Channelbuild question

zhbrass · Post by **zhbrass** » Mon Jul 17, 2006 12:57 pm

Hello,

I am a high school student who is not only very new to NEURON, but also very new to neurology in general. I have the NEURON book and I created the cell described in chapter 1. I have a channelbuild window open that refreshes the voltage axis for the cell i built. However, no matter which values i change on the channelbuild window, the graph stays the same, even though it keeps refreshing. Is this because the cell i built and the cannelbuild are independent of eachc other? If so, how to i make the channelbuild affect the properties of the cell?

-Zach

Post by **ted** » Mon Jul 17, 2006 4:11 pm

That's odd--haven't seen this occur with the Channel Builder. In fact, I thought for a minute
that you were referring to the CellBuilder, and that its Continuous Create button was off.
Can you put the files needed to reproduce this problem into a zip file and send them to me
ted dot carnevale at yale dot edu
? I'll take a look at them and see if you have discovered a bug or if the problem is that we
haven't documented the Channel Builder clearly enough. By the way, what version of
NEURON are you using?

zhbrass · Post by **zhbrass** » Tue Jul 18, 2006 9:26 am

I'm using 5.8 because the computer I am using has Mac OS 10.3.9, and 5.9 will not run because I havent upgraded to 10.4.x (frankly, I am not sure that this 667 MHz computer can handle 10.4.x). I'll send you the .ses file (i think that is what it is). It has the cell builder window and the channel build window open, in addition to a voltage axis.

Thanks,

Zach

Post by **ted** » Tue Jul 18, 2006 11:05 am

The problem was that the current specified by the Channel Builder was not inserted into the
model.

zhbrass wrote:I'm using 5.8 because the computer I am using has Mac OS 10.3.9, and 5.9 will not run because I havent upgraded to 10.4.x (frankly, I am not sure that this 667 MHz computer can handle 10.4.x).

There isn't that much difference as far as CPU load is concerned.

zhbrass · Post by **zhbrass** » Mon Jul 31, 2006 10:34 am

i know you said that the nahh isnt temperature dependent, but is khh? when I have the channelbuildgate window open and change the temperature, the graph changes. Also, how do i modify the potassium inactivation variable? All i can find is the potassium activation variable.

Thanks,

Zach

Post by **ted** » Mon Jul 31, 2006 10:51 am

Does the model you're working with now still have the hh mechanism, and if so, is
gnabar_hh and/or gkbar_hh nonzero?

how do i modify the potassium inactivation variable? All i can find is the potassium activation variable.

The Hodgkin-Huxley model of the squid action potential did not include gk inactivation.
Experimental evidence of gk inactivation was not found until many years later, by other
investigators working on other kinds of potassium currents in other cell classes. I am not
aware of significant gk inactivation occurring in squid axon.

zhbrass · Post by **zhbrass** » Mon Jul 31, 2006 11:17 am

they are both nonzero
gnabar_hh is 0.12
gkbar_hh is 0.036

Post by **ted** » Mon Jul 31, 2006 11:18 am

There's the source of temperature sensitivity.

zhbrass · Post by **zhbrass** » Thu Aug 17, 2006 1:56 pm

hi again,

i am having problems with the nahh, khh, and the channelbuild again. I created a cell with a transient A current (finally), and I wanted to see if I could get the channelbuilder to work with this model. However, when I went into the biophysics part of the cellbuilder, I discovered that it was gone. Where did it go and how can i get the channelbuild working with this model?

Thanks,

Zach

Post by **ted** » Thu Aug 17, 2006 5:05 pm

zhbrass wrote:I created a cell with a transient A current (finally), and I wanted to see if I could get the channelbuilder to work with this model. However, when I went into the biophysics part of the cellbuilder, I discovered that it was gone.

Gone? It wasn't there in the first place. NEURON has to know about the channel before
you can use it, regardless of whether you want to use it via the GUI or by writing hoc
code.

For mechanisms specified with the NMODL programming language (i.e. specified in mod
files), compiling the mod files produces a binary file that NEURON will automatically load
each time you start NEURON in that directory. The compiled mechanisms will then be
available from hoc and will also appear in all the appropriate GUI tool menus.

What about mechanisms that were built with the ChannelBuilder? The description of the
channel's properties are contained in the ses file that you saved after you finished
configuring the ChannelBuilder. If you want to use that channel in a model, start
NEURON, and then load the ses file for that channel's ChannelBuilder. After that, the
channel is available to hoc and to the GUI. If you have 10 diferent channels that were
specified with ChannelBuilders, load all of their ses files, and then you can use them from
hoc or the GUI.

For example, suppose you built a model cell with a CellBuilder and saved it to a session
file called mycell.ses. The CellBuilder's Biophysics strategy page only offers you a
limited palette: Ra, cm, pas, extracellular, and hh.
A couple of days later, you used a Channel Builder to specify the properties of an A
current, and saved that to a session file called ka.ses.
The next day you decide to add the A current to your model cell. So you write a file
called init.hoc, and here's what you put in it:

Code: Select all

load_file("nrngui.hoc")
load_file("ka.ses")
load_file("mycell.ses")

When you double click on init.hoc, NEURON starts and then does the following:
1. reads the standard run library and GUI library
2. reads the ses file that recreates your A current's Channel Builder
3. reads the ses file that recreates your model cell's Cell Builder. However, this time
the CellBuilder's Biophysics strategy page will contain a button for your A current. So
now you can insert the A current into any section or subset you like.

zhbrass · Post by **zhbrass** » Thu Aug 24, 2006 10:03 am

I did that, but there was some sort of parse error. This code comes up:

Code: Select all

loading membrane mechanisms from /Users/lmb/Desktop/NEURON stuff/Zach's New Model/powerpc/.libs/libnrnmech.so
Additional mechanisms from files
 ka.mod kdrf.mod kdrs.mod khhchan.mod nachan.mod
Created Java VM
java version "1.3.1_16"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.1_16-root_1.3.1_050825-11:50)
Java HotSpot(TM) Client VM (build 1.3.1_09-83, mixed mode)
/Applications/NEURON-5.8/nrn/powerpc/bin/nrniv.app/Contents/MacOS/nrniv: parse error
 in init.hoc near line 1
 {\rtf1\mac\ansicpg10000\cocoartf102

I'm pretty sure I did everything right, but usually I am wrong whenever I think that. What am I doing wrong?

-Zach

Post by **ted** » Thu Aug 24, 2006 10:16 am

How did you create init.hoc, i.e. what program did you use? Is init.hoc a plain text (ASCII)
file?

zhbrass · Post by **zhbrass** » Thu Aug 24, 2006 12:18 pm

Aha, it was a .txt file with the extension changed to .hoc. It works finally! Once again, Dr. Carnevale to the rescue!

zhbrass · Post by **zhbrass** » Fri Aug 25, 2006 12:32 pm

My model works great now! But is there any way to edit my imported ka channel in the channelbuild window (i.e. density and things like that)?

Post by **ted** » Fri Aug 25, 2006 12:43 pm

The Channel Builder is for specifying the properties of the channel per se, i.e. its ionic
selectivity, GHK vs. ohmic IV relationship, number of gating states and their kinetics, etc..
Channel density is an attribute of the section into which the channel mechanism has been
inserted, and is managed with hoc statements e.g.
gnabar_myhh = 1/PI
or with GUI tools such as the CellBuilder.