 

Ultimately, y hope ir system could be part of an implant, similar to those worn by some epilepsy patients, that translates wearer's thoughts directly into words.