Deb Aronson

Hear Ye! Hear Ye! How the Brain Processes Auditory Information

First appeared in University of Illinois Department of Molecular and Integrative Physiology newsletter in January 2001 under Profiles

How do frogs in a pond resemble people at a cocktail party? In both cases communication is possible only if frogs and people focus in on a particular voice… or croak… and tune out other, extraneous sounds.

Albert S. Feng, professor of physiology, biophysics, bioengineering and neuroscience, has spent the last decade teasing out the elements of listening in what he terms “a complex auditory environment.” His research has led him to study not just frogs and cocktail parties, but bats, which rely on their keen sense of hearing to gobble thousands of insects per evening.

Male frogs produce calls in the spring to advertise their sexual readiness. Many males typically call together, forming a chorus. Female frogs need to both identify the calls of males of their species and to locate the direction of the call. People in a cocktail party similarly must sort out the chatters; they must filter out all voices but the ones with whom they converse. People with hearing problems do not function well in cocktail parties —they find it difficult to discern individual voices.

“How does the female frog identify the most desirable male in a chorus complex?” asks Feng, when explaining his original interest in frogs. “This activity is critical in the mate-selection process because there are usually many species sharing a single breeding pond. The female has to identify the type of frog calling and its location. That is a complicated process.”

Feng has worked to understand that process by breaking this task into its components of pattern recognition and localization. By observing frog behavior and then looking at individual cells in the brain, Feng was able to understand both what the ear does, and the way in which the brain processes the information from the ear. The key difference in those processes, is that the brain is able to integrate information from both ears, not just one, which is critical for localization of sound. In fact, Feng found that the midbrain was far better at distinguishing two sounds close to one another than a single ear was.

Although he pursued the “frog question” out of pure scientific curiosity, once Feng gained in sight into the computational algorithm leading to sound localization and identification, he recognized that that could also be used to improve the design of hearing aids. Traditional hearing aids receive a signal, process it and amplify it. Some of them can filter out background noises when they are repetitive and stationary (such as mechanical systems sounds), but when background noise consists of people speaking, the hearing aid cannot filter out the competing sounds. In addition, hearing aids typically process information from individual ears, which is useless for localizing sound, and which results in the renowned “cocktail party problem.”

“People over and over again have told us how terribly it is to be socially isolated in a restaurant or cocktail party,” says Feng. “They either sit and smile and pretend they can hear, or they tear their hearing aid out in frustration and go hide in a corner.”

Feng’s research group at the Beckman Institute has built a prototype hearing aid that makes use of information from both ears. Having auditory information from both ears helps the user locate sounds and focus on sounds they want to listen to. The device also incorporates an algorithm that mimics the computational algorithm the brain uses to process acoustic information. Feng’s hearing aid enables the user to select what sounds they hear, rather than letting the device dictate what they hear, further compounding their hearing problem.

“Over millions of years, this magnificent biological system has developed effective solutions to particular problems in hearing in the real world — in an animal with a brain the size of the tip of your pinkie,” observes Feng. “It’s a wonderful to see how the principle from biological systems can be used to do something hearing-aid manufacturers have been struggling to do for decades.”

Feng’s group has developed a “real time prototype” hearing aid and is negotiating with several hearing-aid manufacturers to incorporate this new technology into their product. It will be about two to three more years before hearing aids with this technology are on the market.

In a further effort to understand hearing in a complex acoustic environment, Feng also studies bats, which have some of the most highly developed echolocation systems in the animal kingdom. Not only can bats catch and eat thousands of insects nightly on the fly, they selectively snatch certain insects. In one study, 95 percent of what they caught were a certain type of insect … which comprised only four percent of the insect population in the particular area being studied. Researchers have known for a while that bats use wing-beat rate to identify their favorite bugs, but Feng was curious about how the bat brain extracted that information. In the course of his research — in which he did find neurons that selectively filter that information — he also found something he wasn’t looking for. Feng found that bats are better able to analyze sound amplitude and frequency when sound pulses are part of a series rather than when they are presented in isolation.

Researchers had observed for some time that bats increase the rate of sonar pulses as they approach a target. Feng and his students determined that this increased rate enables bats to get a more accurate perception of its environment. Amplitude of the echo reflects the size of the bug, thus an increase in amplitude resolution improves the perception of target size. “Nothing before had suggested that the increased pulse rate influenced increased acuity,” says Feng.

This finding has major implications for research technique throughout physiology. Typically, experiments are designed to isolate individual sounds. Feng’s findings have shown how important it is to design the acoustic stimuli to closely approximate the subjects’ natural environment.

“To really understand bats’ echolocation performance or response characteristics of auditory neurons, acoustic stimuli should closely approximate what the animal encounters in nature,” says Feng. “Based on our findings, even the basic response properties of central auditory neurons that we have learned over the years cannot predict how neurons respond in ‘real world’ situations where signals occur in rapid succession.”