Objects In The Brain
We know very little about how the brain represents objects. Cognitive neuroscience has focused on finding regions of the brain that seem to respond selectively to objects of different categories, such as faces, buildings, and letters. But identifying such areas is only the first step. How does the visual input get transformed into representations of objects? At the earliest stages of processing, an object like a letter is represented in terms of elementary features (e.g. oriented lines). How do these oriented lines become letters? To address this, we combined crowding with neuroimaging. By adding clutter, we can make an object unrecognizable without actually changing the object itself. Thus, crowding is a novel tool for specifically manipulating object processing, and we used crowding to uncover a novel neural signature of object processing: correlated activity between early and higher, letter-selective areas of the visual processing pathway (presented at the Society for Neuroscience conference, 2008, manuscript submitted). The results suggest that the computation mediating object processing takes place via interactions between early, feature-selective neural populations and higher object-selective neural populations. We are now characterizing the stimulus dependency of this effect, and whether it depends on attention. This work is in collaboration with Tobias Donner and David Heeger.