Abstract
It is difficult to detect and recognize a peripheral face in a crowd compared to when the face is presented isolated, a phenomenon known as crowding. A strong crowding effect can even render a face completely invisible, which is perceived to be indistinguishable from its scrambled version. It is not known to what extent the invisible face can be processed by the visual system. Here, we studied whether invisible emotional (fearful and happy) and neutral faces induced by crowding can influence spatial attention. A Posner cueing paradigm was adopted to measure the spatial cueing effect of invisible faces on an orientation discrimination task. A fearful, happy or neutral face was presented briefly in either the lower left or the lower right quadrant. Meanwhile, a scrambled neutral face was presented in the contralateral position. Both of them were flanked by inverted neutral faces, which rendered the face invisible as confirmed by a face detection task. Then a gabor patch was randomly presented at the position of either the face or the scrambled face. Subjects had to indicate the orientation of the gabor (clockwise or counterclockwise tilted). We found the attentional cueing effect only with fearful faces. That is, subjects' performance was better when the gabor was presented at the position of a fearful face than at the position of a scrambled face. To rule out that the cueing effect was due to low-level feature differences among fearful, happy and neutral faces, we conducted two control experiments in which faces were replaced with their space-scrambled or phase-scrambled counterparts. In both experiments, no attentional cueing effect was observed for all three kinds of faces. These results demonstrate that fearful faces could survive crowding and be profoundly processed by the visual system, consequently, direct the distribution of spatial attention.
This work is supported by the National Natural Science Foundation of China (Project 30870762, 90920012 and 30925014).