A) On who controls whom in an interacting system?
On the conjecture that there are parts of the brain that control other
parts, a number of people posed the following dilemma about determining who
controls whom:
(1)"When glucose levels are low in the blood, some brain structure receives
that message and releases some neurotransmitter that causes a particular
group of neurons to fire, thus the sensation of hunger and the drive to eat.
Does the blood control the brain structure? Does the brain structure
control the behavior of eating?"
(2)"It comes down to the precise definition of the word "control", for which
the issue of different time scales is critical. It is quite proper to say
that I control the the off-on switch on my TV set. But, if I don't like what
I see, I may be influenced to turn it off (eg, partly "controlled" by the
advertisers), but that will take some time."
By engineering definition, a control system consists of a "master or
controlling subsystem" where the "control" signals are generated, and a
"subservient subsystem" where these control signals are used. In a simple
feedback control mechanism, the output of the subservient subsystem is
fedback to the controlling subsystem. So although the overall system can be
seen as interacting subsystems, we call one subsystem the "controller"
because of the "nature or role" of its signals. The fact that the signals
from the controller are based on or influenced by signals from other
subsystems does not change the "nature or role" of these signals that we
call "controlling." So blood is perhaps not the "controller" of the system,
although the system does not work without blood. Analogy: electrical systems
do not work without the connecting wires. And although a person is
influenced by what is on TV, the person is still the "controller" of the
ON/OFF signal. This leads to the next question of what characterizes a
controller and how do we detect a controller in an unknown system.
B) Who is a controller and how do we detect it in an unknown system?
This raises further questions about who can be labeled a "controller" when
there are many interacting subsystems in a system. For man-made systems,
this is not a difficult task, because the system designer knows exactly who
is the controller. But how do we determine/identify the controlling
subsystem in a natural system, an unknown system? I suggest here a means of
doing so, but I welcome other ideas.
Let me raise a few questions here before suggesting a specific mechanism for
identifying the controller of a system. It is perhaps now an accepted fact
that the brain controls the rest of the body. (I hope this is a settled
issue and not a source of fresh debates.) Thus we think in terms of motor
control and other controlling functions of the brain. My question is, how
did the scientific community arrive at this conclusion, that the brain
"controls" other organs of the body? How and why were other subsystems like
the heart and the liver and the lungs ruled out? Any insight on the history
of this question would be appreciated.
C) On the test for a controller in any system
>From an engineering point of view, a controller is distinct and separate
from the basic system. The basic system should function in some manner
without the involvement of the controller. For example, a car or a plane in
motion can still be a functioning system, although not in a desirable way,
without an active driver or a pilot. Or the economic system can still
function, at least for some time before the crash, without the active
involvement of the Federal Reserve Bank in setting interest rates or
controlling the money supply. So the basic feature of a controller is that
the system can function without the controller, although such "functioning"
can be brief and result in a disaster. So a possible means of testing for a
controller in an overall system might be to isolate each subsystem and
observe the behavior of the resulting system. If the resulting system is
still a functioning system, although briefly perhaps, then the isolated
subsystem might be the source of control in the system.
I think this test can be applied very well to any man-made system to
determine the controlling subsystem in a system. My question is, are these
ideas applicable to natural systems as well? If not, why not? What are the
flaws in this idea? For example, using this test mechanism, it is easy to
determine why the heart or the liver or the lungs do not qualify as
controllers of the system. But a brain-dead person or one in a coma is still
a functioning system, with some severe limitations of course. Does that
qualify the brain as the controller of the system?
My question is, can a test idea like this be used to determine that there
are parts of the brain that control other parts? I quote here from a very
interesting response to my earlier note. I made only some slight
modifications for privacy reasons.
"I am writing in response to your request posted on the
cnero@bbb.caltech.edu news group concerning the removal of neurotransmitter
systems from the brain. The lab of Dr. Richard Palmiter at the University
of Washington has used genetic techniques to generate a number of
"knockout" mouse strains deficient of various neurotransmitters. The basic
approach of the lab has been to "knockout" the enzymes involved in the
synthesis of a particular neurotransmitter, thus producing mice completely
deficient of that neurotransmitter. To date, the lab has mice deficient in
dopamine, norepinephrine, and NPY (neuropeptide Y). A medline search on his
name will provide you with the relevant citations, if you are interested."
"The suprising result of producing many of these mice (with the exception of
the dopamine knockout mice) is that they are quite normal in appearance and
behavior, despite the deficiency of an important neuromodulator. One
article, that I have not myself read but heard about, addresses this issue
in the NPY knockout mice and can be found in a recent issue of "The
Scientist." Of course, each knockout mouse does have an associated
phenotype: the dopamine knockout mice are the most extreme in that they do
not eat or move in the absense of dopamine, the norepinephrine knockout mice
have deficiencies in regulating body temperature, mating, and maternal
behaviors."
It appears that the "test of a controller" framework outlined above is being
used in this study. My questions again are: First, can this test mechanism
for a controller be used on natural systems too? Second, does the above
study point to sources of control in the brain?
With regards,
Asim Roy
Arizona State University