Abstract
Goffaux and Rossion (2006) proposed that holistic face processing is based on low spatial frequencies (SFs) whereas featural processing relies on higher SFs. Since upright faces are supposedly recognized through holistic processing and inverted faces, using features, we can easily take the leap and suggest a qualitatively different SF tuning for both orientations. Two independent studies (Gaspar et al., 2008; Willenbockel et al., 2010) tested this question and found the same SF tuning for both orientations. However, these studies used small faces presented through an elliptical aperture, i.e. stimuli of poor ecological value; differences in the SF tuning for upright and inverted faces were thus possibly missed. The present study revisits the SF tuning for upright and inverted faces using SF Bubbles. This method randomly samples SFs on a trial-by-trial basis and determines which SFs are correlated with accurate face identification. Experiment 1 (n=16) compared the SF tuning for upright and inverted smaller faces (5° of visual angle) when natural facial contour was present or hidden. Experiment 2 (n=14) compared the same conditions as Experiment 1, but using larger face stimuli (9.1° of visual angle). A Pixel test (Chauvin et al., 2005) was applied to the classification vectors to determine significance (Sr=256, FWHM=3.53, Zcrit=3.45). In Experiments 1 and 2, the SF tuning for both orientation conditions did not significantly differ (p>.05). These results bring further support to the hypothesis that the face inversion effect is due to quantitative differences in the efficiency with which information from the same SF band is used in both orientations, regardless of the ecological value of the stimuli.
Meeting abstract presented at VSS 2015