Can You Hear Me Now?
BU researchers created a brain-inspired algorithm called BOSSA that helps people with hearing loss focus on voices in noisy settings.
Hundreds of millions of people — over 5% of the world’s population — have a disabling form of hearing loss, and these numbers are only expected to climb in the years ahead. For these individuals, everyday interactions with other people can be a major challenge because of the substantial communication barriers they experience. In certain crowded spaces, like at dinner parties or in busy public spaces, these problems are multiplied. Picking out voices from a crowd can be impossible, leaving individuals with hearing loss feeling isolated and unable to communicate effectively.
Hearing aids can help to some extent by boosting the volume, but the methods that they use to filter out background noise are not very effective. In fact, both anecdotal experiences and research has shown that the noise reduction algorithms and directional microphones used in traditional hearing aids can actually make the problem worse. As such, there is a significant unmet need in this area at present.
A trio of researchers at Boston University has taken a completely different approach that may finally give individuals with hearing loss some real help in noisy environments. The team developed a biologically inspired sound segregation algorithm known as BOSSA, short for Biologically Oriented Sound Segregation Algorithm. This technology draws from the way the human brain naturally separates sounds using spatial cues like timing and volume, mimicking the brain’s ability to focus on specific voices in a noisy crowd.
Unlike traditional beamforming technology, which often distorts spatial cues by collapsing sound into a single channel, BOSSA preserves natural spatial differences. This means listeners are able to retain awareness of where voices are coming from — a key factor in understanding speech in complex environments.
The algorithm was tested in a series of experiments involving young adults with sensorineural hearing loss. Participants were asked to identify a single speaker among five talkers positioned around them. The study compared BOSSA’s performance to both standard algorithms and unprocessed audio. The results showed that BOSSA consistently outperformed both, even in the most difficult listening conditions.
Beyond hearing loss, the algorithm may also be of use to populations with attention-related difficulties, such as individuals with ADHD or autism, who also struggle to filter relevant auditory information in noisy settings. The team is now exploring a next-generation version of BOSSA that incorporates eye-tracking to help users direct their auditory focus more precisely.
With major tech companies like Apple entering the hearing aid space via their earbuds, the researchers believe that traditional hearing aid companies will be wiped out in the near future if they do not start innovating. Algorithms such as BOSSA could help to serve unmet needs and keep traditional hearing aids relevant in the world of tomorrow.