Mechanisms of hearing question

Didier A. Depireux didier at
Wed Dec 16 09:34:13 EST 1998

First, I apologize for saying "The 7 green books"... 7 was the original
plan, but by now they are at volume 11... Anyway, these books, in the
Springer Handbook of Auditory Research series, are the state of the art in
hearing (and of course many are already outdated!).

F. Frank LeFever (flefever at wrote:
: find the time to look systematically at the distinctionns between
: retina and cochlea that Didier describes, but a nagging question about

Now, there are differences between the cochlea and the retina, but they may
not be relevant. As a matter of fact, you can go a long way thinking of the
cochlea as a one-dimensional retina that looks at the time-windowed Fourier
transform of the acoustic waveform along a log-frequency axis.  That's the
point of view of our lab, doing recordings in inferior colliculus and
auditory cortex. I.e., the idea is that once you have gone beyond the
sensory epithelium, hearing and vision use similar principles to decompose
and analyze perception (for instance, we use auditory gratings to study
single units in auditory cortex). The big difference btw cochlea and retina
is the local iris provided by the outer hair cells, and the timescales
involved (auditory nerve fibers lock to a stimulus up to 2 kHz, and have
steady firing rates of up to 200 Hz).

: to the idea that neural noise enhances detection of extremely
: weak signals: is this relevant to the point Dider makes about cats
: detecting signals below the level of Brownian movement?  Would this
: make a sufficiently "non-passive" cochlea?

I can only say this with certainty: stochastic firing enhancing signal
detection has only been shown in batracians, where you have the advantage
that you can change the noise level (by changing the temperature of the
animal, therefor changing the thermal noise) and therefore study
transmission as a function of the noise. I don't think (but I don't know
for sure) that it's been shown in mammals. It could be used.

However, the point about the OHC is that if you measure the pressure in the
cochlea at a given frequency (at rest, with no stimulation), you find a
gaussian type of noise. Nothing exciting there. But once in a while, we
(mammals) all have these high pitch sounds in one ear, that last about 10
seconds to a minutes, and that sound like a TV is being turned on. These
otoacoustic emissions occur all the time, and we don't hear most of them.
But if you record the pressure at the frequency of one otoacoustic
emission, you will find a U-shaped distribution of the amplitudes, just as
you would get from a sinusoidal oscillation. We know that these emissions
are from the OHC... Draw your conclusions!


Didier A Depireux                              didier at
Neural Systems Lab       
Institute for Systems Research          Phone: 301-405-6557 (off)
University of Maryland                                -6596 (lab)
College Park MD 20742 USA                     Fax: 1-301-314-9920

More information about the Neur-sci mailing list