Exploring Sensory Networks: How Physics Models Explain Our Senses

[0rthodontist]

I saw this link at the right. This is incredibly interesting to me. It is posted here because I am not interested in the specific biology.
http://www.sciam.com/article.cfm?chanID=sa003&articleID=000CCE51-4865-1449-886583414B7F0000&ref=rss

Now physicists have shown how the mathematical models that describe phase transitions in physical systems might also explain our capacity to hear, see, smell, taste and touch. Mauro Copelli and Osame Kinouchi of the University of Sao Paulo in Brazil used a mathematical formula to show how a random network of "excitable elements," such as neurons or axons, have a collective response that is both exquisitely sensitive and broad in scope. When subtle stimuli hit the network, sensitivity is improved because of the ability of one neuron to excite its neighbor. When strong stimuli hit the network, the response is similarly strong, following what are known as power laws--mathematical relationships that do not vary with scale.

Where can I learn more about this type of model?

 

[eeka chu]

Call me simple but I thought this was pretty much a done and dusted discovery?

Sensory neurons all (well, for all intents all of them) flow back into synaptic junctions and burst neurotransmitter into the cleft. If you have more than one axon summing into that cleft and a fixed level for the neurotransmitter to reach before the output fires, it seems pretty obvious that you have just increased the sensitivity of that pathway.

Say you have one junction with one input and one output. It takes four impulses from the input to discharge enough neurotransmitter into the junction to make the output fire once. You don't change the neurotransmitter requirements but add another input neuron, identical to the first, to that same junction. Each need now only contribute two impulses for the output to fire once, which will occur more rapidly than with just one input provided both neurons are being stimulated simultaneously (nerves send signals very, very, very, very slowly with comparison to a transistor, about 500 - 1000Hz max for a neuron and over 600,000,000,000Hz for the best transistors).

Seems the only thing they've done is describe this by a mathematical law, that adding inputs increases sensitivity.

If you want to read about neurology, try Cell publishing's Neuron. I have a subscription to it. It's heavy stuff!

 

[Mk]

Also, if you have any more neurology related questions see physicsforum's M&B and Biology boards.
https://www.physicsforums.com/forumdisplay.php?f=149
https://www.physicsforums.com/forumdisplay.php?s=&daysprune=30&f=82

 

[0rthodontist]

Is it really as simple as that? The article said it had something to do mathematically with physical state transitions and there was a biologist in the article who replied and said that the physicists' assertions were untested. Also increased propensity to fire isn't necessarily the same as increased perceptual sensitivity. A neuron that fires all the time isn't sensitive at all.

 

[eeka chu]

Also increased propensity to fire isn't necessarily the same as increased perceptual sensitivity. A neuron that fires all the time isn't sensitive at all.

Well, it wouldn't necessarily fire all the time just because it had more than one other neuron feeding into it, but it's noise floor probably would be raised. So that output neuron would be more prone to 'twitching' impulses.

It's just like stacking light sensors together in series, you only need a small amount of light to fall on each for output to be high enough to trigger an event. But any background light might also set off the event, as well as any transients in the self produced noise the sensors create during operation. I'm sure a few neurons probably fire every now and again just because some process within the neuron has accidentally triggered an impulse. Probably a lot more than I think given the number of things going on inside a human body that could affect the membrane voltage across a neuron.

When you think about a lot of neurons all coming together at one synaptic cleft to trigger another it's easy to think that the sensory neurons are in parallel, since they're not wired how they would be if they were in series electrically (e.g. one after the other and with the two ends of the series chain going into the synapse), but the 'answers' of each neuron are able to sum onto each other in the cleft which means their output is being handled in a serial manner; each connection entering a cleft doesn't produce it's own unique neurotransmitter, it's one of a small selection. Of coarse, they can sum positively or negatively depending on which particular neurotransmitter is released.

I'm pretty sure this is just a mathematical model of how the process occurs, it's a very well accepted idea that neurons can function in groups to alter each other's sensitivity to some extent via the make-up of the synaptic clefts. The axons themselves can even hyperpolarise themselves so as to make it harder for an impulse to be triggered. Hmmm... example... if you get hit particularly hard the surrounding area might temporarily go a bit numb.

I suspect our nervous system is up to more interesting things than this. For instance, I'm reasonably sure large sections of our nervous system are being 'multiplexed' as it produces the most versatile system when a dynamic environment is present to it.

Your hearing is a good example. Try listening to a tune really carefully. Now try singing the lyrics to another tune. Simultaneously, try tapping the beat to another. It's virtually impossible to do because our auditory system is trying to find one 'program' or thread to run, it can't really parallel process multiple threads simultaneously (neither can a single CPU really as CPU's require blocks of program to run on, but they are much better at switching between them). That seems to occur for a lot of our sense and is one of the main reasons why illusions and magic work. Going further on again, sometimes events conflict even when they don't seem to be directly competing with each other for the same areas of brain activity. Trying to do multiplication whilst moving around for instance. Which suggests that areas are much more shared than at first seems obvious and are helping each other in some way.

Multiplexing in this manner would help free up our brain from pointless information processing. While your sitting reading this, your brain probably doesn't need to know what the precise temperature across your foot is, so you just have a general feeling of how hot or cold it is. Whereas if I was sitting under your desk holding a lighter under part of your toes, you'd instantly start paying attention to your foot. As you start adding more and more serious events happening simultaneously, it get's hard to focus on them all. For example, when holding a lighter under you foot and simultaneously stabbing another part of your body with a needle, it doesn't necessarily become twice as detailed. Infact, quite often, the pain of one seems to distract or detract from the lesser. This is quite a clear sign of a.) not enough spare processing room in the brain and b.) the input to that processing is being multiplexed depending on it's level of apparent severity.

 

[0rthodontist]

What does that have to do with state transitions, and why was the biologist saying the physicists' model was untested?