Abstract
Malach et al. (1995) discovered that the lateral occipital cortex and the posterior fusiform gyrus, cortical areas that they termed the lateral occipital complex (LOC), yielded greater fMRI BOLD responses when viewing intact images of familiar objects than their scrambled versions (resembling texture). Malach et al. discounted a role of familiarity by showing that unfamiliar “abstract” Henry Moore sculptures also activated LOC more than its scrambled versions. Although such a comparison does indicate that intact images produce greater LOC activation than their scrambled versions, it is not clear, without control for lower-level stimulus features, whether there is, in fact, no effect of familiarity. There is strong evidence that LOC represents objects in terms of their parts (Hayworth & Biederman, 2006). We put this issue of object familiarity to test by comparing cortical activation to 72 familiar objects and their novel counterparts, produced by rearranging their simple geon-like parts, thus holding the part ensemble of an image constant while varying familiarity (Panel A of Fig). Each object was composed of three or more geons, with each geon corresponding to a simple part of the object. The intact minus scrambled versions of each object was used to define LOC itself. With subjects performing an orthogonal task, greater activation in LOC was found for the familiar compared to the novel objects (Panel B). There was no effect of object symmetry. Although novel objects produce greater BOLD activity in LOC than their scrambled versions, there was still greater activation for familiar objects than their novel arrangements.
Meeting abstract presented at VSS 2015