Abstract
Visual short-term memory (VSTM) capacity is larger for faces than other complex objects. Inversion reduces capacity for faces more than nonfaces (Curby & Gauthier, 2007). These findings suggest that VSTM is not influenced simply by object complexity, but also by the encoding processes employed by face experts.
Previous research (Scolari, Vogel, & Awh, 2008) found that perceptual expertise enhances the resolution but not the number of representations in working memory. In other words, people have a more detailed memory, instead of a larger WM capacity, for faces than nonfaces. Since we are more expert at recognizing own-race than other-race faces, we investigated whether the own-race advantage is due to a higher resolution of own-race face representations.
Six study items (Chinese and Caucasian faces as well as shaded cubes) were simultaneously shown on screen on each trial. After a short delay, a single image was presented. Participants were asked to judge whether this image was the same or different from the item that originally appeared in that location. Performance of both cross-category changes (i.e. face to cube, cube to face) and within-category changes (i.e. face to face, cube to cube, color to color) was measured.
Neither own-race nor other-race faces showed an inversion effect when stimuli changed between categories. However, an inversion effect was found for both own-race and other-race faces when stimuli changed within a category. These results suggest that both own-race and other-race faces are stored with high resolution in working memory.
This study was supported by a grant from the Hong Kong Research Grants Council (HKU744209) to William Hayward.