Finding an emotional face in a crowd and the role of threat-biased attention in social anxiety

Sammanfattning: The evolutionary perspective on the functionality of prioritizing efficient threat detection (Öhman, 1986, 1993) provided the foundation for a hypothesis of an attentional bias for preferential detection of angry faces in a group of faces. Furthermore, an evolutionary analysis implied that such a threat bias would be enhanced in socially anxious individuals, whose bias for detecting threat in faces would be potentiated by their increased sensitivity for facial signals of social dominance. In Study I, we used a standardized set of photographed real faces in examining visual search for happy, angry, and fearful target faces among neutral distractor faces in three separate experiments. Contrary to the hypothesis about prioritized threat detection, there was a strong happy-advantage and no consistent effects of social anxiety or social phobia, even when it was activated by a fearinduction procedure. However, in the final experiment, with perceptually controlled schematic faces, in support of the hypotheses, we obtained more effective detection of angry than happy faces, which was most obvious for highly socially anxious individuals when their social fear was experimentally activated. In Study II, we developed and tested a theory to reconcile the opposing findings with real and schematic faces in Study I. This theory was based on three theoretical concepts: (a) Rauschenberger and Yantis s (2006) theory of stimulus redundancy, (b) Lavie s (2005) concept of perceptual load, and (c) the differential ease of recognizing facial emotions of anger and happiness in the two genders (Becker et al., 2007). The results were predicted on the presumptions that redundant (i.e., familiar ) distractor faces should leave more perceptual processing resources for finding the target stimulus, which would lead to a larger facilitation of the processing of angry, than happy, faces, because their recognition seems to require a configural facial analysis. Our results demonstrated that two interacting conditions were required to obtain the angry face advantage: (a) a small stimulus set (i.e., they should be highly redundant); and (b) a male target face. In Study III we addressed the uncertainty about which attentional process was involved in the enhanced threat bias that had been observed in Study I. Thus, we recruited high and low socially anxious individuals for two different tasks. The first one was designed to reveal a potential association between social anxiety and (enhanced) attentional shifts to socially threatening stimuli; the second task examined the (dis-) ability to disengage attention from a socially threatening stimulus. Consistent with Study w found the superior orienting towards social threat in a visual search task to be unaffected by self-reported social anxiety. The socially anxious participants performed worse than the low-anxious control group in trials where a socially threatening task-irrelevant facial stimulus was presented. A prolonged threat dwelling explanation for the socially anxious participants slowed RT performance was undermined by the recorded eye-movement patterns, which suggested that they rather had problems with controlling attention. Thus, while not being preoccupied with the threatening stimulus in itself, their unstable attentional focus seemed to reflect the draining of attentional resources available for the cognitive task, suggesting that endogenous attentional control efforts were compromised.