| |
Abstract:
The purpose of this study was to develop a reliable protocol
for the objective spatiotemporal characterization of language
processing. Magnetoencephalographic data were collected from three
subjects performing two language tasks. In the lexical decision
task, four-letter words and consonant strings were randomly
intermixed and presented at fixation. In the object-naming task,
randomly mixed simple line drawings of familiar objects or the same
objects distorted beyond recognition were presented. In both tasks,
participants pressed one of two buttons depending on whether the
stimulus was a word/nonword or a recognizable/unrecognizable
object, respectively. Participants were encouraged to covertly
articulate words and recognizable objects. For all subjects in both
conditions, multiple visual sources were evident in the interval
between 80-150ms post-stimulus. Temporoparietal sources near
Wernicke's area and left-lateralized sources near Broca's area
(inferior frontal cortex) were found. An example onset time for
these sources was ~150ms although, there was variability in the
existence of these sources across subjects within conditions and
within subjects across conditions. These data will be compared with
functional Magnetic Resonance Imaging data from similar tasks.
Despite variability of the results across subjects, there appears
to be consistency with classical theories of language organization
and these results demonstrate the utility of multidipole
spatiotemporal modeling of magnetoencephalo- graphic language
data.
|
