Which way does the dancer spin?

[Factory101]

Ah - the alcohol method...cold science looks much warmer through a half-full glass...

[MacDoc]

Why wouldn't it be random?

Because if was random then any one person would experience one direction or the other randomly each time it was viewed.

No persistence of direction repeatedly.
That's not the case except for a few people that can reverse easily.

Instead you have some people persistence in one direction.
Some in the other.

Some aspect of the brain is biasing one way or the other, some quite markedly.

That is clearly the case.

it was also presented as a left/right phenomena.

If you have an alternative that explains the distributed bias we're all ears.

[DavidMcC]

Some aspect of the brain is biasing one way or the other, some quite markedly.

That is clearly the case.

it was also presented as a left/right phenomena.

If you have an alternative that explains the distributed bias we're all ears.

As Factory101 said, the left-right bias is probably the directional bias in the V5 neurons, but not the entire brain. I infer from his remarks that it is the relative populations of V5 neurons associated with the left-to-right and right-to-left motion detection and with the various parts of the visual cortex that determines each individual's response to the illusion. Right, Factory?

[MacDoc]

Those were also situated I believe on either side of the corpus callosum so the two ideas are mutually exclusive and again it was presented as a left right phenom so somewhere someone must have done some research to justify that.

[msv]

Why wouldn't it be random?

Because if was random then any one person would experience one direction or the other randomly each time it was viewed.

It's random, only persistant in each case. Even though I can switch, there is a small amount of effort required. It requires equal amount of effort to switch to either direction.

Instead you have some people persistence in one direction.
Some in the other.

Some aspect of the brain is biasing one way or the other, some quite markedly.

That is clearly the case.

Not so clear to me, to me it's way more probable that it's an arbitrary decision to wich you cling. That's the most important thing here - consistency. For example, if it weren't persistent people would see it switching at random intervals, which would be annoying and impractical.

it was also presented as a left/right phenomena.

That doesn't say much now does it?

If you have an alternative that explains the distributed bias we're all ears.

The 3d-interpretation of the brain chooses one of two possible directions since it already assumes the object your viewing is in 3 dimensions. So your brain has options, either not calculate the 3d object's form because it can't know which direction the dancer is going, or make a quick abitrary (or simple guess) as to which way the dancer is spinning and parse the object's form.

[Factory101]

Some aspect of the brain is biasing one way or the other, some quite markedly.
That is clearly the case[…]it was also presented as a left/right phenomena.
If you have an alternative that explains the distributed bias we're all ears.

As Factory101 said, the left-right bias is probably the directional bias in the V5 neurons, but not the entire brain. I infer from his remarks that it is the relative populations of V5 neurons associated with the left-to-right and right-to-left motion detection and with the various parts of the visual cortex that determines each individual's response to the illusion. Right, Factory?

David McC - re: "...infer from his remarks that it is the relative populations of V5 neurons associated with the left-to-right and right-to-left motion detection and with the various parts of the visual cortex that determines each individual's response to the illusion."

I didn't mean for that inference to be drawn (apologies) so I will try to make it more distinct: rather than 'left-right' motion detection, there are additional specific neurons for clockwise and counter-clockwise rotation. In perception of visual events there exists a different type of specialist cell for each type of movement or phenomenon - rather than generalist cells. For example: cells specialised in sideways movement, rotation, etc. (but sorry that's a whole new subject).

Those were also situated I believe on either side of the corpus callosum so the two ideas are mutually exclusive and again it was presented as a left right phenom so somewhere someone must have done some research to justify that.

MacDoc - you are quite right and I should have stated earlier that the V5 is situated in each hemisphere:

Also I haven't established whether the receptor-neurons in each V5 are identical or different. This is important as: one might have clockwise receptors and the other counter-clockwise; from this there is the possibility that another 'executive' region decides which V5 to process first (reverse rotation then being a switch to the other V5).

The simple explanation I have given accounts for the phenomenon and provides the mechanisms needed to do this (without an executive mechanism) and this explanation is validated by subjective comments from observers.

The question I have asked myself is: How can we be sure that there is not an executive area of the brain that mediates the direction we see the dancer rotate - rather than the preference of the receptor-neurons alone?

From this diagram we can see that the field of vision crosses and overlaps passing through the optic chiasm:




As the central field of vision overlaps, direct perception of the dancer will involve: an equal distribution of signalling through: first, each eye independently (photon to electric impulse via retina), then the signal is mixed in the optic chiasm, then separates again - part of the signal eventually reaching each V5.

If each V5 has only one type of (central) receptor, either clockwise/counter-clockwise, then the signal would be identical (in terms of motion detection) and merged; an executive mechanism is not needed for the process to operate.

If each V5 has a different predisposition i.e. one is clockwise, the other counter-clockwise, then it seems plausible that a separate 'executive' region decides which V5 to process first (reverse rotation then being an executive switch to the other V5) – but this cannot be as we already know that the switch is related to peripherality of vision (at each end of the Bell curve). Also, an executive mechanism has not been identified yet.
However, in ambidexterity – the ability to freely perceive the dancers rotation (directly, not peripherally) – what is happening? If there is no executive – why isn’t the dancer seen moving in the manner perceived by both sets of V5 signals? Some mechanism must be determining direction of rotation, not necessarily another module but a consequence of architecture and strength of signal. More later.


Much of the understanding in neuroscience comes from efforts to comprehend and alleviate the problems associated with neurological damage or endogenous conditions; it would be interesting to hear accounts of the dancers rotation from participants with damage or dysfunction the V5 region e.g. stoke victims, dyslexics, etc.

My copy of ‘Physiology of Behaviour’ is the 7th edition so I’m sure there is more up-to-date research out there…

[mlewis]

I was able to add a bit of information (in the form of grey lines) to the ambiguous frames where her legs cross each other in order to make it easier to see her spinning in *either* one direction *or* another. Check it out!
Spinning dancer deconstructed
matt

[Factory101]

I was able to add a bit of information (in the form of grey lines) to the ambiguous frames where her legs cross each other in order to make it easier to see her spinning in *either* one direction *or* another. Check it out!
http://www.randominc.net/spinninglady
matt

(Given that: my left eye is receiving information directly from the left image and peripherally from the right image; and vice versa): If I focus my attention directly at the left dancer then both rotate clockwise; if I look directly at the right dancer then both rotate counter-clockwise. If I look peripherally at the left dancer then both rotate counterclockwise; and vice versa.

[DavidMcC]

I was able to add a bit of information (in the form of grey lines) to the ambiguous frames where her legs cross each other in order to make it easier to see her spinning in *either* one direction *or* another. Check it out!
Spinning dancer deconstructed
matt

I assume you're aware that your grey lines negate the illusion completely, because they leave nothing for the visual cortex to invent in order to make 3-D sense of the images.

[Factory101]

I was able to add a bit of information (in the form of grey lines) to the ambiguous frames where her legs cross each other in order to make it easier to see her spinning in *either* one direction *or* another. Check it out!
Spinning dancer deconstructed
matt

I was able to add a bit of information (in the form of grey lines) to the ambiguous frames where her legs cross each other in order to make it easier to see her spinning in *either* one direction *or* another. Check it out!
Spinning dancer deconstructed
matt

I assume you're aware that your grey lines negate the illusion completely, because they leave nothing for the visual cortex to invent in order to make 3-D sense of the images.

mlewis - I notice that your animations are in two separate windows (as opposed to previous ones in the blog that are directly ajoined). Is it possible for you to produce the same separate animation frames but without the grey lines?

If it's not too much trouble this would be of interest, but not if it's too much work for you. Many thanks.