Seeing eye-to-eye: Social gaze interactions influence gaze direction identification

We tested whether gaze direction identification of individual faces can be modulated by prior social gaze encounters. In two experiments, participants first completed a joint-gaze learning task using a saccade/antisaccade paradigm. Participants would encounter some 'joint-gaze faces' that would consistently look at the participants saccade goal before participants looked there (Experiment 1) or would follow the participants gaze to the target (Experiment 2). 'Non-joint-gaze faces' would consistently look in the opposite direction. Participants then completed a second task in which they judged the gaze direction of the faces they had previously encountered. Participants were less likely to erroneously report faces with slightly deviated gaze as looking directly at them if the face had previously never engaged in joint gaze with them. However, this bias was only present when those faces had looked first (Experiment 1) and not when the faces looked after participants (Experiment 2). Comparing these data with gaze identification responses of a control group that did not complete any joint-gaze learning phase revealed that the difference in gaze identification in Experiment 1 is likely driven by a lowering of direct gaze bias in response to non-joint-gaze faces. Thus, previous joint-gaze experiences can affect gaze direction judgements at an identity-specific level. However, this modulation may rely on the socio-cognitive information available from viewing other's initiation behaviours, especially when they fail to engage in social contact.