Abstract
The extent of our ability to ignore distraction from salient objects has long been the subject of considerable controversy. One possible resolution could be that observers vacillate between periods of time when attention capture is relatively strong and periods when capture is negligible. The present fMRI study was designed to investigate this possibility. Observers were scanned while searching static displays for a target circle among nontarget squares; an irrelevant color singleton distractor appeared on 50% of trials. Results typically reveal a “capture effect,” in which RTs are slower on distractor-present than distractor-absent trials (Theeuwes, 1992, Percept Psychophys). The current analysis sought to identify neural predictors of variability in this capture effect. To this end, a whole-brain data-driven approach was used to determine whether fMRI activity preceding each trial, or pretrial signal, could predict RT (Leber, Turk-Browne, & Chun, 2008, Proc Nat Acad Sci). Results uncovered several brain regions in frontal and parietal cortices in which pretrial signal predicted attention capture effects. Specifically, on trials when pretrial signal was low, capture was substantial; yet, when pretrial signal was high, capture was negligible. This confirms that, within single sessions, individuals do indeed vacillate between periods of susceptibility and resistance to capture. Moreover, additional analysis established a striking link between the capture-predicting regions and occipital early visual areas: as pretrial activity in the frontoparietal regions increased, the trial-evoked visual response to singleton distractors, when compared to non-singleton items, was almost completely attenuated. This result demonstrates frontoparietal control over attentional prioritization, which varies on a moment-to-moment basis. Taken together, the present findings illuminate a visual processing system that is dynamic, and they offer a parsimonious resolution to the contentious debate over attention capture.