Abstract
Previous fMRI studies have identified a region in the left mid-fusiform gyrus that selectively responds to visual words. This region is termed as the visual word form area (VWFA) based on a general consensus that the function of this region is domain specific (i.e. engaged by visual words, regardless of the type of processing). Here we used fMRI to test an alternative hypothesis that this region is engaged in a specific process that can be applied to stimulus class other than words. Specifically, we tested the sensitivity of the VWFA to body and face stimuli that were visually dissimilar to visual words and yet invoked the same cognitive process as visual words in terms of communication through body gestures or facial expression. Consistent with this process-specific hypothesis, we found that the VWFA responded significantly higher to both body parts and faces than to objects or scrambled objects. In addition, a conjunction analysis demonstrated that the joint representation of visual words, bodies, and faces was only observed consistently across subjects in the left fusiform gyrus, consistent with the lateralization of language processing. Second, we used fMRI adaptation to test whether VWFA response was sensitive to differences between bodies (or faces) or it was instead due to low-level properties of the stimuli. We found that VWFA response monotonically decreased as repetition frequency of body stimuli increased. This adaptation effect was not found in a set of novel objects that were visually similar to bodies but did not contain information for communication. Taken together, these findings indicate that the function of the VWFA is not domain specific for visual words forms but process specific possibly for translating the form of a symbol to the meaning of that symbol.