Human see, human do? Viewing tool pictures evokes action-specific activity in visual hand-selective occipitotemporal cortex

Every day we have countless experiences with different types of objects. Amongst all the objects we encounter, tools are unique because they are tightly linked to predictable actions. Neuroimaging studies have consistently reported selective visual responses in occipitotemporal and parietal cortices for viewing pictures of tools or hands. The specificity of these responses, however, is unknown. Does viewing a tool picture evoke a response specific to the action associated with that tool? Here, we used fMRI and multivoxel pattern analysis to ask whether viewing tool pictures associated with specific actions can be decoded from these regions, and if so, are the patterns of neural activity elicited by viewing pictures and performing the corresponding actions similar? Participants (N=18) viewed tool pictures (while performing an orthogonal task) and, in separate runs, pantomimed tool use actions (without tool in hand) in response to tool names. Our stimulus set consisted of familiar tools (e.g., screwdriver or tongs) that differed in the tool use action (rotate or squeeze) but were carefully matched for low-level differences (e.g., grip, colour, orientation). We found that merely viewing tool pictures triggered automatic retrieval of action information associated with the depicted tools, within both ventral occipitotemporal and parietal areas, and even primary somatosensory cortex. However, the patterns of activity elicited by tool pictures and corresponding pantomime tool use actions were unrelated, thus suggesting a clear neural distinction between perceived tool action affordances and pantomimed tool use actions. Interestingly, visual hand-selective occipitotemporal cortex was the only region where tool use action decoding was higher than tool identity decoding for both viewing and pantomiming, suggesting that its response cannot be simply explained by shape differences. These data show that visual hand-selective cortex represents action-specific information associated with tool use, and that these representations are triggered by simply viewing pictures.