Explaining the face-inversion effect: the face–scheme incompatibility (FSI) model

Abstract

The face-inversion effect (FIE) can be viewed as being based on two kinds of findings. According to the face (UI) effect, perception and recognition are better for faces presented upright (U) than for faces presented inverted (I). According to the face/object(UI) effect, inversion impairs the processing of faces more than the processing of nonfacial objects (e.g., buildings or cars). Part I of this article focuses on the face(UI) effect and the configural-processing hypothesis, which is considered the most popular explanatory hypothesis of the FIE. In this hypothesis, it is proposed that inversion impairs the processing of configural information (the spatial relations between features) but hardly (if at all) impairs the processing of featural information (e.g., eyes, nose, and mouth). Part II of the article starts from the conclusion reached in part I, that the configural-processing hypothesis has not succeeded in explaining a substantial number of the findings and in resolving certain theoretical problems. The part then goes on to outline a new alternative model, the face-scheme incompatibility (FSI) model, which contends with these theoretical problems, accounts for the configural-processing hypothesis, succeeds in explaining a considerable portion of the empirical findings related to the face(UI) effect, and proposes a relatively new research program on the concept of the face scheme. The basic assumption of the FSI model is that schemes and prototypes are involved in processing a visual stimulus of a face and in transforming it to a "meaning-bearing" face, and that different schemes are involved if the face is presented upright or inverted.

Developmental Origins of the Face Inversion Effect

“…However, more recently, researchers have begun using the term "face inversion effect" to refer only to differences between upright and inverted faces in recognition as well as other aspects of face processing (for review, see Rakover, 2013). Because we will be evaluating the different ways inversion has been shown to affect face perception in infants, we follow the latter use of the term "face inversion effect."…”

“…Based on Yin's classic study, the "face inversion effect" historically has referred to the fact that inversion disproportionately affects face recognition relative to nonface stimuli (Rakover, 2013;Valentine, 1988). However, more recently, researchers have begun using the term "face inversion effect" to refer only to differences between upright and inverted faces in recognition as well as other aspects of face processing (for review, see Rakover, 2013).…”

“…Results have shown that adults are better at recognizing and discriminating upright faces than inverted faces (e.g., Valentine, 1988;Yin, 1969); they use more sophisticated processes, such as holistic and second-order configural processing, for upright faces than they do for inverted faces (e.g., Diamond & Carey, 1986; for reviews, see Maurer, Le Grand, & Mondloch, 2002;Rakover, 2013); they categorize them at different levels, treating upright faces at a subordinate level (i.e., as individual faces) (Tanaka, 2001). They even show specialized event-related potentials (ERPs) and functional magnetic resonance imaging (fMRI) responses to upright faces (e.g., Kanwisher, Tong, & Nakayama, 1998;Rossion et al, 2000; for review, see Rossion & Gauthier, 2002).…”

“…This effect is known as the "Thatcher illusion." It has been suggested that the Thatcher illusion is the result of adults' using second-order configural information to process upright, but not, inverted faces (e.g., Bartlett & Searcy, 1993;Diamond & Carey, 1986; but see Rakover, 1999Rakover, , 2013.…”

Infants' perception of faces in face-like and ambiguous images.

“…It is believed that processing a face configurally is a marker of expertise in face processing (Maurer, Le Grand, and Mondloch (2002). Adults process upright faces configurally but process inverted faces featurally (e.g., Diamond & Carey, 1986;Maurer, Le Grand, & Mondloch, 2002;Rakover, 2013).…”