David Mountain, who has been teaching at ENG for twenty years, begins a sabbatical this fall but won't be far from the lab: he's working with a Children's Hospital researcher on -- what else? -- hearing.

Reverse engineering is common: get a competitor's product, take it apart, figure out how it's made, then build your own version. When you're dealing with computers or other manufactured products, that's often difficult enough. But reverse engineering Mother Nature? Much harder, indeed. Yet that's what many biomedical engineers are trying to do. Case in point: Biomedical Engineering Professor David Mountain and fellow researchers at the College of Engineering.

Mountain's work centers on hearing: how do humans take in sounds, amplify and clarify them, and give them meaning? Much of his research has focused on cochlear biomechanics, and other applications abound, from industry to the battlefield.

How Sound Sounds

But start at the source, Mother Nature. In all mammals, sound comes in through the external ear, causing vibrations in the middle ear bones, which in turn create pressure changes in the inner ear fluids. These vibrate the basilar membrane and create waves that travel down it. "One of the breakthroughs in the last twenty years is that it appears that some of the cells in that organ are like little motors and actually amplify this wave as it travels down the length of the membrane, increasing our hearing sensitivity by a factor of at least a hundred," Mountain says. "I'm interested in understanding how they do that job, and how that amplifier works." What's striking is that the amplification is a combination of electrical and mechanical processes. "The cells respond to this wave-like motion with a change in voltage within the cell, and that change in voltage causes the cell to change its length -- a combination mechanical-to-electrical transduction process and electrical-to-mechanical transduction process."

Understanding how the process works in healthy ears is part of the goal; another objective is to see what happens when the system is damaged. "Loud sounds, for example, are obviously bad for your hearing, and if they're really loud, they kill off the cells altogether," says Mountain. "Profound deafness is usually the result of losing the sensory cells in the inner ear, but moderate deafness can be the result of losing the cells that do the amplification, meaning that you've still got the inner hair cells, which send the information up to the brain."

How, then, to discern the division of labor between two cell groups? Using a variety of methods, Mountain's research group studies the hearing system in gerbils. Then, together with Electrical and Computer Engineering Associate Professor Allyn Hubbard, a longtime collaborator, they build computational models of the hearing system. The models are "the glue that brings all this together," Mountain says. "It's sort of a closed cycle. We have these different kinds of experiments that lead us to form hypotheses. But they're difficult to test directly within the animals, so we test the hypotheses on the mathematical models."

An Earful of Applications

Using what they have learned about sound and hearing, Mountain and his colleagues are now developing applications that range from battlefield surveillance systems to automated milling equipment monitoring devices. That may sound a far cry from inner ear research, but they all employ the same general technology. Take, for example, a system that could identify different vehicles that are out of line of sight. "The Army has this thought that it would deploy microphone arrays all over a battlefield, and they would all be sending back information such as, 'I've got a tank at 90 degrees, I've got something else at 30 degrees,' and it would go back to a central command station where there'd be a big screen that would show everything moving around," Mountain says. It's easier to imagine than build: after all, the vehicle's sound would differ depending on distance, weather, and terrain -- crackling over dry leaves in a forest, whooshing over sand, or crunching across snow. Not to mention the difficulty in battle conditions of filtering a single tank from all the other sounds abounding.

Another potential application, for industry, is simpler, and probably will make it off the drawing board sooner. Computerized milling machines frequently make specialized machine parts, such as for the aircraft industry. Predicting machine failure could eliminate breakage of expensive parts. "It turns out an experienced machinist, by feeling the vibrations or by listening, can pick up on things that are going wrong," Mountain says. The idea is to develop an automatic system that would constantly monitor the machine, a more efficient process than always having skilled machinists around. The challenge is to do the calculations in real time. That means coming up with a custom integrated circuit, which Hubbard is also working on, that could be in a box mounted on the milling machine, ready to raise the alarm when the machine started to fail.

"The thing I've enjoyed the most about this [research] is that it's such a humbling experience. You see something that fairly simple animals can do quite easily, and we're having such a hard time duplicating their performance with all these resources and fancy computers," says Mountain. "It gives me a lot of insight into biology, too. I see where we're failing on the engineering side, and it makes me want to go back and look at the biology some more and try to understand better how Mother Nature has come up with a solution to this problem."