There are 18 topographic subunits within fronto-parietal cortex that are known to direct the locus of top-down attention by generating weights in favor of the contralateral visual field. However, it remains unclear how each region individually contributes to the complex process of selection. In a previous neuroimaging experiment, we investigated how the attentional control network signals fluctuations in stimulus position within a visual field. These results supported a functional divergence between subunits: FEF and aIPS are likely involved in integrating information across a hemifield, whereas pIPS is likely involved in fine spatial selection within a hemifield. While the last experiment manipulated stimulus parameters as behavioral goals were fixed, the current experiment is designed to do the opposite. Here we investigate how the control network signals parametric fluctuations in the behavioral relevance of a stimulus by manipulating the validity of a spatial pre-cue. During each stimulus block, gratings appeared at a fixed position on both sides of fixation. A pre-cue indicated whether targets would appear within the left or right stimulus with a pre-specified probability level (50%, 75% or 100% valid), giving five levels of probabilities for each location (0%, 25%, 50%, 75% and 100%). As expected, a contralateral bias was observed across the network when the pre-cue was 100% valid. However, control subunits were differentially modulated by cue validity: while activation within pIPS monotonically increased across all probabilities, aIPS, FEF, and preCC exhibited peaks of activation when the pre-cue was 75% valid. These results suggest that pIPS signals the amount of attention directed to a location, whereas aIPS and FEF are likely involved in attentional disengagement from a high-probable cued location to a low-probable uncued location. This study contributes to the growing body of research which suggests that individuated attentional control subunits are involved in separable aspects of top-down selection.
Meeting abstract presented at VSS 2014