Abstract
The classical Thatcher effect (TE) is experienced when global inversion of a face makes it difficult to notice the local inversion of its parts (Thompson, 1980). The TE can be quantified by comparing the ease with which observers compare a normal and locally-transformed image, when both images are shown upright vs. inverted. Here we compared the classical TE for images of adult faces to a wide variety of other categories, including grimacing faces, baby faces, animal faces, buildings, scenes, and various types of letterstrings. If the TE reflects a special form of configural processing for faces, faces should show a much larger TE than all other categories. Because a ceiling on the TE may arise if the accuracy with which changes are detected is very high for inverted pairs, we used two levels of transformation, allowing us to match conditions in terms of performance with inverted pairs. Faces showed a larger TE in RT than all other matched categories, apart from high frequency words. However, several nonface categories, including building and close-up scenes, yielded robust TEs in RTs, and in accuracy, the TE was comparable for faces and cars. A separate analysis compared TEs for low and high frequency words and their scrambled versions. Error rates revealed larger TEs for words than for scrambled words. Response times showed larger TEs for low frequency strings, regardless of whether they were words or scrambled versions of the same strings. Familiarity may influence the perception of the global orientation whereas word frequency may facilitate detection of local changes. Our results suggest the TE is not exclusive to faces - it does not appear to uniquely depend on factors such as expertise or the grotesque appearance of the transformation (as measured by independent ratings), although the string results suggest that familiarity is important.
This work was supported by a grant from the James S. McDonnell Foundation to the Perceptual Expertise Network