Davie Yoon
Graduate Student

When we look at brains that are looking at faces, we find several distinct face-selective regions. I want to know whether these regions are connected in a face processing network, with different regions specialized for different functions. This model is motivated by the idea that visual recognition requires the perceiver to detect variance on relevant dimensions and discard variance on irrelevant dimensions. A problem arises when dimensions that are irrelevant for recognizing one aspect of a face may be critically relevant for recognizing something else. For example, when I smile or frown, my face looks different. An expert face identifying mechanism should not be fooled by these changes. However, an expert expression recognizing mechanism should be pick up on this change, even though many other aspects of my appearance are not changing. The need to represent different information for different functions creates a conflict that may drive different regions to become functionally distinct over time. I therefore expect representations that require prolonged experience and learning to be computed by regions that are increasingly tuned to changes in relevant dimensions and decreasingly sensitive to changes in irrelevant dimensions across development. My dissertation work will address this hypothesis with fMR-adaptation and DTI fiber-tracking experiments in children and adults.