Semantic representations in the visual cortex of blind and sighted humans

Abstract

In blind humans, the "visual" cortex responds to linguistic stimuli, such as words and sentences. This is sometimes taken as evidence that this brain region supports starkly different computations in blind and sighted individuals. Here, we challenge this view and show that, during word processing, the visual areas in these two populations represent the same semantic dimension - the knowledge about physical properties of word referents. Using analysis of fMRI activation patterns, we found that the visual cortex in both congenitally blind and sighted participants represented differences between individual words. In both groups, the activation patterns for words in the visual cortex reflected physical, but not conceptual similarity between word referents. Furthermore, the between-group correlations in these activation patterns were comparable to within-group correlations. Finally, during word processing, the visual areas in both groups showed greatest "representational connectivity" to the occipitotemporal areas. Overall, our findings suggest that responses to linguistic stimuli in the visual cortex of blind individuals are driven by representational mechanisms that are functional also in the sighted adult brain. In sighted individuals, information about physical properties of word referents might be backprojected to visual areas, from the occipitotemporal cortex, to support visual predictions, imagery, and visuospatial thinking. In blind individuals, this mechanism might be preserved and, combined with increased excitability of the blind visual cortex, drive strong responses of this region to linguistic stimuli.