Why do the dots disappear?

September 2, 2014 • 8:10 am
This is one of the most baffling illusions I’ve ever seen. Take a look at the gif below.  First, look at any yellow dot as the figure moves. The yellow dot remains present and stationary. If you concentrate on all three yellow dots, they remain there as well.
But now concentrate on the central green dot. You will see one or more of the yellow dots disappearing and then reappearing sporadically. They are not—this is an optical illusion. The dots remain and your brain simply doesn’t register their presence from time to time. Weird, eh?


An article by John Whitfield in Nature News, “A brain in doubt leaves it out,” explains the phenomenon:

Yoram Bonneh, of the Smith-Kettlewell Eye Research Institute in San Francisco, and colleagues have been showing people a swirling pattern of blue dots superimposed on some stationary yellow dots1. [JAC: for some reason the reference isn’t given.]

The yellow dots seem to wink in and out. But the erasing happens in the mind, not the computer. Nearly everyone tested saw the effect.

The brain seems to have internal theories about what the world is like. It then uses sensory input – which tends to be patchy and disorganized – to choose between these. In some sensory situations, different theories come into conflict, sending our perceptions awry.

The illusion, which Bonneh’s team calls motion-induced blindness, catches the brain ignoring or discarding information. This may be one of the brain’s useful tricks, a deficiency – or perhaps both, says Bonneh.

The researchers suggest this may (and I suggest that it certainly must) happen in daily life:

The researchers speculate that this phenomenon could happen in everyday life without us noticing it. A highway at night, with drivers staring dully at a mass of moving lights, might recreate the kind of conditions used in the experiments, says Bonneh, causing objects – the tail lamp of the car in the next lane, for example – to temporarily vanish.

Jack Pettigrew, a neuroscientist at the University of Queensland in Brisbane, believes that the illusion results from a tussle for supremacy between the left and right halves of the brain.

He has found that applying a pulse of magnetism to the brain to temporarily disrupt its function affects the occurrence of motion-induced blindness. When the pulse is applied to the right hemisphere (leaving the left dominant) the dots disappear; zapping the left brings them back2.

The left hemisphere seems to suppress sensory information that conflicts with its idea of what the world should be like; the right sees the world how it really is. Some people with paralysis caused by injuries to their right hemisphere will deny that they are disabled.

My only question is why it takes motion to generate this illusion. Is that because motion is associated with visual confusion?

Source of gif: Professor Michael Bach at the University of Freiburg, via reader Grania. Bach has a page with 113 of these damn things!


103 thoughts on “Why do the dots disappear?

    1. That doesn’t happen for me but if I stare at one yellow dot, the others also randomly blink in and out of existence like the green dot.

    2. More: cover the center dot with your finger, the disappearing effect still works. The dot and its blink are either a red herring or something that keeps your attention focussed on the proper field of view.

      The outer dots also sometimes fade rather than vanish entirely.


      1. The target in the center is intended to keep your eye fixed, as the illusion goes away when you make a saccade (a quick eye movement to another part of the image).

    3. The yellow turning green was the first thing I saw. Immediately. In fact, I thought that was the illusion until I read some text. I too am red/green color blind.

    4. When I stare at one of the 3 yellow dots, the green dote becomes yellow! hard to check but im pretty positive so ur right 🙂

    5. When you stare at one of the three yellow dots on the picture on facebook (without any movement), the other two disappear as well. So the movement is here not essential, it seems to me.

  1. These types of illusions are useful for drivers to consider when thinking just how easy it can be to accidentally hit a pedestrian or cyclist.

    1. When driving at night on a country road, it is often hard to tell the distance of something approaching you. Many times I’ve thought I’ve seen a motorcycle but it was a car with one burned out light or I thought I saw a car far away & it’s been a motorcycle up close!

      1. That’s what I think all the time. A pedestrian or cyclist should always assume no one sees them, ever. When we commute we should assume we are anti-important, until we get to our destination, then we can resume whatever illusions of superiority.

  2. This should be available for use in trials involving witnesses who claim to see things that perhaps they do not????

    1. At this point artificial vision is very poor. Think about it. You can do CAPTCHA’s much better than computers can.

      I kind of always hope it stays poor. I don’t the drones to start deciding by themselves who to kill.

      1. I’m sure the dronemasters won’t let the lack of decent machine vision slow them down.

        “Damn civilian, he was just asking to be blown away, not displaying his barcode properly. Let’s just put him down as an insurgent, OK?”

    2. I work in the field of computer vision. In some very narrow applications the algorithms can outperform humans (e.g. recognizing faults in manufacturing steps). But in general I wouldn’t trust those algorithms in safety critical systems.

      Vision is a complex beast. And since computers are still unable to comprehend what they are seeing they need to rely on heuristics and statistics. Heuristics can easily be fooled and the whole algorithm goes horribly wrong.

      I always like to look at Google StreetView glitches. Many of them occur because the computer has to assemble a big image from a collection of little ones. Usually that works well enough but sometimes that can go wrong (and especially when the content changes between two images due to motion)

      1. But who needs vision when you’re a machine. It’s clearly flawed and other sensors may yield better results.

        1. As Steven Wright’s apprehension-inducing pilot said in “So I Married an Axe Murderer”: “That’s the artificial horizon…which is better than the actual horizon.”

        2. True, but while other sensors might be better suited for a certain task, the problems with them are very much alike. No sensor gives flawless raw data. Data is always noisy and potentially ambiguous. They also require mathematical models to make sense of the data.

          And since the machine also has no understanding of what it is sensing those models are also usually based on statistics and heuristics.

          1. Sure but that’s now. What about the singularity or computer enhanced brains one day? I don’t want my MacBook Air telling me what to do but if we figure out the hard problem of consciousness or even before that maybe we’ll end up with some cool/villainous/powerful stuff?

  3. It’s more than a bit of a stretch to generalize the very unnatural animation in this image as “motion”, and guess that actual motion in the real world might cause the same effect.

  4. Does it matter what the colors are? If so, which ones work and which don’t?

    This one seems to work even if I adjust the image so only two yellow + the green show. And if I cover it so only two yellows show, and I stare at it longer, they start disappearing too.

  5. The illusion works at the normal distance from the monitor, but when I step back to about four feet away, it no longer works. I discovered this accidentally.

  6. I’ve seen a similar illusion, without motion, where the non-focus objects disappear and do not reappear.

    I’d guess the motion is causing the yellow dots to reappear (somewhat). And I’d guess the explanation is that motion is invoking a bit more updating of peripheral information than would a stationary field.

  7. I noticed it also does not depend on all three dots, or how you look at the picture. Take any dot away and tilt your head and it still happens. Also there doesn’t seem to be any pattern as to whether one, two or all three dots disappear or reappear, any of the combinations can happen at any time. (At least in my brain). Perhaps with further study there are statistical patterns to dots coming and going that I don’t see. Anyway amateur experimentation. Interesting.

  8. This would be great for explaining Lawrence Krauss’s Something from Nothing via quantum fluctuation. Both give you a headache, though.

  9. I’ve read that blind spots can sometimes envelop things in your field of vision that your brain has deemed to be totally static. In an attempt to turn a phrase, one cognition researcher (Robert Jourdain, who is also a musician) oversimplified it with the phrase “our brains are ultimately only interested in change”. I think Dan Dennett has made a similar point on a number of occasions. I wonder if this might be an alternative explanation.

    I mentioned that Jourdain is a musician because the blind spot phenomenon has an interesting repercussion for music. It is, claims Jourdain, the reason dynamic manipulation in performance is stressed so much by teachers. If you’ve ever taken lessons on an instrument you’ve undoubtedly had your teacher admonish you for “playing the notes all equally”. Phrases should normally blossom and taper. The techniques of vibrato and messa di voce are also meant to combat stasis.

  10. For me, when the yellow dots would re-appear, they usually seemed to be a very dark yellow color, instead of the bright yellow that they actually are.

  11. I’m not getting it. I stared at the green dot for a minute straight and the only dot that came and went was the green one.

    Magic Eye pictures never worked for me either.

    I guess I’m one of those people with a weird brain that doesn’t see the effect.

    1. I’ve also never been able to “see” Magic Eye pictures, but this effect works very well for me. In fact, I’m amazed at how long the dots stay invisible. It’s like an exercise to keep them gone as long as possible.

      1. Also too magical (post nystagmus surgery at 5yo) for Magic Eye, here, and this illusion works spectacularly for me with both eyes or either one. It’s conclusively not based on stereoscopic vision.

    2. I can’t not see it.

      I don’t know if the effect is exacerbated by the local processing bias associated with my Asperger’s – that seems to have something to do with information in my brain not passing easily over distances.

  12. I’ve read several articles in recent years about the unreliable nature of eye witness testimony. Apparently the parts of our brains that process vision are “playing tricks on us” far more often than we realize, as this experiment suggests.
    All this serves to make me uneasy about prison inmates who’ve been convicted on the basis of eye-witness testimony alone.

    1. If my memory serves me, every time you recall a memory, you brain actually re-evaluates/re-processes the information. When you stop thinking about that memory, your brain may save a slightly different version than what you originally recalled. In fact you may subconsciously make a good memory better or a bad one less bad each time you recall it.

      Do this over and over again with a particular memory and it may end up like playing the ‘telephone’ game with yourself. What you get at the end may hold little resemblance to the original memory or the actual events.

      There is a technique for treating PTSD (Post Traumatic Stress Disorder) that uses this. While under medication that makes you susceptible to suggestion and guided by a psychiatrist, recall the traumatic memory. Talk about it with the shrink and they help you to tone it down a bit. When you stop recalling that memory, your brain saves it a little less bad than it was. Rinse and repeat until the memory is no longer traumatic.

      Now the question is, did I just want to remember it that way or is that what really happened while I was seeking help for my PTSD?

  13. If I look with both eyes it ‘works’ that is the yellow dots blink in and out. If I close the right eye it no longer works. Closing the left eye has no effect. I am pretty right eye dominant especially for close up. (I’m using an IPad)So I tend to go along with the idea that the right side of the brain is imposing its preconceptions. But why should it have that preconception?

    1. Unless you are a rabbit, deer, …
      one eye doesn’t correspond to one half the brain. In frontal-eyed animals: carnivores and primates (and you) both eyes project to both henispheres: the left half of each retina to the left hemisphere and the right half of each retina to the right hemisphere. So your idea doesn’t wash. Similarly, I doubt Prof. Pettigrew’s idea. The key thing here is that the neural activity in response to the peripherally viewed dots adapts over time and then is obliterated by the moving field. You can abolish the disappearance/reappearance of the dots by rolling your head back and forth slightly so that the peripheral dots are continually stimulating fresh bits of retina — this abolishes the fading. As others have pointed out, motion isn’t strictly necessary, one can get peripheral fading without motion, and as the Cheshire cat effect shows one can use motion with binocular rivalry to make someone’s face disappear.

  14. Speaking of illusions, I’ve found that if I really concentrate, I can see the two lines of the Mueller-Lyer illusion momentarily as the same (correct) length! Anyone else notice that?

    This one is particularly interesting. Do the colours matter?

  15. There are a number of known illusions in which an object can be blanked out (or filled in by the background) when the eye is held in a static position. As soon as a saccade is made, the object reappears. The typical illusion of this sort is often based on an image with regions of low-contrast. It’s possible that the motion (what a visual psychophysicist might call an “optic flow”) induces a similar blurring of local contrast, causing nearby objects to blend into the background.

  16. The left hemisphere seems to suppress sensory information that conflicts with its idea of what the world should be like; the right sees the world how it really is. Some people with paralysis caused by injuries to their right hemisphere will deny that they are disabled.

    V.S. Ramachandaran has written about this asymmetry in Phantoms of the Mind.

    1. A fabulous book about left and right sides of the brain- very long,erudite, chewy and compelling is :The Master and his Emissary. By Iain MacGilchrist.

  17. I don’t see anything unexpected here. Yellow and blue are perceived by the same cones. Because of the focal point, the yellow dots are in the parafovea, not the fovea, so there are far fewer cones.

    The movement of the blue lines means that specific blue cones are going on and off and yellow cones are going off and on-ish.

    Attenuation and spreading activation causes us to throw the baby out with the bath water.

    This illusion, while compelling and generally cool, is the perceptual equivalent of saying water is wet. Without neural imaging showing the impact of neurons outside the anterior occipital lobe (responsible for the earliest parts of vision) there is absolutely no evidence of any effect of attention here.

    1. I don’t know much about these things, but my first thought on seeing the illusion was that it was one of those that originated in the eyes in ‘pre-processing’, before the signal ever made it to the brain. But this sentence from the OP makes it seem like the illusion originates in the brain, “He has found that applying a pulse of magnetism to the brain to temporarily disrupt its function affects the occurrence of motion-induced blindness.” Any thoughts from someone who knows more about this?

    2. I don’t think this illusion is dependent on color (though there likely is a peripheral vs parafoveal effect). As for whether the illusion is mediated by retinal or cortical mechanisms, I can’t exactly say. But based on the large degree of adaptation seen in neural firing to static images at the level of the thalamus and higher, and the role of saccades and microsaccades in “refreshing” that neural activity, I’m guessing it’s mainly a cortical effect. (Disclaimer: I am an auditory + vestibular (oculomotor) neurophysiologist, and while I have postdoctoral experience in recording from visual cortical neurons, that was a long long time ago and I’m definitely not an expert in retinal physiology.)

      1. I do agree that it’s not an attentional effect. But any illusion is interesting to a psychologist or physiologist studying a sensory system, because it reveals limits on processing that aren’t necessarily apparent in everyday experience (thus, I take exception to the “water is wet” analogy). One of my favorite visual illusions (which, perhaps, like this one involves local contrast adaptation to motion) is the so-called waterfall illusion. Other cool illusions involve interactions between the visual and vestibular systems.

  18. No, no – it just isn’t happening!

    Fascinating yet eternally annoying knowing some of the numerous limitations to our brain and senses.

  19. So, now that so many have had their say and generally expressed interest if not outright fascination…

    …I think it’s reasonable to suggest that this is the sort of thing Sam is promising with his new book, except to perception of cognition rather than visual perception.

    At least, expectation (hope?) is why I personally ordered the book….


  20. One could say we aren’t actually “seeing” the outside world. What we are experiencing is a model of the world the brain has concocted. In this case, apparently two conflicting models which the brain switches between because it can’t “make up it’s mind” which model is correct.

    This is called Model Dependent Realism. Hawking/Mlodinov use this to explain their theory of the multiverse in The Grand Design. Our area of visual acuity resembles a donut. We can’t see anything in the “hole” of the donut because the wires from our rods and cones (which are inserted backwards in the retina) channel through that hole back into the brain’s visual cortex. So the brain fills in that hole with part of the model of reality that we experience as the outside world. This explains optical illusions in general and why magician’s are able to “misdirect” our attention, making us “see” one thing when something entirely different is actually happening. Fun, huh?

  21. There’s a bit of faulty analysis going on here.

    If you stare at the blinking green dot, the yellow dots all disappear. They will then remain invisible unless and until you move your eyes or blink. But if you keep your eyes firmly fixed on the blinking green dot the yellow dots will all remain invisible.

    If that doesn’t happen for you, that just means that you’re not keeping you eyes fixed on the green dot. Any slight movement of your eyes or blinking your eyes and the yellow dots immediately reappear.

  22. “My only question is why it takes motion to generate this illusion. Is that because motion is associated with visual confusion?”

    Motion is not required, at least not for me.

    I discovered a while ago that if I keep my eyes perfectly still for a while (5 seconds or so) without paying strong attention to particular objects, things outside the center of my vision start to fade and darken. If I can keep it up long enough (and usually I can’t), eventually just about my whole visual field becomes a dark grey blob.

    This illusion seems like a special instance of that phenomenon.

    1. Aspid, not everyone can maintain fixation enough to avoid making large enough saccades (fast eye movements) to “refresh” an image. But I have heard of people that can do it (and apparently you’re one of them!) You are right that motion is necessary for this type of adaptation illusion, as I wrote in another comment. The motion seems to have the same effect as a blurred or greyed background, which is a new phenomenon to me.

    2. I can do this as well – “discovered” this “talent” when I was young and bored in school.

      Is this all not due to the fact our eyes are not cameras and that we only see in the centre of our vision is “real”, the rest being added by the brain based on what it last saw (with rapid eye movement or by making assumptions). Peripheral vision detects movement, and we are drawn to that.

      This is partly why we see unusual things “out of the corner of the eye” – the brain misinterprets clues and creates illusions.

      If you stare at a point and eliminate rapid eye movement the brain begins to “forget” what was in the periphery.

      The movement in the above I think makes it even more difficult for the brain. It has to process and “create” both a still and moving image, and simply cannot do it reliably, occasionally “forgetting” the still yellow dot.

      However you then notice the yellow dot is gone, are drawn to it and so it reappears.

      1. Yes!@johnlbirch, “Peripheral vision detects movement, and we are drawn to that.” When I concentrate on the flashing green dot, my peripheral vision in my left eye sees the yellow dot on the upper left vanish at the same instant that the green dot reappears; however, my right eye’s peripheries still detect the upper left yellow dot, but not the upper right yellow dot (though I continue to see the upper right yellow dot, presumably in the periphery of my dominate left eye). The lower yellow dot remains stationary for me the entire time.

        BUT, if I micro blink just as the yellow dot vanishes, then I can see the yellow dot again, even in my left periphery, while the green dot remains lit, and still as the green turns off, up until the moment the green dot reappears, again, at which the yellow dot vanishes, once again (from my dominate left eye visual field–thus inducing in me the urge to, once again, blink my left eye).

        Strange indeed, and although my perceptual sensing and awareness may not considered entirely normal, it’s nice to know that I’m not all that different than quite a few others here with similar perceptual processing. 😊

  23. my guess is that the brain focuses on the moving stimulants, where the danger came from, as it is evolutionary developed. so if for a while, the eye catches a lot of moving stimulants and does not move on the stationary spots, these are considered by the brain harmless and disappears, to ease the eye’s job to catch the moving parts.

  24. You can cancel the disappearance effect very simply: rock back and forth while viewing, so the distance between you and the screen is constantly changing. All the dots remain visible, as long as you don’t stop moving.

  25. at the end it asks why is motion necessary for the illusion. but on the facebook post i saw of this that was a still image i was able to make the yellow dots dissapear as well… took a few more seconds but it happened

  26. “The left hemisphere seems to suppress sensory information that conflicts with its idea of what the world should be like; the right sees the world how it really is.”

    There’s a brilliant political message in there somewhere.

  27. I noticed the other day while sitting stopped in traffic, on I-5 that all the cars and trucks were really close together and large but when I drive the normal speed it seems like the roadway is really wide and the cars and trucks are smaller and widely spaced across the roadway. Perhaps the faster we move the more wider our field of view appears. But in reality they are much closer together and larger but our minds are tricking us to perceive them differently.

  28. This is simply an example of retinal cell fatigue. The moving field is irrelevant.
    Stare at a small dot of anything and after awhile it will disappear. Stare at a few dots on a piece of paper and one or two of the other dots will disappear. To prove that the moving field is irrelevant, cut 3 small holes in a piece of paper so that the yellow dots shine through and the moving pattern is obscured. Stare at one of the yellow dots and soon one of the other yellow dots will disappear.

Leave a Reply