Manual-aiming bias and the Müller-Lyer illusion: the roles of position and extent information.Exp Brain Res. 2005 May; 163(1):100-8.EB
Several studies have shown that rapid manual aiming movements from one to the other wing of Müller-Lyer illusion figures are biased in a manner consistent with their influence on perceptual judgments. Two experiments examined the role of extent and position information in Müller-Lyer figures in biasing pointing movements. Experiment 1 compared the effect of starting the action either close to the non-target wing or from a position well outside the conventional wings-out (>-<) and wings-in (<->) Müller-Lyer figures. Pointing movements were longer to the target wing of the wings-out than of the wings-in figure but only for actions starting from a position adjacent to the non-target wing. Experiment 2 found no effect of the direction of the target wing on pointing movements between the outer wings of the combined Müller-Lyer illusion figure (<->-<). These findings suggest that perceived extent (of the shaft) and perceived position (of the wings) in Müller-Lyer figures are largely independent of each other, and are consistent with the claim that the illusion is more likely to influence pointing actions if the action is based on extent information than on vertex position information.