The effects of lightness differences and luminance differences on visual search efficiency
Poster Presentation 36.425: Sunday, May 17, 2026, 2:45 – 6:45 pm, Pavilion
Session: Visual Search: Features, scenes, real-world stimuli
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Howard Jia He Tan1, Simona Buetti1, Cathleen Moore2, Alejandro Lleras1; 1University of Illinois Urbana Champaign, 2University of Iowa
According to Reverse Hierarchy Theory, processes like visual search first operate on the end-products of visual processing. One such end-product of the visual processing system is perceptual constancy. Perceptual constancies provide us with invariant representations of objects out in the world. In the current study, we investigate the level of visual representation used in visual search with the phenomenon of lightness constancy and phenomenal scission. Here, we used displays that invoked lightness constancy illusions and controlled for local luminance differences between search stimuli and their backgrounds. We hypothesized that search processes would use lightness differences between items (“post-constancy” representations), when they are informative, to guide search. Subjects searched for the perceptually brightest target semicircle among other semicircles. We controlled luminance differences (“pre-constancy” representations) and defined the target by a uniquely bright luminance (easy) or luminance-matched with that of distractors in a different region of the display (hard). There were two background conditions, one induced a perceived filter (Adelson snake illusion) and the other a flat surface. Set size was manipulated (Target only, 4, 8, 16) and distractors were distributed equally across the high- and low-luminance regions. RTs increased logarithmically as a function of set size, indicating that participants had performed the search in parallel across the display. When perceived lightnesses were equal, search was efficient and no significant differences were found between background conditions. More importantly, when targets were luminance-matched with distractors in the surface condition, search became significantly harder. However, and somewhat surprisingly, in the filter condition, search was just as easy when the target was luminance-matched as when it was a unique luminance, even when the luminance information available should have resulted in a more difficult search. These results suggest that early visual information is only used when necessary and are in line with the Reverse Hierarchy Theory.