Visual persistence of figures defined by relative motion

In order to measure visual persistence of figures that were solely defined by relative motion (motion-defined figures or motion figures), random-dot kinematograms were used to form stimulus figures in the two-frame, missing element task introduced by Di Lollo, V. (1977 Nature, 257, 241-243). Experiment 1 showed that motion-defined figures persisted for about 130 msec after the termination of the stimulus presentation (i.e. after the dots stopped moving). This was similar to but several tens of milliseconds longer than the visual persistence of figures which were defined by a luminance difference (luminance-defined figures or luminance figures) in the same random-dot pattern. Since motion detectors are not found in the retina or lateral geniculate in primates, our results strongly suggest that visual persistence is not only a retinal phenomenon but also a cortical one. Experiment 2 investigated the possible influence of motion aftereffects on the visual persistence of motion figures. The results showed that coherent movement of the dots over the whole display after the stimulus offset did not reduce the visual persistence of motion figures, suggesting that the source of this persistence is not a motion aftereffect. In Experiment 3, visual persistence for the motion-defined figures was shown to be longer than that for luminance-defined figures independently of the contrast of the stimulus figure as long as the stimuli could be seen clearly enough. This suggests that different mechanisms are involved in the visual persistence of motion-defined and luminance-defined figures.