Hearing aids may cut noise by reading brain signals

Published 7 June 2026 · science

In a crowded room, it can be impossible for hearing aid users to focus on what they want or need to hear. Research into auditory

In a crowded room, it can be impossible for hearing aid users to focus on what they want or need to hear. Research into auditory neurotechnology may enhance sounds and voices and reduce noise in real-time. Hearing aids amplify everything — "indiscriminately" — wrote a group of scientists working on a system that may make brain-controlled hearing aids possible. "Current hearing aids are good at amplifying sounds and voices, but they struggle with the classic 'cocktail party problem' — deciding which voice matters to the listener," said Vishal Choudhari*, lead author on their study published in Nature Neuroscience. It can take a lot of effort to focus your attention on one voice in a crowded room. "Hearing is not only about whether words are understood correctly," Choudhari told DW. "Two people may both understand [what they're saying], but one person may need far more mental effort to follow the conversation. That can become exhausting over time." As a result, many people stop using hearing aids just when they need them most — in restaurants, cafeterias, parties, or busy social spaces. So, Choudhari and his colleagues are trying to develop a smart technology that knows what a hearing aid user is listening to. They want to enhance that one sound or voice, while at the same time reducing the volume of any other sounds, voice or background noise. And to do that, they designed a system that reads brain waves and, using artifical intelligence, interprets what a listener is listening to.

"Many hearing aids use beamforming, which enhances sounds coming from a certain direction, usually in front of the listener. But real conversations are dynamic," said Choudhari. "People turn their heads, switch attention, or even listen to someone without directly looking at them." Better hearing — In Good Shape To view this video please enable JavaScript, and consider upgrading to a web browser that supports HTML5 video Brain-controlled hearing from theory to practice Under Nima Mesgarani, a professor and principal investigator at Columbia's Zuckerman Institute, Choudhari and the team developed real-time machine learning algorithms that could examine brain waves and identify the conversations that four normal-hearing test participants were listening to. The researchers called it a closed-loop auditory attention decoding (AAD) system. They wanted to find out whether AAD could be accurate and fast enough to selectively amplify the voice of an individual speaker while suppressing background voices. While the idea sounds amazing, it's far from ready for general use. For now, the concept relies on electrodes being attached to the brain in a clinical setting. In the study, the four test patients were undergoing brain monitoring for epilepsy — so, they already had intracranial electrodes and that was convenient for the researchers. The participants were presented with recordings of two competing sound sources, coming from small speakers, positioned left and right. The recordings featured people of various gender mixes, in conversation about food, travel and exercise.