The role of perceptual cues in the trapezoid oscillatory illusion

The present experiment was designed to test the hypothesis that binocular disparity functions as a veridical cue, and linear perspective as a nonveridical cue in the perception of a rotating trapezoid. Three trapezoids with proportionately increasing slope were viewed by 10 subjects through viewing screens designed to control the availability of binocular disparity. Monocular and binocular viewing conditions were employed. The dependent variable, proportion of perceived oscillation, was analyzed by a factorial analysis of variance and appropriate post hoc tests. The obtained data indicate that increasing the availability of binocular disparity by increasing viewing screen width favors veridical rotation perception. When the binocular depth cue was attenuated by decreasing viewing screen width, the observer relied on linear perspective, and non-veridical oscillation perception was favored. As linear perspective was increased by increasing trapezoid slope, perceived oscillation increased both monocularly and binocularly. However, for any level of linear perspective, monocular viewing invariabily produced more oscillation than binocular viewing, indicating that monocular depth cues (movement parallax and interposition) were not as effective at neutralizing the illusory perspective effect as the binocular disparity cue.