Abstract
LOC, an area that has been shown to be critical for shape perception (James et al. 2003), is comprised of the Lateral Occipital Cortex (LO) and the Posterior Fusiform (pFs). It is defined (localized) as the region that shows greater activation when viewing intact objects compared to their scrambled versions (resembling texture). But scrambling a) reduces the familiarity of the objects, b) destroys the integrity of the parts, and c) leaves the relations among the parts undefined. We assessed the potential effects of familiarity, with strict control for low-level features by computer modeling familiar objects (Fig. 1a), such as a chair or a lamp, and then rearranging the relations to produce novel, intact objects (Fig. 1b). By scattering the intact parts (Fig. 1c)—but not breaking them up (as would be done in scrambling)—we could assess whether the effects of scrambling an image (Fig. 1d) can be attributed to the loss of parts, relations, or both, with different effects perhaps localized to different regions of LOC. Novel and familiar objects both yielded equivalent activation throughout LOC which was markedly greater than that produced by scrambling. In LO, intact objects (both novel and familiar) produced equivalent activation to scattered parts, suggesting that this area of LOC does not code interpart relations (Fig. 2). However, greater activation in pFs was evidenced by intact objects compared to scattered parts, suggesting that this region of LOC is sensitive to relations between parts.
Meeting abstract presented at VSS 2016