Searching for Illusory Motion

In a series of four experiments, standard visual search was used to explore whether the onset of illusory motion preattentively guides vision in the same way that the onset of real-motion is known to do. Participants searched for target stimuli based on Akiyoshi Kitaoka’s classic illusions, configured so that they either did or did not give the subjective impression of illusory motion. Distractor items always contained the same elements as target items, but did not convey a sense of illusory motion. When target items contained illusory motion, they popped-out, with flat search slopes that were independent of set size. Search for control items without illusory motion -- but with identical structural differences to distractors -- was slow and serial in nature (> 200 ms/item). Using a nulling task, we estimated the speed of illusory rotation in our displays to be approximately 2°/s. Direct comparison of illusory and real-motion targets moving with matched velocity showed that illusory motion targets were detected more quickly. Blurred target items, that conveyed a weak subjective impression of illusory motion, gave rise to serial, but faster (< 100 ms/item) search than control items. Our behavioral findings of parallel detection across the visual field, together with previous imaging and neurophysiological studies, suggests that relatively early cortical areas play a causal role in the perception of illusory motion. Furthermore, we hope to reemphasize the way in which visual search can be used as a flexible, objective measure of illusion strength.