Groundbreaking research by American scientists has unveiled a revolutionary device that interprets brain signals to automatically amplify the desired voice. This innovative technology could serve as a vital solution for the 430 million individuals globally affected by disabling hearing loss.
Participants wearing intracranial electrodes engaged in two overlapping conversations, while their neural activity was processed in real time. The system selectively amplified the participant’s chosen speech by leveraging low frequency (LF) and high gamma (HF) features. Image credit: Choudhari et al., doi: 10.1038/s41593-026-02281-5.
Deciphering speech amidst background noise is one of the primary challenges in auditory neuroscience and technology.
In noisy settings, individuals utilize selective attention to concentrate on the target speaker while filtering out competing voices and ambient noise.
Traditional hearing aids often fall short because they lack the ability to understand the listener’s preferences, thus amplifying all sounds indiscriminately. This results in limited effectiveness in real-world environments, leading to decreased user adoption and social isolation.
“Our innovative system functions as a neural extension of the user, harnessing the brain’s inherent skill to filter sounds in complex environments and dynamically isolate the desired speech,” stated Columbia University researcher Dr. Nima Mesgarani.
“This advancement enables us to envision a future beyond conventional hearing aids, moving towards technology that can restore the brain’s sophisticated ability to selectively hear.
In their study, Dr. Mesgarani and his team collaborated with surgeons and epilepsy patients undergoing brain surgery to accurately identify seizure triggers.
Patients who volunteered for the study had pre-implanted electrodes, enabling the team to monitor brain activity while attendees focused on overlapping conversations occurring simultaneously.
The system could detect the conversation the patient concentrated on and automatically adjusted the volume in real time, amplifying the chosen dialogue while reducing others.
One volunteer described the experience of controlling a brain-activated system as astonishing, even questioning if researchers were secretly adjusting the volume.
Others expressed excitement about the potential benefits for their hearing-impaired friends and family members, with one participant remarking, “It’s like science fiction.”
“A key unanswered question was whether brain-controlled hearing technology could evolve from theoretical models to practical prototypes that enhance hearing in real time,” Dr. Vishal Chaudhary of Columbia University noted.
“We demonstrate for the first time that a system leveraging brain signals for selective speech enhancement can provide significant real-time advantages.”
“This innovation transitions brain-controlled hearing from concept to practical application.”
Researchers have developed a fast machine learning algorithm capable of identifying which conversations patients are engaged in by analyzing their brain waves.
Once implemented, the system can swiftly determine which dialogues each listener focuses on to enhance clarity.
This capability was evident both when researchers directed participants to specific conversations and when participants freely selected, mirroring real-world interactions.
“To function effectively in real time, the system must be exceptionally fast, accurate, and stable, ensuring a comfortable experience for the listener,” Dr. Mesgarani emphasized.
Find the team’s research published in the latest edition of Nature Neuroscience.
_____
V. Choudhary et al. Real-time brain-controlled selective hearing enhances speech recognition in multi-speaker environments. Nature Neuroscience published online on May 11, 2026. doi: 10.1038/s41593-026-02281-5
Source: www.sci.news
