Abstract
The study of brain-behavior relations continues to be plagued by controversy regarding the extent to which localized and circumscribed brain regions are content-selective. On the one hand, much neuropsychological and neuroimaging data have argued for a narrow localization of cortical regions and have suggested that domain-specificity is a natural distinction in cortex. On the other hand, there is strong empirical support for the claim that the psychological and neural mechanisms are more general-purpose and that differences in processing certain stimulus classes emerge from differential experience and/or processing demands. To assess the selectivity of two brain regions, one responsive to faces (fusiform face area; FFA) and the other responsive to words (visual word form area; VWFA), we examined the visual performance of adults with profound deficits in either face (prosopagnosia) or word (pure alexia) recognition following lesions to the critical cortical area. Both populations completed several experiments designed to elucidate their ability to do perceptual discriminations between morphed faces and to match faces across vantage point. They also named and performed lexical decisions on words and letter strings. Although each group was maximally impaired in recognition of its own ‘primary’ stimulus class (prosopagnosics on faces and pure alexics on alphanumeric stimuli), both groups also showed a significant decrement, relative to controls, in the other domain too, although the deficit was not as severe as in the primary domain. Thus, a lesion to each area produces a graded rather than a binary outcome, indicating that neither face nor word regions are entirely content-selective. Rather, the data are compatible with an account in which cortical organization is graded and distributed, with words primarily, but not exclusively, represented in the left hemisphere (presumably to constrain proximity to language-related areas) and faces primarily, but not exclusively, represented in the right hemisphere.
This work is supported by a grant from NIH (MH54246) to MB.