Past Lab Research Archive
The accurate perception of the position and orientation of an object in space is the prerequisite for handling it in a purposeful manner. To determine the position of an object, we can make use of several depth cues, such as linear disparity, occlusion, stereopsis. What we do not know exactly is how the brain "sorts out" what cues are important or unimportant in a particular case. One method to study the effects of specific visual cues on action is to examine how people point to, reach for or grasp objects that are part of an illusory display.
For some time we examined the effect of the Ames trapezoidal window illusion on slant perception and aiming movements. We selected the static trapezoidal window illusion because of its compelling use of linear perspective and shading and its proven effect on perception. We test how observers judge the orientation of the window or how accurate they aim with both index fingers towards the sides of the illusionary display. Under monocular viewing, the perspective cues of the display created the illusion of a window slanted in depth and strongly affected slant estimation in both perception and action. Binocular information was not sufficient to recover the veridical orientation of the illusory trapezoid. However, in a control task binocular viewing recovered the true orientation of a non-illusory rectangular window display. Both results indicate that even during binocular vision pictorial cues were not fully ignored for the slant estimation of the trapezoid. Under binocular viewing, the visuomotor system almost completely compensated for changes in orientation. In contrast, rotations of the display were only partially compensated when perceptual judgments were required. Without binocular information, visuomotor error was similar to the perceptual error. We argue that observers used different object features to assess window orientation in the two tasks. While the perceptual judgment of slant relied substantially on perspective cues, determining the targets for bimanual aiming relied more on binocular information. Our data indicate that the basal ganglia as well as the cerebellum are involved in the processing of depth cues, since patients with a dysfunction of these brain structures show an impaired performance in judging depth.
Motor response to the visual illusion
Pointing towards the edges of the window illusion during monocular vision induces a large motor error
Trapezoid window visual illusion from above
View a larger picture of a trapezoidal window display.
To learn more about depth cues visit:
Vision and Art: Use of Visual Information
In cooperation with Albert Yonas, Ph.D., Institute of Child Development, University of Minnesota
In the last two decades neurophysiological and neuropsychological studies provided support for the notion that object attributes are encoded in two separate visual pathways in primate brains.
The ventral visual stream from the primary visual to the inferotemporal area is said to be involved in visual perception. It is concerned with extracting object features from its surrounding and generating an object centered (allocentric) representation of the object.
In contrast, the dorsal visual stream to the parietal areas is said to encode viewer centered (egocentric) object representations. It is involved in extracting object features that are necessary for visuomotor behavior, e.g. for scaling the aperture of the hand according to object size.
One way to study the function of these two visual pathways is to use illusionary displays like the Ebbinghaus illusion.
Researchers found that adults regularly misjudge the physical size of the centre disc, yet scale their hand aperture according to its actual size. That is, the two streams operate somewhat separately in adults.
Visual stream pathways
In this study we examined the ontogenetic development of this dissociation between perception and action in 35 children between the ages of 5-12 years. Our results indicate that grasping behavior in children is subject to an interaction between ventral and dorsal processes. Both pathways seem not to be functionally segregated in early and middle childhood. Our data demonstrate that children were relying on both visual processing streams during perceptual as well as visuomotor tasks. We found that children used egocentric cues to make perceptual judgments, while their grasping gestures were not exclusively shaped by viewer centered but also by object centered information.
In cooperation with the Department of Psychology, Universität Düsseldorf, Germany