| |
Abstract:
Abstract: We investigated the neural correlates of automatic
and controlled semantic priming using event-related potentials
(ERPs). Twenty-two subjects performed two lexical-decision semantic
priming tasks that employed either a 200ms or 1000ms SOA (stimulus
onset asynchrony: period of time from the appearance of prime to
the target presentation). Priming at the 200ms SOA is believed to
be due to automatic processes only, and at the 1000ms SOA due to
further controlled processing that enable cognitive expectancies or
strategies. There were three types of prime-target pairs,
semantically-related, semantically-unrelated and prime-pseudo-word.
Choice reaction times (RT) were made to the target according to
whether it was a real word. ERPs were recorded from 54 scalp sites
referenced to the average of the mastoid electrodes (0.01-100 Hz
filter; 10,000x amplification; 250 Hz acquisition rate).
Behavioural data confirmed a significant RT advantage for related
prime-target pairs. The ERP results showed that the
semantic-priming effect started significantly earlier, and with
greater mean amplitude, during the controlled compared to the
automatic task condition (300 and 360 ms respectively). In both
cases, the semantic-priming effect consisted mainly of the
attenuation of the N400 potential, and had a broad distribution
over the central scalp region. The high temporal resolution of the
ERP method unveiled that cognitive variables may influence the
timing of linguistic processes in the brain.
|
