Abstract
Usually, dot-probe studies find an attentional bias to threatening stimuli in anxious, but not in non-anxious participants. However, some studies applying the same or related paradigms report a significant attentional bias to threatening stimuli also in unselected samples. According to the contingent-capture theory of spatial cueing, attentional control settings are tuned to relevant target features to facilitate target detection. An irrelevant cue can therefore capture attention if it matches a relevant feature of the target. Thus, consistent with this theory, one can hypothesize that an attentional control setting tuned to threat is permanently active in anxious individuals but can also be activated in non-anxious individuals under certain conditions. In Experiment 1, we aimed to replicate typical contingent-capture effects employing temporal and spatial parameters commonly used in dot-probe tasks. Participants had to classify green or red schematic target faces that were preceded by either green or red cues. Consistent with the contingent-capture theory, we found significantly larger cueing effects for matching than for non-matching color cues. In Experiment 2, we changed the critical feature from color to threat. Now, participants had to classify schematic target faces that were preceded by two photographic cue faces, one angry and one neutral. The targets were either defined by their emotional valence (angry expression, i.e., matching condition) or by a non-emotional feature (i.e., non-matching condition). Participants' anxiety was assessed by the state-anxiety scale of the STAI. As expected, we found a non-anxiety-related attentional bias to angry faces when participants had to classify angry targets. However, we also found a non-anxiety-related attentional bias to angry faces of equal size when participants were searching for non-angry targets. This suggests that a bias to angry faces can be found in non-anxious participants, but that this bias is not contingent on attentional control settings induced by current task demands.
Meeting abstract presented at VSS 2016