| |
Abstract:
Based on lesions studies distinct neural cell assemblies have
been postulated for the representation of nouns and verbs. The
present study intended to uncover distinct cortical sources of the
N400 effect dependent on wordclass. In a semantic priming
experiment 16 subjects performed a lexical decision task which
comprised verb-verb and noun-noun word pairs. We recorded reaction
time, error rate and event-related potentials (128-channel, high
resolution ERP). The behavioral and the average ERP data followed
the known picture, in that strongly or moderately primed target
words produced faster reaction times, less errors, and a smaller
N400 amplitude than unrelated word pairs. To identify cortical
source densities combined structural MRI and ERP information was
analyzed with algorithms of the CURRY+ package. The localization
procedure revealed no differences between nouns and verbs when the
grand average ERP data were analyzed. In contrast, fitting current
source models to single subject ERPs uncovered distinguishable
distributions for verbs and nouns. However, these patterns did not
follow a common, intraindividually consistent schema. This suggests
that semantic integration processes may be performed by cell
assemblies whose location is, at least in part, individually
specific.
|
