Logan Wu¹, Jessica Huang¹, Jason Haberman²; ¹Rhodes College

Our visual system recognizes faces across a wide range of depictions. We can readily recognize a familiar individual stylized to look like a cartoon or even a caricature. Although face recognition is invariant across various depictions, we nonetheless exhibit extreme sensitivity to whether a face is real. This raises the question of whether realness is explicitly coded by the visual system, which we can explore using adaptation procedures. In our experiments, we examined how adaptation to either a photographic or cartoonish depiction of a face influenced subsequent realness judgments using the method-of-constant-stimuli. In Experiment 1, we generated interpolated morphs from ‘real’ to ‘cartoonish.’ Participants (N=27) first adapted for 45s to a single identity in either photograph or cartoon form, and then categorized whether a subsequently presented morph (between 20%-80% real) appeared more like a photograph or a cartoon. Each trial was preceded by a 3s top-up stimulus. Participants completed four blocks in random order (two blocks with a photograph adaptor and two blocks with a cartoon adaptor). Results revealed a stark shift in ‘realness’ thresholds — participants were more likely to call a morph real following adaptation to a cartoon, and vice versa. We replicated this pattern in a second experiment (N=23), in which the morph identity participants had to judge was randomly replaced on every trial with one of seven possible different identities. This confirmed that whatever features confer a sense of realness from a face remain consistent across different identities, even though other physical features may differ. Overall, our results show that animacy or ‘lifelikeness’ is an adaptable feature explicitly coded by the visual system.