Monday, April 28, 2008

Lucy in the Sky

Here is a variation of an illusion that I presented at the Neural Correlate Society's annual illusion contest in 2005. The effect was published in the Journal of Vision in 2005.

[The Demo will not be seen in the RSS feed.]

You are looking at rings made up of diamonds with black and white edges; the diamonds change from bright yellow to dark blue and back again. The button makes the black and white edges appear or disappear. If the edges are present, the diamonds dance back and forth, and the rings appear to undulate. If the edges are removed, the diamonds no longer dance, but the bright-yellow to dark-blue modulation appears to circulate around the rings. (Justin Charles and I titled the effect "Lucy in the Sky" because of the diamonds, and because of the illusion of motion).

Brief Comments:

The diamonds appear to be moving because of the contrast with the edges and the background. As the diamonds shift from bright yellow to dark blue, the minimum level of contrast shifts location. This shift in contrast can be described by simple algebraic models.

The Lucy in the Sky demonstration is directly related to the "phenomenal phenomena" of Gregory and Heard (1983). Here is an interactive demonstration of figure 1 from their paper:

[The Demo will not be seen in the RSS feed.]

A compelling demonstration based on the same principles was presented by Michael Pickard in the 2007 illusion contest.

Notice that when a square becomes white, the motion moves away from the white edge, and when a square becomes black, the motion moves away from the black edge (this applies to the diamonds and edges as well). The motion is therefore not produced by a newly created thick edge; my supposition is that the motion is produced by shifting the minimum contrast between the edge, the square or diamond, and the background (not unlike the motion produced in the illusion presented in the previous post).

Although the contrast that drives this type of illusion may be physically present in the stimulus, the resulting appearance of motion may seem surprising (how can there be an appearance of movement if the diamonds and edges are stationary?). Perhaps the motion seems counterintuitive because we tend to represent static objects (in this case, diamonds and edges) much better than we represent shifting relationships between static objects (the changes in contrast between the diamonds, edges, and background).

Next week: a preview of one of my entries from this year's illusion contest, which will be held in Naples, Florida, on May 11.


Unknown said...

I arrived at your site from a Cognitive Daily RSS feed. Very nice new blog! I love the internet just for this reason- for folks to share their passions with one another and to help build upon the wealth of information on-line. One never knows if your site will spark the imagination and creativity of someone and allow them to go and build a better paint scheme for cars or more visible traffic lights, etc. Thanks for taking the time to share your insight.

My hobby is nature photography and writing about the common wildflowers and insects around my house. You can see my blog at It is also hosted by blogspot.

Thanks again! Very cool site on a cool subject!

Anonymous said...

Fabulous site - hope it goes well. I've already emailed details to colleagues working in the world of vision.

Anonymous said...

I stumbled on a similar illusory motion while tinkering with texture boundaries:

It is just four frames cycled, where the two pairs allow apparent motion in just one direction. There is a luminance difference across the texture boundaries, but no explicit edge. I don't understand it.

A bit more info at .

Anonymous said...

Excellent blog! We create the world we experience--understanding how is not only important but extremely fascinating. Reality--what a construct!

Jay Saul

Jun Okumura said...

Thank you, thank you, thank you, Professor Shapiro. Keep 'em comin'!

Anonymous said...

Re: Lucy in the Sky and Gregory & Heard 1983:

I can obtain the precept that the (individual) figures are thin tiles which are rotating, if I assume a lighting source. The apparent direction of travel of the edge is in the direction that would be the case if the figures actually were rotating. I would therefore hypothesize that this illusion is based on expectation of motion, and doesn't depend on contrast considerations.

--peter blicher

/*********google analytics*/ /*********google analytics end */