It seems a variation of this well-known illusion: http://kids.niehs.nih.gov/images/illusions/pinkdots.gif

I don't think motion is required. I remember staring at a patterned ceiling in 8th grade algebra class - if I focused on one point long enough, the further out patterns "disappeared."

That's called Troxler's fading, caused by the gradual adaptation of photoreceptors in the retina to the recent average light level. But in the OP, the yellow dots are high in contrast, disappear in less than 5 seconds, and then seem to blink on and off at random. So I think this might not be the same effect.

http://en.wikipedia.org/wiki/Troxler%27s_fading

You can actually stop the motion in that page. I tried it, and although the dots seem to fade to a light gray, they did not go away completely for me.

This reminds me somewhat of a visual field test, which I cannot pass to save my life.

I expected this to affect me either 1) Not at all, since I have monocular vision and a lot of optical illusions don't work for that reason, or B) ridiculously well, since I have a visual field of only about 15 degrees, a rather significant, centrally located cataract, and severe nystagmus.

However, it seems the reactions I was having to the dots pretty much mirrors others' experiences, which is, to me at least, rather interesting.

I was able to have the dots come on and off while I focused on the center. Turning them on and off at will when i wanted them to. What shocked me is that you can make the yellow dots 5 times there size, and they will still disappear.

I think this shows there are multiple modes of human vision. During the first few hundred milliseconds we are in "situation assessment mode" where we see the outer dots perfectly, during "focus on something" mode, the dots go away. This provides insight into the exact algorithm our visual cortex uses to parse whatever data streams come from the eyes. Amazing!

That's probably a fair general characterization of what is going on, but the core part is that our perceptual systems respond primarily to change - if things in our field of vision are not changing, they produce less of a "signal" in our brains. It's not our perceptual system that is switching modes as much as it is the world changing between active and static.

In the MIB case, most of the field of view is constantly changing, providing a lot of signal, but the yellow dots are low signal, and so our brains happily fill in the static dots with the in-motion background around it. Moving your eyes, blinking, or moving your head around all introduce enough visual change to re-set the full visual field, which is why the yellow dots reappear.

Through top-down attentional control, you can have some influence over whether/which dots disappear - in effect, you are turning up the signal gain on particular receptors in the visual field around the target dots.

You can also see why this affects people with monocular vision just as well - a lot of this perceptual processing happens at a very low level, well before binocular cues get integrated.