Abstract
It is trivial to find a target line segment in an array of distractors that are oriented orthogonal to the target, even when the display is presented for a very brief duration. Here we investigated the effect of temporal integration on this pop-out effect. Specifically, we presented observers with alternating line-segment pop-out arrays that changed in luminance contrast (i.e. black to white) with a fixed change in orientation. Participants were asked to identify the location of the pop-out target while freely moving their eyes. Using eyetracking, we measured accuracy, reaction-time and the time to the first fixation of the target as a function of flicker rate (3.125Hz, 4.16Hz, 6.25Hz, 12.5Hz, 25Hz and 50Hz). This corresponds to individual array presentations of 160ms, 120ms, 80ms, 40ms, 20ms, and 10ms. At the highest rates, the displays take on the appearance of flickering Xs. We found that increasing the flicker rate led to a monotonic decrease in accuracy and monotonic increases in reaction time. Similar to manual reaction times, latency to the first target fixation increased with flicker rate, with the first statistically-significant effect occurring between 120ms and 80ms (4.16Hz, 6.25Hz). Moreover, across all flicker rates (more so at higher rates), observers occasionally fixated the target without detecting it and instead continued their search. On some of these ‘attentive blank stare’ trials, the observer eventually did detect the target, demonstrating that the target was in fact detectable, whereas on others of these these trials, their search timed out in vain. Conclusions: The results provide evidence for a temporal integration window of ~100ms during which the presentation of multiple stimuli can interfere with attentional pop-out. In addition, we find evidence that goal-directed overt attention is not necessarily sufficient to allow the detection of a detectable fixated target.
Acknowledgement: NSF 1632738 and NSF 1632849