Measuring total membrane current in multicompartment model

[tine]

Dear all,

For a demonstration I want to show the attenuation of a tonic currents at some specific locations in the dendrites to the soma.

My plan of action is
- insert Iclamp at a specific location of the dendrites (for example 100 microm) and insert a long step of 10 pA
- measure currents at the soma (as in wetlab experiment with a second voltage clamp electrode).

I have a problem with the second step. How is the total membrane current defined in NEURON. I can find the individual currents (i.e. i_na) but I am not sure if I should add them all up. An earlier post pointed out that the non-specific currents are not taking into account with this method. This post suggested to add an extracellular matrix to the model and take the current flowing through this segment.

I am not familiar with extracellular mechanism.
- How do you define this mechanism for a model that consists of a detailed anatomical model?
- the current of this extracellular mechanism is still defined in xxx A/mm2 and I have to multiply this with the area to get the total current. But what is area in a multi-compartment model?

Any help is highly appreciated

[ted]

I want to show the attenuation of a tonic currents at some specific locations in the dendrites to the soma.

My plan of action is
- insert Iclamp at a specific location of the dendrites (for example 100 microm) and insert a long step of 10 pA
- measure currents at the soma (as in wetlab experiment with a second voltage clamp electrode).

Three points here.

1. Since you will be using a voltage clamp (SEClamp, please, and it would be a good idea to reduce its series resistance rs to a very small value, say 0.001 megohm) to capture the current that reaches the soma, you won't have to fiddle around with surface area of anything--the SEClamp's i variable is in units of nA.

2. If your injected current is 1 nA, SEClamp.i will automatically tell you what fraction of injected current gets to the soma.

3. If your model is completely passive, current transfer from dendrite to soma will be the same as voltage transfer from soma to dendrite. So if you inject a current step at some location X in a dendrite and look at the current captured by an SEClamp at the soma, then in a separate experiment voltage clamp the soma and apply a voltage step command while measuring membrane potential at dendritic location X, you will find that the ratio
Iinjected_at_X_in_dendrite/Icaptured_by_SEClamp_at_soma
equals the ratio
Vstep_applied_at_soma/Vobserved_at_X_in_dendrite
Also, the time course of Icaptured_by_SEClamp_at_soma will be identical to the time course of Vobserved_at_X_in_dendrite (may have to normalize the amplitudes to see this).

If your model is active, but your applied current or voltage step is small, current transfer from dendrite to soma will be approximately the same as voltage transfer from soma to dendrite.

NEURON's Impedance tools
NEURON Main Menu / Tools / Impedance
do many interesting things, including calculating voltage transfer ratios.

How is the total membrane current defined in NEURON.

Sum of capacitive and ionic currents.

I don't see why it would be necessary to insert extracellular or know membrane current if your conceptual experiment is as you describe, i.e. inject a current at one location and use a voltage clamp to record current at a different location.

[tine]

Dear Ted,

Thanks for your quick response. I conceptually understand your solution but I end up with SEClamp.i=0. Can you help me where my error is?

I inserted an Iclamp to insert tonic current step

Code: Select all

objectvar stim300
apic[14] stim300 = new IClamp(0.36)

stim300.del = 50
stim300.dur = 200
stim300.amp = -0.10 // nA

tstop = 68.7

I inserted an SEClamp to measure the current at the soma

Code: Select all

objectvar VCrec
soma VCrec = new SEClamp(0.5)
VCrec.rs=0.001
VCrec.dur1=1e9
VCrec.amp1=0

When I run the simulation

Code: Select all

proc run(){
  stdinit()
  continuerun(tstop)
  Save_SomaVm()
}

VCrec.i is zero for the whole simulation, while I do see a response at the soma (soma.v(0.5)) What do I miss?

[ted]

What happens if the IClamp and SEClamp are attached to the same section, e.g. soma(0.5)?

[tine]

I am not sure if it is possible to have them both simultaneously. If I have the most simple model, namely only soma,

Code: Select all

load_file("nrngui.hoc")

create soma

soma.nseg=1
soma.diam=20
soma.L=20
soma.Ra=123

insert hh

objectvar elec_Iclamp
soma elec_Iclamp= new IClamp(0.5)
elec_Iclamp.amp=0.1
elec_Iclamp.del=100
elec_Iclamp.dur=100

tstop=300

objectvar elec_Vclamp
soma elec_Vclamp=new SEClamp(0.5)

elec_Vclamp.i remains zero for the whole traces while the current clamp step does induce a number of action potentials. Is there a limitation to the number of point processes, or does a VEClamp only gives an i value it has a defined amplitude of its own?

[ted]

I think it's best that we just go back to your post from January 5 that mentioned

I end up with SEClamp.i=0

Can you zip up the hoc, mod, and ses files necessary to reproduce this, and email them to me
ted dot carnevale at yale dot edu
so I will be able to reproduce this myself?