Abstract
It remains an open question in visual neuroscience whether the recognition of written words and visual objects engages distinct neural mechanisms intrinsically, unaffected by confounding factors such as stimulus properties and task demands, and, if so, where these differences are localized. Previous studies comparing these two processes have faced challenges in simultaneously controlling stimulus properties, including low-level visual features and high-level phonological and semantic attributes, as well as task demands. Here, we addressed these issues using Chinese pictographs, visually identical stimuli that can be interpreted either as words (lexical symbols) or as objects (visual depictions) and that were rigorously matched in a visual form, phonology, and semantics. During functional magnetic resonance imaging, 36 male and female human participants performed three language tasks (realness judgment, sound retrieval, and meaning judgment) on pictographs that were contextually recognized as words or objects, with each task applied to both recognition types under identical procedures. Results revealed robust word-object differences in the inferior parietal lobule (IPL), anterior cingulate cortex (ACC), and their associated networks. Compared with object recognition, word recognition elicited stronger activation in the IPL and reduced deactivation in the ACC. Furthermore, both regions exhibited distinct multivoxel activation patterns between the word and object recognition and showed stronger functional connectivity with other brain regions specifically during word recognition. This study provides well-controlled evidence for intrinsic neural dissociations between word and object recognition, highlighting a parietal-cingulate network as a core substrate differentiating these processes.