"The Treachery of Images": How Real Objects Affect Human Brain Processing During Action, Perception, and Cognition Tasks
Jody Culham
University of Western Ontario

For reasons of convenience and experimental control, neuroimaging studies, like most studies in cognitive neuroscience more generally, typically employ two-dimensional images as stimuli. But are images a valid proxy for the real objects that they represent? My lab has developed new apparatus (the Delivery of Real Objects for Imaging Device or DROID) to rapidly present real objects and 2D photographs of those same objects as stimuli in functional magnetic resonance imaging (fMRI) studies. Across several experiments, we have found that the brain does not process photos and real objects as equivalent. First, two fMRI techniques that examine neural coding (fMRI adaptation and multivoxel pattern analysis) have revealed differences in the representations of photos and real objects in brain areas implicated in object recognition. Second, even in a high-level cognitive task of assigning value to stimuli, neural signals are modulated by stimulus realness. Specifically, a neural signal in ventromedial prefrontal cortex shows a stronger valuation-correlated signal for real objects than photos. Taken together, these results suggest that real objects have important features, perhaps including their potential for interaction, that affect brain responses across a range of action, perception and cognition tasks.