A light source seeming to disappear when I look directly at it

Factual Questions
1

The other night I was lying in bed in a dark room with a CD player going, and the machine had this small but rather bright blue LED light on the front that was keeping me awake. So I put a piece of paper over it, so that I could still see the light but it was less harsh and more diffuse, I suppose.

The weird thing now, though, was that when I looked directly at the light, it seemed to disappear; but looking just a little bit to the right or left, or above or below, made it come back. I tried it with either eye, and then with both eyes, and it was always the same.

This wasn’t the blind spot of my eyeballs, either; I actually tested this by closing either eye and looking a bit to the side–yup, blind spot’s still there. But this other effect was when I was looking directly at the light source.

I might be worried, except for the fact that I tried it with other point light sources in the dark, and it didn’t happen; and I have a distant memory of when I was a kid looking at the night sky, sometimes very faint stars would seem to disappear in the same way when I looked directly at them. But it’s never happened since then, and I figured my eyes must have been much more sensitive at that time.

So why now, only with a blue LED light?

2

This phenomenon is known to astronomers, who train themselves to look just to the side of a dim object they want to see.

The centre of your visual area is optimised for colour perception, at the expense of low-light sensitivity - vice versa for the area outside of centre, and the periphery of your vision is optimised for sensing change of brightness (i.e. movement).

Many optical illusions are dependent on these peculiarities.

3

This is normal. The center of the fovea is quite insensitive to blue light, maximum blue sensitivity is about 1 degree off-center. A blue led has a sharply defined spectrum and the eye just can’t focus on it very well. This is why blue led signage is so annoying.

This is not directly related to how we perceive weak stars, the mechanism there uses the rods. They are also not found in the visual center, thus we need to look aside.

The blind spot is much further away from all of these.

4

There are no blue-sensitive, S-cones at all in the foveola, at the very central focus of vision. There are no rods either. Instead it is packed very densely with the red and green sensitive L and M cones. Even in the rest of the fovea (the entirety of which encompasses only about 2º of visual angle (“about the size of your thumbnail, held at arm’s length”) S cones are quite sparsely distributed compared to the L and M ones.

On the other hand, S cones are found, albeit sparsely, far out into the periphery of vision, whereas once you get a degree or two away from the fovea the numbers of L and M cones fall off very sharply and there are none very far into the periphery at all, where most of the receptor cells are rods. The S cones in the periphery do not subserve color vision though (you can’t see blue, or any other color, in the periphery)* because color requires a differential signal from two or more different cone types. The peripheral S cones provide daylight peripheral vision (basically movement detection) when the rods that are also in the periphery are swamped by too much light.

The effect you describe, by the way, probably only occurs because LED light, unlike nearly all naturally colored light or objects, is practically a monochromatic, single-wavelength blue. Most other sources even of very blue looking light would probably have enough of other wavelengths mixed in with them as to be visible even to the foveola.

¬¬¬¬¬¬¬¬

*You may think you can see some faint color in peripheral vision. This is a common illusion. When someone knows what the color of something is (because they have seen it more centrally, before it moved into the periphery of vision) it can seem as though it still retains that color in the periphery. (In a sense, of course, it does.) However, it is quite easy to demonstrate that if something of a yet unknown color comes into their peripheral vision people cannot tell at all what color it is until it comes into foveal vision. In normal, non-laboratory circumstances, we quickly turn our eyes to bring anything detected in the periphery into foveal vision, but if you manage to refrain from doing that, you can tell that something is there (especially if it is moving), and you may have some very vague apprehension of its shape, but you cannot determine its color.

5

I hear the comments about the purity of blue LED light, but I have played with a spectroscope and blue LEDs, and in my experience they have a long visible tail into the green and red. Any reason for this discrepancy? Could it be because I did this around the time they were new, and their output has become more pure over the years?

6

Check out the section on night vision. Click here!

7

We call looking to the side of what you actually wish to study, “averted vision.” The jokers in our club call it perverted vision.

8

This is false.

I just tested it - my daughter randomly handing me coloured crayons from behind my back to my arm, outstretched about 100 degrees to the right - I brought them around towards the front and I can identify their colours before they get to 60 degrees from front. I had trouble with the greens, but I can reliably discern red (2 different shades), orange, yellow, blue (2 shades), purple, turqouise.

9

Previous thread on the topic:

10

And a citation which says that the distribution of cones falls sharply away from the centre of the retina, but not to zero:
http://www.cis.rit.edu/people/faculty/montag/vandplite/pages/chap_9/ch9p1.html

11

I have tested it too and it is not false. I am pretty sure there are laboratory controlled studies showing it to be true, although I am not sure where to find a cite at the moment. Anyway, unless you have a cite to the contrary I put no faith in your informal experimenting. You probably did not maintain full fixation (which is actually very tricky to do, and it is quite hard to recognize when your eyes do a tiny quick flick to the side).

12

Here’s one:

13

Another account of your informal, at home experimentation. Sorry, no, you are deceiving yourself. Anyway (as I should have said in my previous post) 60º is not the far periphery of the visual field. There may be some very sparse L and M cones out that far, giving some weak color vision, but there is plenty of visual field beyond that. The full visual field is about 200º.

I did not say there are no cones in the periphery, I said there are only S cones, and they do not subserve color vision. (And perhaps I should have said “far periphery” since peripheral vision is a rather vague notion. Certainly there is a parafoveal region around the fovea, extending to about 10º or a bit more, and a perifovea extending a few more degrees beyond that, where all three cone types are found, and that can see the normal range of colors, though noticeably less brightly and distinctly than the fovea itself.)

Cites:

Wooten, B. & Wald, G. (1973). Color-vision Mechanisms in the Peripheral Retinas of Normal and Dichromatic Observers, Journal of General Physiology, 61, 125-142.

Curcio, C.A., Allen, K.A., Sloan, K.R., Lerea, C.L., Hurley, J.B., Klock, I.B., & Milam, A.H. (1991). Distribution and Morphology of Human Cone Photoreceptors Stained with Anti-Blue Opsin, Journal of Comparative Neurology, 312 #4, 610-624.

The Curcio et al. paper is also a cite for what was my main point: that the foveloa is entirely devoid of S cones.

14

This really depends on the materials used, but typical blue leds have close to zero emissivity at 500nm, which is considered to be the lowest wavelength sensed by the M cone. See SML sensitivities graphs.

15

Yes, yes - please feel free to simply ignore that.

Neither is it foveal vision. You claimed in an earlier post: “people cannot tell at all what color it is until it comes into foveal vision” - this is not true.

16

Who are you gonna believe, me or your own lyin’ eyes?

(Thanks all for the responses, BTW)