Abstract
Impossible objects are 2D drawings that represent objects that could not exist in real space. Yet, despite being perceived as exceptionally unusual, they still possess fundamental Gestalt attributes, such as closure and volume properties which are known to mediate holistic representation. In Experiments 1 and 2, we used Garner's speeded classification task to test this notion. Results showed that participants were unable to processes one dimension of possible and impossible object's shape while ignoring task-irrelevant dimensions belonging to the same objects. These findings suggest that the shape of impossible objects is processed in a configural manner and are in-line with recent neuroimaging findings (Freud, Ganel, Avidan; 2012) that showed that possible and impossible objects are processed by shared neural mechanisms, alongside differences between the two objects categories that were attributed to late cognitive processes. Experiments 3 and 4 were therefore aimed to test whether the initial perception of impossible objects involves similar perceptual mechanisms to those underlying typical object perception and that any perceptual differences obtained between the two object categories would only emerge later along the processing hierarchy. Experiment 3 utilized the object-based attention paradigm to show that impossible and possible objects are represented similarly such that processing features belonging to the same object was enhanced compared to features belonging to two different objects. Yet, responses for impossible objects were overall slower. Experiment 4 was designed to examine whether this difference could be attributed to late processing stages by manipulating exposure duration. Importantly, differences in accuracy between possible and impossible objects emerged only for long stimulus exposures thus verifying our working hypothesis. Overall, these findings suggest that the visual system utilizes intact shape attributes to create an object-based, configural representation of impossible objects and highlight the importance of these attributes for early stages of perceptual organization.
Meeting abstract presented at VSS 2013