Abstract
When we view images of objects, we can easily recognize the type of object (e.g. a mug) even if we haven't seen the specific exemplar (e.g. that mug) before. The neural mechanisms underlying object recognition are typically studied using 2-D pictures of unfamiliar objects. But how might real-world experience with a specific object affect its representation in the human brain? To address this, we conducted an fMRI study that involved exploration of real-world objects prior to scanning. Participants (N = 30) physically explored one of two sets of 12 ordinary household objects from six categories (clocks, mugs, bags, socks, stuffed toys, shoes) for 30 seconds. Each set contained two unique exemplars from each category. Following real-world exploration, participants viewed photographs of both seen and unseen objects in the MRI scanner (3.0T GE Discovery) in an event-related design. Each of the 24 individual objects were depicted twice in different contexts (indoor, outdoor) and at different viewing angles, for a total of 48 images. A post-scan memory test confirmed that participants correctly identified which objects they had explored from the photographs. Real-world encounters with objects were associated with widely distributed foci of increased activation in response to their 2-D image, and especially with increased activation in medial parietal areas compared to unseen objects. Similarly, region of interest analyses revealed that the object-responsive lateral occipital cortex was also sensitive to whether an object had been encountered in real life, but was insensitive to visual context. Together, the results demonstrate that even brief encounters with objects alter their brain representation. The richness of object representations beyond their photographic image has significant implications for our understanding of object recognition in both the human brain and in computational systems.