Abstract
Facial expressions are important visual cues for affective perception. Previous studies have used schematic faces that were assembled from simple geometric shapes to investigate emotion perception. Moreover, some simple noncontextual shapes, such as a downward pointing triangle, are perceived as threatening and capable of activating the neural circuitry underlying threat detection (Larson et al., 2008). However, recent behavioral and electrophysiological evidence shows that shape features underlying threatening faces can improve visual search performance when and only when they are interpreted as a facial component (Weymar et al., 2011). How may facial contexts modulate the affective perception of simple shapes? We generated a series of simple geometric shapes that varied in roundness (triangle/triangle-like ellipse/circle-like ellipse/circle), orientation of the acute angle (downward/upward), and color (black/white). Being part of a fictitious treasure-hunting scenario, participants were asked to make preference choices of a direction marked with one of the shapes that were described above. The effect of simple geometric shapes on affective perception was replicated: participants tended to avoid sharp angled and downward-pointing shapes, especially when the shapes were presented in black. However, we found no categorical boundary between "good" and "bad" shapes. In a follow-up experiment, we measured participants’ preferences of schematic faces with the same simple geometric shapes placed in the mouth position. In this case, the affective effect of the shapes reversed: participants preferred the schematic face with a downward-pointing and less rounded mouth. No effect of color was found in this context. Additionally, our results suggest a categorical boundary between "good" and "bad" schematic faces. In summary, affective visual perception of simple shapes changes with the addition of a facial context, suggesting that face processing may play a distinct role in affective categorical perception.
Meeting abstract presented at VSS 2012