Abstract
The face inversion effect (FIE) is defined as a disproportionate drop in recognition performance when faces, compared to other objects, are inverted in the picture plane. Recently, the FIE has been attributed to a shrinking perceptual field of view (PFV; i.e., the visual field area in which features can be processed in parallel and integrated in a perceptual whole) as a direct result of inversion (Van Belle et al., 2015). This offers an elegant and testable explanation to the FIE, since shrinking the PFV by inverting faces would limit face processing to fewer (or even one) features, instead of the whole face, as is normally the case when faces are upright. Recent research has shown inversion also reduces efficiency with which horizontal spatial information is processed (Pachai et al., 2013), though spatial frequency use appears unaffected (Royer et al., 2017). The goal of this study was to test whether PVF shrinking is causally linked to the FIE. To this end, we simulated the effect of PFV shrinking in ten participants by revealing faces through a small gaze-contingent Gaussian window. We then measured face processing strategies for upright, inverted, and windowed stimuli using spatial frequency (Willenbockel et al., 2010) and orientation (Duncan et al., 2017) bubbles (520 trials per condition). Size of the Gaussian window was determined on a subject basis to produce FIE-magnitude accuracy loss. For inverted vs. upright faces (i.e., FIE), there was a reduction in horizontal spatial information processing efficiency, but no difference in spatial frequency processing, replicating previous findings. For windowed vs. upright faces (i.e., PFV simulation), there was a reduction in horizontal processing efficiency, but it was not commensurate with the FIE. Furthermore, there was also a change in frequency tuning, which is not expected of the FIE. Thus PFV reduction does not cause the FIE.