Tiny hair cells located in the inner ear help us hear and maintain balance. On top of each hair cell is a hair bundle, a sensory organelle that converts mechanical input from sound or movement into electrical output, which is then passed on to the brain. Previous research has shown that hair bundles aren’t simply passive entities. They actively oscillate to amplify weak audio signals or to tune into specific frequencies. Biologists have also observed bundles oscillating in the absence of stimuli. Models have tried to capture this bundle behavior, but the connection between active oscillation and the audio response has not been made clear. A new thermodynamic model of energy flow within hair bundles suggests that they work like tiny machines [1]. Depending on the stimulus, the bundles either extract power from incoming sound waves or inject power into them—corresponding, respectively, to sensing or amplifying a stimulus.

In the inner ear, an active process called cochlear amplification helps humans (and other mammals) hear the faintest of sounds. When a faint whisper enters the ear, for example, the outer rows of hair cells respond to the weak signal by moving in a way that amplifies the sound waves for the inner hair cells, which are the ones that send a message to the brain. Molecular motors propel the movement or twisting of hair bundles required for these functions.

Previous work has explored how much energy a hair cell consumes to drive bundle oscillations, but the resulting models have typically assumed that bundles are moving spontaneously—that is, in the absence of external stimuli. Roman Belousov from the European Molecular Biology Laboratory in Germany and his colleagues have developed a stochastic thermodynamic model that includes an energy input from sound waves. “Instead of just looking at how a hair bundle moves on its own, we wanted to add what happens when it interacts with sound,” Belousov says.