brp at muttley.eecs.berkeley.edu (Bruce Raoul Parnas) writes:
>The common wisdom on the layout of the basilar membrane says that it
>is logarithmically arranged in terms of `best' frequency.
There are a few classic like Liberman (1984) that look at that question. One
of the most recent and thourough articles on the subject (that I know of) is
Greenwood DD, "A cochlear frequency-position function for several species -
29 years later", J Acoust Soc Am (JASA) 87 (190) 2592-2605.
A few things are also mentioned in "The evolutionary biology of hearing", by
Webster, Fay and Popper (pp 243ff). The most recent research article that I
know of, though, is
Moore BCJ and Sek A, Auditory filtering and the critical bandwidth at low
frequencies", from the proceedings of the 10th international symposium on
hearing. The title of the proceedings is "Advances in Hearing Research",
and should appear any day now from World Scientific.
The question of mapping is not an easy question, and no one has actually
recorded from the apical end of the cochlea. So Moore and company use an
indirect method, looking at the critical bandwidth of human listeners. It is
known that the CB corresponds very well to the mapping in the cochlea;
basically the CB should remain constant if the cochlear mapping becomes
linear. The conclusion of the last article I mentioned is that CB become
smaller as you go down in frequency, even under 100 Hz. Therefore mapping
may not be perfectly logarithmic all the way down the mammalian cochlea,
but at least you can be sure it's not linear until at least frequencies
well under 100 Hz. In other words, for all practical purposes, it's log all
the way down!
I'd be glad to send you a copy of the last Moore article if you so wish. You
may also want to look at Moore BCJ et al, Auditory filter shapes at low
center frequencies, JASA 99 (1990) 132-140, to see the relation CB-mapping
didier at src.umd.edu