[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: mechanical cochlear model

At 8:59 PM -0500 3/9/10, David Mountain wrote:
...  When this energy reaches the beginning of the peak region, some process, usually called the "cochlear amplifier" and involving outer hair cells, takes over and shapes the peak response.

David, I like your definition and description, with one caveat: it is often said that the "cochlear amplifier" is restricted to a narrow region before the peak, or it's described as "taking over", or as a different mechanism from the traveling wave. That's OK, but I think it's perhaps more useful to say that the mechanism operates everywhere, and contributes to the physics that defines the wave equation and its dispersion relation.

The nonlinear properties of the mechanism that adds energy to the traveling wave gives rise to important nonlinear properties, including compression and distortion products. But for the analysis of the traveling wave, it is useful to treat as linear, or approximately linearized at a given operating level. Then its properties can be described simply, in term of how the gain or forces relate to frequency and wavelength. In the region more basal than an octave before best place, the wavelength is quite long, and the energy moves through quite quickly, mostly away from the membrane, so the effect of the cochlear amplifier is necessarily very small. The effect comes on gradually, and is greatest where the wavelength is short and the energy is concentrated near the membrane. The effect can be described very simply as a negative-resistance-like part of the membrane impedance, in impedance-based models.

It's certainly true what you say, that we don't yet understand the micromechanical mechanisms of the cochlear amplifier, though we're a lot closer than we were ten years ago. But there's ample evidence that its effect can be fit into a traveling wave model, and that such models do a credible job of fitting both neural and mechanical data simultaneously. To the extent that there are remaining discrepancies, there are lots of people paying explicit attention to those, and trying to resolve them. For example, if a high-frequency plateau is seen in mechanical data, and not in neural data, what does that mean, and how can it be resolved? The problem is being addressed head-on, several ways, not being ignored as some would suggest. It's a tiny effect, in terms of BM motion and energy, but understanding it is likely to help complete our understanding of the whole system.

Generally speaking, we've seen a convergence of experimental data, neural, mechanical, even psychophysical, and physics-based models that describe how it all fits together, over the last 40 years (measuring back to Rhode's 1971 mechanical measurements that showed the kind of nonlinearity near the peak that was already well known in neural data).

If people hadn't been questioning our understanding, our data, and our interpretations all along the way, we wouldn't be in the relatively good situation we're in today. A few years ago in a talk I criticized Helmholtz and Ohm for their support of the phase-blind spectral analysis point of view of cochlear function; I realized later that I was off-base on this. Like Bekesy and others that came since, they were operating half-blind, with limited apparatus, technique, and outlook, and doing their best in good faith to advance the science. So I sensitized myself to such criticism of our forebears, which is why I sometimes speak up sounding a bit annoyed when I see unwarranted accusations like Martin's "Bekesy and his followers succeeded in neglecting all unwelcome evidence." The historical evidence shows otherwise, and it's not a constructive interpretation of where the field has been or is headed, even if it can be shown to be true that certain individuals and groups were temporarily misled by adhering too long to a particular conception when the data were trying to take them elsewhere.

David, I count you and Jont and some others here as among those who drove progress over the last few decades, and who influenced my own progress in understanding this system, through your work and your generous help. And I'm sure you understand that when I argue with you, it's part of addressing the issues that the data and models bring to our attention. I look forward to more from you, Jont, Reinhart, and others, and hope that those with alternative ideas to contribute will be able to do so without falling into that negativity thing that I fell into myself with Helmholtz and Ohm.