Haley Frey¹, David Whitney¹; ¹University of California, Berkeley

Previous work has demonstrated that changes in perception can be measured by comparing maximum likelihood difference scales (MLDS). A preceding stimulus can alter the appearance of the current stimulus, if the stimuli are similar, leading to a greater perceived similarity of those stimuli relative to others along a continuum. Here, we expand upon these results by using dynamic stimuli to prime perception and compare difference scales to a baseline (no prime) curve. The stimuli are video clips that slowly morph through a subset of colors from the continuum. We anticipated that the dynamic prime would show a similar serial dependence effect as the static prime did, reflected by greater similarity ratings of the stimuli involved in the preceding morph compared to stimuli outside of this range. Indeed, we find evidence of such changes, particularly in the center range of the continuum values. Our results demonstrate a “reverse crispening” that occurs in the MLDS curve in response to dynamic primes in this center range. This “reverse crispening” is interesting because it reflects a pattern in similarity perception that opposes the typical psychometric function. One possibility is that the perceptual effect of primes from the central region of the continuum have a greater impact on the scale. Stimuli that are already perceived as similar (usually at either end of the continuum) may not show an obvious increase in perceived similarity following a prime. However, an increase in perceived similarity among stimuli at the center of the continuum, which observers usually rate as more different, would drastically alter the resulting difference scale, leading to the seeming reverse crispening we observe here.

Acknowledgements: Institutional Training Grant T32EY007043 awarded to the Department of Vision Science, School of Optometry, University of California, Berkeley