Tuesday, December 16, 2008

Rotating Reversals

What to notice: You are looking at two spinning rings. When you look at the yellow dot in the center of the spinning ring on the right, the rings spin toward each other; when you look at the red dot in the center of the spinning ring on the left, the rings spin away from each other.

What is happening? The rings are made up of two components:

1) Six ovals that rotate in one direction

2) Lines inside the ovals that rotate in the opposite direction

When you look directly at the display, you perceive the rotation of the ovals.
When you look toward the red dot or the yellow dot, you perceive the rotation of the internal lines in the ring that is further away from the dot.

Comments: Many illusions “work” because they pit two sources of information against each other (look at the first illusion in this blog for an example). In the rotating reversals demonstration above, the global motion of the ovals is pitted against the internal motion of the lines. To see what I mean, let’s take a look at one ring by itself in the demonstration below.

There are two sources of information.

The global motion rotates counter-clockwise; the internal motion rotates clockwise.

Your visual system has to “choose” how to perceive these conflicting sources of information. In other words, will perception be guided by the motion of the ovals? Or by the motion of the internal lines? Or by a combination of these two? Or will you be able to see both types of motion at the same time, while keeping their signals separate?

When you look directly at the one-ring display, you can discern both sources of information (the ring will spin one way, and the motion caused by the internal lines goes the other way). But when you look at this display peripherally, it becomes difficult to separate the two sources of information, and the internal motion drives the perceived direction of the ring.

(To look at the display with your peripheral vision, focus your eyes on a spot a few inches above the display.)

In the two-ring display, I simply flipped one of the rings so that there would be a conflict in perceived direction of motion when you focus on one dot or the other.

Why is there a difference when you view the display foveally (i.e., directly) and when you view the display peripherally?
The foveal visual system is quite different from the peripheral visual system. We know, for instance, that the world looks blurrier to the peripheral visual system than to the foveal visual system (vision scientists would phrase this as, “The peripheral visual system has poorer spatial resolution than the foveal visual system”). But a blurry peripheral perception alone does not seem to explain why the disks appear to reverse direction. If you blur the display and then look at it in the fovea, the rings do not seem to reverse motion. Well, at least to me and others who have done this experiment in my laboratory, they don’t. I’ll be interested to hear your comments on this topic.

In the Shapiro, Knight, and Lu talk at the recent Society for Neuroscience conference (Nov. 2008), we hypothesized that the machinery of the foveal visual system allows us to represent multiple features simultaneously, but this machinery is absent in the periphery. The peripheral visual system seems to mix up the features that are available in the scene. We called this “feature blur,” and we showed a number of illusions that are consistent with this hypothesis.

At first, the “feature blur” hypothesis may seem counter-intuitive: when you focus on one point, the features in the periphery don’t often appear to jumble together. I think that the reason that some, but not all, displays show strong feature blur is that the effect depends greatly on the contrast with the background. To see this, move the lever in the single-ring display to adjust the background luminance. When the background is brighter or darker than the luminance inside the ovals, the ring no longer reverses when you focus on a spot a few inches above the dot.

Peter Meilstrup, at the University of Washington, points out that the spinning rings also juxtapose motion over different scales. The brain can register “short-range” motion (i.e., motion over a small region) and “long-range” motion (i.e., motion over a large region). Changing the luminance of the background also changes the relative responses to short-range and long-range motion. When the background is gray, the short-range motion signal is strong, but when the background is black or white, the short-range motion signal is weak. As a result, processes that respond to long-range motion energy may predominate against a white or black background, but not against a gray background.

Here are two references (courtesy of Peter) that examine the juxtaposition of long-range and short-range motion processes:

G. Mather, P. Cavanagh, and S. M. Anstis (1985). A moving display which opposes short-range and long-range signals. Perception, 14(2): 163–166.

C. Chubb and G. Sperling (1989). Two motion perception mechanisms revealed through distance-driven reversal of apparent motion. Proc Natl Acad Sci U S A, 86(8): 2985–2989.

* * * * * *

History of the illusion: As indicated on the displays, the illusion has been developed independently in two laboratories, and was presented by both laboratories at the Society for Neuroscience conference in Washington, D.C., in November 2008.

I developed the effect as an extension of the illusions that Emily Knight, Zhong Lin Lu, and I presented at the May 2008 Best illusion of the year contest (here is a link to the pdf of the entry) and as an extension of our work on “feature blur” in the visual periphery.

Peter Meilstrup and Mike Shadlen presented their version of the illusion as part of a continuation of Shadlen and Movshon’s work on motion signals in the brain (specifically, in area MT of the visual cortex). Here is a link to Professor Shadlen’s webpage.

* * * * * *

Here are code snippets for you to post the illusion on your blog or website, if you would like.

Two rings demonstration

<embed pluginspage="http://www.macromedia.com/shockwave/download/" src="http://arthur.shapiro.googlepages.com/RotatingReveralsForIllusionSciences.swf" type="application/x-shockwave-flash" height="535" width="555"></embed>

One ring demonstration

<embed pluginspage="http://www.macromedia.com/shockwave/download/" src="http://arthur.shapiro.googlepages.com/OneRingForIS.swf" type="application/x-shockwave-flash" height="540" width="500">></embed>

* * * * * *
Once again, I feel a need to apologize for the long delay between posts. I have changed the tag line on my blog to read “An illusion whenever I can get around to it.” I find it quite difficult to manage the blog while the semester is in progress.

Thursday, Dec. 18, 12:15
The site was running up against a google-pages bandwidth limitation and so people who visited were not able to see the illusion. I moved the illusion files to a different host site--that should fix the problem for now.

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