Sensory Adaptation and the Relation between Qi and p

[From Bruce Abbott (2017.02.25.1010 EST)]

In my previous post [Bruce Abbott (2017.02.24.1220 EST)] I described how adaptation of sensory receptors to ambient conditions changes the relationship between the environmental controlled variable (Qi) and its perception, p. Adapting the temperature sensors in one hand to cold water and those in the other to hot water gives rise to the surprising perception of two different water temperatures when both hands are plunged into water of an intermediate temperature: The hand previously adapted to cold water now feels warm and the one adapted to hot water now feels hot.

Another example we all have experienced is adaptation to darkness. A change in illumination from bright to dark brings about a gradual increase in the sensitivity of the photoreceptors in the retina. Consequently the threshold intensity for perceiving the light decreases over time following a transition from light to darkness. These changes can be observed experimentally by giving a participant control over the intensity of a light-source. The participant is asked to adjust its intensity until the light is just barely perceptible. In this way the threshold is sampled after various intervals of time spent in darkness. Such experiments demonstrate that the sensitivity of the rods (the photoreceptors that give us night vision) increases up to a million-fold after a change from bright light to complete darkness.

To my mind, the logical way to represent such changes in a control-system diagram is as a change in the input function that converts Qi to p. To conversion formula would compute p from Qi as a dynamic function of time and the recent variations in light intensity to which the eye was being exposed. In this system, Qi remains the intensity of the light – out there in the physical environment – and p is the perception of the light’s intensity.

Rick Marken (2017.02.24.2100 PST) offers an alternative proposal. According to Rick, light sensitivity changes should be viewed as producing a change in Qi. In Rick’s view, Qi is not the physical intensity of light but the physical equivalent of p, obtained by solving the input formula for Qi and plugging in the value of p to obtain the equivalent Qi. According to Rick, Qi is not the intensity of the light, but the adapted intensity of the light.

What Rick’s solution does is to make Qi the combination of environmental variables that is equivalent to the perception. The combinatory rule is not present in the environment, however, but in the system’s input function. This makes Qi equivalent to what the observer would infer is being controlled by the system.

I think it makes more sense to think of Qi as representing one or more physical variables out there in the environment (like light intensity), where they can be observed and measured, and letting the system’s input function determine how these variables get transformed into a perceptual-signal representation. That way, what changes during sensory adaptation is one’s perception, and nothing changes in the environment. Adaptation occurs in the sensitivity of the receptors as computed in the input function – in the gain – and this changes how p relates to these input quantities.

However one prefers to describe and label the input variables and Qi, the same mathematical transformations occur around the loop and the system behaves in exactly the same way. When discussing how sensory adaptation affects the perception of input variables such as skin temperature or light intensity, I think it makes for a clearer discussion to treat Qi as the environmental variable, p as its perception as determined by the input function, and adaptation as involving dynamic changes in the input function. That way it is one’s perception of the environmental input that changes during adaptation, not Qi.

Bruce

[From Bruce Abbott (2017.02.25.1010 EST)]

      I think it makes more sense to think of Qi

as representing one or more physical variables out there in
the environment (like light intensity), where they can be
observed and measured, and letting the system’s input function
determine how these variables get transformed into a
perceptual-signal representation. That way, what changes
during sensory adaptation is one’s perception, and nothing
changes in the environment. Adaptation occurs in the
sensitivity of the receptors as computed in the input function
– in the gain – and this changes how p relates to these input
quantities.

      However one prefers to describe and label

the input variables and Qi, the same mathematical
transformations occur around the loop and the system behaves
in exactly the same way. When discussing how sensory
adaptation affects the perception of input variables such as
skin temperature or light intensity, I think it makes for a
clearer discussion to treat Qi as the environmental variable,
p as its perception as determined by the input function, and
adaptation as involving dynamic changes in the input
function. That way it is one’s perception of the
environmental input that changes during adaptation, not Qi.

Bruce

[From Bruce Abbott (2007.02.11.1130 EST)]

[Martin Taylor 2017.02.26.11.05] –

[From Bruce Abbott (2017.02.25.1010 EST)]

I think it makes more sense to think of Qi as representing one or more physical variables out there in the environment (like light intensity), where they can be observed and measured, and letting the system’s input function determine how these variables get transformed into a perceptual-signal representation. That way, what changes during sensory adaptation is one’s perception, and nothing changes in the environment. Adaptation occurs in the sensitivity of the receptors as computed in the input function – in the gain – and this changes how p relates to these input quantities.

However one prefers to describe and label the input variables and Qi, the same mathematical transformations occur around the loop and the system behaves in exactly the same way. When discussing how sensory adaptation affects the perception of input variables such as skin temperature or light intensity, I think it makes for a clearer discussion to treat Qi as the environmental variable, p as its perception as determined by the input function, and adaptation as involving dynamic changes in the input function. That way it is one’s perception of the environmental input that changes during adaptation, not Qi.

Bruce

MT: Yes. But remember that in the new PCT, Qi is defined as not being in the environment, because the environment consists of only the unknown entities that directly influence the individual sensors. You are talking old (Powers version) PCT, so you and Rick are talking at cross purposes. I don’t think that’s likely to result in agreement.

BA: Ah, give me that old tyme religion . . .

Bruce

[From Rick Marken (2017.02.25.1130)]

···

Bruce Abbott (2017.02.25.1010 EST)–

Â

BA: To my mind, the logical way to represent such changes in a control-system diagram is as a change in the input function that converts Qi to p.Â

RM: You are assuming that Qi is an environmental variable, like light intensity. But Qi is a function of environmental variables (v’s) , like light intensity; Qi = f(v1, v2…vn). It is presumably the same function of environmental variables as is the theoretical variable, p. In your example, p is likely to be a function of just one environmental variable, light intensity. So Qi is the same function of that variable as is p, the function being the light adaptation effect.Â

BA: To conversion formula would compute p from Qi as a dynamic function of time and the recent variations in light intensity to which the eye was being exposed.Â

RM: This is the same as my proposal except that the dynamic function, f(), would be applied to the same physical variable to get both Qi and p. That is, Qi = f(v) and p = f(v). This would be clear to you if you actually ran a control task where a person was to keep the brightness of a target object matching that of a comparison object after varying durations of light adaptation. What you would observe is that the light intensity at which the target object  is controlled varies during the post adaptation period. This would show that the controlled variable, Qi, is not simply proportional to the light intennsity (v) of the target object. That is, Qi <> k*v. This would show us that Qi is a dynamic function of v; Qi = f(v).Â

Â

BA: In this system, Qi remains the intensity of the light – out there in the physical environment – and p is the percerception of the light’s intensity.

RM: I think the problem is, again, that you are taking a “theory first” approach to this. Remember, Qi is the controlled variable as perceived by an observer. It is the behavior of Qi that the observer is trying to explain with the control model; the behavior of Qi being the fact that it is controlled. So a “phenomena first” approach to the adaptation effect would result in studies, like the one I described above, where it would become clear to the researcher (using the TCV) that Qi – the variable being controlled by the control system – is not proportional to light intensity but, rather, is a dynamic function of light intensity. That is, Qi is f(v) so a model of the behavior in this task must have p = f(v).Â

RM: I don’t like to blow my own horn (well, not as much as our so-called president does) but I really think it’s important, if you want to get PCT right, to take Bill’s advice (that I posted a couple months ago):Â

Date: Sat, 4 May 91 09:55:17 -0500
[From Bill Powers]
Â
BP: Phenomena first. Theory second. Hearken to Marken.

BestÂ

Rick


Richard S. MarkenÂ

"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery