ATTENTIONAL BIAS FOR THREATENING FACIAL EXPRESSIONS: MANIPULATION OF STIMULUS SET SIZE

Abstract

When multiple stimuli are presented simultaneously, they compete for limited attentional resources. This competition can be mediated by emotional valence, with angry faces eliciting faster and more accurate responses than neutral or positive faces in visual search and attentional blink studies. The current study was designed to investigate the consequences of this attentional bias on the accuracy of recognition memory for unfamiliar faces. In order to explore this issue, we used an inverse square root function based on the Sample-size model to characterise the relationship between N (set size) and d’ (a signal detection theory measure) for threatening and non-threatening faces. The results revealed a decrease in recognition accuracy for neutral faces as the number of to-be-remembered (TBR) faces increased. When the target-face had an angry expression, however, d' was significantly greater than neutral faces but only when there are two faces in the display. This advantage appears to be largely independent of saccadic sampling and suggests the benefit occurred during maintenance rather than encoding. Importantly, when one of the distractors had an angry expression, sensitivity to neutral faces reduced dramatically. This suggests that angry faces exhaust the resources available to process neutral faces. Importantly, because set size increases the distribution of resources during encoding (indexed by reduction in fixations duration) and maintenance, the observed interaction between set size and threat superiority effect suggests that decreasing the quality of the perceptual information available by increasing set size, is likely to decrease threat superiority because it reduces the information driving the activation of the amygdala.