Speaker/Affiliation: Norma Graham Ph.D. Professo

r Department of Psychology Columbia University

When/Location: 12/3

/07 12:00 PM SEA 4.244

Host: Mary M Hayhoe

Title of Talk: '

'Two Contrast Adaptation Processes - One Old and One New''

Reception
with Refreshments at 11:30 AM

Abstract: We serendipitously found so

me unusual dramatic effects of contrast adaptation in human pattern vision.
Here contrast adaptation changed some simple contrast-defined patterns fr

om impossible-to-identify to easy-to-identify (and vice versa). Test patte

rns that contained contrasts both below and above the adapting contrast (st

raddling test patterns) were extremely difficult to identify which we thou

ght bizarre. It is as if some process (known in our lab by the nickname Buf

fy adaptation) compares the current contrast at each spatial position to a

contrast-comparison level. This contrast-comparison level continually adap

ts to equal the recent time-averaged contrast in some neighborhood of that

position. It is the unsigned difference between the current contrast and t

he contrast-comparison level that largely determines the output of the proc

ess at each spatial position. The resetting of the comparison level occurs

very rapidly most of it at least within 100 milliseconds.

The new p

rocess is quite different from a process hypothesized previously to explain
effects after adaptation to a blank gray field (0% contrast). This old pr

ocess produces Weber-law type behavior. In experiments covering the entire
range of adapt and test contrasts the joint effect of these old and new p

rocesses is as follows: Performance is best for test patterns composed of

contrasts near the current comparison level and then worse on those compose

d of contrasts further away (in either direction) but this is true only fo

r non-straddling test patterns. As mentioned above identification perform

ance on test patterns composed of contrasts straddling the current comparis

on level is generally very bad. And we still think this bad performance ve

ry puzzling.

We wonder whether this bad performance on straddling pa

tterns is of some possible use in natural vision or is it just a not-too-c

ostly side effect of other desirable characteristics of contrast adaptation

. We wonder whether these processes'' functions are relevant to what occur

s within an eye fixation or between fixations or both. We wonder about a n

umber of mechanistic issues as well e.g. at what level(s) in the visual ne

ural system are these processes producing their perceptual effects.