The vertebrate ear is extraordinarily sensitive, able to discern faint whispers but still find meaning in extremely loud noises, and capable of tuning in to an important conversation over the cacophony of a city street. But individual hair cells, the sensory cells that detect sound vibrations, do not seem to be sensitive enough to account for how well the ear works.
Instead, groups of hair cells in the inner ear gain some of their extraordinary sensitivity by their mechanical interactions with their neighbors, according to new research published in the Proceedings of the National Academy of Sciences.
Jérémie Barral and Kai Dierkes and their colleagues at the Institut Curie in Paris, France, and the Max Planck Institute for the Physics of Complex Systems in Dresden, Germany, knew that individual hair cells actively produce force, wiggling back and forth, and that the wiggling could help to selectively and non-linearly amplify faint sounds at a frequency close to the wiggling frequency. But the wiggling also produces noise. For a single hair cell, the noise should swamp out any benefit.
The scientists hypothesized that connections between the hair cells might serve to damp out noise without impeding the non-linear amplification process. Hair cells in a living ear are coupled together by a springy, gelatinous matrix. Perhaps that mechanical linkage is the key to the ear's sensitivity.
To test their hypothesis, Barral and Dierkes built an innovative device that linked an isolated hair cell to a computational simulation of other hair cells. They isolated a group of hair cells from a bullfrog's inner ear, dissolved away the gelatinous matrix, and attached a probe to a single hair bundle. The probe could both apply force and measure the movement of the hair bundle. Then they measured the frequency and amplitude of spontaneous oscillations in the cell, and used these and other measurements to produce a ‘cyber clone’, a computational simulation of a hair cell with the same properties as the real one. The simulations ran quickly enough that they could use their probe to apply forces to the real hair cell as if it were connected by springs to one cyber clone on each side.
Once the cyber clones were running, the group examined the interaction between cells. With no linkage and no sound, the real cell and its cyber clones wiggled back and forth as they normally do, out of sync and fairly noisy. With springs connecting them, the cells rapidly synchronized and became much less noisy. At the same time, the amplification of external vibrations – sound – increased dramatically. For faint sounds, the coupled cells were nearly twice as sensitive as any one alone.
Twice as sensitive is still not very good, at least compared with the performance of the whole ear, which can be nearly 100 times more sensitive than a single hair cell. But the hair cell only benefited from two neighbors. Increasing the size of the group to 9×9 cyber clones pushed the amplification up to a realistic gain of 52 dB – about 60 times that of an individual cell – close to the levels achieved in the mammalian cochlea. Thus, it seems that relatively small groups of hair cells, connected by a springy matrix, can reach the extraordinary performance seen in the whole ear.
