Demos and illusions…
1. Motion integration
Below are movies of stimuli used in psychological experiments and analyzed in scientific articles (see references when appropriate or go o the scientific publications pages). These movies bare on the integration of moving contours into a single coherently moving shape. The scientific issue at stake is to understand when and how apparently independent moving contours (possibly belonging to a single shape) are merged (integrated) into a single « global » shape moving along a circular trajectory. As an introduction to the underlying problematic, refer to Lorenceau & Shiffrar (1992) and to the references cited in this article.
(For a better rendering Right click on videos and enable Video Loop mode)
This movie shows a moving diamond shape whose corners are masked by visible or invisible rectangular occluders. As it can be seen, the perception switches from a fully coherent diamond moving as a whole behind visible occluders along a circular path, to an incoherent motion of its contours when the occluders have the same luminance as the background, despite the motion remaining the same. This can be checked by looking to this movie in peripheral vision (look HERE!). In this case, the circular motion of the diamond is perceived even when the occluders are invisible (See Lorenceau & Shiffrar, 1992, Vision Research, for more information).
In this movie, a Diamond shape is smoothly changed into a Chevron shape. The corners are always hiden by grey occluding rectangles. One should see the Diamond as a rigid shape moving along a circular trajectories, but when the shape is more ‘Chevron like’, the percept shifts, and independent segments are seen to move on their own (along the vertical). This shows that even when the rectangular occluders are visible, the shape influences the capability to group moving contours into a single coherently moving object (here along a cricular trajectory). Global motion is seen with a diamond but not with a chevron shape indicating that Form and Motion processing in the brain must be interacting (see Lorenceau and Alais, Nature Neuroscience, 2001 for more).
Along the same lines, in this movie, each moving « contour » is made of a set of aligned dots that altogether define a diamond shape. The « contour dots » are oscillating along diagonals while their « local » motion is altered by fast « motion noise » whose amplitude varies over time (large to small amplitude). As it can be seen, the percept alternates between a coherent global motion along an anticlockwise path (for a large amplitude of the « motion noise ») and the independent motion of the contours (for a small amplitude of the « motion noise »). With a small noise amplitude, one may feel that the motion is in the reverse -clockwise_ direction (see Lorenceau 1996, Vision Research for a discussion of this phenomenon). As with the first Diamond movie above, the percept is always that of a coherent global motion when the stimulus is seen in the visual periphery.
Here are two examples showing how « virtual » shapes and contrast modulate motion integration. (See Lorenceau & Zago, 1999, for more)
In this example, on can see that grouping motion components into a single shape strongly depends on the presence of the occluding orange diamonds and on the the spatial organization of the segments. The visual system prefers closed well « behaved shapes » -a small diamond embedded in a large one-, although this might not correspond the veridical stimulus… So, the percepts of expansion or translation seen in these 4 patterns flip when the occluders are absent.