We have two population of neurons, population A and population B which have been wired up so that an X stimulus pokes population A neurons more than population B neurons, while a not-X stimulus pokes population B more than population A. But both stimuli poke both populations and there is plenty of noise in the system. Following our first example, the population A neurons might have been wired to the red cones on a retina and the population B neurons might have been wired to the green cones.
An active neuron will fire if it is poked. But if it is poked a lot it needs to take a rest and becomes inactive for a while.
Judgement is the observation of the ratio of A firing to B firing. If A/B is greater than H (H for high) then the subjective, conscious judgement is X. If A/B is less than L (L for low) then the judgement is not-X. Otherwise the judgement is don’t know, neutral.
If we have a lot of X stimulus in a short period of time, then a lot of the A neurons will become inactive for a while and will not fire in the meantime. This will tend to depress the ratio A/B, pushing the judgement in the not-X direction.
If there is no stimulus for a while, then all the neurons will become active again and judgement will return to normal.
Resulting in a system which is sometimes better at detecting a change in something rather than its level.
With a mechanism of this sort, according to reference 1, being the cause of various curious errors of perception.
Numbers
Let XS be an X stimulus. This pokes FA(XS) A neurons and FB(XS) B neurons. FA(XS) is significantly bigger than FB(XS). Say TR=FA(XS)/FB(XS)>H.
Let AC(A) and AC(B) be the proportions of active neurons at the time in question. Functions of firing activity in the recent past.
Then FA(XS)*AC(A) A neurons will fire and FB(XS)*AC(B) B neurons will fire.
Then the judgement ratio JR=FA(XS)*AC(A)/ FB(XS)*AC(B) = (FA(XS)/ FB(XS))*( AC(A)/AC(B)) will differ from the true ratio TR=FA(XS)/ FB(XS) to the extent that the activity ratio AC(A)/AC(B) differs from one. We might even have it that JR<L.
It would not be a big job to model all this in time, in Visual Basic, with the Visual Basic which comes with MS Excel being quite up to this sort of thing. With the snap above showing work in progress, having done enough to show the effect on the perception of the stimulus history. No need to crank up MATLAB again quite yet!
I associate to the lags and leads in models of economic and social systems and all the often untoward complications that result from them. For example, think of what happens if the public's perception of what ought to raised in tax and spent on the police is the aggregate of their experience of crime over the last twenty years. Assuming that spending on police is a reliable way to get crime down, I dare say this will give one some nice sine waves.
A rather different example from a concert hall
The concert hall being Milton Court at the Guildhall School of Music, next door to the Barbican Centre at Silk Street. A concert hall we visited yesterday for a Biss-Padmore Schubert concert.
We were sitting in seats F19 and 20, that is to say bottom left in the snap above and a few seat in from the left-hand aisle.
For a short while during the first half, during the D959 piano sonata, I became conscious of the vertical stripes of wood making up the background throbbing, pulsing a bit, perhaps best described as irregular waves travelling slowly up and down what were supposed to be the joins between successive planks.
Waves which I have tried to suggest in the sketch above, with the waves along successive joins being, in my recollection anyway, roughly 180° out of phase.
Is this a manifestation of the adaptation discussed above?
I remember something similar happening once, late at night, at a suburban railway station, where I was waiting for a train having taken drink. Not unlike that snapped above, with the difference that the planks were running the other way, along the platform rather than across it. Perhaps it was just a shelter, rather than an entire canopy.
Sitting back on a bench, something like the one middle left and looking up, idly, at the canopy above, I seemed to lose vertical hold, with the whole array of planks drifting smoothly backwards and forwards, up and down.
Presumably there is a degree of top-down perception here, with the brain knowing or believing that the stripes are stationary and pushing that knowledge or belief into the perception, but a pushing which can be weakened by tiredness or drink. Perhaps one can have such an illusion while drifting into or out of sleep, although I have never sought or noticed such a thing.
References
Reference 1: Color adaptation induced from linguistic description of color - Liling Zheng, Ping Huang, Xiao Zhong, Tianfeng Li, Lei Mo – 2017.
No comments:
Post a Comment