An intelligent hearing aid that allows the user to focus on specific voices has been created by researchers, who say that it could transform the ability of people with hearing loss to handle noisy environments.
The unit mimics the brain's natural ability to sing out and enhance a voice against background calls. So far, the most advanced hearing aids are also working by increasing all voices simultaneously, which can be perceived as a sound cacophony for the wearer, especially in tight environments.
Nima Mesgarani, who led the latest research at Columbia University in New York, said: "The brain area that treats sounds is extraordinarily sensitive and powerful. It can amplify a voice over others, seemingly simple, while today's hearing aids are still pale comparison. "
This can seriously impede the ability of a bearer to engage in conversations, making busy social occasions particularly challenging.
Researchers have been working for years to solve this problem, called the cocktail party effect. The brain-guided hearing system seems to have cracked the problem using a combination of artificial intelligence and sensors designed to monitor the listener's brain activity.
The hearing aid first uses an algorithm to automatically separate the voices from several speakers. It then compares these audio tracks to the listener's brain activity. Earlier work from Mesgarani's lab found that it is possible to identify which person is attracting attention, as their brain activity traces the sound waves in that voice closest.
The device compares the sound of each speaker to the brain waves of the person wearing the hearing aid. The speaker whose voice pattern fits closest to the listener's brain waves is amplified over the others so that they can easily adapt to that person.
The researchers developed an earlier version of the system in 201
To test the device, the laboratory-employed epilepsy patients already had electrodes implanted in the brain to monitor seizure activity prior to planned brain surgery.
The patients were played sound by various speakers while their brain waves were monitored via the electrodes implanted in their brain.
An algorithm followed the patient's attention when they listened to different speakers that they had not previously heard. When a patient focused on a speaker, the system automatically amplified that voice, with a delay of just a few seconds. When their attention was shifted to another speaker, the volume levels were changed to reflect this shift.
The current version of the hearing aid, which involved direct implants in the brain, would be inappropriate for ordinary use. But the team believes it will be possible to create a non-invasive version of the device within the next five years, which would monitor brain activity using electrodes placed inside the ear or under the hair scalp.
Mesgarani said that the device could also be used as a pair of audio "binoculars" to secretly listen to people's conversations, even though this was not the intended application.
The next step will test the technology of people with hearing loss. One question is whether it will be as easy to match the brain activity of people who are partially deaf with sound waves from speech. According to Jesal Vishnuram, Technology Manager of the Hearing Loss Charity Action, one of the reasons why people find common hearing aids is unpleasant in noisy environments is that their brain is practiced to filter sounds and it makes it less effective.  "One of the reasons people fight is that they often wait a long time before they get a hearing aid and in that time the brain forgets how to filter out the noise and focus on speech," she said. "This is really interesting research and I would love to see the real world consequences of it."
The results are published in the journal Science Advances.