Quantitative prediction of errors in RSVP: Modeling the time course of suppression during the attentional blink

Vul, Konkle, Love, Williams, & Niewuwenstein

Selection of a target in a rapid serial visual presentation (RSVP) stream hinders processing of subsequent items--an effect referred to as the attentional blink (AB). Despite theories to the contrary, recent evidence suggests that the degree to which an item is blinked is a function of time elapsed between targets, rather than the serial position of the item. Wyble and Bowman [2004, in Proceedings of the Sixth International Conference on Cognitive Modelling Ed. C Schunn (Mahwa, NJ: Lawrence Erlbaum Associates)] reported lag-2 sparing in RSVP at 20 items s-1. Furthermore, precuing targets during the attentional blink can substantially improve performance, suggesting that by triggering attention earlier, one can mitigate the AB (Nieuwenstein et al, 2005 Journal of Experimental Psychology: Human Perception and Performance 31 1463 - 1475). We propose a quantitative temporal model of the second stage of processing posited by Chun and Potter (1995 Journal of Experimental Psychology: Human Perception and Performance 21 109 - 127) which accounts for both findings. This model is independently verified with a novel paradigm in which cues and targets are spatially and temporally separated. This method, coupled with analyses of intrusions around the second target, affords a precise measure of the speed of attentional selection at different intervals from the preceding target. Further, this paradigm allows us to manipulate cue onset independently of target onset; thus dissociating the two stages of processing. Our results are best accounted for by a mechanisms of continuous suppression rather than a processing bottleneck.

Vul E, Konkle T, Love A, Williams M A, Nieuwenstein M, 2006, "Quantitative prediction of errors in RSVP: Modeling the time course of suppression during the attentional blink" Perception 35 ECVP Abstract Supplement.