[From Chad Green (2011.12.31.1928 EST)]
Just found out about this journal article on the CHAOPSYC listserv
(thanks, Gavin):
A Tale of Two Stories: Astrocyte Regulation of Synaptic Depression and
Facilitation
"We devised a biophysically plausible computational model to investigate
the characteristics of astrocyte modulation of presynaptic short-term
plasticity. Using the model, we were able to identify the parametric
regime in which the synaptic response to action potential stimulation
can switch from facilitating to depressing and vice versa. This ability
to switch synaptic modus operandi depended critically on the
characteristics of astrocyte-to-synapse signaling. These findings
highlight the new potential role played by astrocytes in defining
synaptic short-term plasticity and could explain contradicting
experimental evidences."
Related news article (model attached):
More than glue: Glia cells found to regulate synapses
"The model provides a 'new view' of how the brain functions. While the
study was in press, two experimental works appeared that supported the
model*s predictions. 'A growing number of scientists are starting to
recognize the fact that you need the glia to perform tasks that neurons
alone can*t accomplish in an efficient way,' says De Pittà."
Happy new year!
Best,
Chad
Chad T. Green, PMP
Program Analyst
Loudoun County Public Schools
21000 Education Court
Ashburn, VA 20148
Voice: 571-252-1486
Fax: 571-252-1633
Web: http://cmsweb1.loudoun.k12.va.us/50910052783559/site/default.asp
There are no great organizations, just great workgroups.
-- Results from a study of 80,000 managers by The Gallup Organization
Bill Powers 12/30/11 12:37 PM >>>
[From Bill Powers (2011.12.30.090-6 MST)]
JRK: The simulation assumes that the offspring of any member of the
population has a phenotype similar to but randomly different from
its parent. (Sexual recombination is another of the many things that
this simulation leaves out.) The random difference is uniformly in
any direction, and is independent of the random difference between
the parent and the grandparent. The likelihood of reproducing is a
function of the phenotype, which I didn't plot, but the contour
lines of the fitness function are concentric circles centred on the
final position of the population blob.
After a few days of mulling this over, I finally got it. Very nice,
very simple.
The gradient is centered in the upper right quadrant. The blob moves
nicely in that direction, shedding individuals that are eliminated by
the limit on population as the population reproduces. The result is
that the most fit lines make it all the way to the goal and those
that develop in different directions are weeded out because they
reproduce less rapidly.
So far this looks much like E. coli except that E. coli doesn't have
to die if it swims in the wrong direction. It just tumbles again. If
we could give your model the same ability (where "it" now refers to
some control system that keeps working across generations) a lineage
would, when it tumbles, change how fast each of its characteristics
changes between mutations. If headed in a less favorable direction,
the variation would change direction or speed, and keep repeating
that change until the change was favorable again. No external weeding
would be necessary, though we could include it in the model.
What we need to make the E. coli version work is some variable that
indicates how far from maximum fitness the organism is. For example,
consider "The beak of the finch" (Rosemary and Peter Grant, Google
Books). Here the construction of the beak varied from one generation
to the next according to the kinds of seeds availa
ble. The standard
explanation is that each new generation would have beaks of some
average size with a random distribution. Those with strong enough
beaks to crack one kind of seeds would survive under one condition;
those with long enough beaks to find deeply buried but not tough
seeds would survive under another condition.
Emacs!
Now look carefully at this picture. The "striking" difference is a
slight variation in the length of the beak and perhaps the vertical
dimension where it meets the head. Otherwise, the birds look the same.
It seems to me that this is a very common situation in the
evolutionary record: a series of small changes in the same variable.
The neck of the okapi-like ancestor retained the same seven vertebra,
but all the vertebrae kept lengthening until the modern giraffe,
still with the same seven bones in its neck, had vertebrae the length
of a man's forearm. What we see is not a random variation in length,
but a systematic variation of length (and diameter).
For the finches, the information about fitness would come not from
the fitness itself but from the immediate effects of selection
pressures. I'm talking about stress-induced mutation, and
specifically about mutation of the direction of change of some
parameter, here the length and stoutness of the beak. The beak
doesn't just jump from long to short to longer still in a random
fashion: it lengthens and shortens by small amounts, and if the
effect on the stress-indicator -- hunger? -- is favorable, the next
generation's beak lengthens again or shortens again, by a small
amount. Those changes are not mutations in the standard sense. They
correspond to E. coli's periods of swimming in a straight line.
The effect of this, if it increases the stress, is to induce a true
mutation. The various dimensions of change are randomly scrambled.
Now natural selection can come into play, because some number of
scrambles that leaves the changes still in the wrong direction can
eliminate that lineage.
But the lines in which the result is to head down the gradient of
stress will change very efficiently in the right direction and keep
changing that way without further mutations until somewhere, down the
road, that favorable direction becomes unfavorable. This is what I
was talking about when I was musing over the idea of a continuous
universe with continuous derivatives.
So what can you do to convert your model to an E. coli version? We
need a stress indicator like hunger, and we need a background kind of
change that is continuous and that continues in the same direction as
long as the stress indicator is decreasing. Only when the blob moves
past the point of closest approach to the center of the gradient does
a random change in the rate of change of each parameter (two in your
model) occur.
If this is done right, I think you can illustrate not only E. coli
evolution, but show that the observational evidence could also be
taken as evidence of the standard picture given by your present
model. The difference between mutating the values of parameters and
mutating the direction of change of parameters may not be very
noticeable until you look for it. This would be a useful result,
because it could explain why a wrong interpretation was adopted by
perfectly intelligent scientists. In the same way, we can show via
the behavioral illusion how stimulus-response theory could have
seemed a reasonable hypothesis.
Best,
Bill P.
···
At 04:55 PM 12/29/2011 +0000, Richard Kennaway (CMP) wrote: