[From Bill Powers (2011.05.09.0820 MDT)]
In the attached article, the
authors say they’re searching for a
“mechanistic model” to explain their results (see the
highlighted sentences
on p. 4664). They even use the term “error signal” (p.
4658).
We’ve got the model.
This stuff all comes out of Samuel’s lab at the Harvard Physics
Department
(sic).
I’m going to write Prof. Samuel and introduce him to PCT. Would anyone
care
to help me? (I’m looking at you, Bill.)
Sure, but you’ll do fine without me. Send him those two PDFs in Warren
Mansell’s post of today. And the other appropriate references. Prof.
Samuel et. al. are looking for input-output relationships, which of
course they will find. They say
"The best understood sensorimotor behaviors in C. elegans are
reflexive avoidance to body touch (Chalfe et al., 1985), its
chemotactic ability to crawl up and down chemical gradients
(Bargmann and Horvitz, 1991), and its thermotactic ability to
descend thermal gradients in what is called cryophilic
movement (Hedgecock and Russell, 1975)."
These are three perceptual-control behaviors: control of the sensation of
touch on the body, control of rate of sensed increase or decrease of
chemical concentration (as in E. coli but more systematic), and control
of sensed temperature (“thermotactic” or warmth-seeking,
“cryophilic” or cold-seeking behavior). In each case, the
behavior used to control a variable is named for the variable being
controlled and its reference level rather than being described in its own
right – turning and swimming, mainly. The “output” in the
input-output relationship is actually a sensory input, a consequence of
behavior rather than an actual behavior.
You may quote this to Prof. Samuel. But prepare to be ignored.
I also attach a paper of mine I found printed out without being able to
find the original file. Finally I scanned it and OCRed it to recreate the
file. Here’s the first paragraph which you may also cite:
Reafference&PCT1.doc (36.5 KB)
···
At 12:01 PM 5/8/2011 -0600, Ted Cloak wrote:
===========================================================================
When E. von Holst, with Horst
Mittelstaedt, studied the subject of reflexes, he arrived at some
conclusions that led to a new concept of how they work. The key insight
came from observing that reflexes were not simple brief input-output
processes, but arose from continuous activity in the nervous system.
“A continuous stream of impulses links the higher and lower centers
even when there is external motor inactivity.” (all citations from
Von Holst, 1950). Furthermore, he noticed a paradox. Postural reflexes
were then thought to maintain “normal postures” by reflex
actions that corrected deviations from normal. However, von Holst had
observed that fish and other creatures could maintain not only normal
postures, but abnormal ones, such as a fish orienting its body vertically
or on its side rather than in the more usual horizontal configuration.
Not only could they do that, but when disturbances caused deviations from
the abnormal posture, the muscles acted reflexively to restore not the
normal posture but the abnormal one. How a simple reflex could act this
way was a mystery.
=========================================================
I doubt that C. Elegans has many levels of control, so it’s not
likely to maintain variable reference conditions, but the PCT description
certainly works and doesn’t require describing behavior in non-behavioral
terms.
Best,
Bill P.