Abstract

Traditional theories assume that amodal representations, such as feature lists and semantic networks, represent conceptual knowledge about the world. According to this view, the sensory, motor, and introspective states that arise during perception and action are irrelevant to representing knowledge. Instead the conceptual system lies outside modalit-y-specific systems and operates according to different principles. Increasingly, however, researchers report that modalit-y-specific systems become active during purely conceptual tasks, suggesting that these systems play central roles in representing knowledge (for a review, see Martin, 2001, Handbook of Functional Neuroimaging of Cognition). In particular, researchers report that the visual system becomes active while processing visual properties, and that the motor system becomes active while processing action properties. The present study corroborates and extends these findings. During fMRI, subjects verified whether or not properties could potentially be true of concepts (e.g., BLENDER-loud). Subjects received only linguistic stimuli, and nothing was said about using imagery. Highly related false properties were used on false trials to block word association strategies (e.g., BUFFALOwinged) . To assess the full extent of the modalit-y-specific hypothesis, properties were verified on each of six modalities. Examples include GEMSTONE-glittering (vision), BLENDER-loud (audition), FAUCET-turned (motor), MARBLE-cool (touch), CUCUMBER-bland (taste), and SOAP-perfumed (smell). Neural activity during property verification was compared to a lexical decision baseline. For all six sets of the modalityspecific properties, significant activation was observed in the respective neural system. Finding modality-specific processing acros...

Citations