Page 1 of 1

Modeling Gap Rectifying Junction

Posted: Tue Sep 28, 2010 1:39 pm
by nianwosuh
Please could you explain to me the difference between the Gap.mod modeled in NEURON Book listing 10.2 and that modeled in the on-line course under "linear circuits, rectifying junction"

The PARAMETER and BREAKPOINT sections of both mod files are different. Please could you explain the reason for these differences?

In building the Linear circuit in the 0n-line course under the "linear circuit rectifying junction", the Gap-junction was built using the channel builder as a density mechnasim, though you explain the reason was to make it possible for the file to be ran on-line in all machines. The question then is why was the Gap built as a density Mechanism and not as a point process which can also be built using channel builder?

I was able to reproduce the "linear cicuits, rectifying junction" in the on-line course without error.

Rather than using the rGap.ses, I tried to reproduce the linear circuit rectifying juction, using the Gap.mod file of listing 10.2. in two different ways

In the first approach, I created the "left and Right sections" using cell builder, and saved each section as a separate hoc file (called right and left sections .hoc each)
I wrote the hoc code for the instrumentation for inserting the Gap as in section 10.1.2.1 (usage), without the "creating a & b and their segements" but starting with "objeref" replacing a & b with right and left.

I wrote the initalization hoc code that opens the hoc file of the two sections created in cell builder, and the instrumentation.hoc file, I have this error message;

point process not located in a section
in Instrument_Gap.hoc near line 7
}
^
xopen("Instrument_Gap.hoc" )
execute1("{xopen("Instrument_Gap.hoc")}" )
load_file("Instrument_Gap.hoc" )
0
nrniv: Segmentation violation
in Instrument_Gap.hoc near line 9
Right g[0].loc(0.0001)


The second method was to repeat what you have under section 10.1.2.1 (Usage) of the NEURON book, only changing a & b to dend[0], dend[1]. I have a similar error message:
point process not located in a section
in /cygdrive/C/Documents and Settings/_Irene_/Desktop/NEURON_Modeling/NEURON_BOOK/Chapter_Ten/Model_Gap_Junction.hoc near line 16
}
^
Please could you explain these error messages and how to fix them.

Thank you
Irene

Re: Modeling Gap Rectifying Junction

Posted: Tue Sep 28, 2010 11:01 pm
by ted
nianwosuh wrote:Please could you explain to me the difference between the Gap.mod modeled in NEURON Book listing 10.2 and that modeled in the on-line course under "linear circuits, rectifying junction"
Before the advent of NEURON's LinearMechanism class, the only way to implement a gap junction was with a pair of point processes. One point process would be attached to the "postsynaptic" compartment, where it would inject a current equal to ggap*(vpre - vpost). It used a POINTER to discover the vpre voltage. The other point process would be attached to the "presynaptic" compartment, where it injected a current of equal magnitude but opposite sign; it used a POINTER to discover the vpost voltage. This works, but it is awkward to set up.

The LinearMechanism class allows one to add linear algebraic equations to the family of current balance equations that describe a model. The terms in these equations can be used to set up resistive coupling between different compartments. This can be done either by hoc statements, or by using the LinearCircuitBuilder. It's an improvement over the old method because it's more direct.

That's good as far as it goes, making it easier to set up ohmic (linear) gap junctions, but what if you want to create a nonlinear gap junction?

The exercise in the on-line course shows how to take advantage of the LinearMechanism class (via the LinearCircuitBuilder, but it can also be done via hoc statements) to represent nonlinear gap junctions. The details are explained in the exercise.