Saltatory Conduction: single AP or not?

[somasimple]

this what?

which actually seems to move backward.

so it can't be a sound wave becaise it decays? hmmmm.

Sound is a wave and it strength decays with distance.
A wave is a wave, so...

 

[somasimple]

If the node is purely passive as expected you must have a backward effect:

Here is a simple working hypothesis:
The movie is not completed.
www.somasimple.com/flash_anims/node_01.swf

 

[granpa]

oh. yes, that is what I was referring to.

I'll try to watch your animation but my competer tends to freeze whenever I do.

 

[granpa]

but the backward effect is the opposite of what you would expect. it returns the axon to its resting state.

 

[somasimple]

oh. yes, that is what I was referring to.

I'll try to watch your animation but my competer tends to freeze whenever I do.

Update your flash player;
http://www.adobe.com//downloads/

 

[somasimple]

but the backward effect is the opposite of what you would expect. it returns the axon to its resting state.

Hmmm, no, it implies an effect we do not see on curves.

 

[granpa]

thanks. I did get to see it. I watched for about 20 seconds. it just repeats doesn't it?

 

[somasimple]

Update your flash player;
http://www.adobe.com//downloads/

or download it on your computer.

 

[granpa]

that is exactly what I would have expected but like I said the backward effect is exactly the opposite.

 

[somasimple]

thanks. I did get to see it. I watched for about 20 seconds. it just repeats doesn't it?

Yes it repeats. Just a working hypothesis to see the transition phases.

 

[somasimple]

that is exactly what I would have expected but like I said the backward effect is exactly the opposite.

If you do not see the expected prediction of a theory, change the facts or theory.

 

[granpa]

If you do not see the expected prediction of a theory, change the facts or theory.

hmm. facts or theory. hmmm

 

[granpa]

http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

imagine this: four nodes. nodes 1,2,3 and 4. nodes 3 and 4 would be locked (all nodes would normally be locked). meaning they won't pass signals. the first signal that reaches node 3 is from node 1. it has the effect of unlocking the node. this would take time so there would be a delay before node 2 fires. node 2 then fires and the ap passes through node 3 (without delay) to node 4 which it unlocks. there is a delay while node 4 is being unlocked. then node 3 fires.

just before each node fires it locks itself again. so when it does fire the ap can only go in one direction.

so the sequence for each locked node would be:
1 receive a weak signal from 2 nodes away that unlocks it
2 pass, without delay, a stronger signal from 1 node away to unlock the next node.
3 lock itself and then fire its own ap.

the point of course would be to prevent misfiring. this would explain why there is such a long and seemingly needless delay at each node.
and after each node locks itself the previous internode can begin to return to its resting state. hence the backward moving anti-action-potential I mentioned. (anti not because it is moving backword but because it returns the axon to its resting state)

 

[atyy]

We disagree how is the passive section and what is happening.

What is the disagreement? Model #2 in your post 21, and DaleSpam's cable equation seem to be basically the same, and Huxley and Stämpfli seem to mention both for the internode:

HS p329 cf Model #2: "length of a fibre which does not include a node is closely similar to ... a resistance and capacity in parallel...".

HS's p333 cf cable equation: the bottom equation has a second order derivative in space, and first order derivative in time.

The models are two low pass filters and only the cutoff frequency will be changed without any phase change for such a signal. NO DELAY.[/url]

Don't RC circuits have frequency dependent phase shifts? Also, the time course of a signal depends on the presence of all its Fourier frequencies, so wouldn't a frequency dependent reduction in amplitude (even without a phase shift for each Fourier component) produce a shape change in the time course anyway?

The passives sites must exhibit what the active are doing in both directions because they are passive.

after the delay, the beginning of the peak at one node coincides with the beginning of the downstroke of the previous internode. which actually seems to move backward.

If the node is purely passive as expected you must have a backward effect

HS discusses how the backward effect is seen in their data (p323): "surprising feature that the descending phase occurs earlier at the distal than at the proximal end of the internode ... Graph C in one internode and Graph B in the next more distal internode represent different aspects of the same disturbance spreading symmetrically from the node separating them.

 

[somasimple]

What is the disagreement?

Delay in the internode propagation.

a resistance and capacity in parallel...

This does not tell us how they are connected.
Why the capacity is omitted since it is 40 time greater than at node?

Don't RC circuits have frequency dependent phase shifts? Also, the time course of a signal depends on the presence of all its Fourier frequencies, so wouldn't a frequency dependent reduction in amplitude (even without a phase shift for each Fourier component) produce a shape change in the time course anyway?

Phase shift is not equal to delay.
BTW, the better way to test such models is sinusoidal signals.

surprising feature that the descending phase occurs earlier at the distal than at the proximal end of the internode

Graph it! It seems normal with a decay. (see the pictures I provided)

Granpa,
Make a drawing: I'm lost.

 

[somasimple]

http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=9#page
This comes from page 323:

This spread takes place with a finite velocity (not necessarily constant) so that graph B becomes later, and graph C earlier towards the distal end of each internode.

Ditto!
An attenuation that occurs with distance must take time.

 

[somasimple]

Granpa,
Make a drawing: I'm lost.

imagine this: four nodes. nodes 1,2,3 and 4. nodes 3 and 4 would be locked (all nodes would normally be locked).

Hmmm, no: It is better to lock something to ensure the transmission. Locking something before the transmission will be the best source of problem and no transmission.

 

[Dale]

A theory must describe all facts and make logical links between them. ... Why are you reducing the field of discussion?
Sub-threshold activity, as the whole cable theory, describes a facet of a thing that has many others.

This is completely wrong. Each scientific theory has some limited domain of applicability. For example, Maxwell's equations do not model the orbits of planets nor does it model the photoelectric effect, such things are outside of its domain. Maxwell's theory of EM does not need to "describe all facts", it only needs to describe and link facts within its domain. The search for a "Grand Unified Theory" or "Theory of Everything" is ongoing, and even if such a theory were available it would likely be too cumbersome to apply to neurons.

The domain of cable theory is sub-threshold activity, it always has been. Did you not understand that? The HH model describes supra-threshold activity and does not include any propagation mechanism. The cable model includes a propagation mechanism, but does not describe supra-threshold behavior. I thought that was understood, perhaps this is the real problem.

Please choose the electric model that mimics this activity and give us some values?
I understand why you continuously refuse...
Langauge barrier? I do not think so.

I have told you the model I support several times: the standard HH model and cable equation. Since these are the standard models used my mainstream scientists there are plenty of references describing in exhaustive detail their use and their experimental validation. You know that perfectly well since you have read many of these papers. What could I possibly put in a post than would be more informative that what you already have read? I have neither the time nor the inclination for such a pointless pursuit.

 

[somasimple]

I have neither the time nor the inclination for such a pointless pursuit.

That is not a problem. I have the same inclination.

 

[somasimple]

The HH model describes supra-threshold activity and does not include any propagation mechanism.

http://en.wikipedia.org/wiki/Hodgkin-Huxley_model

The Hodgkin–Huxley model is a scientific model that describes how action potentials in neurons are initiated and propagated. It is a set of nonlinear ordinary differential equations that approximates the electrical characteristics of excitable cells such as neurons and cardiac myocytes.

(sic)!

 

[granpa]

Hmmm, no: It is better to lock something to ensure the transmission. Locking something before the transmission will be the best source of problem and no transmission.

the idea is to prevent misfirings. so yes the idea is to create no transmission at all until everything is ready and the first few nodes are unlocked.

at anyone time no more than 3 nodes would be unlocked

 

[granpa]

every ap is initiated at one node (which had just locked itself so the ap con only go in one direction) passes without delay through the next node (which then gets ready to fire) and finally the by now much weakened ap strikes the third node which is locked so the signal ends there. but in striking the third node it unlocks it so the next ap can pass.

when each node fires it first locks itself so the normal state of all nodes is to be locked

 

[somasimple]

Suppose a node remains locked => no transmission.
We have seen that APs are overlapping.
If a locking mechanism exists (high probability) it may function when a node is already firing and initiating an AP at the next node. It may avoid node interactions.

 

[somasimple]

when each node fires it first locks itself so the normal state of all nodes is to be locked

contradiction: nodes are locked at firing and at rest.

 

[granpa]

contradiction: nodes are locked at firing and at rest.

to signals coming from one direction.

each node is unlocked by one signal
it then passes a second signal
at then locks itself and then fires. it doest need to be unlocked to fire because it isn't passing a signal. its creating a signal which only goes one way from the node.

here is a very good picture of what I am talking about:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

 

[somasimple]

here is a very good picture of what I am talking about:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

Elaborate!

 

[granpa]

contradiction: nodes are locked at firing and at rest.

locked means it won't pass a signal. not that it can't initiate one. but if it does initiate one then the ap can only go in one direction.

 

[granpa]

curve B after the delay at the node fires. this is the firing of the node. it travels to the next node (which does not yet fire) and passes straight through without delay becoming the horizontal part of curve A and finally ending at the second node which is locked. while that node is being unlocked there is a delay. then the next node fires and repeats the process. as each node fires a backward propagating anti-action-potential (curve C) restores the previous internode to its resting state. that is why the node must lock itself before firing.

 

[granpa]

so each node receives 2 signals. a weaker one that unlocks it and a stronger one that causes it after a short delay to fire.
a single signal alone isn't enough to cause it to fire. this also helps to prevent misfiring.

 

[somasimple]

curve B after the delay at the node fires. this is the firing of the node. it travels to the next node and passes straight through without delay becoming the horizontal part of curve A and finally ending at the second node which is locked. while that node is being unlocked there is a delay. then the next node fires and repeats the process. as each node fires a backward propagating anti-action-potential (curve C) restores the previous internode to its resting state. that is why the node must lock itself before firing.

The A curve shows rising phases of APs
The B curve shows peaks
The C one: falling phases.

Since there is decay in internode then all is OK.
Your process is too complicated and threshold + delay explains fully the process.

 

[somasimple]

a single signal alone isn't enough to cause it to fire.

No, a single is largely sufficient.

 

[granpa]

No, a single is largely sufficient.

the first signal is small (decayed). and in fact is usually just considered to be part of the rising phase of a single ap which is in fact 2 aps. the older decayed one followed rapidly by a newer more intense one. so it would appear that one signal was sufficient to trigger it but its actually 2

 

[granpa]

The A curve shows rising phases of APs
The B curve shows peaks
The C one: falling phases.

Since there is decay in internode then all is OK.
Your process is too complicated and threshold + delay explains fully the process.

I know what ABand C are supposed to show. but I think they are wrong. A is the older decayed signal from further away. B is the actual signal from a freshly opend gate. and C is the backward propagating anti-action-potential.

I don't understand why you think that's complicated. is it any more complicated than a human eye? or the human brain? seems well within the capability of evolution. moreover it explains the seemingly needless delay at each node. it has to wait for the node down the line to become unlocked. and the whole process is to prevent noise and misfirings.

seems pretty reasonable to me.

not to mention supported visually by this:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

 

[somasimple]

taking two signals implies a delay and AP that is larger in duration.
Since there is a delay, it would be very easy to verify your hypothesis: axon elongation => enlarges or shrinks the AP duration...
I keep the single signal processing.

 

[somasimple]

I know what ABand C are supposed to show. but I think they are wrong.

Facts are facts: They were recorded and these curves are results of simple computations.

 

[granpa]

taking two signals implies a delay and AP that is larger in duration.
Since there is a delay, it would be very easy to verify your hypothesis: axon elongation => enlarges or shrinks the AP duration...
I keep the single signal processing.

none of that makes any sense. axon elongation wouldn't have any effect on the ap duration. you have apparently misunderstood something i said. and what delay are you talking about. the only delay is the one that is apparent in the graph. about 0.1 ms at each node.

 

[granpa]

Facts are facts: They were recorded and these curves are results of simple computations.

I don't contest the facts. those curves are the basis of everything i am saying. why would i contest them?

look at one node on the axon. first it receives the A signal then the B signal then then it fires then receives the C signal. i think the A signal is the ap from 2 nodes away. the B signal before the delay is the ap from 1 node away. the B signal after the delay is the node itself firing. the C signal is the backward propagating anti-action-potential which is resetting the previous internode to its resting state.

 

[granpa]

taking two signals implies a delay and AP that is larger in duration.
Since there is a delay, it would be very easy to verify your hypothesis: axon elongation => enlarges or shrinks the AP duration...
I keep the single signal processing.

i'm not sure i understand you but you appear to be forgetting that each ap passes through 2 internodes before being stopped by a locked node. so a single node con receive 2 signals just as fast as before. there is no delay.

the first signal is decayed and doest cause the node to fire. it just unlocks it. the second one causes it to fire (after a short delay).

 

[somasimple]

Facts are facts: These curves are results of simple computations.

Computations may distort the facts in that case. But I cant' how you compute your two signals => They must be added and there is a delay.

 

[granpa]

Computations may distort the facts in that case. But I cant' how you compute your two signals => They must be added and there is a delay.

i already told you in post 87. the only delay is the one everyone already agrees on. there are 2 signals because the ap passes through the first node it reaches (which after a short delay fires) and continues on without delay to the next node which it unlocks.

i don't understand what it is that you don't understand. please ask more specific questions.

 

[granpa]

this is just an idea that I am presenting. it isn't established fact. the facts are here:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

nodes 1234. nodes 3 and 4 and all further nodes are locked. meaning that they will not pass signals.
ap=action potential

t0 node 1 fires producing ap1 which moves at the speed of sound in water. 1500 m/s
t1 ap1 almost instantly reaches node 2 and passes through WITHOUT DELAY
t2 ap1 almost instantly reaches and ends at node 3 and unlocks node 3 (which takes some time)
t3 after 0.1 ms node 2 fires producing ap2 which moves at the speed of sound in water. 1500 m/s
t4 ap2 almost instantly reaches node 3 and passes through WITHOUT DELAY
t5 ap2 almost instantly reaches and ends at node 4 and unlocks node 4 (which takes some time)
t6 after 0.1 ms node3 fires producing ap3 which moves at the speed of sound in water. 1500 m/s

t0 node 1 fires producing ap1
t1 ap1 almost instantly reaches node 2
t1-t3 delay of 0.1 ms at node 2 before it fires
t3 node 2 fires producing ap2
t4 ap2 almost instantly reaches node 3
t4-t6 delay of 0.1 ms at node 3 before it fires

there is therefore only one delay and it is the 0.1 ms one that everyone already agrees on. so it takes 0.1 ms for an action potential at one node to create an action potential at the next node which is typically 1 or 2 mm away. that gives a net speed of 10-20 m/s. if there were no delay at each node then the signal would move at 1500 m/s. the speed of sound in water. (thats just a guess but its certainly at least a good fraction of that speed)

just before each node fires its ap it relocks itself so the ap can only go in one direction and the previous internode can immediately begin to return to its resting state. this is seen in curve C:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

 

[somasimple]

https://www.physicsforums.com/showpost.php?p=1897614&postcount=66

This spread takes place with a finite velocity (not necessarily constant) so that graph B becomes later, and graph C earlier towards the distal end of each internode.

finite => delay
+ delay to initiate the next AP since there is a decay in internode.

So I reject, one more time, your point of view. Sorry.

 

[granpa]

i thought we agreed that the speed of the ap through the internode was around 1000 m/s and all the delay came at the node as is suggested by the data here:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8

 

[somasimple]

i thought we agreed that the speed of the ap through the internode was around 1000 m/s and all the delay came at the node.

It gives 1~2µs for the internode and since the decay is quite 1/3 => total delay around 20 µs.

 

[granpa]

according to this:
http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8
the delay at the node is about 0.1 ms or 100 microseconds. a 20 microsecond delay would be almost negligible

 

[granpa]

travelling at 1000 m/s an ap will travel the 1-2 mm from node to node in about 1 or 2 microseconds. so i guess we agree on that. whether decay results in a delay i don't know but i don't see what it matters. the inherent delay at the node swamps it out anyway.

i fail to see what any of this has to do with the idea that nodes lock and unlock

 

[somasimple]

whether decay results in a delay i don't know but i don't see what it matters.

A lot, since the threshold will be reached later at next node.

 

[atyy]

HH model for unmyelinated axon:
-Describes AP at a point and its propagation.
-Considers active and passive circuit components uniformly distributed along the axon.
-Is a wave equation with a well-defined propagation velocity which matches experiement.
-Reduces in a certain limit to the linear passive cable equation which does not have a well defined velocity.

HS model for myelinated axon
-Nodes active, internodes passive
-Internodes considered as resistor and capacitor in parallel (I think I know what they mean, but agree with somasimple it's not obvious), and apparently equivalently with an equation that resembles the linear passive cable equation.
-Expected backwards propagation from node into internode is apparently seen in the data and discussed.
-No explicit calculation of internode velocity, but heuristic and dimensional arguments are given for its form.

FH model data for myelinated axon
-FH model is standard reference for myelinated axon
-Only FH 1964 seems to be available to me, and does not describe propagation, but there may be other FH papers.

 

[granpa]

not much later. not enough to make any difference. the delay at the node is already 100 microseconds anyway. i don't see what internode delay has to do with anything at all much less whether nodes lock or unlock.

 

[granpa]

-Expected backwards propagation from node into internode is apparently seen in the data and discussed..

http://www.pubmedcentral.nih.gov/pagerender.fcgi?artid=1392492&pageindex=8
but the backward propagation that is seen (at least in the data i saw) isn't a backward propagating ap. its an anti-ap. it doesn't depolorize the axon. it returns it to its resting state.

but that's not the issue at the moment.