Abstract
In visual perception, human faces are detected efficiently: we look at them preferentially, while is it difficult to ignore them. While such preferential processing has been established for faces, little is known about potentially similar processing of body parts such as hands. Oculomotor and related control mechanisms can be quantified by measuring eye movements toward and away from hands and faces (prosaccades and antisaccades). The study aimed to measure whether a bias for hands would be manifest in antisaccades with high error rates and long latencies, as reported for faces. This study used an eye-tracking method to compare saccades towards and away from images of hands, faces, bodies (without heads) and cars. Each trial presented a target stimulus (for example, an image of a hand) to the left or right of a central fixation mark (9 degree eccentricity), and a square on the opposite site. Depending on instructions, participants had to move their eyes either towards the target or away from the target (towards the square). 200 prosaccades and 200 antisaccades were tested. The difference between antisaccades and prosaccades in terms of latency and error rate was quantified as antisaccade cost. Antisaccade cost on latencies showed a significant main effect of stimulus type, with the cost for hands being higher than for whole bodies or cars, but lower than for faces. Similarly, a significant main effect of stimulus type was found for antisaccade cost on error rates. For hands, cost on error rates was higher than for whole bodies or cars, but lower than for faces. Because the antisaccade cost (on latency and error rate) is higher for hands than for whole bodies or cars, these findings suggest that hands benefit from similar preferential processing as faces (although to a lesser extent), reflecting a bias for hands in social perception.