Abstract
Learning to categorize visual objects causes a selective enhancement of visual discriminability along category-relevant dimensions. This perceptual advantage for relevant dimensions can be observed outside of active categorization using both behavioral and neural measures and persists days after the category learning task is over. These data suggest that although category learning certainly requires attention, its perceptual effects remain in the absence of specific attention to category relevant features. In all previous studies, however, the relevant dimension advantage is measured with the stimulus presented at an attended location. A stronger test of the hypothesis that the relevant dimension advantage is not driven by attention would be to neurally assess this advantage while the stimulus is presented at an unattended location. The first step in this research is to demonstrate behaviorally that the relevant dimension advantage can be obtained when stimuli are presented off of fixation, allowing us to direct attention away from the stimuli in future studies. We therefore trained subjects to categorize a two dimensional space of morphed cars according to a single relevant dimension of variation. Visual discriminability between cars that differed along the relevant or irrelevant dimension was measured before and after category learning. Importantly, the cars were presented 1.6 degrees above fixation during both category learning and discrimination testing. Subjects’ eye movements were monitored during the task. If, during either category learning or discrimination testing, they broke fixation and moved their gaze to the car stimulus, the car disappeared immediately. Replicating our previous studies conducted with foveated stimuli, subjects’ ability to discriminate between stimuli that differed along the category-relevant dimension improved significantly more than those varying along the irrelevant dimension. Future studies will measure neural adaptation effects between off-fixation stimuli with category relevant vs. irrelevant differences while subjects perform an attentional-demanding task at fixation.
Meeting abstract presented at VSS 2013