Foraging for famous faces : exploring the interaction between facial expression and facial identity processing

Abstract

This thesis aimed to investigate two ongoing debates within the face perception literature. The first debate revolves around the question of whether expression and identity processing interact when making decisions about faces. The second debate concerns the possibility of a valance bias during the processing of positive or negative facial expressions of emotion. We adopted a foraging paradigm to explore whether categorising an emotional expression is independent of facial identity in the context of multiple-target search. Stimuli consisted of AI generated images of Caucasian male celebrities, each providing 4 exemplars to fill a 2 (valance) x 2 (arousal) expression space. Specific expressions were characterised as: a slight smile, full laughter, a slight frown, or extreme anger. In separate trials, participants (N=12) searched on an iPad for positive amongst negative facial expressions (or vice versa) irrespective of the level of arousal (high or low) depicted in the images. Each display contained an array of 40 faces: 20 target expressions, 20 distractor expressions (4 images per identity). The crucial manipulation was whether all faces within a trial shared the same identity or whether identity varied, with images being sampled from the full set of celebrities. Data were analysed using a 2 (Identity: Same/Different) x 2 (Expression: Positive/Negative) factorial design. A near identical follow up experiment (N= 12) was conducted, the main change being the stimuli. Additional stimuli for each expression space were created to reduce stimuli repetition. Results from both experiments clearly indicated slower response time for multiple identity trials. Results also revealed a significant processing bias for positive valence facial expressions, with reduced response time for positive target trials compared to negative target trials.