The boundary flow cue is present when the motion of a surface and its edge are equal in both velocity and direction. This surface is then seen as an object. The other surface in contrast would not share common motion with the boundary line, given they do not belong to the same surface. This surface is referred to as the occluding surface. Thus, the relation between the texture of the surfaces and the motion of the boundary line formed by the edge between them determines which surface will be seen as an object and which surface will be seen as occluding. The videos below demonstrate this effect. The only change between the two displays is the relative motion of the boundary line.
We are currently in the process of testing this with 4-month-old infants. A prior study done with 3 to 5 year olds yielded significant effects, indicating the toddlers were able to pick out the object when given only the boundary flow cue. To test this in infants, we will use a habituation-dishabituation paradigm in which the babies are first presented with one of the two displays pictured below, and then are presented with both the objects (pictured below), while their looking time is recorded.
Generic Viewpoint Assumption
The “generic viewpoint” theory suggests that when there is a lack of visual information given to us, our brains tend to make assumptions on what were seeing based on our previous interactions with the world around us. In the top picture to the left, it looks as if there is one square closer to you as the viewer, overlapping the square in the background. While it is possible something else is going on, your brain makes this assumption based on your past interactions with your environment. It is possible, of course, that the two shapes are really a square and 'pac-man' perfectly lined up (middle image)to create the top image, but this coincidence is not likely in your everyday environment. It is much more likely that when the shapes are pulled apart, the top image would become two complete squares. Therefore, your brain takes this “generic” or “non-accidental” viewpoint when you are given little information. We want to know at what age infants begin to use this assumption about objects.
Wire Cube Illusion
Display from infant's viewpoint.
We believe that one aspect of perceptual learning and development consists of the creation and use of representations or models. These representations are stimulated or recognized by input from early visual processes. These representations or models, a cube for example, provides top-down information to an earlier stage in vision, and functions like a hypothesis that is tested against input. If the data matches the hypothesis, then we experience a 3-D cube when only 2-D contours of a cube are presented to the retina (see Figure A). Our question is, when do infants begin to form and use such hypotheses?
To explore the development of this process we constructed a wire half cube and presented it with the vertex pointed away from 22 7-month-old infants. We observed the reaching behavior of the infants when they viewed the display with one eye covered and with two eyes. The vertex of the display appears closer than the outer edges when the real display is viewed with one eye, but when viewed with both eyes the actual layout is perceived.
See the video from YouTube:
Inverted Face Illusion
Same as the wire cube but with a face instead.
We found significant results on the Einstein project and hope to continue research come this fall.
Transfer across depth cues
Is there a representation in the developing human brain that allows children to go from one type of depth cue to another, not only as they experience it perceptually, but also as they transfer from one type of depth information to another?
- surface contours
- bulge or dent
Transfer across pictorial depth cues:
It has been well documented that infants are interested in novel stimuli, i.e., things that are new. We showed 4-7 month old infants a series of images from the four above. Shading and surface-contour cues are pictorial information of the depth of an object. We presented matching pair of either surface-contour or shading cue images with either the center ring being concave or convex to the infants for a period of time so that they would become habituated to the shape. Then we presented another pair of mismatching concave-convex images of the opposing pictorial cue, i.e., if habituation began with a pair of shading cue images, the test trial would follow with a pair of surface-contour cue images. Since we knew that infants preferentially look at novel stimuli, we therefore expected infants to spend more time looking at whichever type of test image, either concave or convex, that wasn’t the same type that they had previously seen. The real question was whether this effect would be observed across pictorial depth cues. The results indicated that 4-5 month old infants did not transfer pictorial depth cues, but that 6-7 month old infants preferentially looked at the novel stimulus, even as it appeared through a new pictorial cue.
Horizontal-Vertical Illusory Contours
This is the study about how infants perceive lightness and darkness.
We are interested in whether or not infants respond to a pictorial depth cue called familiar size. At what point do infants’ knowledge of the size of faces begin to influence their responses to faces in terms of depth. We plan to present infants with faces of various sizes and measure looking responses to these faces to see if infants are sensitive to familiar size cues.