Speech Therapy News

Display News Filters

Brain-machine interface can produce synthetic speech

Friday 24th May 2019
New technology has allowed brain recordings to be transformed into synthetic speech, representing a major step towards helping those who have lost the ability to speak.
Latest Jobs

Several different conditions - from strokes to traumatic brain injuries to degenerative diseases like Parkinson’s - result in patients being unable to speak. In some cases this ability can be regained, but for some patients it is lost forever. 

However, a new machine has been able to generate natural-sounding speech from a patient’s brainwaves, which could potentially give people back the voices they have lost.

Developed by a team of neuroscientists at UC San Francisco, the technology required the creation of a virtual vocal tract that simulates the movements made by the lips, tongue and vocal chords to produce speech. This was then combined with an algorithm that translated brainwave patterns into the movements they were trying to produce.

The result is synthetic speech that is much faster and more natural than other options currently available. For example, devices that produce speech using small eye or facial muscle movements to spell out words letter-by-letter average around ten words a minute, which is up to 15 times slower than ordinary speech.

However, the technology is some way off being perfect. In crowdsourced transcription tests, participants were able to identify 69 per cent of synthesised words and 43 per cent of synthesised sentences correctly when they were given lists of 25 words as options to choose from. 

However, when the list of available words was increased to 50, the accuracy rate dropped to 47 per cent of words and 21 per cent of sentences.

Josh Chartier, one of the leaders of the study, said: “We still have a ways to go to perfectly mimic spoken language. We're quite good at synthesizing slower speech sounds like 'sh' and 'z' as well as maintaining the rhythms and intonations of speech and the speaker's gender and identity, but some of the more abrupt sounds like 'b's and 'p's get a bit fuzzy.

“Still, the levels of accuracy we produced here would be an amazing improvement in real-time communication compared to what's currently available."

Written by Martin Lambert

Mediplacements is a genuine specialist medical recruitment company. Contact us to see the latest speech therapy jobs.

Image Credit: 7activestudioADNFCR-1780-ID-801850273-ADNFCR

Related News