| |
Abstract:
Recent experiments have examined the initiation and control
of visual spatial attention using event-related fMRI in cued
attention paradigms (Hopfinger et al., 2000; Corbetta et al.,
2000). In order to avoid hemodynamic response overlap, these
studies used long cue-to-target intervals, making it difficult to
relate the findings directly to the large body of behavioral and
ERP work which has used much shorter cue-to-target intervals.
Moreover, these experiments did not use a randomized passive- or
neutral-cue condition to ensure that the putative attentional
control activations were not partly the result of non-specific
arousal effects on sensory processing. Using short cue-to-target
intervals, a passive-cue-event control, and a randomization
approach allowing removal of overlapping responses, we employed
fast-rate event-related fMRI to study cued visual spatial attention
with parameters similar to previous cognitive and ERP designs. Our
preliminary analyses (n=10) indicate that, relative to passive
cues, instructional cues (attend left, attend right) activated
parietal, frontal, premotor, and extrastriate areas, which were
similar to those previously reported. Targets produced additional
extrastriate and SMA activity, while uniquely activating motor
cortex and cerebellum. These results argue against working memory
or sensory processing activations being the source of the
cue-related attention effects previously reported. The results also
demonstrate the feasibility of separating overlapping cue- and
target-related responses in paradigms with short cue-to-target and
intertrial intervals.
|
