Posted on 04/29/2025 9:45:33 PM PDT by Red Badger
People with hearing loss often struggle with the "cocktail party problem." (Photo by StudyFinds on Shutterstock AI Generator)
**************************************************************************
In a nutshell
* A new brain-inspired algorithm called BOSSA could dramatically improve how hearing aids handle noisy environments. It could help users focus on a single voice in a crowded room, something conventional hearing aids often struggle with. * BOSSA mimics how the brain processes sound, using just two microphones to preserve natural spatial cues and separate overlapping speech, outperforming both unprocessed hearing and today’s industry-standard hearing aid technology in lab tests.
* While still in development, BOSSA shows strong potential to enhance social connection and cognitive health for people with hearing loss, and future versions could be steered dynamically, possibly even by tracking where a listener looks.
**************************************************************************
BOSTON — When standing in a crowded room, multiple conversations blur together into an incomprehensible wall of sound. For people with hearing loss, this scenario, dubbed the “cocktail party problem,” isn’t just annoying; it’s a barrier to social connection that can lead to isolation and cognitive decline.
But researchers at Boston University have developed a solution that could transform how hearing aids work in these challenging environments. According to a study published in Communications Engineering, their brain-inspired algorithm, called BOSSA (Biologically Oriented Sound Segregation Algorithm), has demonstrated remarkable success where conventional hearing technologies often fail.
The Global Hearing Loss Crisis
Nearly 50 million Americans are believed to have hearing loss today, and worldwide, the number of people affected is expected to reach 2.5 billion by 2050, according to the World Health Organization.
“The primary complaint of people with hearing loss is that they have trouble communicating in noisy environments,” says study author Virginia Best from Boston University, in a statement. “These environments are very common in daily life, and they tend to be really important to people—think about dinner table conversations, social gatherings, workplace meetings. So, solutions that can enhance communication in noisy places have the potential for a huge impact.”
For individuals with hearing loss, the difficulty of selective listening can be so extreme that it seriously impedes communication and participation in daily life. The consequences extend far beyond just missing parts of conversations, potentially leading to declines in cognitive health, with associated societal and economic costs.
How BOSSA Works
The BOSSA technology draws inspiration from how our brains naturally process sound. The algorithm is based on a hierarchical network model of the auditory system, in which binaural sound inputs drive populations of neurons tuned to specific spatial locations and frequencies, and the spiking responses of neurons in the output layer are reconstructed into audible waveforms.
In simpler terms, the technology uses just two microphones (like our two ears) and mimics how our brain distinguishes between speakers based on their location. This allows a person to focus on a single voice in a crowded room, something current hearing aids struggle to accomplish.
The research team tested their algorithm on adults with sensorineural hearing loss, the most common type of permanent hearing loss that affects the inner ear or nerve pathways. In challenging tests with multiple speakers talking simultaneously, BOSSA consistently outperformed both unprocessed hearing and the industry standard technology currently used in hearing aids.
Most previous hearing aid solutions sacrifice spatial information by reducing multichannel inputs to a single-channel output. This means that individual sound sources are not heard at their original locations, which can be disorienting and disrupt the brain’s ability to separate different voices. BOSSA, however, preserves these crucial spatial cues, allowing users to maintain a natural listening experience while significantly improving speech intelligibility.
Participants saw their ability to understand speech improve by anywhere from 0.3 to 11.3 decibels, depending on the test. The higher the number, the greater the improvement. Importantly, no one did worse using BOSSA than hearing without any special processing.
Meanwhile, the standard beamforming technology used in today’s hearing aids (called MVDR, or Minimum Variance Distortionless Response) showed no significant benefit in the study’s multitalker scenarios. It only proved effective in tests with steady background noise, not the complex overlapping speech that creates the most difficulty for people with hearing loss.
With Apple and other tech companies entering the hearing aid market, traditional manufacturers face increasing pressure to innovate. The BOSSA technology could represent the kind of breakthrough needed to stay competitive in this rapidly evolving field.
The researchers tested two versions of their algorithm—DiffMask and RatioMask—both of which consistently outperformed unprocessed sound and today’s hearing aid technology. The tests were conducted using sentences spoken by multiple female talkers arranged at different spatial locations around the listener, creating a challenging environment where voices were highly confusable.
Unlike other technologies that require many microphones, BOSSA uses only two input signals and produces two output signals that preserve natural spatial cues. It is also well-suited for low-power real-time applications, making it practical for wearable hearing-assistive devices where power consumption and device size are critical considerations.
By enabling a clearer understanding of speech in noisy environments, this technology could help people maintain social connections and potentially reduce the cognitive decline associated with hearing loss.
There’s still work to be done before this technology reaches consumers. Future versions of BOSSA could incorporate user input to dynamically change which speaker is being enhanced, potentially using eye-tracking technology to determine which person the listener is looking at.
“In the long term, we’re hoping to take this to other populations, like people with ADHD or autism, who also really struggle when there’s multiple things happening,” says study author Kamal Sen from Boston University.
Biologically inspired algorithms like BOSSA could significantly improve the quality of life for the millions of people who struggle with hearing in complex acoustic environments. As our population ages and hearing loss becomes increasingly common, such technologies could help bridge the gap between hearing and understanding, keeping people connected to conversations, communities, and cognitive health.
Paper Summary
Methodology
The researchers recruited eight adults (ages 20-42) with bilateral sensorineural hearing loss to test the performance of their BOSSA algorithm compared to unprocessed sound and an industry-standard MVDR beamformer. Participants completed two experiments where they had to identify words from target sentences mixed with competing sentences from different locations. In Experiment 1, each word in a sentence was spoken by a different female talker and were time-aligned to increase difficulty. In Experiment 2, words in each sentence were spoken by the same talker with natural timing to ensure results weren’t dependent on specific experimental choices. Sentences were simulated to originate from five different locations around the listener, with the target located at 0° (straight ahead) or ±30° (slightly to the left or right). The researchers tested four processing conditions: Natural (unprocessed), two versions of BOSSA (DiffMask and RatioMask), and the standard MVDR beamformer used in hearing aids. Performance was measured by calculating the percentage of correctly identified words at different target-to-masker ratios (TMRs).
Results
Both versions of BOSSA consistently outperformed both unprocessed hearing and the standard MVDR beamformer across all experimental configurations. Participants experienced improvements in speech reception thresholds ranging from 0.3 to 11.3 decibels, with no participant performing worse with BOSSA than with unprocessed sound. In contrast, the MVDR beamformer provided no significant benefit in the challenging multitalker scenarios, though it did perform as expected in a control experiment using steady background noise instead of competing speech. The researchers found that BOSSA’s approach of preserving spatial cues was particularly beneficial for helping listeners separate speech in complex acoustic environments where multiple people are speaking simultaneously.
Limitations
The current study had several limitations. The sample size was small with only eight participants, and only four completed the second experiment. The algorithm currently requires the number and location of spatially tuned neurons to be fixed according to known locations of competing talkers, which would need to be addressed for real-world implementation. Additionally, while BOSSA performed well with fluctuating speech maskers, it performed poorly with steady noise, suggesting that future versions would need optimization for different acoustic environments. The researchers also noted that their implementation included linear gain to compensate for hearing loss, while commercial hearing aids typically use compressive gain, which might further enhance benefits.
Funding and Disclosures
The research was supported by grants from the National Institutes of Health (Award No. R01 DC013286), the National Science Foundation (Award No. 2319321), and the Demant Foundation. The authors declared no competing interests.
Publication Information
The paper titled
“A brain-inspired algorithm improves ‘cocktail party’ listening for individuals with hearing loss”
was published in Communications Engineering (2025, Volume 4, Article 75) by Alexander D. Boyd, Virginia Best, and Kamal Sen from Boston University’s Hearing Research Center, Department of Biomedical Engineering, Department of Speech, Language and Hearing Sciences, and Neurophotonics Center
So true... (Sigh)
Do they have an app with an equalizer that ets you adjust the level of each frequency?
That *sounds* pretty frustrating.
Yes, you can do that with your smartphone app and save it.
The only thig about it is it’s not saved in the device itself, just on your phone. So every time you turn off your aids they revert back to their permanent program from the audiologist’s data. You can reload any of 4 programs you store for different locations, like ‘WORK’, ‘RESTAURANT’, SPORTS GAME’ etc...........
And each ear is separate.................
Why hearing aids are 96 frequencies. I get a detailed hearing test and then the audiologist adjusts the individual frequencies according to the results of what I can hear at those individual frequencies.
So in answer your question it’s done by the audiologist
I’ve been working with the same guy for 8 years and I think there’s a bit of an art to the adjustment
That’s a dumb design. Why can’t they store the setting in the device?
Probably has something to do with the available space.
When I go to the audiologist to check up she has to take them into another room to permanently program them with the data chart of my hearing test frequency response. So I don’t know how it’s done..............
You should be able to adjust it (tweak it) yourself for the particular environment you’re in.
If certain frequencies are coming in too loud you lower them, if too low you raise them.
You don’t need an audiologist.
“Probably has something to do with the available space.”
Are you kidding? It’s probably no more than a thousand bytes. With the gigabytes you can store on a tiny chip that’s nothing.
These aids also have 12 levels of volume controlled by buttons on the top of each aid as well as button for changing up to four programs that are permanently programmed. Plus see my post 20 for more info...............
It needs the ability for the user to choose which voice to focus on. Otherwise, you’ll be hearing the wrong person well, and the right person poorly.
I don’t go to cocktail parties for many reasons. One of them is that most of the conversations that are going on are not anything I care to join in on.
If there is more research on improvements to current technology then great, proceed with my blessing.
However, some currently available hearing aids do a good job filtering out unwanted sound with multi-mike arrays and software. Jabra and Rexton, to name two.
I called the tinnitus help line, but it just kept ringing.
LOL!....................
*** I called the tinnitus help line, but it just kept ringing.***
So funny!
My hubby needs hearing aids. I have been wondering what is a good option currently available?
______________________________________________________
I was watching for suggestions, too. What served me well for many years until my hearing got a little worse was Unitron. Then I read in Consumer Reports that Oticon was great, but I don’t like them. So the search continues...
My hearing aids are almost 8 years old. They are programmed for quiet or noisy environments and I can switch between those two settings. I can also adjust the sensitivity, the sound volume by six decibels in either direction.
I have no desire to sit in a restaurant and play with dozens of different frequencies in my hearing aids when the reality is my discrimination is so bad it doesn’t help much anyway
The cheapest hearing aids with the best quality come from costco because the best models create costco brands at costco prices from the same factory.
Costco is my source now. Incredible savings.
If it requires binaural hearing, I am SOL.
Only one ear sort of works,
The background yammering is a very real thing.
The aids just make everything loud, not more comprehensible.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.