www.neuron.yale.edu

Posted: **Tue Sep 05, 2006 4:05 pm**

Hi,

i have created my cell (soma and dendrites) and with "hh" and "pas" inserted mechanism. to insert a point process , i 'm goinig through

Tools/Point Processes/Maneger/ Electrod

that contains ICalmp, Vclamp, Clamp Family.
I like to use this optioon becuase it has (in clamp Family ) 3 interesting option (conditionig ,test and return ) level.

I have read in privious post that we are recommended to use SECalmp insted of VClamp.

Finally this is my question, i like to insert a SEClmap in my cell and also have conditioning, test, return level.

How can we do that? any sugestion?

Posted: **Tue Sep 05, 2006 4:36 pm**

Read this under Hot tips:
How to change parameters during a simulation
But why bother, if what you're doing works, is convenient, and isn't causing any problems
that you can see in your simulations? As they say, "if it ain't broke, don't fix it."

Posted: **Wed Sep 06, 2006 12:22 am**

Dear Ted,

Thank you for your advise. I'm trying to undrestand the code in the "How to change parameters during a simulation" post.

I have some voltage clamp protocols that i like to applay in my simulations. The voltage clamp consisted in:

- conditioning the cell at various voltage level (from -125 mv to -60 mv) for 1 sec and then stepping the voltage to a fixed value (-30 mv) (-45mv),(-55mv),(-65mv)

and

-- conditioning the cell at -105 mv for 1 sec and then stepping to (-100mv to 10 mv)

I guess for first one is "vary conditioning level" and for second is
"vary test level" options in Vclamp Family(I/V Clamp Electrode).right?

where is series resistance option? where can i change this parameter?
(I see this parameter is SEClamp not in VClamp)

what's you suggestion?what is the best way to applay these voltage clamp protocol?

Thanks in advance.

Posted: **Wed Sep 06, 2006 10:28 am**

Good questions.

reza_rzm wrote:first one is "vary conditioning level" and for second is
"vary test level" options in Vclamp Family(I/V Clamp Electrode).right?[/quite]
Absolutely correct.
where is series resistance option? where can i change this parameter?
(I see this parameter is SEClamp not in VClamp)
VClamp is itself a model of a negative feedback voltage clamp circuit. The source resistance
of such a circuit is approximately equal to the electrode resistance divided by the open loop
gain of the circuit. For the VClamp mechanism, these parameters are called rstim (units are
megohms, default value 1) and gain (dimensionless, default value 1e5). Default source
resistance of VClamp is therefore 10 ohms.

To see these parameters in the PointProcessManager, you may need to expand the
height of that tool by clicking on its bottom edge and dragging down.

what's you suggestion?what is the best way to applay these voltage clamp protocol?

Depends on how you intend to use your model. The VClamp Family option in the GUI's
Electrode tool is convenient for interactive, exploratory use, and it's great for educational
applications. But if you need to generate dozens of families of voltage clamp data, and
especially if you need to do this day after day, having to fiddle with the GUI in order to
change parameters quickly becomes tedious. You can do that with a VClamp or
SEClamp and a couple of for loops that change whatever amplitudes you need to
change. This example assumes that you're using an SEClamp called clamp, and that
you're capturing the clamp current to a Vector.

Code: Select all

proc grind() { local vcond, dvcond, ncond, vcmd, dvcmd, ncmd, i, j
  vcond = $1
  dvcond = $2
  ncond = $3
  vcmd = $4
  dvcmd = $5
  ncmd = $6
  for i = 0, ncond-1 {
    clamp.amp1 = vcond+i*dvcond
    for j = 0, ncmd-1 {
      clamp.amp2 = vcmd+j*dvcmd
    }
    run()  // standard run system's command that initializes and runs a simulation
    postprocess()  // your proc that does whatever you want to the recorded clamp results
  }
}

grind(-65, 10, 3, -60, 10, 4)

Posted: **Thu Sep 07, 2006 12:05 pm**

Dear Ted,

Thaks so much!
according to your suggestion i have used your code. in this code i like to write a commande to save current in a file. Indeed, the current that appears in VClamp.i graph. which varibale show this current?

And other question, as i know in voltage clamp experiment
1- we inject a electronic current that compensate ionic current , eg. I_inj=sum(I_na,I_k,I_ca,...) to keep membrane potential constant.

2-and when we are gonig to record a specific current,for example ca current, we block other ionic channels, then we say I_inj=I_ca.

right?

But, in numerical simulation (NEURON) we have directly access to any current in every palce, for expamle soma.ica(0.5) that denotes ca current in soma.

The problem: the unit of soma.ica(0.5) is microamper whiles current injected in voltage clamp (Vclamp.i ) is nanoamper.

so, for report a ionic current behavior during a voltage clamp, what we should consider, e.g soma.ica(0.5) or VClamp.i ?

I know, i miss some concepts here. could you please help me to find out?

Reza

Posted: **Fri Sep 08, 2006 10:20 am**

reza_rzm wrote:i like to write a commande to save current in a file. Indeed, the current that appears in VClamp.i graph. which varibale show this current?

If you created the graph with the GUI, the name of the variable appears on the graph's
canvas.

in voltage clamp experiment
1- we inject a electronic current that compensate ionic current , eg. I_inj=sum(I_na,I_k,I_ca,...) to keep membrane potential constant.

True.

2-and when we are gonig to record a specific current,for example ca current, we block other ionic channels, then we say I_inj=I_ca.

Yes, pharmacological isolation of other currents is necessary for experiments on
real cells.

The problem: the unit of soma.ica(0.5) is microamper whiles current injected in voltage clamp (Vclamp.i ) is nanoamper.

False and true. In NEURON, transmembrane currents have density units--soma.ica is
in milliamps per square centimeter. A voltage clamp is what NEURON calls a "point
process," and point process currents have absolute units--nanoamps. For more
information about units in NEURON, see
What units are used in NEURON?
in the Hot tips section of this forum
https://www.neuron.yale.edu/phpBB2/viewforum.php?f=28

for report a ionic current behavior during a voltage clamp, what we should consider, e.g soma.ica(0.5) or VClamp.i ?

If the model has a single compartment (only one section, and nseg is 1), and no other
currents are present, the voltage clamp current will be the sum of total capacitive current
and total calcium current. My personal preference would be to focus on calcium current
density because it does not depend on the size of the cell, and it is not contaminated by
capacitive charging artifact.

Posted: **Fri Sep 08, 2006 12:02 pm**

Dear Ted,

Thanks.

In model, I have soma and a 275 dendrits. for every section i set nseg=1. Since, in experimental result, pepole report current, e.g. ca-current, in unite of (micro,nano,pico)amper not as a density unite,mA/cm2. i like to have the same result in my simulation.

In my model,the unite of soma.ica(0.5) is mA/cm2 and electronic current injection in voltage clamp is nA.

To compare VClamp.i with soma.ica(0.5) at least numericaly, i set g_na,g_k in soma equall zero, and even g_l=0 for whole cell. so theoretically the injection current during voltage clamp should be equall to ca-current.right?(or still there is other thing?)

Total calcium current can be obtained from

soma.ica(0.5)[mA/cm2]*area(0.5)[um2].

again is in order of microamper. not nanoamper. even if i add to above formula sum of calcium current in every dendrtic segment times to area of dendritic segamnt i will obtain bigger number.

please give me your points,

I await!!

Posted: **Fri Sep 08, 2006 1:54 pm**

reza_rzm wrote:In model, I have soma and a 275 dendrits.

Then it is very likely that
1. the voltage clamp does not have complete control of membrane potential throughout
the model cell
and
2. the voltage clamp does not capture all of the membrane current

Also, it almost certain that membrane current density is not uniform throughout the model.

for every section i set nseg=1

Did you have a particular reason for choosing that value?

Since, in experimental result, pepole report current, e.g. ca-current, in unite of (micro,nano,pico)amper

Some do, some don't. If you read Biophysics Journal you will see many papers that
report membrane currents in density units mA/cm2, and many others that report
currents in mA/uf (milliamps/microfarad). That's what experimentalists do if they can
measure cell surface area or cell capacitance, and they know that their voltage clamp is
able to control membrane potential over the entire surface of the cell and can capture all
of the cell's membrane current. Absolute current units are used by experimentalists who
don't know cell surface area, or can't establish good space clamp (i.e. can't control
membrane potential over the entire cell, and can't capture all membrane current).

To compare VClamp.i with soma.ica(0.5) at least numericaly, i set g_na,g_k in soma equall zero, and even g_l=0 for whole cell. so theoretically the injection current during voltage clamp should be equall to ca-current.right?(or still there is other thing?)

Good question. 20 years ago, experimentalists started trying to use channel blockers to
get better voltage clamp results from neurons. Some of them thought that, if they could
block all the channels that they weren't interested in, their voltage clamp would be able to
have better control of the membrane potential of the entire cell, and they'd be able to
make much more accurate recordings of current that passes through channels way out
on the dendrites. It didn't work. The problem is that cytoplasmic resistance and
membrane capacitance sit between the clamp electrode and distant membrane, and
act like a low-pass filter. Fast changes of membrane potential at the location of the
voltage clamp electrode are slowed and attenuated as they spread toward distant
membrane. Fast current changes at distance membrane are slowed and attenuated as
they spread back toward the clamp electrode. Even in the steady state, there is still
cytoplasmic resistance between the clamp electrode and most of the membrane of the
cell, which prevents perfect control of voltage.

Total calcium current can be obtained from

soma.ica(0.5)[mA/cm2]*area(0.5)[um2].

again is in order of microamper. not nanoamper. even if i add to above formula sum of calcium current in every dendrtic segment times to area of dendritic segamnt i will obtain bigger number.

Yep. The voltage clamp can only capture the charge that is in the near neighborhood of
the electrode. Calcium current that enters distant dendrites needs time to flow down the
dendrite to where the clamp's electrode is located. It's like a river. You're sitting at the
mouth of the river, where it empties into a lake, and you're watching how many liters are
flowing per minute into the lake. Upstream, for miles and miles along the river, hundreds
of people start dumping buckets of water into the river, but you won't see any immediate
change in the flow rate where you're sitting.

Posted: **Sat Sep 09, 2006 4:02 pm**

Dear Ted,

Thanks so much, for your attention and your patience!!!

Ok. let me ask other thing;

assume, we have a 1-compartment model ,e.g. a soma, that conatins leak, Na, k and Ca currents. And we like to study a current during voltage calmp.

in contrast to experimentalist, i don't need any blocker. i can record every current in any palce (That's numerical simulation!!)

so, for our case, in presence of Na, K and leak current (g_na, g_k and g_l are not zero),i can recorde Ca current (soma.ica(0.5)) for every step of voltage clamp. right?

But theoretically for this 1-compartment model (and nseg=1),without poor any space clamp error, i expect to have equal current for "soma.ica(0.5)" and "VClamp.i" when we set g_na=g_k=g_l=0. Since we expect VCalmp.i should compansate ica(0.5) during voltage clamp to keep membrane potential constant.

But, in my simulation, i have different number for soma.ica(0.5)[mA/cm2]*area(0.5) [um2] as total calcium current in soma and VClamp.i (nA).

I miss one important (or even maybe very simple) point, that i have this problem!!!!

as you suggest me before i will focus on current density for my simulation projects, instead of total current. Do you know any referece that they report calcium current in thalamocortical cell in unite of density?

Yours,
Reza

Posted: **Sat Sep 09, 2006 4:20 pm**

You might still see a capacitive current (dis)charging the membrane. If this is the only cause the currents should become equal after a short transient.

Other possible currents are the passive leak current (for which you should set g_pas=0 to block it) and a calcium pump. I don't know your model, but if you are reusing an existing model that is explicitly modelling calcium there is a fair chance that a calcium pump was included in the model.

Posted: **Sat Sep 09, 2006 6:37 pm**

Dear Raj,

Thanks for reply,

In my 1- compartment model, i have leak, Na, K and Ca current and(yes!) a calcium pump. for calcium pump i 'm useing the following code:

Code: Select all

TITLE Fast mechanism for submembranal Ca++ concentration (cai)
:
: Takes into account:
::	- increase of cai due to calcium currents
:	- extrusion of calcium with a simple first order equation
:: This mechanism is compatible with the calcium pump "cad" and has the 
: same name and parameters; however the parameters specific to the :pump  are dummy here.
:: Parameters:
::	- depth: depth of the shell just beneath the membran (in um)
:	- cainf: equilibrium concentration of calcium (2e-4 mM)
:	- taur: time constant of calcium extrusion (must be fast)
:	- kt,kd: dummy parameters
:: Written by Alain Destexhe, Salk Institute, 1995
:
INDEPENDENT {t FROM 0 TO 1 WITH 1 (ms)}

NEURON {
	SUFFIX cad
	USEION ca READ ica, cai WRITE cai
	RANGE depth,kt,kd,cainf,taur
}

UNITS {
	(molar) = (1/liter)			: moles do not appear in units
	(mM)	= (millimolar)
	(um)	= (micron)
	(mA)	= (milliamp)
	(msM)	= (ms mM)
}

CONSTANT {
	FARADAY = 96489		(coul)		: moles do not appear in units
:	FARADAY = 96.489	(k-coul)	: moles do not appear in units
}

PARAMETER {
	depth	= .1	(um)		: depth of shell
	taur	= 5	(ms)		: rate of calcium removal
	cainf	= 2e-4	(mM)
	kt	= 0	(mM/ms)		: dummy
	kd	= 0	(mM)		: dummy
}

STATE {
	cai		(mM) 
}

INITIAL {
	cai = cainf
}

ASSIGNED {
	ica		(mA/cm2)
	drive_channel	(mM/ms)
}
	
BREAKPOINT {
	SOLVE state METHOD euler
}

DERIVATIVE state { 
	drive_channel =  - (10000) * ica / (2 * FARADAY * depth)
	if (drive_channel <= 0.) { drive_channel = 0. }	: cannot pump  :inward
	cai' = drive_channel + (cainf-cai)/taur
}

( ica: is T-type Ca channel current)

in this "mod" file , i don't see any current ! just a diffrential eq. for internal calcium concentration. right? so i don't have pump current!

I set g_na=g_k=g_pas=0 and then, Current Graph/ plot what?/ in symbol name i have entered

soma[0].ica(0.5)-soma[0].i_cap(0.5)

and still i have big difference between this graph and Vclamp.i !!!

Posted: **Sun Sep 10, 2006 7:18 am**

Which leaves us with a fundamental question about the internal workings of NEURON. As far as I know the membrane potential is indeed calculated directly from the currents so you are right the pump current in your mechanism doesnot contribute to the ica current and therefore not to the clamp current needed.

Which brings me back to your dimension analysis:

Total calcium current can be obtained from

soma.ica(0.5)[mA/cm2]*area(0.5)[um2].

again is in order of microamper. not nanoamper.

assuming ica in these units is about 1 the current will be microampere if your area is for example 30um by 3000 um which is on the large side for a soma but not unreasonable for a pyramidal cell with the dendrites included. For the soma itself dimensional analysis gives a nA current:
[mA/cm2]=(1e-3/1e8)[ A/um2]=[10pA/um2] with typical somatic diameters of about 10 to 20 um this give a nA current.

So are you sure you got your conversion factors right?

Posted: **Sun Sep 10, 2006 10:23 am**

Dear Raj,

thanks,

In model, 1-comaprtmental model, the area of soma is 24058.317 um2.

during voltage calmp maximum current amplitude for "Vclamp.i" is about -12 nA, while in current graph that shows "soma[0].ica(0.5)-soma[0].i_pas" maiximum amplitude is -0.06 mA/cm2. (when g_na=g_k=g_pas=0)

so -0.06 [mA/cm2]*24058.317 (um2)=144.3444 mA !!!!

we can even test a demo file that "Karlo" had sent a post under "demo" topic.

you can find the demo here

Low Threshold Calcium Currents in TC cells (Destexhe et al 1998)
http://senselab.med.yale.edu/senselab/m ... ?model=279

from that demo file, select "Burst behavior in single compartment model"

1- tstop=1200
2- inseted of IClamp in I/VCalmp electrod secelt VClamp
according to their paper:
dur0=1000
dur1=200
amp0=-105
amp1=-80

3- open VClamp.i graph, and from a current graph that showes "soma[0].ica(0.5)-soma[0].i_pas"

4-set g_na=g_k=g_pas=0
5- celcius=24 // as the paper

6- in VCalmp family click on "Test Level"

That's it. you will see the graph, and other parametres.

However, parameters in this paper should be logical,(of course better than my simulation test).

I await to hear you!

Thanks,

Reza

Posted: **Sun Sep 10, 2006 1:35 pm**

Dear Reza,

I played around with the model you pointed out. There the difference between the clamp current and the capacitive current disappears if I go to smaller time steps 0.0025 ms in this case works.

If I plot SEVClamp.i and soma[0].i_cap*24058.317/100 together in one plot there is a pronounced difference when dt=0.025 which disappears with a smaller timestep, e.g. dt=0.0025 ms.

I set
gkbar_hh2=0
gnabar_hh2=0
g_pas=0
pcabar_itGHK=0

After putting : pcabar_itGHK=-2e-3
I still see good agreement between the currents.

So it seems to me that your ghost current is due to numerical errors. You might want to move to variable timestep, but you will have to reimplement or find a new calcium pump to be able to do that. If all your other mechansism are CVode proof that is. SEVClamp isn't it should be replaced by SEClamp for use with CVode.

Regards,
Raj

Posted: **Sun Sep 10, 2006 2:18 pm**

Dear Raj!

Thank you very much for your time!

I didn't understand, you set pcabar_itGHK=0 aslo!! so in this case while gna_hh2=gk_hh2=g_pas=0, => dVm/dt=0 , then there is no capacitive current!!!
so, what does it mean " difference between SEVClamp.i and i_cap"?

both are equal zero!

and other thing,as Ted wrote:

the voltage clamp current will be the sum of total capacitive current
and total calcium current.

so when pcabar_itGHK=8e-5 (the defult value in demo file),
then i expect to have this relation

SEVClamp.i=(soma[0].ica(0.5)+soma[0].i_cap(0.5))*area(0.5)

right?

but the right side is more bigger, 1000 times!!!!!!!!!!! even when dt=0.0025.

what do you think?

Thanks in advacne,

Reza

www.neuron.yale.edu

I/V clamp electrod

I/V clamp electrod