An analysis of the time course of attention in preview search

Abstract

We used a probe dot procedure to examine the time course of attention in preview search (Watson & Humphreys, 1997). Participants searched for an outline red vertical bar among other new red horizontal bars and old green vertical bars, superimposed on a blue background grid. Following the reaction time response for search, the participants had to decide whether a probe dot had briefly been presented. Previews appeared for 1,000 msec and were immediately followed by search displays. In Experiment 1, we demonstrated a standard preview benefit relative to a conjunction search baseline. In Experiment 2, search was combined with the probe task. Probes were more difficult to detect when they were presented 1,200 msec, relative to 800 msec, after the preview, but at both intervals detection of probes at the locations of old distractors was harder than detection on new distractors or at neutral locations. Experiment 3A demonstrated that there was no difference in the detection of probes at old, neutral, and new locations when probe detection was the primary task and there was also no difference when all of the shapes appeared simultaneously in conjunction search (Experiment 3B). In a final experiment (Experiment 4), we demonstrated that detection on old items was facilitated (relative to neutral locations and probes at the locations of new distractors) when the probes appeared 200 msec after previews, whereas there was worse detection on old items when the probes followed 800 msec after previews. We discuss the results in terms of visual marking and attention capture processes in visual search.

Bridging the gap between physiology and behavior: Evidence from the sSoTS model of human visual attention.

“…A preview presented for 350 ms prior to a search display, without any imposed secondary task, is suboptimal for search (Humphreys, Jung Stalmann, & Olivers, 2004;Watson & Humphreys, 1997). Presenting a preview for this duration and then withdrawing attention through a secondary task, while leaving the Figure 28.…”

“…Note that the parameters were not changed further, and so they did not vary across the exposure durations. Human data reported originally by Humphreys, Jung Stalmann, and Olivers (2004) are shown in Figure 12. Figure 13 presents the results for sSoTS.…”

“…There is also evidence for effects of old items appearing to have reduced contrast compared with new stimuli (Allen & Humphreys, 2007). Importantly, this evidence for suppression of old distractors occurs only under conditions of attentional prioritization to new stimuli; there is little difference in the detection of probes on new and old stimuli when probe detection is the primary task (Humphreys, Jung Stalmann, & Olivers, 2004;Watson & Humphreys, 2000). Watson, Humphreys, and Olivers (2003) proposed that participants actively inhibit the old items to enhance any competitive advantage for new stimuli when they appear.…”

“…Human visual search for a conjunction target can be made more efficient if the temporal exposure of the distractors is staggered so that half the distractors appear as a preview before the other half (plus the target, when present; Watson & Humphreys, 1997). Interestingly, this preview benefit for search has a long time course-search is optimal only when there is an interval of 500 ms or so between the preview and the new search display (Humphreys, Jung Stalmann, & Olivers, 2004;Humphreys, Olivers, & Braithwaite, 2006;Watson & Humphreys, 1997). Given that observers can temporally segment and group stimuli appearing within 100 ms of each other (Yantis & Gibson, 1994), it is not clear why attentional benefits have such a long time lag.…”

“…However, in contrast to the earlier study, we introduced a top-up preview of 350 ms prior to the secondary task starting. This duration preview typically does not optimize search performance (Humphreys, Jung Stalmann, & Olivers, 2004). We assessed whether, even though insufficient in its own right, the top-up preview "pushed" processing of a prolonged preview into its adaptation state, to generate maximum efficiency in the preview task.…”