# Optical illusions: what your brain still sees, even when you know better

**Canonical address:** https://optimistra.com/en/blog/optical-illusions-and-the-brain
**Authors:** `michel-wozniak`, https://optimistra.com/team/michel-wozniak; `nathalie-de-marce`, https://optimistra.com/team/nathalie-de-marce
**Category:** `Beliefs and Change`
**Published:** 2026-11-18
**Updated:** 2026-11-18
**Reading time:** 8 min, 1572 words

## In brief

An optical illusion isn’t a malfunction: it is the brain interpreting what it sees, correcting for shadows or filling in what’s missing. The most striking part is that the illusion holds even once you know how it works. Three famous illusions show this, and research explains why.

---

There is a one-minute experiment that changes the way you hold your own certainties. You look at an image, you learn its secret, you check it, then you look again, and it fools you exactly as it did before. This article presents three of these images, explains what research knows about them, and shows what you can do with that on an ordinary day.

## What an optical illusion is

An **optical illusion** is an image that the visual system interprets differently from what is actually drawn. It corrects for a shadow, fills in a gap, or combines two cues that contradict each other. This is neither a disorder nor an exception: it is perception working as it normally does, made visible by an image built to catch it off guard.

In our courses, we call them **benign hallucinations**. The brain sometimes adds what isn’t there, and sometimes removes what is: the keys you look for everywhere that were right in front of you. Both of these go on all the time, and most of the time they serve us well.

## Adelson’s checker shadow: two squares, one gray

The best-known image was created in 1995 by Edward Adelson, a researcher at MIT. A checkerboard, a green cylinder standing on it, and the cylinder’s shadow falling across the squares. Two squares carry a letter: A, outside the shadow, looks clearly darker than B, which is inside the shadow. Take a look at them.

![Adelson’s checker shadow illusion: a gray checkerboard, a green cylinder standing on it, and its shadow; square A, outside the shadow, looks dark, and square B, inside the shadow, looks light.](/medias/blog/95d838e050a10ccc.webp)

*Adelson’s checker shadow illusion. © 1995, Edward H. Adelson, reproduced freely as the author permits.*

They are the same gray. We measured it pixel by pixel on the author’s file: both squares come out at **120** on a scale that runs from 0 (black) to 255 (white). Adelson’s proof image connects A and B with two vertical bars of that same gray, and the difference disappears:

![Adelson’s proof: the same checkerboard, crossed by two vertical bars of uniform gray that connect squares A and B; connected this way, the two squares appear to be the same gray.](/medias/blog/4d821fb6c5af8221.webp)

*The proof: two bars of the same gray connect A and B. © 1995, Edward H. Adelson.*

Now scroll back up to the first image: the difference returns.

On his lab’s web page, Adelson explains what is happening. The visual system is trying to work out the color of surfaces, and a shadow changes the light they reflect: a white square in shadow can reflect less light than a black square in full sun. So the system compensates. A square that is lighter than its neighbors is probably light, shadow or no shadow; and a shadow with soft edges reads as a shadow, not as paint. His conclusion is the finest sentence you can read on the subject: *“As with many so-called illusions, this effect really demonstrates the success rather than the failure of the visual system.”* The visual system is not a light meter, and that isn’t its job. Its job is to tell us what things are made of.

## The Hermann grid: spots that don’t exist

Let your eyes wander over this grid without settling on any one point.

![A Hermann grid: black squares separated by white streets; at the intersections, nothing is drawn.](/medias/blog/01ef164bb9792c4e.webp)

*A Hermann grid, drawn for this page: at the intersections, there is only white.*

In 1870, the physiologist Ludimar Hermann described this grid of black squares separated by white streets. At the intersections you are not looking at directly, gray spots appear. As soon as you look at one of them, it goes away and comes back somewhere else.

For decades, textbooks explained it by the wiring of the cells in the retina, which were said to respond differently at the center of an intersection and in the middle of a street. The explanation is elegant, and it falls short. Grids with wavy lines, proposed by the Hungarian researcher János Geier, keep exactly the same intersections. When the lines are sinusoidal, write Piers Howe and Margaret Livingstone (2007, *PLoS ONE*), *“the illusion completely disappears.”* A teaching page from Peter Schiller’s lab at MIT sums it up: the cells of the retina should behave the same way at the intersections of these different grids, and yet the spots are almost absent there.

Even the explanations of illusions can be illusions of certainty. The same goes for a very common idea, that we see in black and white at the edges of our vision. Christopher Tyler (2015, *i-Perception*) calls it a *“widespread misconception”*: barely 1% of the roughly five million cones, the cells in the retina that see color, lie in the central five degrees of vision. The rest are spread across the periphery, which sees color far more than we think.

## Knowing doesn’t change what you see

This is the most surprising point, and it is what makes the experiment worth doing. You know that A and B are the same gray. You have checked. And you still see them as different.

Chaz Firestone and Brian Scholl, in a review published in 2016 in *Behavioral and Brain Sciences*, point out that these lightness illusions *“persist in the face of countervailing knowledge”*: the correction happens within perception itself, as a reflex, before any thinking takes place. They add that there is probably no better way to feel the difference between seeing and thinking than to *“visually experience the world in a way you know it not to be.”*

## The ear isn’t spared: the McGurk effect

Illusions are not reserved for the eyes. In 1976, Harry McGurk and John MacDonald played people one syllable while showing them the lips of someone pronouncing another. Many people then hear a third syllable, which is neither the one in the sound nor the one on the lips: a “ba” that is heard with a “ga” that is seen often produces a “da.” With your eyes closed, the illusion disappears; with your eyes open, it comes back.

The best thing is to try it: [the BBC video](https://www.youtube.com/watch?v=G-lN8vWm3m0) (three minutes) plays the same syllable over two different mouths. Listen once with your eyes closed, then once with your eyes open.

The spectacular figure from the original study, 98% of adults, deserves a closer look. A follow-up study published two years later found only 64% for the same combination. In 2015, Debshila Basu Mallick, John Magnotti, and Michael Beauchamp tested **165 adults** on 12 different recordings. Depending on the person, the effect ranged from **0% to 100%**: some never heard it, others heard it every time. And for any one person, the tendency was still stable a year later. So the McGurk effect is both very robust and very uneven from one person to the next.

## Not everyone sees the same thing

In 2015, [a photograph of a dress](https://en.wikipedia.org/wiki/The_dress) divided social media: blue and black for some, white and gold for others. Look at it before you read on: what color do you see? Rosa Lafer-Sousa, Katherine Hermann, and Bevil Conway asked **1,401 people**: 57% saw it as blue and black, 30% as white and gold, and 11% as blue and brown. Their hypothesis matches Adelson’s: each of us corrects for the lighting we assume. Someone who pictures the dress under a blue sky subtracts blue and sees it as white; someone who pictures it under warm light sees it as blue.

So two honest people, looking at the same image, can see two different things, and each be right about what they see. That is exactly what we wrote about [the map and the territory](/en/blog/the-map-is-not-the-territory): each of us perceives the world through our own map, and two different maps can both be accurate.

## What this changes in your day

“I only believe what I see” is a less solid sentence than it sounds, since what we see is already a construction. That doesn’t make our eyes unreliable. It invites us to hold a first certainty a little more lightly, especially our own.

Take a moment to picture Monday’s meeting. A colleague describes the situation, and it isn’t what you saw. You feel your certainty, sharp and clear, like the two squares on the checkerboard. And you hear yourself ask, in a steady voice, “What did you see?” Your colleague answers. A detail comes up that you hadn’t noticed, and the discussion takes a different turn. How does the rest of that meeting go, once everyone has been able to show what they saw?

This simple question builds on two other articles: the one on [the invisible gorilla](/en/blog/invisible-gorilla-inattentional-blindness), where attention lights up only one spot at a time, and the one that shows how to [observe without jumping to conclusions](/en/blog/observing-without-jumping-to-conclusions).

## An exercise to see it for yourself

The most convincing thing is still to experience it. Take an illusion whose secret you know:

- **First, just look.** On the checkerboard, note what you see: A dark, B light.
- **Then check.** Cover the rest of the image so that only A and B show, or look at the proof image.
- **Look again, now that you know.** Note what holds. This is the moment that matters: knowing has changed nothing about what you see.

Then do it with someone else. Let them bet on the two squares, show them the proof, and watch their face when they go back to the image.

We are Michel Wozniak and Nathalie de Marcé, the world’s first two French-speaking Master Trainers named by Richard Bandler himself, co-creator of Neuro-Linguistic Programming (NLP). We show these illusions in our courses: they come on the second day of the [NLP Practitioner course](/en/nlp-practitioner), in the work on perception.

*Article written by Michel Wozniak and Nathalie de Marcé, and optimized by artificial intelligence.*

---

## Frequently asked questions

### What is an optical illusion?

It is an image that the visual system interprets differently from what is drawn, because it corrects for a shadow, fills in a gap, or combines several cues. It is not a flaw in your eyesight. The researcher Edward Adelson writes that his checkerboard demonstrates the success of the visual system rather than its failure.

### Why do I still see the illusion when I know how it works?

Because these corrections happen within perception itself, before any thinking. Firestone and Scholl (2016) point out that lightness illusions persist in the face of what we know, and that there is probably no better way to feel the difference between seeing and thinking.

### Are squares A and B in Adelson’s checker shadow illusion really the same gray?

Yes. Measured pixel by pixel on the author’s image, both come out at 120 on a scale from 0 to 255. Adelson’s proof image shows it by connecting the two squares with two bars of the same gray.

### Does the McGurk effect work on everyone?

No. Among the 165 adults tested by Basu Mallick, Magnotti, and Beauchamp (2015), some never perceived it and others perceived it every time. For any one person, the tendency was still stable a year later.

### Is peripheral vision black and white?

No, that is a common misconception. According to Tyler (2015), barely 1% of the roughly five million cones, the cells in the retina that see color, lie at the center of vision; the rest are spread across the periphery, which therefore perceives color.

---

## Sources

- [Howe & Livingstone (2007), The use of the cancellation technique to quantify the Hermann grid illusion, PLoS ONE](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0000265)
- [Schiller, The Neural Control of Vision — Illusions (Hermann grid), MIT Schiller Lab](http://web.mit.edu/bcs/schillerlab/research/A-Vision/A15-7.htm)
- [Firestone & Scholl (2016), Cognition does not affect perception: evaluating the evidence for top-down effects, Behavioral and Brain Sciences](https://perception.yale.edu/papers/16-Firestone-Scholl-BBS.pdf)
- [Basu Mallick, Magnotti & Beauchamp (2015), Variability and stability in the McGurk effect, Psychonomic Bulletin & Review](https://pmc.ncbi.nlm.nih.gov/articles/PMC4580505/)
- [Lafer-Sousa, Hermann & Conway (2015), Striking individual differences in color perception uncovered by 'the dress' photograph, Current Biology](https://pmc.ncbi.nlm.nih.gov/articles/PMC4921196/)
- [Tyler (2015), Peripheral color demo, i-Perception](https://pmc.ncbi.nlm.nih.gov/articles/PMC4975120/)
