Abstract
Studies of race aftereffects show that adaptation biases responses away from an adapting stimulus. However, it remains unclear if shifts in response frequencies result from changes in perceptual representations or in decisional mechanisms supporting race classification. General recognition theory (GRT) provides a single modeling framework within which we investigated adaptation-induced changes on perceptual and decisional mechanisms. Phase 1 measured participants' white-black discrimination thresholds on facial feature and skin tone dimensions. Results were used to construct a two-dimensional stimulus set for a GRT experiment tailored to each participant's race thresholds. Phase 2 constituted a set of typical race adaptation tasks, in which participants made 2AFC (white, black) responses to faces varying on a single stimulus dimension (features, skin tone) under each of four adapting conditions (white/black features, light/dark skin tone). We replicated previous findings that adaptation shifts perceived features away from the adapting stimulus (adapting to white features made faces appear more black and vice versa), and we provide new skin tone adaptation results showing a similar effect on this dimension. Phase 3 included five tasks requiring complete identification (CI) feature and skin tone responses to four faces in feature-skin tone space (derived in Phase 1). The no-adaptation CI task provides participants' baseline models of perceptual race space, revealing positive correlations between features and skin tone within and across stimuli. As face features were perceived as more black (white), skin tone was perceived as darker (lighter). CI completed under the four adaptation conditions enabled modeling of adaptation-induced changes in participants' race space. Fitted models reveal shifts in perceptual representations away from adapting stimuli, variability in the within-stimulus correlations, and shifts in the decision bounds toward the adapting stimulus. Additionally, equal numbers of self-identified Caucasian and African-American subjects allow us to explore potential race group differences in adaptation aftereffects.
This research was funded by NIMH grant 5R01 1MH057717 to J.T.T.