Abstract
Neuroimaging studies have revealed a number of regions in the human brain that respond to faces. However, the way these regions are defined and how they interact is a matter of current debate. Here, we used fMRI to define face-selective regions in the human brain, and functional connectivity to determine how these regions interact. First, we compared the response to faces with the response to a range of non-face images. The pattern of selectivity across the brain varied for each contrast (face > bodies, face > objects, face > places, face > scrambled). Nevertheless, the core face regions in the occipital and temporal lobe (FFA, OFA, STS) all showed significant face selectivity, with a larger response to faces compared to each non-face image condition. Face-selectivity was also evident in other regions that have previously been incorporated into models of face processing, such as the amygdala, inferior frontal gyrus and intraparietal sulcus. However, we also found face-selectivity in the precuneus and superior colliculus. Next, we determined how the face-selective regions interact. We found significant correlations between the time-courses of the core face-selective regions (FFA, OFA, STS). To determine whether this coactivity between the core regions reflected functional connectivity, we removed the stimulus-driven component in each time-series. Although the non-stimulus driven time-courses in the OFA and FFA showed a significant face-selective correlation, we found no evidence for functional connectivity between the OFA and STS or between the FFA and STS. Interestingly, the connectivity between corresponding face regions in different hemispheres (e.g. rFFA-lFFA) were higher than between different core regions in the same hemisphere. These findings challenge models of face processing that propose the OFA projects to both the FFA and STS. These data also imply that inter-hemispheric connections between corresponding face-selective regions may play an important role in face processing.