On Tue, 5 Oct 1999, Eugene Leitl commented on my comments:
>
> Unless we're talking about ultrarapid freezing you'll get phase
> transition in the lipid bilayer, possible poration and aggregation of
> membran-integral proteins with subsequent possible loss of function.
I stand corrected in light of Eugene's clearly greater experience in these areas.
>
> "large enough to physically move the synapses around" is a remarkable
> understatement. One of these days I should really scan some of these
> freeze-substitution etch EM slides vs. controls. Bad, bad, bad.
>
This would certainly be useful (or even citing some references that point out what happens to lipid membranes or the protein structures embeded in lipid membranes during normal cooling processes).
>
> Of course this assumes that they know what the proper locations
> are. It is _impossible_ to predict the original shape of color clouds
> in a glass of water after stirring it a little.
It will be interesting if you consider the brain an "approximation" engine. -- Doctor tells the nanobots to do a reasonable job and asks the patient "how does that feel", patient says "I'm having problems pronouncing my t's", doctor tweeks the location of a few neurons and says "try again", patient says "thats better".
[Of course this is grossly oversimplified and getting to the point where the patient knows what a "t" is is problematic...]
>
> > well, then you should be able to do the reanimation without nanotech.
>
> If you could redo Suda's stunt using vitrification and modern signal
> processing techniques on the EEG that would really be impressive news.
>
I'm lost on this ref. Who is Suda?
> > If it turned out that the physical location of the receptors within
> > a synapse (rather than just their overall "density") was significant
> > things would be more difficult. But I doubt very much that the cells
> > have mechanisms for locking receptors at specified locations. Most
>
> Freely diffusable ion channels are a pathological case. Pathological
> both in restricted to certain types/locations and ischaemic trauma.
>
Perhaps I wasn't clear. There may be mechanisms for locking receptors in regions, but within those regions they may be relatively mobile. I suspect it comes down to whether or not specific proteins are connected to the cytoskelaton. Some of the membrane proteins probably are connected to maintain the shape of the cell. My guess would be that the membrane-cytoskeleton links create barriers that confine the diffusable proteins within specific regions of the cell surface. You have got to maintain the shape of a synaptic junction and presumably you want to confine the neurotransmitter receptors to the synaptic junction and not have them wandering all over the surface of the neuron. But I really doubt there is an "addressing" and/or "locking" system for the receptors at specific locations within a synaptic junction (that would involve a huge amount of information to be managed).
>
> Recently (as in yesterday) I learned there is a rare type of brain
> trauma (20-30 cases/year) which wipes out the entire memory
> irreversibly, leaving only algorithmic memories. This leads to
> frequent divorces, because the original personality is wiped out. Fun,
> huh?
>
Ah, Eugene, but you cannot say for certain whether the memory is really lost or the retreival mechanism is broken. If I lose the DSP to the read head on my disk drive, I haven't lost any data, I've just lost the ability to read it. Now, thats *really* interesting because it may mean there may be a central switch region of the brain through which all memories must be processed.
Anders could probably comment on this.
Robert