Previous studies investigating attentional biases in social anxiety disorder (SAD) have yielded mixed results. Recent event-related potential (ERP) studies using the dot-probe paradigm in non-anxious participants have shown that the P1 component is sensitive to visuospatial attention towards emotional faces. We used a dot-probe task in conjunction with high-density ERPs and source localization to investigate attentional biases in SAD.

Twelve SAD and 15 control participants performed a modified dot-probe task using angry–neutral and happy–neutral face pairs. The P1 component elicited by face pairs was analyzed to test the hypothesis that SAD participants would display early hypervigilance to threat-related cues. The P1 component to probes replacing angry, happy or neutral faces was used to evaluate whether SAD participants show either sustained hypervigilance or decreased visual processing of threat-related cues at later processing stages.

Compared to controls, SAD participants showed relatively (a) potentiated P1 amplitudes and fusiform gyrus (FG) activation to angry–neutral versus happy–neutral face pairs; (b) decreased P1 amplitudes to probes replacing emotional (angry and happy) versus neutral faces; and (c) higher sensitivity (d′) to probes following angry–neutral versus happy–neutral face pairs. SAD participants also showed significantly shorter reaction times (RTs) to probes replacing angry versus happy faces, but no group differences emerged for RT.

The results provide electrophysiological support for early hypervigilance to angry faces in SAD with involvement of the FG, and reduced visual processing of emotionally salient locations at later stages of information processing, which might be a manifestation of attentional avoidance.