| |
Abstract:
We investigated the neural correlates of phonological and
orthographic priming using event-related potentials (ERPs). Twenty
subjects performed a lexical- decision priming task that employed
five conditions: homophone (phonological), anagram (orthographic),
pseudo-homophone and pseudo-anagram priming, and a prime-letter
string baseline condition. Primes were always a real word. For each
of the priming conditions there were unrelated and related
prime-target pairs. Choice reaction times (RT) were made to the
target according to whether it was a real word. ERPs were recorded
from 60 scalp sites referenced to the average of the mastoid
electrodes (0.01-100Hz filter; 10,000x amplification; 250Hz
acquisition rate). Behavioural data confirmed a significant RT
advantage for the homophone and pseudo-homophone related pairs, but
not for the anagram and pseudo-anagram related pairs. The ERP
results showed that the phonological- priming effect consisted of a
large midline posterior negative component (240- 600ms, i.e. N400
potential) concurrent with a left frontal positivity. The
pseudo-homophone-priming effect consisted only of attenuation of
the N400 potential. In contrast, anagram and pseudo-anagram target
processing consisted of an early midline posterior positive effect
(140-180ms), followed by a significant right frontal negativity
concurrent with a left posterior positivity (260-360ms). The ERP
findings complemented the behavioural data and revealed differences
between real and pseudo homophone targets. The high temporal
resolution of the ERP method unveiled the differential time-course
and distribution of the effect of phonological and orthographic
processing in the human brain.
|
