| |
Abstract:
The identification of hemispheric dominance for language is a
crucial aspect of presurgical planning of tumor resection.
Functional Magnetic Resonance Imaging (fMRI) promises a viable
"non-invasive" strategy for assessing hemispheric dominance. Using
fMRI, the efficacy of 6 tasks with increasing linguistic complexity
was tested for their ability to produce hemispheric laterality.
Using a 1.5T GE scanner, fMRI activation maps were acquired from 17
right-handed subjects (6 female). The experimental paradigm was an
alternating boxcar of 20s stimulus/rest cycles and activation was
defined as 5 contiguous activated pixels correlated to the boxcar
function, r = .40. A laterality index (LI=L-R/L+R) was computed
from the sum pixel activation for hemispheres (HLI), temporal
(TLI), and frontal lobes (FLI). Tasks included one passive
listening, two attending (pitch and vowel discrimination), and 3
covert word generation tasks: word repetition,
Letter/noun-generation (LNG), and word/verb-generation (WVG). All
observed HLIs fell between .02 and .44 and increased with
linguistic complexity. Word generation tasks produced stronger HLIs
(.23-.44) with WVG producing the strongest HLI. In general, tasks
which increased left frontal activation led to stronger HLI
(corr.=.79). This was particularly evident in WVG, FLI=.72,
TLI=.18. FMRI is a sensitive indicator of hemispheric
lateralization of language with the ability to assess different
levels of linguistic processing. This data provides evidence that
semantically salient tasks produce the strongest HLI.
|
