Figure 2. Perceived jump direction in Experiment 1 (error bars represent bootstrapped 95%-confidence intervals of the model estimates).

Negative values represent backward jumps, positive values forward jumps (N = 12). In case of the Spatiotopic and Retinotopic trials, inducer duration refers to postsaccadic inducer duration (see Fig. 1 and Table 1). We observed a strong High Phi illusion with Full match trials (dark blue bar at 1066.7 ms of inducer duration). Moreover, the High Phi illusion could be stored retinotopically across saccades, similar to motion after effects (brown bar at 33.3 ms of inducer duration). The retinotopic effect faded out at 1066.7 ms of inducer duration, because after saccade execution the Inducer was no longer retinotopically matched with the Transient (see Fig. 1 for description of the trial sequence). Interestingly, the accumulated motion information can also be updated spatiotopically (orange bar at 33.3 ms of inducer duration). The spatiotopic effect was larger than what was observed in Full match trials (dark blue bar at 33.3 ms of inducer duration) and could not be explained as a long range or decisional bias, since the effect was larger than in the Long Range condition (light blue bar at 33.3 ms of inducer duration).