Conflict (was Re: Good Corporate Citizen

[From
Bjorn Simonsen (2005.10.28,22:25 EuST)]

[From
Rick Marken (2005.10.28.1000)]

Bjorn Simonsen
(2005.10.28,15:30 EuST) –

From Rick Marken (2005.10.27.0900)

A person can be tolerant with respect to some goals and intolerant

with respect to others. In PCT terms, this means that a person

controls for some perceptions with low gain and for others with

high gain, respectively.

Could it also be with respect to the reference value? A person controls for

some perceptions with a not so high reference value for some goals and with a

high reference value for other goals.

Yes,
I think this could work if the reference were set to zero, which in a

one
way control system would mean that the system is no longer controlling

for
that variable when there is no perception of that variable. So setting

the
reference to zero would be like ceasing to control the conflicted

variable,
which is equivalent to leaving the conflict.

Agree,
but original you talked about tolerance. And my question was about tolerance.

Can we
without more ado use the same model for conflict and tolerance? Isn’t there a
difference between “tolerant – intolerant” and
“not conflict – conflict”?

But, in
general, changing the reference signal from high to low will not

reduce
conflict (at least, in terms of the PCT model of conflict). This is

because
the value of the reference signal just specifies the intended state

of
the controlled variable; it doesn’t determine how hard the system will

work
to get that variable to the reference.

Agree,
but will not changing the reference signal from high to low reduce “the degree
you tolerate another man”? Will not a reference signal near zero specify that
the intended state of tolerance is near zero?

I
think the easiest and clearest way to think about this is in terms of the

rubber band
demo. Assume that there are two different
people, each with

their
finger in a different loop of the knotted rubber bands and each trying

to
keep the knot over different dots (targets). The value of the reference

signal
in each person determines where each person wants to keep the knot –

the
distance of the knot relative to the target dot. Assume two way control

systems
so a zero reference setting results in error (and, thus, action)

when
the knot is some distance in either direction from the target. The

result
will be conflict, with each person pulling their rubber band end in

opposite
directions to oppose each other’s disturbance, regardless of the

setting
of each person’s reference (as long as the references specify even

slightly
different reference positions for the knot).

I am
sorry. If this is a good explanation, I am very slow.

You say:
“The value of the reference signal in each person determines where each person
wants to keep the knot – the distance of the knot relative to the target dot”.

The way I
understand this is that there is a dot. One person will keep the knot 3 o’cloc
– 5 cm. This is a certain distance from the dot and you say that this distance
constitutes the reference value. The
other person will keep the knot on the dot. The distance from the dot is zero.
The reference value is zero.

I neither
see the problem that the one with reference zero would not be able to control
his knot on dot as well as the other knot on 3 o’clock – 5 cm.

I would understood
it better if there were two different rubber band tied together, one strong
rubber band and one weak rubber band. Then the one with his finger in the
strong rubber band would not need to move as much as the other, because the
gain was greater in the output function(?).

The intensity
of the conflict, at least in terms of the initial size of the

response
to disturbance, is determined by the size of the difference

between
the references setting in the two parties; the greater the

difference
the more intense the initial response to disturbance because the

error
signal in one or both parties start out very large.

Is it correct to say “The intensity of the conflict is determined by
the size of the difference
between the reference settings in the two parties;…” Isn’t it enough to say
“The conflict is determined by the size of ……”. “The greater the difference,
the greater conflict”.

Bjorn

[From Rick Marken (2005.10.28.1440)]

Bjorn Simonsen (2005.10.28,22:25 EuST)

Isn�t there a difference between �tolerant � intolerant� and� �not conflict �
conflict�?

Yes. I think you can be tolerant (or intolerant) whether or not you are in a
conflict. To me, tolerance just means putting up with error in a control
system, whether that error is the result of an active attempt by someone to
change the state of something you care about or whether it is a passive
result of a disturbance. For example, you can be tolerant or intolerant of
missionaries who are trying to convert you to their belief (a conflict) and
you can be tolerant or intolerant of a person who accidentally steps on your
foot in a crowd (no conflict). The tolerant person listens to the
missionaries, thanks them and bids them adieu. The intolerant person starts
yelling at the missionaries and throws them out the door. The tolerant
person grimaces and nods politely to the foot stomper; the intolerant person
replies with a swift jab to the stomper's kidneys.

Agree, but will not changing the reference signal from high to low reduce �the
degree you tolerate another man�? Will not a reference signal near zero
specify that the intended state of tolerance is near zero?

Not the way I see tolerance. Tolerance, to me, is related to the gain of a
control system. A low gain control system is tolerant, a high gain one is
not. The setting of the reference is irrelevant to tolerance, except to the
extent that a zero setting completely removes the person from the error
creating situation by stopping control of the error creating perception.

You say: �The value of the reference signal in each person determines where
each person wants to keep the knot -- the distance of the knot relative to the
target dot�.
...
I neither see the problem that the one with reference zero would not be able
to control his knot on dot as well as the other knot on 3 o�clock � 5 cm...

Maybe I didn't explain it clearly. I think if you just loop two rubber bands
together and ask a friend to keep the knot on a dot on the table while you
try to keep the same knot over another nearby dot, you will see the problem
pretty quickly.

I would understood it better if there were two different rubber band tied
together, one strong rubber band and one weak rubber band. Then the one with
his finger in the strong rubber band would not need to move as much as the
other, because the gain was greater in the output function(?).

Yes, the gains of the two control loops would be different in this case but
the conflict would still exist even if the two rubber bands were equal in
size.

Is it correct to say �The intensity of the conflict is determined by the size
of the _difference_ between the reference settings in the two parties;..�
Isn�t it enough to say �The conflict is determined by the size of ���. �The
greater the difference, the greater conflict�.

I don't think so. When you do the rubber band demo of conflict, try
increasing the distance between the two target dots. I think you'll see that
the size of the distance between the dots, which corresponds to the size of
the difference between the references in each person, makes a difference in
terms of how quickly the conflict accelerates.

You can also demonstrate tolerance with the rubber bands. Just try to be a
little more easy-going about where the knot is relative to your target.
You'll find that you are still in conflict with your friend but you're
willing to put up with more error so that the conflict doesn't escalate.
This works best if the target dots are not too far apart. Putting the
targets way far apart will create error that would test anyone's
tolerance;-)

Best regards

Rick

···

--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400

--------------------

This email message is for the sole use of the intended recipient(s) and
may contain privileged information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply email and destroy all copies
of the original message.

[From Bjorn Simonsen
(2005.10.30,09:15 EuST)]

[From Rick Marken
(2005.10.28.1440)]

Bjorn
Simonsen (2005.10.28,22:25 EuST)

Isn’t there a difference
between “tolerant ­ intolerant” and “³not conflict ­

conflict”?

Yes. I think you can be tolerant
(or intolerant) whether or not you are in a

conflict. To me, tolerance
just means putting up with error in a control

system, whether that error is
the result of an active attempt by someone to

change the state of something
you care about or whether it is a passive

result of a disturbance.

It’s the first time I have seen
such a technical definition for the concept “tolerance”. I liked it. But I
tried to use your definition on two examples.

Here is the first. “I wish to
perceive that honesty is a policy for all my control of perceptions”.

If the gain is high in this
output function (and lower levels output functions ?), it will send a high
reference value to the Program level. I will then have strong program in my
circle of acquaintances having beliefs or practices differing
from or conflicting with my own. Her I agree.

Here is the other
example. “I wish to perceive that law-abidingness is a policy for my control of
perceptions”.

If the gain is high
in this output function, I will send a high reference value to the Program
level. I will have a strong program that chastises offenders. This is not
“sympathy or indulgence for beliefs or practices differing from or conflicting
with one’s own”. This is not tolerance.

If High Tolerance
is a Relationship perception and Low Tolerance is another Relationship
perception, this definition will account for my two examples.

What do you say?

Not the way I see tolerance.
Tolerance, to me, is related to the gain of a

control system. A low gain
control system is tolerant, a high gain one is

not. The setting of the
reference is irrelevant to tolerance, except to the

extent that a zero setting
completely removes the person from the error

creating situation by
stopping control of the error creating perception.

If you have a low gain at the
principle level “I wish to perceive that honesty is a policy for all my control
of perceptions”, you will get programs executing this control with lower
references. Is this Tolerance?

Bjorn

Here is the first. �I wish to
perceive that honesty is a policy for all my control of
perceptions�.

If the gain is high in this output function (and lower levels output
functions ?), it will send a high reference value to the Program
level.
[From Bill Powers (2005.10.30.0759 MDT)]
Bjorn Simonsen
(2005.10.30,09:15 EuST) –
You’re leaving something out here. High gain does not mean sending a high
reference value to the program level, because the output is the error
signal times the output gain. If the error is zero, the output will be
zero no matter what the gain is. If the error is positive, the output
will be positive and send a positive reference level to lower systems. If
the error is negative, the output will be negative and send a negative
reference level to lower systems. Note that in a model constrained by
properties of neurons, we would have to says “sends a high reference
reference signal to positive-acting lower systems,” or “sends a
high reference signal to negative-acting lower systems.”
The effect of increasing the output gain in a control system is not to
produce significantly more output, but to make the error in that control
system smaller. Were you able to run that “live block diagram”
program I sent you? A high gain in a control system means that it will
produce more output for the same amount of error signal, but that
will have the effect of making the error signal smaller, which will
reduce the output. The output will end up only a little bit larger than
before if the loop gain is (for example) 10 or more.

You have to reduce the gain of an output function to a very small value
before the result will be a large change in output. By that time, the
error signal will be perhaps five or ten times its normal size. Check it
out using the model.

Best,

Bill P.

[From Bjorn Simonsen (2005.10.30,21:30 EuST)]

[From Bill Powers (2005.10.30.0759 MDT)]

Bjorn Simonsen (2005.10.30,09:15 EuST) –

Here is the first. “I
wish to perceive that honesty is a policy f

or all my control of perceptions”.

If the gain is high in this output function (and lower levels

output functions ?), it will send a high reference value to the Program
level.

You’re leaving something out here. High gain does
not mean

sending a high reference value to the program level, because

the output is the error signal times the output gain. If the error

is zero, the output will be zero no matter what the gain is.

Maybe I did. But so far this is as also I think.

If the error is positive, the output will be positive and send a

positive reference level to lower systems.

Still as I think.

If the error is negative, the output will be negative and send a

negative reference level to lower systems. Note that in a model

constrained by properties of neurons, we would have to says

"sends a high reference reference signal to positive-acting lower

systems," or "sends a high reference signal to negative-acting

lower systems."

Here I would express myself different. But I am open for your way of
thinking. I see you and Rick are right. But let me explain my thinking. Maybe
it is still wrong.

The error is an error signal with a frequency like the value of the
error. A frequency can’t be negative. Therefore I don’t think upon a negative
error.

I remember your BCP (second ed.) page 83. “The reference signal is
actually a composite of many converging neural currents, some negative and some
positive, but since the other input to the neuron is inhibitory, there can be
no effect on the motor neuron unless the reference current is positive.”

I remember also
your earlier teaching: “Positive
feedback doesn’t require that perceptual signals be positive. A comparator can
work equally well if the reference signal is inhibitory (-) and the perceptual
signal is excitatory (+) at the comparator.[From Bill Powers (2005.02.26,1201
MST]

Therefore
I picture to myself that the composite effect of many converging neural
currents has a positive value, but it has an inhibitory effect because the
comparator where this output signal/reference signal goes is a Renshaw cell (Here
the perceptual signal must have an excitatory (+) effect).

Now back to my >>Bjorn
Simonsen (2005.10.30,09:15 EuST) –

Here is the other example. “I wish to
perceive that

law-abidingness is a policy for my control of perceptions”.

If the gain is high in this output
function, I will send a high

reference value to the Program level. I will have a strong

program that chastises offenders. This is not “sympathy or

indulgence for beliefs or practices differing from or conflicting

with one’s own”. This is not tolerance.

If the reference signal has an inhibitory
effect to the Program level it will have not be a strong Program that
chastises offenders. This is Tolerance.

The effect of increasing the output gain in a
control system is

not to produce significantly more output, but to make the

error in that control system smaller.

Yes because the
reference signal has an inhibitory effect.

Were you able
to run that “live block diagram” program I sent you?

Yes a wonderful
program, I am working to make an equivalent in PowerSim. Is that OK?

……. Check it
out using the model.

Yes I will.

Am I correct when I
say Rick’s thinking

[From Rick Marken
(2005.10.28.1440)]

“ Not the
way I see tolerance. Tolerance, to me, is related to the gain of a

control system. A low gain
control system is tolerant, a high gain one is

not. The setting of the reference
is irrelevant to tolerance, except to the

extent that a zero setting
completely removes the person from the error

creating situation by stopping
control of the error creating perception”

is correct if we think the
reference signal to the level below has an inhibitory effect?

I have studied your
¨Emotions” in second ed. BCP this week-end. Interesting!

Bjorn

[From Rick Marken (2005.10.30.1330)]

Bjorn Simonsen (2005.10.30,09:15 EuST)

Rick Marken (2005.10.28.1440)

Yes. I think you can be tolerant (or intolerant) whether or not you are in a
conflict. To me, tolerance just means putting up with error in a control
system, whether that error is the result of an active attempt by someone to
change the state of something you care about or whether it is a passive
result of a disturbance.

It’s the first time I have seen such a technical definition for the concept “tolerance”. I liked it. But I tried to use your definition on two examples.

Here is the first. “I wish to perceive that honesty is a policy for all my control of perceptions”.

If the gain is high in this output function (and lower levels output functions ?), it will send a high reference value to the Program level. I will then have strong program in my circle of acquaintances havingbeliefs or practices differing from or conflicting with my own. Her I agree.

Here is the other example. “I wish to perceive that law-abidingness is a policy for my control of perceptions”.

If the gain is high in this output function, I will send a high reference value to the Program level. I will have a strong program that chastises offenders. This is not “sympathy or indulgence for beliefs or practices differing from or conflicting with one's own”. This is not tolerance.

I'm sorry, I just don't see how those examples relate to my idea of tolerance.

If High Tolerance is a Relationship perception and Low Tolerance is another Relationship perception, this definition will account for my two examples.

I agree that we perceive different degrees of tolerance. But what are we perceiving when we perceive different levels of tolerance? I think what we perceive is the degree to which people (or ourselves) are putting up with things that we know they (or we) can't stand. In PCT, something we can't stand is a perception that doesn't match the reference for that perception. Putting up with something we can't stand is putting up with error. So when we see people (including ourselves) being tolerant (like when we see Christians who put up Jews, by not trying to convert of kill them) we are seeing control systems that are intentionally controlling poorly; that are putting up with error.

Control systems that control poorly are low gain control systems. So in order to control for the perception of being tolerant a control system has to be willing to lower the gain on the systems that are controlling for the perception that is being "tolerated". So, for example, the tolerant Christian has to lower the gain on the system that wants to see people accepting Jesus as their savior. When a Jew comes along and says Jesus is not my savior this control system will experience error. The natural inclination is to remove the error by bringing the perception to its reference -- by converting or eliminating the Jew. But the tolerant Christian has lowered the gain on the system that wants to see people accepting Jesus as their savior, so there is no action to reduce the error. The tolerant Christian will put up with the error created by people who don't accept Jesus as their savior. There will be error tolerated in one control system (the one controlling for people accepting Jesus as their saviour) in order to reduce the error in another (the one controlling for being tolerant).

Best

Rick

···

---
Richard S. Marken
marken@mindreadings.com
Home 310 474-0313
Cell 310 729-1400

[From Bill Powers (2005.09.30.1440 MDT)]

The error is an error signal
with a frequency like the value of the error. A frequency can�t be
negative. Therefore I don�t think upon a negative
error.

If you stick with a strict neural model, you have to double the number of
control systems involved with any controlled variable. One control system
would compare an inhibitory perceptual signal (reaching the comparator
through a Renshaw cell) and an excitatory reference signal, so the error
signal would indicate a perceptual signal that is too small. The other
control system would compare the same perceptual signal P(without a
Renshaw cell) with an inhibitory reference signal (which now involves a
Renshaw cell), so that error signals would indicate a perceptual signal
that is too large. The error signal representing perceptions that
are too large would drive output functions that make the perceptual
signal smaller. The other error signal would drive output functions that
make the perceptual signal larger. This gives us one comparator that
outputs an error signal to one output function when the perceptual signal
is greater than the reference signal, and a second comparator that
outputs an error signal to a different output function, whose external
effects are opposite to the first one, when the perceptual. signal is
smaller than the reference signal.

When you consider these two sets of comparators and output functions
together, they form a single bi-directional comparator and output
function that can correct both positive and negative errors, measured in
terms of the controlled variable. I prefer to treat this as the general
case because it covers many situations and requires only a single
bidirectional control model.

In some cases, control is only in one direction – the controller reacts
only when the perceptual signal is too small, or only when it is too
large, with the opposite case being ignored. I think these are much less
common cases.

I ask again – did you get the “live block diagram” program I
sent you with the book chapters? This will answer your questions about
the effects of output gain.

Best,

Bill P.

From [Marc Abrams (2005.10.30.1910)

In a message dated 10/30/2005 4:58:54 P.M. Eastern Standard Time, powers_w@FRONTIER.NET writes:

[From Bill Powers (2005.09.30.1440 MDT)]
An interesting post and like Rick, Bill introduces a biological entity, the Renshaw cell into an argument and tries to shoe horn it into a PCT structure. Here I present a counter argument. Please notice the references at the end, Motor control has been understood for a very long period of time, long before the birth of PCT.

Why hasn’t Bill cited some of these authors, especially Stark and Granit?

Copyright Notice:

First published [v1.0] 08:28 BST 1st April 2003

Earlier versions of this material appeared in

Biological Cybernetics This material was written and published in Wales by Derek J. Smith (Chartered Engineer). It forms part of a multifile e-learning resource, and subject only to acknowledging Derek J. Smith’s rights under international copyright law to be identified as author may be freely downloaded and printed off in single complete copies solely for the purposes of private study and/or review. Commercial exploitation rights are reserved. The remote hyperlinks have been selected for the academic appropriacy of their contents; they were free of offensive and litigious content when selected, and will be periodically checked to have remained so. Copyright © 2003, Derek J. Smith (Chartered Engineer).Smith (1997; Chapter 4). It is simplified here and supported with hyperlinks.
** Although this paper is reasonably self-contained, it is best read as a extension of our paper on Control Theory. To go directly to the earlier paper,** click here** , and to see the author’s homepage,** click here.

1 - Biological Cybernetics in General

···

We turn now to the topic of control in biological systems, that is to say, to the search for the control hierarchies and servomechanisms in the body’s skeletomuscular system. Wiener himself, in the years following the publication of “Cybernetics”, was personally active in this very field. As early as 1948 he was discussing feedback circuits in the body’s motor control systems, concluding that the complexity of biological control processes demanded a “multiplicity of feedback”, very little of which ever became conscious (Wiener, 1948). During the 1950s and 1960s, another pioneer, Larry Stark, made reasonable progress analysing the neural circuitry controlling the intrinsic and extrinsic muscles of the eyes (eg. Stark, 1968), and Ragnar Granit analysed the role of alpha and gamma motor neurons in skeletal muscle control (eg. Granit, 1970; see Sections 2 and 3 below). Walter (1961) even went so far as to accord the phenomenon of biological thermostasis - that is to say, homeostasis of body temperature - the greatest evolutionary significance. It was, he argued, a supreme event in natural history, because it made possible the survival of mammals on a cooling globe. With similar enthusiasm, Bennett (1979) has called feedback “the twentieth-century metaphor”, and Gosling (1994, p7) calls control theory “the art of being and doing”. Nevertheless, real advances have actually been few and far between. By and large, cybernetic analyses of biological systems have been too simplistic, and have failed more or less totally to explain the “vital spark” of life, as manifested in such unpredictable behaviours as play, curiosity, and creativity, and (above all) in the phenomenon of consciousness. Here, with an inbuilt glossary of terms, are some of the success stories.

2 - Alpha-Gamma Muscle Control


Granit (1970) has shown that the body’s skeletomuscular system includes sophisticated spinal level control technology. The alpha-gamma muscle control system is a good example of this. This is a major biological servosystem involving widespread structures of the central and peripheral nervous systems, and responsible for managing posture and locomotion. Here is the basic vocabulary:

Mini Glossary - Muscles and Muscle Control Systems

Action Potential, Muscle: When a neuromuscular junction is repeatedly stimulated by incoming neural action potentials, the end-plate potentials eventually exceed the threshold for an action potential to be initiated within the muscle fibre instead. The active ions in this new type of action potential are calcium ions instead of sodium, and the resulting calcium ion influx is the key enabling factor in excitation-contraction coupling.

Alpha-Gamma Control System: An important biological servosystem involving structures of the central and peripheral nervous systems. Helps to manage bodily posture and locomotion from the level of the spinal segment, thus (a) making for very rapid muscular response when problems are encountered, and (b) generally reducing neural traffic up and down the spinal cord. (See alpha motor neuron and pyramidal tract.)

Alpha Motor Neurons: These are large lower motor neurons, situated in the ventral spinal grey. Their axons form the bulk of the motor root of the spinal nerve.

Annulospiral Endings: These are the sensory receptors of the intrafusal muscle fibres. They are wound spirally around the belly of each such muscle fibre, and are thus admirably placed to detect the “fatness” of that fibre at any moment in time. Their sensory information is conveyed back to the spinal segment by the sensory branch of the spinal nerve.

Antagonism: Type of muscle organisation where muscles groups oppose each other in moving a limb. When controlling such muscles it is necessary for one group to relax while the other is pulling. (See extensor muscle and flexor muscle.)

Corticobulbar Tracts: These are fibres derived from upper motor neurons whose lower motor neurons are situated in the cranial nerve nuclei of the medulla. (Compare corticospinal fibres.)

Corticospinal Tracts: These are fibres derived from upper motor neurons whose lower motor neurons are situated in the spinal grey. (Compare corticobulbar fibres.) This is therefore the major class of fibres found in the pyramidal tract.

Electro-Myograph (EMG): Electromyography is the technique of detecting the synchronised discharge of all the muscle fibres in a motor unit. This is made possible by the fact that this discharge creates an electrostatic field which can be picked up from some distance away by either an implanted or a skin surface electrode (the former being very sensitive, the latter less so but quicker and more convenient to apply). EMGs are much used in research and clinical medicine, as well as being available in kit form for personal amusement and biofeedback.

End-Plate Potential (EPP): The potential in a motor end-plate following the arrival of a neural action potential**.** This is not yet a muscular action potential, rather it is the equivalent to the EPSP in neurotransmission**.** (It is “graded” activity in other words, rather than all-or-none.) Typically, the EPP takes longer to decay (ca 10 msec) than do the traces of the acetylcholine which caused it (3 msec). It may therefore require the temporal summation of a succession of incoming impulses to produce a full muscular action potential.

Excitation-Contraction Coupling: Term used to describe the link between the action potential of muscle fibres, and the contraction of the myofibrils which then follows. This phenomenon results from sensitivity of the protein filaments within the sarcomere to a sudden influx of calcium ions.

Extensor Muscle: In an antagonistic muscle pair, the one which straightens the limb in question rather than bending it. (Compare flexor muscle.)

Extrafusal Muscle Fibres: These are the main muscle fibres making up a skeletal muscle. They provide the main contractile power for bodily posture and locomotion.

Extrapyramidal Tract: This is the ancillary spinal motor tract. It arises in a variety of basal ganglia, midbrain, and hindbrain locations, descends more ventrally in the spinal white than does the pyramidal tract, and carries involuntary muscle instructions to the gamma motor neurons.

Fibre Bundle (Muscle): The first level of muscle organisation above the muscle fibre. Consists of bundles of individual muscle fibres. These bundles are sometimes referred to as fasciculi, and consist of individual muscle fibres encased within a layer of connective tissue called the perimyseum.

Flexor Muscle: In an antagonistic muscle pair, the one which bends the limb in question rather than straightening it. (Compare extensor muscle.)

Gamma Motor Neurons: These are small lower motor neurons, situated in the ventral spinal grey. Their axons join those of the alpha motor neurons to form the motor root of the spinal nerve.

Golgi Tendon Organs: These are the sensory receptors of the muscle tendons. They detect the amount of tension in the tendons, and therefore of the entire muscle. Their sensory information is conveyed back to the spinal segment by the sensory branch of the spinal nerve. Excitation of the Golgi tendon organs excites inhibitory interneurons
within the spinal grey which inhibit the alpha motor neurons which caused the muscle contractions in the first place. In the absence of any other effect, therefore, muscle stretch inhibits itself, thus serving as a valuable safety mechanism against cramp, tetany, etc.

Interneuron: One of several neurons in a neuronal circuit, usually relatively small and with only a short axonal travel. Frequently inhibitory by nature, so that they act as negative feedback devices (as, for example, with Renshaw cells). Inhibitory interneurons in the spinal grey are involved in the alpha-gamma control system, where they are responsible for reducing the output from a given alpha motor neuron
whenever the muscle being controlled by that neuron is determined to be contracting too vigorously.

Intrafusal Muscle Fibres: An individual muscle fibre which is smaller and thinner than normal. These are small muscle fibres contained in the muscle spindles, and wound about by the annulospiral endings. These fibres are excited by the gamma motor neurons, and respond via the annulospiral endings. Functionally, these are the control fibres, whereas the extrafusal muscle fibres are the strength fibres. See under muscle spindle for mode of operation. Also optionally known as spindle fibres.

Lower Motor Neuron: Synonym for alpha motor neuron. (Compare upper motor neuron.)

Motor End-Plate: The motor axons of a spinal nerve branch into numerous telodendria as they approach the muscle they innervate. Where a single telodendrion touches the muscle fibre it flattens out to make a better contact. This enlargement is known as the motor end-plate, or neuromuscular junction. It is thus the nerve-muscle equivalent of a synaptic button.

Motor Root: One of the neural fibre bundles which emerge from the ventral horn of a given spinal segment and make up the motor element of that segment’s spinal nerve. (Compare sensory root.)

Motor Unit: The bundle of muscle fibres innervated by a single motor end-plate. Can be of widely differing size: in eye muscles there are seven muscle fibres per nerve fibre, but in leg muscles this rises to 1700 muscle fibres per nerve fibre.

Muscle Contraction, Overview: There is a six-stage sequence of events during the initiation of a striate muscle contraction. These are (a) arrival of the neural action potential at the neuromuscular junction, (b) release of the acetylcholine neurotransmitter, (c) creation of the partial depolarisation in the muscle fibre (the end-plate potential), (d) triggering of the muscular action potential, (e) triggering of the calcium ion action potential within the tubule membranes of the sarcoplasmic reticulum, and (f) contraction of the sarcomeres making up each myofibril.

Muscle Fibres: Muscle cells. Striate muscle fibres are of two types. Red muscle fibres contain plentiful supplies of myoglobin, a variant of haemoglobin, because they have to satisfy the metabolic demands of large muscle movement. White muscle fibres contain less myoglobin and fatigue more easily. The proportion of each depends upon biological need (eye muscles are mainly white, for example), but can be affected by training. The main subcellular components of a muscle fibre are the myofibrils and the sarcoplasmic reticulum. (Not to be confused with nerve fibres!)

Muscle Spindles: These are small bundles of intrafusal muscle fibres scattered in moderate numbers throughout the main bulk of a muscle. They make no real contribution to the power of that muscle’s contraction, but serve instead as contraction sensors. The information they provide is one of the key factors in the alpha-gamma control system.

Nerve Fibres: A collection of axons en route from their source neurons to their appointed destination. (Not to be confused with muscle fibres!)

Neuromuscular Junction (NMJ): The junction between an efferent nerve axon and a skeletal muscle. Much like a synapse, but involving an alpha motor neuron and a motor unit rather than two neurons. Also known as the motor end-plate.

Perimyseum: The connective tissue wrapped around a bundle of muscle fibres.

Pyramidal Tract: This is the direct corticospinal tract. It arises in the primary motor cortex, as a bundling together of axons from the upper motor neurons. By the time it reaches the medulla, it contains a million or so fibres, 90% of which decussate (cross over) and travel down the contralateral (or “crossed”) lateral corticospinal tract of the spinal cord. The pyramidal tract is conventionally believed to carry voluntary
muscle instructions to the alpha motor neurons. (See also corticospinal fibres, and compare corticobulbar fibres.)

Renshaw Cell: An inhibitory interneuron found in the spinal grey, and associated in two ways with an alpha motor neuron. Firstly, it receives an excitatory collateral from the alpha neuron’s axon as it emerges from the motor root, so that is “kept informed” of how vigorously that neuron is firing. Secondly, it sends its own inhibitory
axon to synapse with that alpha neuron. The rate of discharge of the Renshaw cell is thus broadly proportional to the rate of discharge of the associated motor neuron, and the rate of discharge of the motor neuron is broadly inversely proportional to the rate of discharge of the Renshaw cell. Renshaw cells thus act as “limiters”, or “governors”, on the alpha motor neuron system, thus helping to prevent muscular damage from tetanus.

Sensory Root: One of the six or so fibre bundles which emerge from the dorsal horn of a given spinal segment and make up the sensory element of that segment’s spinal nerve. (Compare motor root.)

Spinal Grey: The mass of neuron cell bodies making up the central core of the spinal cord throughout most of its length. Shoes a distinctive butterfly shape if transected, the tips of the wings of which being known as the dorsal horns (dorsally) or the ventral horns
(ventrally). The alpha motor neurons are located in the ventral horns. (Compare spinal white.)

Spinal Nerve: A bilaterally symmetrical pair of peripheral nerves arising from each spinal segment, and responsible for the body’s sensory, skeletomuscular and visceral operation.

Spinal Segment: A section of spinal cord corresponding more or less to a single vertebra, from which a single bilateral pair of spinal nerves originates.

Spinal White: The spinal cord’s ascending and descending axon tracts. So called because the myelin sheathing of each component axon gives the tissue a white colouring upon dissection. In spinal cross-sections, the white matter can be seen to be channeled into the “flutings” provided by the horns of the spinal grey. (Compare spinal grey.)

Spindle Fibres: See intrafusal muscle fibres.

Tetanus (1): State of constant muscle tension caused (a) by incoming excitations arriving so quickly that their individual twitches overlap and merge (“fused tetanus”) (see end-plate potential), or (b) pathological state such as tetanus (2) or strychnine poisoning.

Tetanus (2): Name of disease (commonly known as “lock-jaw”) characterised by severe and life-threatening tetanus (1) caused by action of toxins produced by the bacteria responsible.

Tetany: Muscle spasm (uncontrolled overcontraction) caused by insufficient calcium ions available at the muscle end-plate.

Tremor: Minute oscillations of a muscle which accompany its contraction. Their frequency is in the range 8 - 12 Hz, and their amplitude about 1% that of the main muscle contraction.

Twitch: A sudden but unsustained contraction of one or many muscle fibres.

Upper Motor Neuron: A relatively large neuron in Layer V of the primary motor cortex (Brodmann’s Area 4). The axons pass down through the corona radiata to the internal capsule, then down through the cerebral peduncles of the midbrain to form the pyramidal tract.

3- The Alpha-Gamma Spinal Servomechanism


The essence of the alpha-gamma system is that the spinal cord simultaneously conveys two sets of descending instructions to the ventral grey of the spinal segment from which the muscles in question are innervated. The first of these - the power signal - descends from the neocortex via the pyramidal tract, and synapses with the alpha motor neurons. The axons of these neurons - the alpha fibres - emerge via the ventral root of the spinal nerve, travel to the destination muscle, and cause contraction of the extrafusal muscle fibres (which form the bulk of the muscle). Unfortunately, alpha control of this sort is not very sophisticated, and can go wrong in one of two ways. At one extreme, the muscle tension might turn out to be insufficient for the job, and at the other extreme it might turn out to be too much. In the first instance your limb either stays where it is or moves too slowly, and in the other instance it moves away too violently; and in both instances you are at risk of injury, if not of loss of life.

The second type of descending instruction helps remove this risk. These signals - the control signals - descend from the extrapyramidal system via the extrapyramidal tract, and synapse with the gamma motor neurons at the appropriate spinal level. From the brainstem downwards, therefore, they run in parallel with the pyramidal alpha system. Their axons too - the gamma fibres - emerge via the ventral root of the spinal nerve, and travel to the destination muscle. Here, however, instead of exciting the extrafusal muscle fibres, they enter the muscle spindles and excite the intrafusal muscle fibres. This arrangement is important, because it provides a means of sensing how appropriately the alpha muscle contractions are proceeding. The argument at this juncture goes as follows:

Question:

****Are the extrafusal muscle fibres and the spindle fibres contracting at the same speed, and, if not, are the spindle fibres contracting slower or faster than the others?
Possible Answers: There are three mutually exclusive possible answers to this question:

(a) Answer = YES: If the two types of fibre are contracting at the same speed, then the muscle as a whole is deemed to be contracting at the right speed, and no remedial action is necessary. But keep checking at regular intervals just in case.

(b) Answer = NO, with gamma slower: If the intrafusal muscle fibres are contracting slower than the extrafusal, then the extrafusal are contracting too quickly. This means that the limb loading has been overestimated, and that it would be wise to slow the movement down. This is done by feeding back inhibition from the Golgi tendon organs to the alpha motor neurons, thus subtracting locally from that being fed down the pyramidal tract.

(c) Answer = NO, with gamma faster: If the intrafusal fibres are contracting faster than the extrafusal, then the extrafusal are contracting too slowly. This means that the limb loading has been underestimated, and that it would be wise to speed the movement up. This is done by feeding back excitation from the spindle fibre to the alpha motor neuron, thus adding locally to that being fed down the pyramidal tract.

Thus, notwithstanding the fact that it is the alpha system which says “go” to a given muscular contraction, it is the gamma system which says how strongly to go. Moreover, it is the gamma system which is then best placed to bring about automatic changes to those contractions. If the gamma signal became rhythmic, for example, then the muscle contractions would also become rhythmic even if the alpha signal remained unchanged. All of which is totally consistent with the conventional neuroanatomical opinion that the neocortical pyramidal system serves voluntary movement whilst the extrapyramidal system concerns itself with involuntary fine coordination and synchronisation.

There is also evidence for a hunting mechanism (see earlier this chapter) in the tremor which accompanies the normal contraction of muscle. Lippold (1971) found a slight oscillation at approximately 10 Hz superimposed upon the main contractions of skeletal muscles, and argued that this was a valuable property of the stretch reflex arc because it allowed more rapid response than would have been the case if this particular circuit were more effectively damped. As he puts it, reflex action must involve “a necessary compromise between speed of response on the one hand and a certain degree of overshooting, or inaccuracy, on the other” (p70).

The anatomical features of the alpha and gamma systems are shown diagrammatically in the associated paper on the Pyramidal and Extrapyramidal Motor Systems.

4 - Head/Eye Muscle Control


The oculomotor control system serves a variety of biologically essential behaviours such as food search and predator avoidance (Galiana, 1990). It therefore needs to be every bit as complicated as the skeletomuscular system it is helping to guide. This functionality is provided by having a complex of feedforward, predictive, and feedback control loops at work. To start with, there are mechanisms controlling the automatic focussing of the lens, binocular vergence, and the automatic stopping down of pupillary aperture. There are then additional mechanisms to control the automatic positioning of the eye relative to the head as the head moves relative to both the body and the external world. These latter mechanisms place heavy information processing demands on the vestibular system. This is the system which processes the information provided by the semicircular canals of the inner ear (the “labyrinth”), the body’s balance detectors. Information from the semicircular canals travels to the brainstem down the vestibular branch of the vestibulocochlear nerve (CN VIII). Here it links in via the vestibular nuclei of the lower pons to the cerebellum and a host of other components of the extrapyramidal system.

5 - Stammering


Lee (1951) pioneered a technique of replaying a person’s speech to that person’s own ears, subject to a variable time delay. He found that there were two types of common effect. Either subjects slowed down and raised their voices, or else they began to speak haltingly, repeating syllables in a form of “artificial stutter”. The same phenomenon emerged with skilled tympanists reading a drum-beat, and for the key presses of skilled morse operators. Lee gives the following specific examples:

aluminum…

** degrades to aluminum-num…
ten-nine-eight-seven… degrades to ten-nine-nine-eight-seven…

Lee interpreted these findings as evidence of a multiple loop control hierarchy, with four levels of feedforward and feedback. The top control level releases individual thoughts for action, and then monitors that action for successful progress and completion. The second control level does the same for the words expressing the thought. The third does the same for the syllables making up each word. And the fourth does the same for the phonemes making up each syllable. It is confusion at the hand-over between the second and the third level which presumably causes the aluminum-num syllable repetition. There were no single-phoneme repetitions.

6-OtherControl System Pathologies


Other clinical syndromes where control system deficiencies have been suggested from time to time are:

(a) Parkinsonism: The motor disorders which characterise Parkinson’s disease are conventionally attributed to disorders of muscle control circuitry. Wiener himself likened Parkinsonian tremor to the oscillations of under-“damped” control loops (Wiener, 1950), Flowers (1978) blames lack of prediction, Harrington and Haaland (1991) blame “central processing deficits”, and Dinnerstein, Frigyesi, and Lowenthal (1962) blame slower than normal proprioceptive feedback for a variety of the standard Parkinsonian symptoms, such as rigidity, slowness, and lack of coordination.

(b) Dyspraxia: This is an inability to initiate
voluntary movements despite intact muscles and motor systems, and is a common correlate of CVAs. The suspicion here is that a major feedforward pathway or mechanism is unserviceable.

(c) Dysarthria: This is an inability to deliver voluntary movements cleanly once initiated. Thus when speech becomes slurred after a CVA, even to the extent of falsely suggesting drunkenness. The suspicion here is that the initiating feedforward is intact, but the subsequent fine control and sequencing is faulty.

(d) Dyslexia: This is an inability to process visually presented text efficiently. Whilst this is at first sight a perceptual
problem, the very complexity of the oculomotor control system (see above) makes it a motor problem as well. You cannot read if you cannot control the movement of your eyes. When reading this text, for example, your eyes will be fixating after every eight characters (about every one and a half words) (Rayner and Pollatsek, 1989), and many authorities (typically Pavlidis, 1981/1985) believe that developmental dyslexia can be explained by defects in sequencing these fixations for maximum information uptake. Developmental dyslexics do appear to have eye movement patterns which differ from those of normal readers (Rayner and Pollatsek, 1989). However, this factor per se has not been strongly confirmed. Indeed, Rayner and Pollatsek place greater store in Stein and Fowler’s (1982, 1984) findings of “vergence control” problems in dyslexics. Vergence movements are those which keep both eyes pointing at the same centre of attention. In normal readers, the two eyes move “conjugately”, that is to say, they track at the same speed and in the same direction. Stein and Fowler’s data suggests that about one in six cases of developmental dyslexia can improve reading performance with treatment of this problem in isolation.

(e) Learning Difficulties: Many categories of learning difficulty present with an inability (amongst other things) to communicate effectively at a pragmatic level. This can be alleviated to a greater or lesser extent by training at what Williamson (1992) describes as “backchannel” skills. These include a wide variety of both vocal and nonvocal responses, such as nods, shakes, grunts, facial expressions, etc., whose function is to feed back to a speaker the extent to which his/her utterances are being understood.

References

Bennett, S. (1979). A History of Control Engineering, 1800-1930. London: Institute of Electrical Engineers.

Carpenter, M.B. (1991). Core Text of Neuroanatomy (4th Ed.). Baltimore, MD: Williams and Wilkins.

Dinnerstein, A.J., Frigyesi, T., and Lowenthal, M. (1962). Delayed feedback as a possible mechanism in Parkinsonism. Perceptual and Motor Skills, 15:667-680.

Fairbanks, G. (1954). A theory of the speech mechanism as a servomechanism. Journal of Speech and Hearing Disorders, 19:133-139.

Flowers, K. (1978). Lack of prediction in the motor behaviour of Parkinsonism. Brain, 101:35-52.

Gosling, W. (1994). Helmsmen and Heroes: Control Theory as a Key to Past and Future. London: Weidenfeld and Nicolson.

Granit, R. (1970). The Basis of Motor Control. London: Academic Press.

Harrington, D.L. and Haaland, K.Y. (1991). Sequencing in Parkinson’s disease. Brain, 114:99-115.

Lee, B.S. (1951). Artificial stutter. Journal of Speech and Hearing Disorders, 16:53-55.

Lippold, O. (1971). Physiological tremor. Scientific American, March 1971, 224(3):65-73.

Pavlides, G.T. (1981). Do eye movements hold the key to dyslexia? Neuropsychologia, 19:57-64.

Pavlides, G.T. (1985). Eye movement differences between dyslexics, normal, and retarded readers while sequentially fixating digits. American Journal of Optometry and Physiological Optics, 62:820-832.

Rayner, K. and Pollatsek, A. (1989). The Psychology of Reading. London: Prentice-Hall.

Smith, D.J. (1997). Human Information Processing. Cardiff: UWIC.

Stark, L. (1968). Neurological Control Systems. New York: Plenum.

Stein, J.F. and Fowler, S. (1982). Ocular motor dyslexia. Dyslexia Review, 5:25-28.

Stein, J.F. and Fowler, S. (1984). Ocular motor problems of learning to read. In Gale, A.G. and Johnson, F. (Eds.), Theoretical and Applied Aspects of Eye Movement Research. Amsterdam: North Holland.

Walter, W.G. (1961). The Living Brain (2nd Ed.).
Harmondsworth: Penguin.

Wiener, N. (1948). Cybernetics. Cambridge, MA: MIT Press.

Williamson, G. (1992). Conversation stoppers. Therapy Weekly, 19th November 1992, 4.

**** ** [ISBN:** 1900666081**] **

The error is an error signal with a frequency like the value of the error. A frequency can’t be negative. Therefore I don’t think upon a negative error.

If you stick with a strict neural model, you have to double the number of control systems involved with any controlled variable. One control system would compare an inhibitory perceptual signal (reaching the comparator through a Renshaw cell) and an excitatory reference signal, so the error signal would indicate a perceptual signal that is too small. The other control system would compare the same perceptual signal P(without a Renshaw cell) with an inhibitory reference signal (which now involves a Renshaw cell), so that error signals would indicate a perceptual signal that is too large. The error signal representing perceptions that are too large would drive output functions that make the perceptual signal smaller. The other error signal would drive output functions that make the perceptual signal larger. This gives us one comparator that outputs an error signal to one output function when the perceptual signal is greater than the reference signal, and a second comparator that outputs an error signal to a different output function, whose external effects are opposite to the first one, when the perceptual. signal is smaller than the reference signal.

When you consider these two sets of comparators and output functions together, they form a single bi-directional comparator and output function that can correct both positive and negative errors, measured in terms of the controlled variable. I prefer to treat this as the general case because it covers many situations and requires only a single bidirectional control model.

In some cases, control is only in one direction – the controller reacts only when the perceptual signal is too small, or only when it is too large, with the opposite case being ignored. I think these are much less common cases.

I ask again – did you get the “live block diagram” program I sent you with the book chapters? This will answer your questions about the effects of output gain.

Best,

Bill P.

[From Bjorn Simonsen (2005.10.31,14:00 EUST)]

From Bill Powers (2005.09.30.1440 MDT)

I cannot se the connection between my argument and yours. But I understand and
agree with every sentence in your mail. I could have written them myself if I
could explain myself better.

Your mail is placed in the correspondence file for files of great
value.

In some cases, control is only in one direction – the controller

reacts only when the perceptual signal is too small, or only

when it is too large, with the opposite case being ignored. I

think these are much less common cases.

Your first sentence is OK, but
I am not sure if I understand your last sentence. I understand it the way that
usually the control is in both
directions. The controller reacts at both systems at the same time.

Transferred to Rick’s example

From Rick
Marken (2005.10.28.1000)

But, in
general, changing the reference signal from high to low will not

reduce
conflict (at least, in terms of the PCT model of conflict). This is

because
the value of the reference signal just specifies the intended state

of the
controlled variable; it doesn’t determine how hard the system will

work to
get that variable to the reference.

My way of
thinking is that if I go from situation 1 to situation 2, the controller will
react fairly equal in Bjorn as in Rick. The q (o) value is reduced more if we
lower the reference value than if we lower the gain.

Bjorn –changing
the reference value

p1 =13.6,
r1 =15, gain is 10, slow is 1–q1(t+1) = q1(t) + s[g(r-p)–q(t)]=q1(t) + 1[10(15-13.6)-*q1(t)]

q1(t+1) =
14

p2= 9.2,
r2 = 10, gain is 10, slow is 1—q2(t+1) =q2(t)+s[g(r-p)-q(t)] = q2(t)+1[10(10-9.2)-
1*q2(t)

q2(t+1) =
8

Rick – changing the gain

P1=14.2,
r1=15, gain is 20, slow is 1 — q1(t+1)=q1(t)+s[g(r-p)-q1(t)]=q1(t)+1[20(15-14.2)-1*q1(t)

Q1(t+1) =
16

P2=13.6,
r2=15, gain is 10, slow is 1 — q2(t+1)=q2(t)+s[g(r-p)-q2(t)]=q2(t)+1[10(15-13.6)-1*q2(t)

Q2(t+1) =
14

Conclusion.

Changing
the reference signal from high to low will change the output signal in the same
degree (in this example) as if the gain is changed from high to low.

Maybe
these two examples are the same.

I
understand that the gain has something to do with the number of dendrites coming
from different neurons. If we will control a perception stronger to morrow than
we do today, we can include more systems tomorrow. If we conclude more systems,
we both change the reference value and the number of connections from dendrites
(we change the gain). Of course we can’t plan the value of the change.

I ask again – did you get
the “live block diagram” program I sent you

with the book chapters? This will answer your questions about the

effects of output gain.

If you go back to my mail you
will find

Were
you able to run that “live block diagram” program I sent you?

Yes a wonderful program, I am working to make
an equivalent in PowerSim. Is that OK?

The two examples above is done on your “live
block diagram”, the values are taken from the diagram. It is really a nice
program and it tells me almost everything.

Bjorn

[From Bjorn Simonsen (2005.11.02,10:10 EUST)]

From [Marc Abrams (2005.10.30.1910)

Why hasn't Bill cited some of these authors, especially Stark and Granit?

He did.
I think Derec J Smith can tell us much. But I think Bill did tell us about
the Spinal Reflex in 1973 more convincing than D J Smith did in the
ninetieths.

I hope Richard (Kennaway) can tell us about Derec J Smith and his negative
feedback applications.

Bjorn