| |
Abstract:
Abstract: Category-specific semantic deficits suggest
specialization in the representation or processing of living things
and man-made objects. Although many patients' impairments can be
understood in terms of stimulus confounds (e.g., Funnell, 1992),
other patients appear to have true, amodal semantic deficits
preferentially affecting one domain of knowledge. Proposed
explanations include neuroanatomical specialization for either
categories (e.g., animals/fruit; Caramazza & Shelton, 1998) or
types of information (e.g., perceptual/functional; Warrington &
Shallice, 1984). An alternate proposal suggests that these deficits
arise from damage interacting with the structure of conceptual
knowledge and thus argues there is no need to postulate
neuroanatomical specialization (Moss, Tyler, et al., 1998). Recent
neuroimaging studies with healthy adults have investigated this
question directly by looking for regional differences in brain
activity when processing natural kinds and artifacts. Although some
differences have been identified, the findings are rarely
consistent across studies and are thus difficult to interpret. We
present two PET experiments, using lexical decision and semantic
categorization tasks, and an fMRI replication of the semantic
categorization, which revealed no regional specialization for
domains of knowledge but instead implicated a single, distributed
lexico-semantic system. We claim that these results, as well as
those from previous neuroimaging studies, are consistent with the
conceptual structure hypothesis in which a single, unitary semantic
system gives rise to preferential impairments of categories or
domains as damage interacts with the functional organization of
semantic memory.
|
