Abstract
To maintain visual stability, we use information from both retinal and extra-retinal sources (Currie et al., 2000). One component of retinal information is non-target objects, which can act as "landmarks" to influence target localization across saccades (Deubel, 2004; Deubel et al., 2010). The current study further investigates how people utilize non-target information, specifically whether more non-targets can provide more information, and which reference frame it is based on. In Experiment 1 participants made a saccade to a target object, followed by a blank screen, and judged whether or not the target stayed at the same location after the blank. We varied the number of non-targets present (0, 1, or 2), and whether they appeared in the same locations relative to the target across saccades ("relative" condition), or the same absolute spatial locations ("absolute" condition). Compared to when there were zero non-targets, we found a benefit on response accuracy for the relative condition with two non-targets, but not for any other conditions. Participants also had an overall stronger bias to report that the target was at the same location when non-targets were present. In Experiment 2 participants directly reported the remembered target location with a mouse click. In general, non-targets improved participants' performance in terms of both smaller response error and variance, but performance was equally good in relative and absolute conditions, as well as in one and two non-target conditions. Additionally, subjects' responses were biased toward both the original fixation location (opposite direction of saccade) and the location of the non-target(s), and these two sources of bias seemed to counteract each other when in the opposite directions. Taken together, both experiments showed effects of non-targets on target localization across saccades, but non-target number and spatial location seemed to contribute differently in the two tasks.
Meeting abstract presented at VSS 2016