In <35kksg$2sr at portal.gmu.edu> herwin at mason1.gmu.edu (HARRY R. ERWIN) writes:
>Has anyone studied how the brain does the log transformation involved in
>scaling sensory input and adapting to the background level?
That is a big question, a lot of work there. A great deal of the "scaling"
takes place at the receptor level, as well as adaptation. If you look
for instance at "tuning curves" (i.e. intensity needed to get a criterion
response as a function of frequency) in the auditory system, they do not
change that much from hair cells of the organ of Corti to Herschl gyrus.
Fechner functions hold for sensory receptor cells at least within a certain
range of stimulus intensities (or even other stimulus properties), as
shown for instance by the S-shaped Naka-Rushton functions for photoreceptors.
The sensory cortex does a bit of fine tuning of these, and adds ofther
features in some cases. The tuning curves in Herschl gyrus are narrower than
those of the receptor cells, but one cell in herschl may have in fact two
curves, i.e. two different best frequencies at different octaves.
This has also been examined at the level of the thalamus, but LGN neurons do
not respond that much differently than retinal ganglion cells, although
you probably need the LGN :) Same for the MGN, their tuning curves are a
bit narrower than those of the fibers of the auditory nerve. I think
almost any reasonable textbook in sensory neurophysiology goes into this,
if this is what you are addressing.
>Internet: herwin at gmu.edu>Just a dumb graduate student working on Katchalsky network models....
>(And if you believe that, I have a ranch in California to sell you.)
How much do you want for that ranch? :)