Visual adaptation to reflectance-specific image motion

Abstract

Recently, we described several surface reflectance-specific motion characteristics that the visual system may use to determine whether a rotating object appears shiny or matte (Doerschner et al., 2011). We used an adaptation paradigm to test whether there exist neuronal populations that are sensitive to such reflectance-specific image motion characteristics. Stimuli were computer-rendered movies of 5 unfamiliar, rotating objects. Eleven stickiness levels for each object were created by morphing between 'sticky' (diffusely reflecting) and 'slipping' (100% specularly reflecting) renderings of a given object with different mixing values, resulting in a total of 55 movies. For familiarization purposes observers were first shown a sequence of movies of an object transitioning from sticky to slipping though all 11 levels. In the pre-test observers rated the apparent shininess for each movie on a scale from 1 (very matte) to 5 (very shiny). The order of presentation was randomized. During adaptation, observers first adapted to a sticky movie for 120 s. This was then followed by a 2 s test in which observers rated shininess. Every fifth trial was preceded by a 24 s top-up adaptation period. Importantly, in order to prevent low-level motion adaptation we randomly selected a new rotation axis (out of 6) for each 2 s interval for the adaptor during adaptation periods. We compared the shininess ratings of all movies in pre-, and post-test. Overall, we found that, across observers and objects, adaptation to a sticky movie significantly affected the perceived shininess of subsequent stimuli (All observers: F(1,1098)=10.4781 p<0.002). Post-hoc analysis revealed that an increase in perceived shininess occurred mainly at higher levels of stickiness. These results support the notion of cortical mechanisms sensitive to reflectance-specific image motion patterns.