Effect of motion smoothness on the flash-lag illusion

Two flash-lag experiments were performed in which the moving object was flashed in a succession of locations creating apparent motion and the inter-stimulus distance (ISD) between those locations was varied. In the first (n=10), the size of the flash-lag illusion was a declining non-linear function of the ISD and the largest reduction in its magnitude corresponded closely to the value where observers judged the continuity of optimal apparent motion to be lost. In the second (n=11) with large ISDs, we found the largest illusions when the flash initiated the movement, and no effect was observed when the flash terminated the movement. The data support motion position biasing or temporal integration accounts of the illusion with processing predominantly based on motion after the flash.