Abstract
Purpose: To determine if attentional signals in early visual cortex are transient and hence associated with shifting of attention or sustained and hence correlated with maintenance of attention. Background: Previous work in our lab (Ress et al., 2000) has shown that performance of a visual detection task produces large BOLD responses in visual cortical areas V1, V2, and V3, even in the absence of a visual stimulus. In these experiments, the beginning of each trial was signaled by an auditory cue. After a one second delay, a low contrast pattern was shown on half the trials, and no pattern was presented on the remaining trials. Subjects reported whether they thought the target was present or absent. These responses were correlated with task performance (subjects were more likely to respond correctly on a given trial if the BOLD response was large for that trial), indicating that the response is probably a top-down attentional signal. Methods: In the present study, the beginning of each trial was signaled by an auditory tone, followed by a variable, fully randomized delay period (2–16 s). A low contrast pattern was then presented on half the trials, and no pattern was present on the remaining trials. A final auditory tone cued subjects to report whether they thought the target was present or absent. On the target-present trials, the target was always shown one second before the response cue. Thus, for all trials, subjects were required to maintain attention during the entire delay period before generating a behavioral response. Results: The BOLD response in primary visual cortex is not transient, and its time course reflects the sustained attention required to perform the task. Conclusion: Presumably, this sustained delay-period activity reflects attentional signals from higher cortical areas that enhance sensory processing in early visual cortex. Experiments are underway to identify candidate areas in parietal cortex that may be the source of these top-down inputs.