Abstract
While configural processing has been observed for faces and words, it remains unclear whether similar neural correlates support such processing for these visually distinct categories. Specifically, since selectivity for faces and words is often observed in different focal regions in the occipitotemporal cortex (e.g., right fusiform ‘face’ area (RFFA) for faces, and left visual word form area (LVWFA) for words), to what extent configural processing of the specific category is observed in the respective category-selective regions? In Experiment 1, Chinese participants (N=20) viewed faces and Chinese characters in either intact or exchanged formats (with top and bottom parts switched). FMRI results revealed that in RFFA1, RFFA2, and LVWFA, defined at the individual-level in a separate localizer, RFFA1/RFFA2 were sensitive to configural information of faces, with significantly stronger activations for intact than exchanged formats (p’s<.001). In contrast, sensitivity to configuration of Chinese characters was observed in LVWFA, but with weaker activations for intact than exchanged formats instead (p=.003). In Experiment 2, native English speakers who were non-Chinese readers (N=20) viewed English words and Chinese characters in either intact and exchanged formats (with left and right parts switched for English words). Similar to the results with Chinese participants viewing Chinese characters, weaker activations were found in LVWFA when native English speakers viewed intact than exchanged English words (p<.001), but no differences were found across the two formats of Chinese characters (p>.27). Additionally, RFFA1 showed stronger activations for intact than exchanged English words (p<.05), similar to the effects of faces in right FFA1 in Experiment 1. Together, both experiments revealed different patterns of configural effects for RFFA1/RFFA2 and LVWFA. Moreover, LVWFA appears to support expertise-related configural processing of words only, suggesting a domain-specific mechanism, while RFFA1 supports configural processing of faces, and of words in the absence of faces, suggesting a domain-general mechanism.