long term potentiation

kkollins at pop3.concentric.net kkollins at pop3.concentric.net
Wed Oct 21 23:25:19 EST 1998


First, thank you for your thought-filled reply. Using your post as stimulus, I'll
briefly discuss the way I've approached the problem of all neural activation,
including "LTP", and it's integration.

Hemidactylus at my-dejanews.com wrote:

> In article <362D2136.DD2A4A84 at pop3.concentric.net>,
>   kkollins at concentric.net wrote:
> >
> >
> > Sturla Molden wrote:
> >
> > > There is no evidence that LTP is a naturally occuring
> > > phenomenon.
> >
> > Would you care to square this with what I posted in respons
> to Walter, above.
> > The research report I read did clearly state that Ca(2+)
> dynamics above a
> > threshold have "staying power" (my words)... that is, their
> effects continue
> > after the sub-threshold ionic flow has ceased. What, if
> anything but an
> > LTP-like effect, is this?
> >
>
> There should be a distinction made between facilitation (due possibly to
> residual CA++ (or inositol mediated release from ER stores?)) and varying
> degrees of potentiation, based on the time interval involved. "staying power"
> is a relative term. Tetany might bring us into the interval known as
> post-tetanic potentiation (PTP). I have a curiosity about whether LTP is
> nothing more than an artificially induced phenomenon with no correlate in
> natural systems. I think critique of LTP is good.

Critique of =everything= is good, and necessary if we are to stay on track toward
understanding.

> >
> > > It araises from repeated tetanisation of
> > > specific neural pathways, eg. in the hippocampus. It is
> > > also not a homogenous phenomenon, there are several
> > > sub-types of LTP depending on different receptors
> > > (eg the NMDA receptor, opioid receptors, mGluRs)
> > > and with different durations.
> > >
> > > LTP depend on the same mechanisms that seem to be
> > > involved in memory formation, but there is no evidence
> > > that LTP is the substrate for memory.
> >
> > Yeah, but show me "memory" without the involvement of LTP...
> only way one can
> > is to show me a "corpse" (brain dead).
> >
>
> I am presently in search of references that critique the use of LTP as a
> model for vertebrate memory processing. Anybody know where I can find some
> good ones? Perhaps LTP is a good model for memory processes, but I would like
> for someone here to give a serious critique.

In my view, one must step back a bit, and consider the system as a whole before
one can grasp the roles of any mode of neural activation. The easiest way to do
this is to work with a constant awareness of the two "fixed" points... the
primary, sensory and motor activations, and of these, begin with motor
activation. When one grasps a pencil and writes, for instance, only that
activation which drives the effectors of the trunk, arm and hand so that a
particular word will be written is "acceptable"... all other activation must be
"stripped away", relative to the "acceptable" activation.

In this, one sees that things are vastly-inhibited... a starting point... the
activation that underpins any discrete activation of the effectors must be
minimal.

Next "level" requires the study of the effector geometry. When one does this, one
finds that some effectors must be activated relatively-more than others, and that
the "waveform" of the activation is very-dynamic.

Next thing is that in order for the effectors to be activated in a way that
accomplishes a task within the 3-D external environment, all activation that
occurs must preserve an "awareness" of the external 3-D geometry. I briefly
discussed how this occurs within our nervous systems earlier in another thread
(the topology of decussation), and all of that serves as powerful constraint upon
anything else that can be said about anything, anywhere within the nervous
system.

To sum so far, activation states must be minimal, they must preserve a
geometrical correspondence with the external 3-D environment, and because the
effectors are asymmetrically-distributed, the activation that's necessary at the
effectors will have to reflect that asymmetry while preserving the necessary 3-D
geometrical congruence.

I know this doesn't sound like "neuroscience", but it constitutes =precious=
constraining stuff that, literally, permeates the nervous system,
powerfully-delimiting the set of more-detailed things one must consider. So
grappling with it, until it's mastered, greatly facilitates our going further.
(The Thermodynamics of the external environment does more of the same, but I'll
save that for another discussion.)

So, if one is to say anything about "memory", all of the above stuff =must= be in
what one says about "memory". Otherwise, one is confronted with what will
obviously be a necessarily-incomplete "solution"... simply because one will be
left having to postulate a "black box" (a set of things where, like that famous
cartoon depicts, "Something magic happens", which, although the 3-D geometry has
not been carried through, never-the-less, activates the effectors in a way that's
commensurate with the 3-D geometry inherent in the external environment.) Do you
follow? It's not sufficient to say anything about anything within the nervous
system without placing it solidly within this global 3-D-geometry-preservation
necessity... in the past, folks have scratched their heads with respect to such,
and muttered, "humunculous"... and there it was, the "black box" where the
"magic" happens unexplainedly... and so nothing "fit together".

The "stepping-back-to-go-forward" exercise, above, is an absolute prerequisite
for everything else.

> > > Theoretical models of associative memory are very
> > > robust with respect to the choise of plasticity
> > > mechanism, and a number of different mechanisms
> > > will do the job equally well. Neurons do also
> > > express different types of plasticities, not just
> > > linear scaling of synaptic weights.
> >
> > Of course, it's kind of hard to "draw a picture" using only
> "straight lines".
> >
>
> I'm not sure what is meant here.

I was referring to the stuff of the 3-D-geometry-preservation... for instance,
linearly-graded activations would allow only very-crude effector actions. The
effector actions that we observe require myriad types of "brush strokes" to
"paint". Do you follow? Wiggle your fingers... manipulate an object with one hand
while gently exploring that hand via touch (with the other hand) and vision... I
know this, too, doesn't sound like Neuroscience, but it's all a necessary prelude
to grasping the nature of the spectrum (the storm, really) of activation that
occurs internally. (In these things, I'm carrying the discussion all the way back
to fundamentals that were pointed out by Torsten and Wiesel in their
investigations of vision... and addressing the question of why, when one monitors
visual cortex, for instance, one does not necessarily witness activation that
"looks like" an image of the momentary visual stimulus set... rather, one sees a
"patch work"... it's relatively the same everywhere within the nervous system,
and the problem reduces to explaining how such readily-observed "patch works"
rigorously correspond to the 3-D geometry of the external environment with
respect to which the effectors must act rigorously.

=Only= after getting all of the above firmly in grasp, can =anything= be said
about "memory".

And, although I've barely discussed sensory dynamics, with what's here, the
nature of the realm in which "memory" is constrained follows concisely. Whatever
the molecular mechanism, the existence of which, because of the fact of memory's
progressive-establishment, we can postulate without explicating completely, it
=must= posses the physical capacity for causing the ongoing activation state to
be minimal, to preserve the 3-D geometry, and to reflect the asymmetry of the
necessary effector activations.

So in this, activation of individual neruons that are observed to be "LTP"-like
are just particular harmonics within the overall 3-D-geometry whole... they are
not, at all, any more-"special" than are momentarily relatively-quiescent neural
activations.

It's just that one cannot begin to approach "memory" without carrying all of this
"lower-"evel" stuff right through one's discussion.

> There is a great deal of complexity
> associated with the LTP phenomenon. I've read discussions of everything from
> NMDA receptors, calcium/calmodulin dependent protein kinases (which might
> impact synapsin presynaptically and have other effects postsynaptically), NO
> synthase (which interestingly is related to an epithelial enzyme and also to
> an enzyme that mediates neutrophil/macrophage activity), and metabotropic
> glutamate receptors that could act via an inositol (or diacylglycerol?)
> pathway. I've posted about a study that looks at cell adhesion molecules.
> With all these various components implicated, it is hard (for me) to arrive
> at a good signal/noise ratio wrt LTP. Perhaps this will subside as I delve
> deeper into all central and ancillary issues. With  an integrated approach
> that includes all the putative components in LTP (a phenomenon marginally
> related perhaps to memory processes in the hippocampus) it is quite hard to
> draw straight lines. It is easier to pick out specific components (e.g.-NMDA
> receptors and protein kinases) and ignore the other stuff. I am interested in
> the cell adhesion molecule aspects though.

Perhaps because I earn my sustenance as a computer programmer, in which field,
I've witnessed the transition from magnetic-core "memory" and various
magnetic-disk "memories", to bubble "memory", to on-chip "memory", and because my
view has been from the perspective of energy-content and configuration which
enable the storage of energy which translates into a computer's capacity for
doing the physical work that memory-"addressing" entails, I see the
molecular-embodiment, while of high-significance, as being "moot"... any memory
=must= be as the capacity for for exerting physical force, at a particular locus
within the ongoing global activation state, that embodies the necessary 3-D
geometry. Check it out, any proposed mechanism of "memory" =must= fall within
this constraint, or the global system cannot process information. So, with these
additional constraints which order everything rigorously, it's as you say,
above.>

> > > There is no reason to believe that Altzheimer patients
> > > have impaired memory because they lack LTP. If they
> > > do lack LTP, one should rather say that they are
> > > impaired on synaptic plasticity.
> >
> > This can't be said... all that can be said is the hypothesis
> must be tested
> > if anything's to be said.
> >
>
> The transgenic study I've looked at ties an amyloid protein expression to
> deficits in LTP.



More information about the Neur-sci mailing list