How optical illusions work?

Researcher Mark Changizi of Rensselaer Polytechnic Institute in New York believes he may have found the cognitive trick that allows all optical illusions to work — humans can see into the future. Allow me to explain, because it’s actually not nearly as silly as it sounds.

The complete neural pathway from light hitting the eyes to the formation of a visual perception of the world in the higher parts of the brain takes about a tenth of a second. Researchers have long wondered how humans manage to be as accurate as we are, considering how much can happen in a tenth of a second (think of how far a pitched baseball traveling at 80mph will move in 0.1 seconds, for example). Mark Changizi believes that our visual system extrapolates about a tenth of a second into the future to make up for the delay (using dead reckoning, I guess?). Thus, we aren’t actually seeing the world as what it is, but as what our highly honed visual system thinks it should have been based on an extrapolation from a fraction of a second prior. The eyes aren’t relaying images directly to the brain; there is some processing going on in between. And for the most part, this solution works just fine.

Except in the case of optical illusions. Optical illusions trick our brains into falsely extrapolating what an image will look like in the very near future. Optical illusions are thus a continual cycle of our visual system predicting something that won’t actually happen, then constantly getting confused about it. This explains how static images can appear to be moving, etc. It’s a really elegant explanation, and combined with the previous knowledge that there is a delay in our visual system, it just feels right to me. I wouldn’t be surprised at all if this explanation ended up being correct.

One Response to “How optical illusions work?”

  1. arensb Says:

    Doesn’t this apply mainly to a subset of optical illusions, the ones with some kind of temporal dimension?

    I liked the explanation Steven Pinker gave in How the Mind Works: the problem of converting two 2-D arrays of color values into a 3-D model, the problem that the visual system tries to solve, is in fact an unsolvable problem: any given dataset has multiple solutions. On top of this, the visual system has to solve this problem in a finite amount of time, making it doubly-unsolvable.

    So the brain makes assumptions about the world that, if true, make the problem tractable: Objects are usually three-dimensional. Adjacent spots on an object tend to be of the same color and brightness. Objects are lighter on the side the sun shines on, and darker on the other side. Objects retain roughly the same shape from one moment to the next. An object moving from A to B will pass through the intervening points. And so on.

    Illusions, says Pinker, are drawings cleverly designed to violate one or more of these assumptions, and thus fool the visual system into seeing something that isn’t there.