Abstract
Adult observers are sensitive to the difference between real and artificial face appearance. Real faces are distinguished from artificial faces easily, and relative to real faces, artificial faces are remembered more poorly (Balas & Pacella, 2015) and trustworthiness is perceived less accurately (Balas & Pacella, 2017). These results, among others, suggest that artificial faces represent an "out-group" of faces, like other-race or other-age faces, based on the extent to which they deviate from faces that dominate experience. How does this sensitivity to artificial appearance develop? Other-race and other-age effects are malleable during childhood, which suggests that children's sensitivity to artificial faces could differ from that of adults. Presently, we chose to examine this question using neural responses (e.g. the N170 ERP component) to real vs. artificial face appearance as a means of characterizing face processing in school-age children. We recruited a total of 36 children between the ages of 5-10 years old (18 5-7 years old, 18 8-10 years old) to participate in this study. We recorded ERPs from each participant while they viewed randomly presented images of upright and inverted faces that either depicted real people or dolls. Specifically, we examined the sensitivity of the P100 and the N170 to real vs. artificial appearance and face orientation. We found that unlike adults (Balas, van Laamswerde & Saville, 2017), the mean amplitude of the N170 did differ as a function of both orientation F(1,34) = 8.45, p = 0.006) and face animacy (F(1,34) = 10.98, p = 0.002). Thus, while we usually think of face processing becoming more fine-tuned with continuing development, these results suggest that children have more sensitivity to the real/artificial distinction at early ERP components than adult observers do. Face animacy may be a unique case where some stages of face processing encode category differences more weakly as development proceeds.
Meeting abstract presented at VSS 2018