Abstract
Images depict specific objects (e.g., a specific dog), yet are named with categorical labels (e.g., “dog”). We examined how semantic representations activated by images may be influenced by implicit labelling. Participants saw images of familiar objects and generated words associated with each image while undergoing transcranial direct current stimulation over the posterior superior temporal gyrus. Additional participants judged how representative generated associates were of the picture category and guessed the category based on the associates. Anodal stimulation was predicted to up-regulate labelling and thereby increase the extent to which participants produced associate that were more representative of the pictured category. Associates generated by anodally stimulated subjects were found to be more representative and enabled more accurate guessing of the category from which they were generated. The general pattern of results was replicated in a follow-up study using words rather than picture cues. Together these results suggest labelling may help stabilise semantic representations, leading to more robust representation of category-relevant information.
