From Tom Bourbon [940429.1047]
[from Mary Powers 9404.29]
Bob Clark to Bill Powers:
your description of the process [predicting], above, ... becomes
an excellent summary of much of the DME's operation.But Bill wasn't talking about the DME. He was talking about
principles and strategies.
I agree, Mary. This is the point I tried to make in my own short reply to
Bob.
* * *
Bob:
excessive attention to time integrals can create other problems.
It can resemble depression, with a generally "sluggish"
response. This response contrasts with the "quick and nervous"
types you suggest for those with excessive attention to time
derivatives. ... it seems to me that we, including McFarland,
discussed this way back when.
Mary:
This is what I was mentioning in my post on Prozac, in case
anyone is interested. The idea is that antidepressants and such-
like psychoactive drugs affect the way control systems do their
controlling. One case discussed in the book was a woman who was
"sluggish in mind and speech", though not depressed according to
the usual diagnostic criteria. After agonizing over whether to
give a drug to someone who was simply a certain kind of person,
and not technically "ill", the psychiatrist did give it to her,
and she became alert and articulate. This gets into all kinds of
fascinating questions - is the way one's control systems function
inborn, or a consequence of experience or trauma?. Should a
symptom like sluggishness be treated if there is no cause (known
at the moment)? If the sluggishness were a consequence, say, of
lead poisoning, there'd be no question about treatment. What
about a way of functioning, a type of control system (too much
integral feedback, or whatever), that is perhaps not
intrinsically off base, but is poorly suited to the culture of
20th century America?
Something along this line is the purpose of work I've just started at the
medical school and our afflilated hospital. All neurologically impaired
patients who go through neuropsychological assessment by our department now
perform a brief battery of control tasks. I am looking to see if the
descriptive statistics and model parameters from PCT control tasks can
discriminate among the various problems that bring patients into the
hospital.
My biggest challenge has been to direct the attention of the professional
staff to real-life problems described by the patients and those close
to them, rather than to the scores from various "assessment instruments." I
believe measures of control, and parameters of best-fit PCT models, will
more readily identify people who came to the hospital because they can't
finish anything because they seem to forget what they are doing, or who
mess things up because they overshoot, or alternatively never quite catch
up, than those measures will identify people who score this or that on
so-and-so's scale of who knows what.
The real issue here is that there seems to be immense confusion
in the psychological/psychiatric community about what is going on
with these drugs - mainly because they are focussed on the
variety of symptoms - behavioral consequences - of more or less
dopamine or serotonin or whatever, and have no idea what is being
affected. In fact a lot of diagnosis is after the fact, in terms
of what drug works, rather than the other way around. For sure,
it's a wild guess on my part that they are affecting the
characteristics of control systems, and it will be a cold day in
hell before any research in that direction is done, if ever, but
I don't see that any current ideas are any better.
Well, maybe it will be a cold day in Houston, rather than in hell --
assuming Houston in the summer time and hell any old time are not one
and the same place :-). If the early results of the project I described
above are good, then we will propose including the control measures in the
set of behaviuoral measures used to study drug effects.
Later,
Tom Bourbon
Department of Neurosurgry
University of Texas Medical School-Houston Phone: 713-792-5760
6431 Fannin, Suite 7.138 Fax: 713-794-5084
Houston, TX 77030 USA tbourbon@heart.med.uth.tmc.edu