Spreading the fluid word

Article metrics

In Fantastic Voyage, the 1966 sci-fi film classic, Racquel Welch and her intrepid team of brain surgeons are shrunk to nanometre dimensions and sent on a journey through the human body. Escaping killer lymphocytes, they get entangled within the cochlea of the inner ear and are buffeted by the sounds as they enter the fluids of the duct. What the leading lady, although appropriately clad in a wetsuit, failed to do was to take any measurements of the fluid flow in the cochlea. Students of cochlear mechanics have regretted the lost opportunity ever since. Over 30 years later, reporting on page 526 of this issue, Elizabeth Olson1 has managed to make just such biophysical measurements, with implications for the way we think about models of hearing.

The organs of hearing, on either side of the human head, are coiled compartments within the temporal bones. If the coils were straightened out, each cochlea would be 34 mm long in humans and about one-third of that size in small mammals2. The basilar membrane — a collagen-fibre membrane no more than 0.5 mm wide — divides the cochlea along its length. This membrane, together with the sensory hair cells of the organ of Corti that also span the length of the cochlea, transmits information about incoming sounds into a pattern code sent to the brain via the auditory nerve. The idea is that each component sound frequency is mapped by the mechanics onto a unique pattern of displacements of the membrane.

Sound waves travelling in air are transmitted to the fluid in the cochlear duct with little loss of energy. But how is the pattern of hair-cell excitation then set up along the basilar membrane? Modelling the vibration pattern of the basilar membrane has been a problem for nearly 140 years — ever since Helmholtz3 first tackled it but ignored the hydrodynamics. The guiding principle since then has been to treat the cochlea as a one-dimensional transmission line4. The hydrodynamics implied in a three-dimensional cochlea have been difficult to add in, both computationally and experimentally. Olson's work1 indicates that three-dimensional fluid flow may have to be included in any complete and realistic description of hearing.

Sound waves propagating in cochlear fluid travel, as in water, at around 1,550 m s−1. Drop a stone into a pond and the sound it creates propagates to the bottom at this speed, although the ripples travel across the surface much more slowly (a Rayleigh wave) at a velocity that depends on the depth of the pond and the surface tension of the water. The basilar membrane is a two-dimensional surface with variable stiffness along its length that separates the two fluid compartments (Fig. 1). It effectively propagates a 'surface' wave — known as the Bekesy travelling wave — with a velocity of about 15 m s−1 (ref. 5). This wave starts at the base of the cochlea, nearest the middle ear, and propagates with decreasing wavelength but increasing amplitude towards a position of maximum amplitude beyond which it rapidly decays. In the cochlea of a living animal, the amplitude of the travelling wave at this peak is further increased about 100 times by the action of the outer hair cells that lie along the length of the cochlea6,7.

Figure 1: The cochlea uncoiled.

The basilar membrane supports a travelling wave. The membrane divides the cochlear duct into two fluid-filled compartments, where sound propagates roughly 100 times faster. Only a central, longitudinal section is portrayed. The vertical excursion of the wave is exaggerated about a million times. The fluid velocity vector at X, perpendicular to the membrane, is the measurement reported by Olson1.

Olson1 offers a rare experimental view of the dynamic pressure changes in the cochlea. It is technically very difficult to make mechanical measurements in the cochlea owing to the extremely small size and delicate nature of the structures, especially at the basal end where high frequencies are processed. This has slowed progress considerably, and there are few direct measurements of pressure waves. Olson has overcome these problems by developing a small but sensitive hydrophone that can be placed into the gerbil cochlea and measure pressure changes within the cochlear fluids.

The new data show that fluid flow perpendicular to the basilar membrane falls away within 15 μm of the membrane. This result seems paradoxical. It has been supposed (in the so-called ‘long wave’ limit4) that fluid should be carried with the basilar-membrane wave throughout the height of the cochlear duct at the basal end of the travelling wave. Because water is incompressible, conservation of mass may imply that fluid is moving in a radial rather than perpendicular direction to the long axis of the cochlea. Alternatively, this is an indication that previous models of the cochlea may have ignored subtle features of the mechanics at the basal end of the cochlea.

Olson's result is consistent with the idea that the cochlea is a three-dimensional, dynamic structure. Not only does it have length, but the cochlear duct also has width and a variable height that determines the mass of fluid moving with the basilar membrane5. There are also indications that the organ of Corti on the membrane must be modelled as a complex matrix that includes not only hair cells, but also other cell types with varied mechanical properties. Modern computing power allows us to show that the structure may alter the fine detail of the local vibration pattern in the basilar membrane8. This pattern can involve higher radial vibrational modes, implying that surrounding fluid would be both pushed and pulled across, as well as along, the membrane9,10. It is a huge challenge to design a small enough pressure sensor, compared with the membrane width, that could resolve the dynamic detail required for the cochlea (Olson's is 160 μm in diameter).

We have yet to construct physiologically based models of the cochlea that convincingly describe all of its real-time abilities to analyse complex sounds. Such models should now probably include local interactions between the compartment fluids and the basilar membrane. It is hard to avoid thinking that cochlear models will get more complex before they become simpler.


  1. 1

    Olson, E. Nature 402, 526–529 (1999).

  2. 2

    Fay, R. R. & Popper, A. N. Comparative Hearing: Mammals (Springer, New York, 1994).

  3. 3

    Helmholtz, H. Die Lehre von den Tonempfindungen (Vieweg, Brunswick, 1863).

  4. 4

    de Boer, E. in The Cochlea (ed. Dallos, P.) 258–317 (Springer, New York, 1996).

  5. 5

    Lighthill, J. J. Fluid Mech. 106, 149–213 (1981).

  6. 6

    Ruggero, M. A. et al. J. Acoust. Soc. Am. 101, 2151–2163 (1997).

  7. 7

    Nobili, R. et al. Trends Neurosci. 21, 159–167 (1998).

  8. 8

    Kolston, P. Proc. Natl Acad. Sci. USA 96, 3676–3681 (1999).

  9. 9

    Russell, I. J. & Nilsen, K. E. Nature Neurosci. 2, 848–853 (1999).

  10. 10

    Cooper, N. P. J. Physiol. 520, S114 (1999).

Download references

Author information

Correspondence to Jonathan Ashmore.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ashmore, J., de Boer, J. Spreading the fluid word. Nature 402, 476–477 (1999) doi:10.1038/44985

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.