| |
Abstract:
This study investigated the functional neuroanatomy used in
hearing words. The purpose was to validate the method of functional
magnetic resonance imaging (fMRI) in auditory word processing by
replicating a PET study (Howard et al., 1992) that localized the
lexicon for spoken word recognition to an area in the posterior
superior temporal gyrus (STG). Unlike many previous fMRI studies of
language, subjects were asked to talk as part of the task. Eight
native speakers of English were presented with two conditions (word
repetition and reverse speech). In the word repetition task,
subjects repeated single words aloud with instructions not to move
their jaws. In the reverse speech condition, they heard a digitally
reversed speech token and said the word "crime." To minimize
artifacts from talking in the scanner, head movement was restrained
and movement correction algorithms were applied. Two subjects were
eliminated due to excessive movement. Cross-correlation
thresholding was used to analyze the two task conditions. To
compare the fMRI study to the PET results, the images were
transformed into standard stereotactic space, averaged across
subjects and smoothed with a Gaussian filter. The fMRI results
showed activation in the same STG region as in the PET study. The
individual subject data revealed variation in the location of the
activation along the STG. The results provide evidence for
traditional neuropsychological models of auditory lexical
processing and demonstrate that it is possible to use tasks that
involve overt speech with fMRI. Supported by NIH DC
R01-3378-01.
|
