The effect of spatial frequency on faster detection of emotional expressions in visual search

Abstract

Facial expressions are adaptive social and communicative tools. In some early investigations of expression detection, both anger superiority effect and happiness superiority effect were reported. Several attempts were made to resolve the contradictory reports of emotion superiority effects in visual search and a well-supported explanation is that preferential detection of emotion can be attributed to simple visual features of faces. The current research focused on the spatial frequency (SF) of facial expressions to test if the reported inconsistencies are related to SF differences between the expressions. To do so, emotional faces were manipulated at the level of SF content by replacing original Fourier magnitude spectrum and four kinds of faces containing new SF information were created as stimuli. An online visual search task was conducted to examine preferential emotion detection. The results revealed the anger superiority effect for faces containing only main SF features that separate emotional expressions, faces with average magnitude spectra per expression category, or faces with 1/f noise frequency. However, ASE did not reveal when main SF features that separate emotional expressions were removed from faces. Such findings suggest that SF content can produce the emotion superiority effect when it can provide emotion-related important information in the detection of emotion.