| |
Abstract:
We used PET to investigate the 'textbook' description of
speech production based on the areas of Broca and Wernicke. The 8
subjects used propositional speech (PS) and two forms of
non-propositional speech (NPS), counting and reciting rhymes. The
results revealed bilateral or midline regions common to all forms
of speech: the supplementary motor area (SMA), primary sensorimotor
cortex, the thalami, and the cerebellar hemispheres. In addition,
three left lateralised regions were observed: the posterior
supratemporal plane (Tpt cortex), the anterior insula and posterior
Broca's area. Both recitation and PS, but not counting, activated
more anterior Broca's area. The difference in the response of two
adjacent regions in the left inferior frontal gyrus and the
posterior activation confined to Tpt cortex defined functional
sub-regions within 'classic' Broca's and Wernicke's areas. PS alone
activated a distributed network which included the left superior
frontal gyrus, anterior left temporal lobe and angular gyrus.
Comparison with a previous study of verbal fluency showed that
neither NPS nor PS activated the dorsolateral prefrontal cortex, a
region often observed in functional imaging studies of verbal
working memory and problem-solving verbal tasks. Therefore,
executive control of PS depends on rostral prefrontal cortex, in
accord with studies on patients with left anterior cerebral artery
territory infarction. We conclude that therapies designed to
improve 'top-down' processing in aphasic stroke should target
rostral prefrontal cortex.
|
