AI hearing aid uses brain activity to amplify specific voices

An experimental hearing aid powered by artificial intelligence essentially reads a user's mind to separate out and amplify the voices they want to hear, according to a report published in Science Advances on May 15.

A group of researchers from Columbia University in New York City developed a system that analyzes brainwaves to single out individual voices in loud environments. Though the unimpaired human ear naturally performs this "auditory attention decoding," hearing aids can currently only distinguish between voices and general background noise, not between individual voices.

In the study, electrodes were implanted on the brains of three voluntary subjects to track neural responses to the sounds of various voices. A decoding algorithm analyzed those responses, then amplified only the sound to which a subject was paying the most attention. In contrast to previous attempts to create a similar device, this one was not only successful in accommodating users' shifts in attention, but could also recognize new voices without previous training.

The device is still in its earliest stages of development: Though it did accomplish the intended decoding, its use of electrodes implanted on the brain is less than practical for widespread use. The study's authors noted that further advancements will need to be made to combine such a high-powered algorithm with the current limitations of external wearable technology.

Still, they wrote, the experimental device "represents a feasible solution for a major obstacle in creating a brain-controlled hearing device, therefore bringing this technology a step closer to reality."

More articles about AI:
AI can predict patient response to popular Crohn's treatment
Machine learning tool detects Alzheimer's in brain tissue
Elon Musk's neurotech startup falls short in funding round

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months