An abundant and varied set of studies has established that attention can be directed to a particular region of space, such that visual input at an attended location is preferentially processed over input at unattended locations (space-based attention or SBA), and this processing bias is manifested in visual cortex. More recently, it has been explored how SBA modulation of sensory signals is controlled via higher-order cortical networks. Based on early patient observations that have since been refined by contemporary imaging studies, topographic subunits within fronto-parietal cortex have been implicated as a source of control. This control network is best described as a gradient of attention across space, wherein the two hemispheres operate in concert by generating attentional weights in favor of the contralateral visual field (interhemispheric competition account). When a relevant item appears in one visual field, the weighting sum is biased in the corresponding direction. We hypothesized that the magnitude of the spatial bias should dictate not only the attended hemifield, but also the eccentricity at which attention is focused. Using high-resolution fMRI, we estimated interhemispheric competition (via a contralateral bias index) for each topographic subunit of fronto-parietal cortex to determine how the control network signals fluctuations in stimulus position. Subjects either attended to a single flickering grating appearing at one of four eccentricities from fixation (2°, 5°, 8°, or 12°), or to an RSVP letter stream at fixation. As expected, SBA effects were observed in all visual (V3v-V7), parietal (IPS0-5, SPL1) and frontal (FEF, PreCC) areas. The contralateral bias showed no systematic patterns across eccentricities in visual or frontal cortex. However, the bias tended to increase with eccentricity in parietal regions. This is consistent with previous hypotheses that frontal cortex is primarily involved in general, goal-directed components of attention, whereas parietal cortex is primarily involved in stimulus-driven components.
Meeting abstract presented at VSS 2013