Dissociating Task Performance from fMRI Repetition Attenuation in Ventral Visual Cortex

↵a Our scene task required the observers to attend to and comprehend the scene images as a whole, whereas our image task required the matching of scene images pixel by pixel. Thus, more holistic and global processing would be necessary for the scene task and more featural and local processing would be necessary for the image task. As a result, observers might have used different strategies in the two tasks. For example, they might have attended more to the edges of the scene photographs to help them perform pixel-by-pixel matching of the scenes in the image task, and attend more to the overall scene photographs in the scene task. Although observers were explicitly instructed to fixate at the center and the stimulus presentation was brief to prevent eye movements, we did not measure eye movements and therefore could not know exactly where observers looked during the two tasks and whether they changed their direction of gaze across trials/tasks. Despite these uncertainties and differences associated with the two tasks, we obtained comparable overall PPA responses for the two tasks. If anything, the overall PPA response was slightly higher in the image than in the scene task, despite more in-depth scene processing in the scene task. Thus, differences in eye gaze and strategy in the two tasks, if they existed, did not seem to affect PPA response in a significant way.