How hair-cell bundles in the inner ear not only sense sound, but also amplify it

By combining biological experiments, AI, physics, and mathematical equations, an international team of researchers has gained new insights into how hair-cell bundles in the inner ear sense and amplify sound. By emitting energy through oscillations, certain hair-cell bundles amplify sound waves so they can be picked up better by other hair-cell bundles. The study by Utrecht University researchers Yanathip Thipmaungprom, Florian Berger, and international colleagues was published today in the scientific journal PRX Life.

We are able to hear faint sounds thanks to tiny structures in our inner ears called hair-cell bundles. They convert sound vibrations in the fluids of the inner ear into electrical signals that are transferred to the brain by the auditory nerve. These hair-cell bundles not only sense sounds, but they can also actively amplify them, as has been experimentally shown in amphibians. Previous studies suggest that hair-cell bundles are not passive sensors, as they consume energy and oscillate spontaneously. However, how this helps to improve hearing has remained unclear.

Remarkably, the model suggests that the hair-cell bundles are also able to extract energy from the surrounding, which turns them into tiny refrigerators.

The interdisciplinary team now shows that by oscillating, hair-cell bundles put energy into incoming sound waves, thereby amplifying them so other hair-cell bundles can pick them up. “This energy comes from inside the hair-bundles, ” Berger explains. “Myosin motors, also known to play a role in muscle contractions, probably drive the motion. They are fueled by ATP, a molecule that provides energy in living cells.”

Heaters and refrigerators

PhD candidate Thipmaungprom, first author of the paper, and Berger worked together on the project with researchers from the Rockefeller University (New York City), the European Molecular Biology Lab (Heidelberg), and Abdus Salam International Centre for Theoretical Physics (Trieste). The team found that, besides sensing sound and amplifying it, hair-cell bundles have two other “modes”. Berger: “Our model suggests that they can also dissipate heat, which effectively turns them into microscopic heaters. And, remarkably, the model suggests that they are also able to extract energy from the surrounding, which turns them into tiny refrigerators.”

The function of these two modes is not yet clear. A cooling effect could potentially have something to do with reducing noise. “We only see the refrigerator effect at higher amplitudes, with louder sounds, ” says Thipmaungprom. However, it still needs to be experimentally tested if the heating and cooling effects are actually there in real life, and what the consequences of these effects are.

Bullfrogs, thermodynamics, and AI

The experimental data for the study came from bullfrogs. “Bullfrogs have very big hair cells,” Berger says. “We collected the layer of cells that have the hair bundles on top, glued this on a piece of aluminum foil with a half-millimeter hole, and then mimicked the conditions of the inner ear using very small amounts of different ionic fluids, which are liquid salts. When done right, the hair cells start oscillating, which you can record under a microscope with a high-speed camera.”

You can think of these hair cells as small engines that convert and control energy.

The researchers then used machine learning, a form of AI that “learns” patterns from data, to fit the experimental data to a mathematical model, which they then analyzed using a relatively new concept from physics called stochastic thermodynamics. “You can think of these hair cells as small engines that convert and control energy,” Berger explains. “Now, classic thermodynamics, which is the branch of physics that deals with the relations between heat, work, temperature, and energy, was developed in the 1850s for steam engines. It works well for big systems, but when you apply it to smaller systems, like cells or molecules, these concepts break down. Stochastic thermodynamics allows the study of the flow of energy in small, unpredictable systems, such as cells, where randomness plays a big role.”

Thipmaungprom adds: “We have mathematical equations from stochastic thermodynamics that can describe the flow of energy in these hair cells, but we needed to figure out the right values to put into these equations. So, we used machine learning to extract those values from the experimental data. Once we had them, we could apply all our thermodynamic tools to analyze how the system behaves.”

Dedicated to Jim Hudspeth

The researchers dedicated the study to Jim Hudspeth, an American researcher who passed away last year. “He did groundbreaking work on the biophysics of hair cells,” says Berger. “He was the one who introduced us to this fascinating system.”