How to use Altavista

From: Stevan Harnad (harnad@cogsci.soton.ac.uk)
Date: Fri May 09 1997 - 19:01:01 BST


At the bottom of this message there will be some important and useful
information for everyone in year 2:

> From: "Liz Lee" <EAL195@psy.soton.ac.uk>
> Date: Fri, 9 May 1997 09:03:19 GMT
> Subject: Night skies
>
> Can you run the "why the night sky isn't white" by me once more
> please, I was thinking about it last night and surely the point is
> that although there are billions of stars, there is infinitely more
> space inbetween them, so we see the space?

I was describing Olber's Paradox to help me explain about visual
receptive fields and excitation/inhibition in the brain.

Olber's Paradox ("Why is the sky mostly black?") is a problem for the
theory that the universe is infinitely large and homogenous (i.e.,
the distribution of stars and galaxies is about the same everywhere.

If that theory were correct, then the sky would be completely white
with stars. The reason is that although from further and further away
the light is fainter and fainter, infinity is "big" enough to
compensate for that. So in between any to bright spots -- say, 2 white
stars with some black space between them -- there will be infinitely
many further away stars, which together make the entire gap white.

There are several ways to resolve Olber's Paradox: One is to assume
that the universe is finite and expamding; another is that the universe
is not homogenous. For more see:
http://www.onysd.wednet.edu/sciweb/astronomy/astrophysics/Olbersp.html

I brought up Olber's Paradox to explain something about receptive
fields: if you have a cell that has an "excitatory centre and an
inhibitory surround" this means that when light is shone on the centre
of its visual field, it gets very excited, and when it is shone on the
doughnut that surrounds the centre, it will be very quiet.

If you shine homogeneous white light on the entire retina, then
the excitation will match the inhibition for this sort of cell,
the way the sky would be according to Olber's paradox, because there
would be a cell that had its on centre, say, at the location of the
"T" in the word "CENTRE", but other cells would have their inhibitory
surrounds at the same point, so the excitation and inhibition would
just about cancel out.

The same would be true if the field were all black (it's just that
the average level of the balance between excitation and inhibition
would be at a different level: the white field looks white and the black
field looks black).

But if the white field is put beside the black field, so that
there is a vertical line that forms the border or "zipper" between the
two fields, then something special will happen along that border because
the cells on just the other sides of the border -- the ones lined
up on the last column of white and the ones lined up on the first column
of black -- will fire very differently from the ones that are further
into the white or black field: The inhibitory fields of the on-center
off-surround cells would add together instead of cancelling each other
out at the boundary.

The effect is even better visualise with gratings. See the
amazing resources there are on the Web. I got the ones at the
end of this message by just going to :
http://altavista.digital.com
and doing an advanced serach on:

lateral near inhibition
receptive near field

The top one goes in the upper box and the lower one in the bottom box.
It should bring back about 200 hits, many of which are excellent.

Try it with every question you have! Lecturers all over the world are
putting their sometimes superb instructional material on the Web.
Every time you don't understand something from one explanation, you can
jump to another one, from a different angle. (Pick up the
URLs with a mouse to minimise the chances of misspelling them.)

Cheers, Stevan

http://server.esc.cquest.utoronto.ca/psych/psy280f/ch3/rf/rf.html#rf_cs
http://www.yorku.ca/research/vision/eye/toc.htm
http://zeus.rutgers.edu/~ikovacs/SandP/prep.html#menu
http://zeus.rutgers.edu/~ikovacs/SandP/prep.html#menu
http://www.socsci.uci.edu/cogsci/courses/psych9b/lec8/lec8notes.html
http://psych.hanover.edu/Krantz/receptive/
http://serendip.brynmawr.edu/~rtimberl/neuro/nbbintro.htm



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:24:17 GMT