Abstract
Theories of attention rely on specific assumptions regarding the order of events mediating attentional modulation of perceptual processing. Yet, behavioral investigations of the timecourse of attentional operation remain largely absent. The mutations paradigm comprises a modified flanker task that assesses the time window of distractors' processing, with high temporal resolution. We report main patterns consistently observed across five experiments. A central target was flanked by two identical distractors, which were either disruptive or neutral. While the target remained unchanged, distractors mutated once per trial, at a random time. There were three trial types; disruptive distractors that mutated to neutral ones, vice-versa, or neutral distractors that mutated to different neutral distractors (control). These manipulations revealed the time window during which disruptive distractors must appear in order to impair performance. Results revealed four main patterns: (1) Presentations of disruptive distractors within the initial 8 ms were sufficient to delay responses (distractor identification). (2) Disruptive distractors presented later than ~50 ms became effectively ignored (distractor suppression). (3) Enhancements of inter-target discriminability from low (targets M and N) to high (M and O) produced faster RTs, earlier distractor identification, earlier distractor suppression and less distractor interference. These patterns support dual-process claims that initially all stimuli are identified and then distractors are inhibited (e.g., Driver & Tipper, 1989) and are inconsistent with (a) early-selection claims that stimuli identification occurs only after attentional mechanisms modulate perception; (b) late-selection claims that attentional modulation is post-perceptual; (c) load theory claims that distractors are processed only under low load conditions, by surplus resources, after target processing (Lavie, 1995). (4) Priming of recurrent targets' features produced earlier distractor suppression, later distractor identification and less distractor interference, as predicted by contingent capture claims (Folk, Remington, & Johnston, 1992). We suggest that the temporal relationship between bottom-up identification and top-down suppression modulates selection efficiency.
Meeting abstract presented at VSS 2014