Valentina, On 7/9/08, Valentina Poletti <[EMAIL PROTECTED]> wrote: > > Could you specify what do you mean by synaptic response curve? If it is > what I think it is it is far from linear, at least from the textbooks I > read, so I am probably not following you.
As I recall, the excitatory synapses they found were pretty linear, but they also captured two inhibitory synapses, which were very different from one another and neither was at all linear. I later published a paper at the first NN conference (in San Diego) explaining that one of the inhibitory responsecurves was EXACTLY what you would expect from an AND NOT function on the logarithms of the probabilities of assertions being TRUE. Note that at True=1, the NOT of True=1 is False=0, and the logarithm of 0 is infinite/discontinuous. The inhibitory synapse had exactly this same discontinuity near the upper end of its operating range!!! While conjectures abound, no solid proposed explanation has been advanced for the other inhibitory synapse. Note that integration and differentiation across synapses is well known and often observed, but has never been carefully observed and measured, so by the perverse "ethics" of neuroscience, it has yet to be documented in the literature. I also have a theory that seems to be consistent with the commonly observed differentiation, which (according to my theory) is a sort of time base restoration, that eliminates much of the computational time delays (at a cost of some computational noise), so that the output from a group of neurons does NOT exhibit much if any delay from the input, at least for gradually changing inputs. This is similar to the way that "peaking" capacitors were used in (now antiquated) RTL (Resistor-Transistor logic). Of course, verification of this theory awaits quantitative analysis, which in turn awaits improved laboratory equipment and methods. Steve Richfield ================ > On 7/9/08, Steve Richfield <[EMAIL PROTECTED]> wrote: >> >> Mike, et al, >> >> When you look at the actual experiments upon which what we think we know >> is based, the information is SO thin that it is hard to come to any other >> rational conclusion. I could describe some of these, where for example a >> group of people spent a year putting electrodes into every one of the >> neurons in a lobster's stomatogastric ganglion so that they could >> electrically quiet all but two of them. This so that they could plot the >> synaptic response curve between those two cells. Why did it take a year? >> Because the neurons would die before they could make a recording. It took a >> year or trying and failing before blind luck finally worked in their favor. >> >> OK, so what does a synaptic response curve look like in our brain? Is it >> linear like many people presume? NO ONE KNOWS. Everything written on this >> subject is pure speculation. >> >> I see only one possibly viable way through this problem. It will take two >> parallel research efforts: >> 1. One effort is purely theoretical, where the optimal solutions to >> various processing problems is first exhibited, then the best solutions that >> can be achieved in a cellular architecture are exhibited, then potentially >> identifiable features are documented to guide wet-science efforts to >> confirm/deny these theories. >> 2. The other effort is a wet science effort armed with a scanning UV >> fluorescence microscope (or something better if something better comes >> along) that is charged with both confirming/denying identifiable details >> predicted by various theories, and with producing physical diagrams of >> brains to guide theoretical efforts. >> >> At present, there is not one dollar of funding for either of these >> efforts, so I expect to stay at the 2 micron point for the foreseeable >> future. >> >> Steve Richfield >> ================ >> > ------------------------------ > *agi* | Archives <https://www.listbox.com/member/archive/303/=now> > <https://www.listbox.com/member/archive/rss/303/> | > Modify<https://www.listbox.com/member/?&;>Your Subscription > <http://www.listbox.com/> > ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=106510220-47b225 Powered by Listbox: http://www.listbox.com
