“Spine drift” illusion
from Michael’s Visual Phenomena & Optical Illusions
In the image above you see an array of spindly thingies (“spines”), with a gap demarcating a square in the middle. Does the central square appear to float, to move relative to the background?
If so, you are perceiving seeming movement, namely Kitaoka’s “Spine Drift” illusion. It is strongly affected by eye movements, thus perceived differently by different people.
To enhance the illusion: shake your display a little (easy, of course, with a notebook or iPad). Or scroll a bit, or shake you head, or glance around the display, or use the “shake” button.
If you press the “rotate 90°” button you will find: Whenever the central and peripheral spines are parallel or at 180°, there is no illusion of differential movement; when they’re at 90° –whichever way– the illusion occurs. Whith the wheel you can try out any angle.
The slider allows contrast change. I had a little pet theory that the illusion is stronger with low (≈50%) contrast as opposed to 100%, but that does not seem to be the case, or the difference is weak.
With slight modifications this is Akiyoshi Kitaoka’s Spine Drift Illusion, used here with kind permission. I propose the following explanation, which is similar to the one for the Ouchi illusion: Eye movements cause motion blur on the retina. When applying motion blur to the original spine object (below, left), the resultant retinal image differs strongly in contrast between blur directions:
| original | after applying motion blur in various directions | ||||||||
| 0° → | 22.5° | 45° ↗ | 67.5° | 90° ↑ | 112.5° | 135° ↖ | 157.5° | 180° ← | |
 |  |  |  |  |  |  |  |  |  |
Thus when the retinal motion is at 45° (or –45°), the retinal image is very poor in contrast. It is well known that reduced contrast reduces perceived speed (Thompson 1982), so when the movement is in the 45°-direction the image seems to lag. When the central square and the surround differ in their spine directions by 45°, eye-movement induced retinal motion blur (in one of the two), causing differential contrast reduction, and thus a seeming shift between central square and surround.
To me this is effect is fleeting, and rather beautiful. In his 2010 congress poster, Dr. Kitaoka makes a good point in drawing connections to the Fraser and the Café Wall illusions: there is a second illusion here, all horizontal/vertical arrangements seem a little tilted.
Kitaoka A (2010) The Fraser illusion family and the corresponding motion illusions. Perception 39 ECVP Abstract Supplement, p178
Kitaoka A (2010) Spine Drift Illusion (scroll down, a little below half down)
Thompson P (1982) Perceived rate of movement depends on contrast. Vision Res, 22:377–380