| |
Abstract:
The neural representation of conceptual knowledge remains a
controversial area. Some researchers argue for specialization as a
function of category, domain, or underlying feature type. However,
others argue for an undifferentiated conceptual system.
Neuroimaging techniques may help us address these differing
accounts. Our aim in this study is to use an event-related,
functional magnetic resonance imaging (erfMRI) design to
investigate whether there are separable neural substrates
associated with living and nonliving domains. Previous neuroimaging
studies have used 'block' designs. The advantages of event-related
designs are that the pseudo-random nature of event presentation
reduces the use of strategic processes and produce studies similar
to existing behavioral designs, which can then be compared. In this
study volunteers were asked to perform a semantic categorization
task in which events were words representing either living or
non-living concepts. Significant areas of activation for semantic
processing were found in a left-lateralised network specifically in
the temporal lobe (inferior temporal gyrus and temporal pole) and
frontal regions (including Broca's area). These are areas, which
have been identified in previous neuroimaging studies of conceptual
knowledge. However, there were no significant differences between
living and non-living concepts. These findings are consistent with
our previous imaging data and with a unitary, distributed system
undifferentiated by domain of knowledge.
|
