| |
Abstract:
Abstract: The N400 component of the event-related potential
has been used to investigate semantic relatedness among several
different types of stimuli: (1) words within a sentential context
(Kutas & Hillyard, 1980), (2) words in a prime-target paradigm
(Holcomb & Neville, (1990), (3) the semantic relatedness of
faces (Barrett & Rugg, 1989), and (4) the relatedness of pairs
of pictures (McPherson & Holcomb, 1999). The present authors
are using the N400 to index relatedness among members of an
artificial category taught during experimental sessions. This
methodology involves teaching arbitrary relations among elements of
various types of stimulus sets using matching-to-sample (MTS)
methodology. Using pre-experimentally meaningless printed words,
Warren and McIlvane (1998) reported that the N400 resulted from
presentation of arbitrarily unrelated but not related word pairs.
The present study used the paradigm to model category formation by
creating several "stimulus equivalence" classes, i.e., categories,
using non-representative picture stimuli. Following MTS training,
stimuli were presented to subjects using a prime-target protocol.
Subjects were required to make a covert relatedness decision. ERPs
reveal an N400 component to target-stimuli that are not related to
prime-stimuli suggesting that the N400 is sensitive to 'mismatch'
in a non-linguistic domain.
|
