Abstract
Observers are good at determining whether a face is real or not, and animacy is perceived categorically in human faces (Looser & Wheatley, 2010). Presently, we used high-level adaptation to determine if face animacy is represented in a category-specific fashion, or if the perception of animacy across face categories is supported by a common neural mechanism. We created four sequences of morphed adult faces that spanned a continuum between real and artificial appearance in increments of 10%. We recruited 56 observers to rate perceived face animacy before and after adaptation to (1) Adult faces, (2) Child faces, and (3) Dog faces. Pre-adaptation, participants viewed the test images in a randomized order for 1000ms each and provided ratings on a 1-7 Likert scale. In the adaptation phase, participants were adapted to an adult face (N=24), a child face (N=16), or a dog face (N=16). The use of real/artificial adapting images was balanced within each group, and adapting images were presented at twice the size of the test images to minimize low-level adaptation. Each participant’s animacy ratings in both phases were fit to a logistic function, and after-effect magnitude was assessed using shifts in the point of subjective equality (PSE). We obtained significant aftereffects following adaptation to adult faces (t(23) = 4.04, p = .001) and child faces (t(15) = 6.96, p <.001), but not following adaptation to dog faces (t(15) = -.48, p = .64). In a follow-up task, we confirmed that within-species animacy aftereffects obtain when morphed dog images and dog adaptors are used (t(15) = 2.15, p = .05), suggesting that the lack of transfer between dog and human faces is not a function of the selected dog images. We conclude that the perception of animacy in face images is species-specific, but generalizes across age categories.
Meeting abstract presented at VSS 2013