Abstract
Face perception is often characterized as depending on holistic, rather than part-based, processing. However, attempts to find neural correlates of holistic processing in “face-selective” regions using functional magnetic resonance imaging (fMRI), relying primarily on stimulus inversion, have had mixed results. Here we report a set of experiments which extend the study of holistic processing in two ways: through use of a different stimulus manipulation, and by measuring how such processing is modulated by familiarity. Our stimuli utilize binocular disparity to create a percept of a face either behind or in front of a set of stripes (Nakayama et al., 1989). While the first case will be “filled in” by the visual system and perceived as a normal face occluded by stripes, the latter cannot be completed amodally and thus is perceived in terms of its constituent parts. We demonstrated this behaviorally using the “whole-versus-part superiority” effect (Tanaka & Farah, 1993): when whole faces were presented in front of the stripes, recognition performance dropped to the same level seen for isolated face parts. Therefore, these “stereo faces” provide a powerful means of assessing holistic and part-based processing. Using these stimuli, we examined the sensitivity of the “face-selective” fusiform face area (FFA) to holistic versus part-based processing in familiar and unfamiliar faces. Existing behavioral results suggest that while familiar faces are processed holistically, unfamiliar faces may be recognized by individual parts. Consistent with these prior results, preliminary data from an adaptation paradigm show an interaction between depth and familiarity in the right FFA, with greater adaptation in the Back (holistic) depth condition relative to Front (parts) for familiar but not unfamiliar faces. This suggests that the FFA is not dedicated to a single type of processing. Rather, activity in this one cortical region may encompass multiple mental operations with distinct behavioral signatures.