The research team, led by Kamal Sen, an associate professor of biomedical engineering at the BU College of Engineering, has developed a computational model that mimics the human brain’s natural ability to isolate specific sounds. Known as the Biologically Oriented Sound Segregation Algorithm, or BOSSA, this technology has demonstrated a remarkable 40-percentage-point increase in word recognition accuracy compared to the current industry-standard algorithms found in commercial hearing aids. The findings, recently published in the journal Communications Engineering, mark a potential turning point in audiological technology and the treatment of sensorineural hearing loss.
The Science of the Cocktail Party Problem
The "cocktail party problem" is a term coined in the 1950s by British scientist Colin Cherry to describe the difficulty of focusing on a single talker in a noisy environment. For individuals with healthy hearing, the brain employs a sophisticated network of neurons to filter out background noise and "lock on" to a specific frequency or spatial location. However, for those with hearing loss, the neural pathways responsible for this filtering are often compromised, and standard hearing aids typically amplify all sounds—including the background noise—making it nearly impossible to follow a single conversation.
Traditional hearing aids rely on two primary technologies to combat noise: noise reduction algorithms and directional microphones, often called beamformers. Beamformers are designed to prioritize sounds coming from directly in front of the wearer while suppressing sounds from the sides or back. While effective in some controlled settings, these systems often fail in "cocktail party" scenarios where multiple voices are coming from different angles and reflecting off surfaces.
"We decided to benchmark against the industry standard algorithm that’s currently in hearing aids," says Kamal Sen. "That existing algorithm doesn’t improve performance at all; if anything, it makes it slightly worse. Now we have data showing what’s been known anecdotally from people with hearing aids."
A Biological Blueprint for Sound Isolation
The BOSSA algorithm represents a departure from traditional signal processing. Rather than using purely mathematical filters to reduce noise, Sen and his team looked toward the architecture of the auditory cortex. For over two decades, Sen has studied how the brain encodes and decodes sound, specifically focusing on the inhibitory neurons that act as a form of internal noise cancellation.
In the human auditory system, different neurons are tuned to specific locations and frequencies. When the brain decides to focus on a particular sound source, inhibitory neurons are activated to suppress the neural firing associated with competing sounds. This "gating" mechanism allows the brain to sharpen the signal of interest while muffling the interference.
The BOSSA algorithm mimics this process by utilizing spatial cues—such as the subtle differences in volume and timing (interaural level and time differences) between the two ears—to identify and isolate a target speaker. By simulating the way the brain segregates sound sources, the algorithm can effectively "unmix" a jumble of voices, allowing the listener to hear a single talker with much greater clarity.
"It’s basically a computational model that mimics what the brain does," explains Sen. "And actually segregates sound sources based on sound input."
Clinical Testing and Unprecedented Results
To validate the efficacy of BOSSA, the BU team collaborated with clinical researchers to conduct behavioral studies. Virginia Best, a research associate professor of speech, language, and hearing sciences at BU’s Sargent College, played a crucial role in designing tests that reflect real-world challenges.
The study involved young adults with sensorineural hearing loss, a condition typically resulting from damage to the hair cells in the inner ear or the nerve pathways leading to the brain. Participants were placed in a simulated "cocktail party" environment using headphones that recreated the experience of multiple people speaking from various locations. The researchers compared the participants’ ability to recognize words under three conditions: with no algorithm, with a standard beamforming algorithm, and with the new BOSSA algorithm.
The results were stark. While the standard hearing aid algorithm offered little to no benefit in the complex soundscape, the BOSSA algorithm led to a 40-percentage-point gain in word recognition. Such a significant jump in performance is rare in the field of audiological research, where improvements are typically measured in much smaller increments.
"We were extremely surprised and excited by the magnitude of the improvement in performance," says Sen. "The primary complaint of people with hearing loss is that they have trouble communicating in noisy environments. These environments are very common in daily life—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."
Global Context and the Economic Burden of Hearing Loss
The implications of this breakthrough extend far beyond the laboratory. According to the World Health Organization (WHO), approximately 2.5 billion people worldwide are expected to have some degree of hearing loss by 2050. In the United States alone, nearly 50 million people currently live with hearing impairment.
Hearing loss is not merely a physical condition; it is a significant public health issue with profound social and economic consequences. Untreated hearing loss is linked to increased rates of social isolation, depression, cognitive decline, and dementia. In the workplace, individuals with hearing loss often face barriers to promotion and productivity, leading to an estimated global cost of $980 billion annually due to lost productivity and healthcare expenses.
Despite the prevalence of the condition, only a fraction of those who could benefit from hearing aids actually use them. Common reasons for non-use include the high cost of devices, social stigma, and, most importantly, dissatisfaction with performance in noisy settings. By addressing the "cocktail party problem," the BOSSA algorithm could significantly increase the adoption and effectiveness of hearing assistance technology.
Market Disruption and the Rise of "Hearables"
The timing of the BU team’s discovery coincides with a major shift in the hearing health industry. For decades, the market was dominated by a small group of specialized manufacturers. However, recent regulatory changes and technological advancements have opened the door for consumer electronics giants to enter the fray.
In 2022, the U.S. Food and Drug Administration (FDA) approved the sale of over-the-counter (OTC) hearing aids, allowing consumers to purchase devices without a prescription or a visit to an audiologist. More recently, Apple announced that its AirPods Pro 2 would feature a clinical-grade hearing aid function, complete with hearing test capabilities and active noise cancellation tailored for hearing assistance.
Sen notes that this influx of "hearables"—consumer devices with hearing enhancement features—is putting pressure on traditional hearing aid companies to innovate. "If hearing aid companies don’t start innovating fast, they’re going to get wiped out, because Apple and other start-ups are entering the market," Sen warns. He has already patented the BOSSA algorithm and is actively seeking industry partners to license the technology for integration into the next generation of hearing devices.
Future Horizons: Eye-Tracking and Neurodevelopmental Disorders
While the current version of the BOSSA algorithm provides a massive boost in performance, the researchers are already looking toward future enhancements. One of the most promising avenues is the integration of eye-tracking technology.
In a natural conversation, humans typically look at the person they are listening to. By incorporating eye-tracking into a hearing aid or "hearable" device, the algorithm could automatically prioritize the sound source located at the user’s focal point. This would allow for a more intuitive and seamless transition between different speakers in a group setting.
Furthermore, the team believes the science behind BOSSA could help populations beyond those with hearing loss. Because the algorithm is built on the fundamental principles of neural attention, it could be adapted to assist individuals with Attention-Deficit/Hyperactivity Disorder (ADHD) or Autism Spectrum Disorder (ASD).
"The [neural] circuits we are studying are much more general purpose and much more fundamental," says Sen. "It ultimately has to do with attention, where you want to focus—that’s what the circuit was really built for. 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."
Conclusion: A New Standard for Auditory Assistance
The development of the Biologically Oriented Sound Segregation Algorithm at Boston University represents a significant milestone in the intersection of neuroscience and engineering. By looking to the brain’s own architecture for inspiration, researchers have managed to overcome a hurdle that has plagued the hearing aid industry for decades.
As the global population ages and the prevalence of hearing loss continues to rise, the demand for effective communication tools will only intensify. The 40-percentage-point improvement demonstrated by BOSSA suggests that the "cocktail party problem" may finally have a solution, offering millions of people the opportunity to re-engage with the vibrant, noisy world around them without the fear of being left out of the conversation. With the potential for commercial licensing on the horizon, the next few years may see a dramatic transformation in how we hear, listen, and connect.
