Abstract
In scene perception, there is evidence that the observers are able to extract the "gist" (i.e., semantics) from images presented as briefly as 13 ms (Potter et al., 2014). Although it is well known that faces are perceived holistically, how the holistic face gist is encoded has not been extensively explored. To test for the presence of gist processing, in Experiment 1, we flashed a study face for either 17, 50, 250 or 500 ms, followed by a 500 ms dynamic white-noise mask. Recognition of face features (eyes, nose, mouth) was then tested in isolation or in the whole face using Tanaka and Farah's parts/wholes paradigm (Tanaka & Farah, 1991). We found that eyes were better recognized in the whole face than in isolation when presented at an exposure duration of 17ms and mouth was better recognized in the whole face when presented for 50 ms and 250 ms. In Experiment 2, participants were asked to recognize the eyes and mouth in upright and inverted faces presented at 17, 50 and 250 ms using the same parts/whole paradigm as Experiment 1. The main effect of orientation was found where recognition in the inverted condition was significantly worse than the upright condition. Moreover, a significant Parts/Whole by Orientation interaction effect was found where recognition of features was better in the whole face than in isolation when faces were upright, but not when faces were inverted. Importantly, for upright faces, at 17 ms exposure duration, eyes were better recognized in the whole face than in isolation. For inverted faces, no differences were found between the whole face and isolation conditions. Collectively, these results present the striking evidence that the in face processing, holistic gist can be encoded at an exposure duration as short as 17ms.
Meeting abstract presented at VSS 2018