Units of currrents in shape plot

[Danko Georgiev]

Dear Ted,

I would like to ask whether all currents are plotted by default in units of [mA/cm^2]? I have searched through The NEURON book and it seems that these are the assigned units for sodium, potassium and calcium currents within the various source codes.

Regards, Danko

[ted]

Good question. First, a bit of background information. Suppose you have a mechanism that WRITEs an ionic current, e.g. the NMODL code that describes the mechanism has a
USEION x . . . WRITE x
statement in its NEURON block. If you attach this mechanism to a section, that section automatically gets another mechanism called x_ion. It is the x_ion mechanism that keeps track of all x currents that are generated by mechanisms attached to that section. The x_ion mechanism adds these up, and reports their sum as ix_x_ion. There are three important facts about ix_x_ion:
1. it is a range variable
2. it has density units--mA/cm2--regardless of whether the mechanisms that WRITE ix are density mechanisms, point processes, or some combination of density mechanisms and point processes.
3. although you can refer to it by its full name "ix_x_ion", it has an alias which is more convenient: ix

So ina, ik, ica, icl, iwhatever--if they exist--all have density units.

In a RangeVarPlot, the units of the dependent variable are the units that belong to that variable. Examples:
--diam is in um (yes, diam can be plotted as a function of position)
--voltages are in mV
--concentrations are in mM
--ionic currents that belong to density mechanism are in density units (mA/cm2). Examples include ica, il_hh, i_pas.

"Suppose my model has many instances of a synaptic mechanism point process called GABAb that WRITEs ik, and these are distributed over the surface of the cell. This is the only mechanism that WRITEs ik. Can I create a RangeVarPlot that shows the potassium currents generated by multiple instances of the GABAb mechanism?"

You could try, but I'm not sure how useful it would be. You can't plot GABAb.ik because hoc will complain that no such variable exists. The variables that do exist will have names like GABAb[0].ik, GABAb[1].ik, etc., so in principle you could make one plot that shows GABAb[0].ik, another that shows GABAb[1].ik, etc.. For each of these plots, the dependent variable would be in units of nA. But each plot would be nonzero at only one location: the segment that contains that particular instance of the GABAb mechanism.

You could instead make a RangeVarPlot of ik, and that would be 0 in all sections that lack GABAb and nonzero in the sections that contain one or more instances of GABAb. But now the units of the independent variable would be mA/cm2, and the magnitude of the independent variable would be inversely proportional to the surface area of each segment (compartment). Also, in any segment that contained multiple instances of GABAb, the plot would show the sum of the potassium current densities generated by all of the GABAb instances at that location.

[Danko Georgiev]

Dear Ted,

Thanks for your prompt reply. In regards to currents only now I have noticed in The NEURON Book the remark that the value given in mA/cm^2 is equivalent to the one given in nA provided that the compartment has a surface area of 100 um^2 (indeed 1 mA/cm^2 x 100 um^2 = 1 nA). Now is clear that in the shape plots are plotted not currents but current densities, so it makes perfect sense. It is good that you provided also a list with other parameters and their default units.

[ted]

There are many gaps in the documentation. To find the answers for some questions it is necessary to run tests, as you did. And even when the answer seems clear, sometimes a test produces a surprising result.