The importance of phase data and model dimensionality to cochlear mechanics.

The vulnerability of the mammalian cochlear amplifier to surgical trauma hinders observations of its behaviour in vivo. This produces a greater need for realistic models to aid the interpretation of the experimental observations. The emphasis in most modelling studies has been to simulate the gain of the response of the basilar membrane. This paper argues that matching the phase behaviour of the response should be given at least equal importance. When it is, many of the models used to justify hypotheses regarding the operation of the cochlear amplifier cannot simulate the response even of the dead cochlea. This discrepancy is due to oversimplification of the mechanics of the cochlear fluids. It is argued that three-dimensional fluid behaviour should be regarded as a bare minimum in any quantitative description of cochlear mechanics. Furthermore, it is shown that a three-dimensional model is consistent with experimental data from a healthy cochlea only when the main effect of the cochlear amplifier is to inject mechanical energy into the basilar membrane. The injection of mechanical energy is fundamentally different to modifying the stiffness of the basilar membrane. This means that existing models which possess cochlear amplifiers that effect large changes on the stiffness of the basilar membrane may not be accurate representations of the real organ.