Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURRENT

NMODL and the Channel Builder.
Post Reply
shailesh
Posts: 104
Joined: Thu Mar 10, 2011 12:11 am

Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURRENT

Post by shailesh »

I recently learnt that it is more appropriate to make use of ELECTRODE_CURRENT instead of NONSPECIFIC_CURRENT when extracellular mechanism is present. I went through the NEURON book and this is what I understood:

ELECTRODE_CURRENT
> Positive values of current would depolarize the cell
> Above suggests that ELECTRODE_CURRENT is injecting current into a cell
> Able to cause changes in the extracellular potential vext

NONSPECIFIC_CURRENT
> Positive values of current would hyperpolarize the cell
> Above suggests that NONSPECIFIC_CURRENT is membrane current

In the NEURON book, gap junctions have been developed using NONSPECIFIC_CURRENT. I was advised to use ELECTRODE_CURRENT for the same in the presence of extracellular mechanism. I was wondering why this change is required. Can't NONSPECIFIC_CURRENT cause changes in the extracellular potential vext? And if this change is required, why is it not a better idea to always have gap junctions, and possibly other mechanisms, implemented using ELECTRODE_CURRENT always?

Thanks,
Shailesh Appukuttan
shailesh
Posts: 104
Joined: Thu Mar 10, 2011 12:11 am

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by shailesh »

Found the following post with a whole lot of info on ELECTRODE_CURRENT & NONSPECIFIC_CURRENT:
http://www.neuron.yale.edu/phpbb/viewto ... 2475#p9807

The following, I believe, relates to a part of my query:
Facts which apply to any section that has NEURON's extracellular mechanism:
1. For a NONSPECIFIC_CURRENT, v refers to vinside-voutside_next_to_the_membrane, i.e. the true transmembrane potential.
2. For an ELECTRODE_CURRENT, v refers to vinside (i.e relative to ground, that is, the sum of transmembrane potential and any radial voltage drop across the extracellular mechanism).
Would appreciate it if someone could explain how this links up with my post and resolve my query.
ted
Site Admin
Posts: 6286
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by ted »

The concepts are skew to each other

By definition, a mechanism that generates a current will affect charge balance. NEURON distinguishes between two kinds of mechanisms that affect charge balance:
those that do so by producing a transmembrane ionic flux, e.g. by active transport or flux through ion channels,
and those that do so by injecting current into a cell via an electrode.

An NMODL specification of a mechanism that produces a transmembrane ionic flux will contain a USEION x WRITE ix statement if you want it to affect mass balance of some ionic species x, or NONSPECIFIC_CURRENT i if you don't want it to affect mass balance. Occasionally it is useful for a NEURON block to contain both a USEION and a NONSPECIFIC_CURRENT statement, e.g. a sodium transport mechanism that WRITEs ina will be electrogenic unless it either {WRITEs some other ionic current that counters the effect of the Na flux} OR {produces a NONSPECIFIC_CURRENT i where i = -ina}.

ELECTRODE_CURRENT is used for representations of instrumentation that inject current into a cell, hence the name "electrode". Think "sharp microelectrode" or "patch clamp electrode" or "clorided silver wire down the interior of a squid axon." Don't use it for gap junctions or synapses or any other current that is generated by transmembrane ion flux such as active transport or flux through an ion channel.
I recently learnt that it is more appropriate to make use of ELECTRODE_CURRENT instead of NONSPECIFIC_CURRENT when extracellular mechanism is present.
. . .
In the NEURON book, gap junctions have been developed using NONSPECIFIC_CURRENT. I was advised to use ELECTRODE_CURRENT for the same in the presence of extracellular mechanism.
That's news to me, as in "I would almost eat my hat if this proves to be the case."
ted
Site Admin
Posts: 6286
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by ted »

On further reflection, I realize I was perilously close to having to eat my hat.

If one is interested in the extracellular field produced by neuronal activity, one would not want the current that flows through a gap junction to be reckoned as being dumped into the extracelluar space. After all, it's merely moving from one intracellular location to another, so it should make absolutely no direct contribution to the extracellular field. For that reason it shouldn't contribute directly to the extracellular mechanism's i_membrane.

So there is the conclusion: NMODL code that specifies a gap junction should produce an ELECTRODE_CURRENT if one is interested in the i_membrane reported by the extracellular mechanism.

Now, what have modelers actually been doing?

In ModelDB I found two different mod files for gap junctions that declare the current to be an ELECTRODE_CURRENT.

par_ggap.mod which is used in
http://senselab.med.yale.edu/modeldb/sh ... r_ggap.mod
http://senselab.med.yale.edu/ModelDb/sh ... r_ggap.mod
http://senselab.med.yale.edu/modeldb/Sh ... r_ggap.mod
http://senselab.med.yale.edu/ModelDB/Sh ... r_ggap.mod

and halfgap.mod which is used in
http://senselab.med.yale.edu/ModelDb/sh ... alfgap.mod
http://senselab.med.yale.edu/modeldb/Sh ... alfgap.mod

Notice that these mechanisms calculate the current by a statement of the form
i = (vgap - v)*something
where something is either gapconductance or 1/gapresistance
and the voltage drop across the gap junction is written as (vgap - v) where v is the local membrane potential and vgap is the membrane potential on "the other side of the gap junction" because a positive ELECTRODE_CURRENT has a depolarizing effect.

There are many more mod files for gap junctions that declare the current to be a NONSPECIFIC_CURRENT.

These differ only in the value of the coupling resistance:
http://senselab.med.yale.edu/modeldb/sh ... ouple5.mod
http://senselab.med.yale.edu/modeldb/sh ... et\gap.mod
http://senselab.med.yale.edu/modeldb/sh ... ll\gap.mod
http://senselab.med.yale.edu/modeldb/sh ... FO\gap.mod
http://senselab.med.yale.edu/modeldb/Sh ... 09\gap.mod
http://senselab.med.yale.edu/modeldb/Sh ... ON\gap.mod
Ditto for the gap.mod that comes with NEURON's source code (see nrn/share/examples/nrniv/nmodl/gap.mod).

These specify a coupling conductance:
http://senselab.med.yale.edu/modeldb/sh ... n1\gap.mod
http://senselab.med.yale.edu/ModelDb/sh ... od\gap.mod
http://senselab.med.yale.edu/modeldb/Sh ... od\gap.mod
http://senselab.med.yale.edu/modeldb/Sh ... rk\gap.mod
http://senselab.med.yale.edu/modeldb/Sh ... 12\gap.mod

This one implements voltage dependent coupling
http://senselab.med.yale.edu/modeldb/sh ... apvdep.mod

This one implements time varying coupling (but works only for fixed time steps):
http://senselab.med.yale.edu/ModelDB/Sh ... apser2.mod

Notice that these mechanisms calculate the current by a statement of the form
i = (v - vgap)*something
where something is either gapconductance or 1/gapresistance
and the voltage drop across the gap junction is written as (v - vgap) where v is the local membrane potential and vgap is the membrane potential on "the other side of the gap junction" because a positive NONSPECIFIC_CURRENT has a hyperpolarizing effect.


So why did some use ELECTRODE_CURRENT while most used NONSPECIFIC_CURRENT? I'm not sure. I may have missed something in my rapid perusal of these ModelDB entries, but I don't think any of these papers addressed the electrical field produced by neuronal activity. The ones whose gap junction currents are implemented as ELECTRODE_CURRENTs were parallelized, which means that there is no guarantee that the coupled sections both exist on the same host--but I'd be surprised if that requires the current to be declared an ELECTRODE_CURRENT.
shailesh
Posts: 104
Joined: Thu Mar 10, 2011 12:11 am

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by shailesh »

Thanks Ted. That list of gap junction implementations is surely going to be handy for many of us.
As for, NONSPECIFIC_CURRENT & ELECTRODE_CURRENT, I suppose it would be fair to conclude that the major difference results in the membrane current (i_membrane) and rest being more or less similar.

The only question that remains might be whether there are any circumstances where NONSPECIFIC_CURRENT would infact be more appropriate than ELECTRODE_CURRENT for modeling gap junctions. The latter having the advantage of being accurate even in the presence of extracellular mechanisms.
ted
Site Admin
Posts: 6286
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by ted »

shailesh wrote:As for, NONSPECIFIC_CURRENT & ELECTRODE_CURRENT, I suppose it would be fair to conclude that the major difference results in the membrane current (i_membrane) and rest being more or less similar.
The membrane current and rest being more or less similar to what?
The only question that remains might be whether there are any circumstances where NONSPECIFIC_CURRENT would infact be more appropriate than ELECTRODE_CURRENT for modeling gap junctions.
I think it's largely a matter of personal preference. I would use NONSPECIFIC_CURRENT for all models that do not involve extracellular, simply because I think of gap junctions as being ion channels and I'm used to the convention that "positive ion channel current hyperpolarizes."

The only issue arises when a model combines gap junctions and the extracellular mechanism. In such cases, a "NONSPECIFIC_CURRENT gap junction mechanism" would still be OK as long as the contribution of the gap junction's current to i_membrane can be ignored. Note that this encompasses two situations:
1. When the gap junction does not make a significant contribution to i_membrane, i.e.
| i_gap | << | segment area * i_membrane |
2. When xg is so large (e.g. left at its default value of 1e9 S/cm2) that membrane current has no effect on extracellular potential. This is the case with models in which e_extracellular is used to control the potential on the outslde of the cell and the effect of the cell's own membrane current on extracellular potential is ignored.

So the cases in which ELECTRODE_CURRENT is preferable are those in which the gap junction current
1. makes a significant contribution to total membrane current
or
2. has a significant effect on extracellular potential.
shailesh
Posts: 104
Joined: Thu Mar 10, 2011 12:11 am

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURR

Post by shailesh »

That pretty much sums it up. Thanks.
The membrane current and rest being more or less similar to what?
I was referring to rest (as in calculated value of v, vext) being same whether gap junctions were implemented as NONSPECIFIC_CURRENT or ELECTRODE_CURRENT, and that only i_membrane would change. But your last response clarifies that there will be differences in the values of vext, given that the gap junctional current contribution to membrane current cannot be ignored.

Could I request you to take a look at the following post (its gone unnoticed before and so raking it up):
http://www.neuron.yale.edu/phpBB/viewto ... 654#p11748
Might take a few moments - unless I am headed in the wrong direction.
Kvogt
Posts: 14
Joined: Mon Mar 25, 2019 5:34 pm

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURRENT

Post by Kvogt »

Dear Ted

When you were referencing the NONSPECIFIC_CURRENT, you mention that it affects transmembrane ionic flux, while the ELECTRODE_CURRENT does not affect this ionic flux. What is this ionic flux referencing and how is it accounted for in the software?

Our confusion about this is because the transmembrane ionic flux is described as active transport through active channels but it doesn't affect the mass balance equations.

Thank you in advance
ted
Site Admin
Posts: 6286
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURRENT

Post by ted »

Declaring that some variable x is a NONSPECIFIC_CURRENT means that x is to be treated like any ionic transmembrane current, in that a negative value of x has a depolarizing effect. However, the charge carried by a NONSPECIFIC_CURRENT is not to be attributed to any particular ionic species. In other words, a NONSPECIFIC_CURRENT affects the charge balance equations but does not affect any mass balance equation.

Declaring that some variable y is an ELECTRODE_CURRENT means that it affects charge balance equations but with the sign convention used by experimentalists for currents injected through an electrode: positive current depolarizes. It is left to the readers of this thread to discover whether a USEION statement can be employed to attribute an ELECTRODE_CURRENT to a particular ionic species, so that it will affect mass balance equations, and to pursue all relevant previous discussions in this thread and elsewhere about ELECTRODE_CURRENT.
Kvogt
Posts: 14
Joined: Mon Mar 25, 2019 5:34 pm

Re: Differences between NONSPECIFIC_CURRENT & ELECTRODE_CURRENT

Post by Kvogt »

Dear Ted,

thank you for your reply.
Post Reply