Initialization for steady states - custom

[Meena]

Hi,

I am trying to model a neuron with 4 dendrites.
The soma of the neuron has several active ion channels like sodium and potassium and leak channels too. The soma also has exchangers and pumps.

The dendrites of the neuron also have active properties such as
sodium and potassium ion channels, which have conductances (S/cm^2) that are lesser in value than their counterparts in the soma.
The dendrites, besides this also exchangers and pumps and a Na and K leak channels.

My aim to create a model in which both the soma and dendrites have active properties.

Now, my main issue is getting everything to steady state at resting potential. Basically, I started with adjusting the leak conductances in the soma such that the resting potential is about -60mV. Next, I ran the model
for about 9000ms, till it was at steady state. I took all the values of the steady state and used these values as initial condition. This is so that, when I do the "run control", the voltage is not oscillating in anyway but is the resting potential.

This as you can guess...is an extremely tiring and very inefficient. When my model had just the soma and no dendrites this was a lot easier.

I have been reading Chapter 8 of the Neuron textbook and saw the SaveState method, where the states are saved and then called again.
While I see how this works for just one section like the soma, how do I make use of this when I have multiple dendrites with multiple ion channels and multiple states ? Please do note, the dendrites have the same ion channels and exchangers and pumps, only with different conductances.

Please do advice ? Do you have suggestions on how I can efficiently do the steady state initialization?

Thanks,
Meena

[ted]

While I see how this works for just one section like the soma, how do I make use of this when I have multiple dendrites with multiple ion channels and multiple states ?

The same way you'd use it for a model with a single section. Please read the first paragraph
of the documentation on SaveState in the Programmer's Reference
http://www.neuron.yale.edu/neuron/stati ... state.html

[Meena]

Ted,

Thank you so much for the quick reply. Just wanted to clarify something else too.

Assuming I have 1 soma and 4 dendrites, thats a total of 5 sections. Assuming all these 5 sections have sodium and potassium channels (however the conductance for these channels vary in soma/dendrites).

The states for the sodium channels are m and h. The state for the potassium channels is n. Will this "SaveState" method save all the states, ie 3 states per section, and 15 unique states for all the sections at the end of the run?

Will I need give the m and h states in the dendrites different names?
Will there be a conflict of names, where there may be information
over-written?
My main concern is that I cannot open the saved "states.dat" file.
So I cannnot view the contents of the saved binary file ?
Is there a way to view the contents of the file??

The other thing is the exchangers and the pumps? How can do I best deal with them ? ( I know this last question is rather broad, but I would appreciate
your advice)

Thanking you inadvance for your time,
Meena

[ted]

Will this "SaveState" method save all the states, ie 3 states per section, and 15 unique states for all the sections at the end of the run?

Yes. But why don't you test to verify this for yourself?

Will I need give the m and h states in the dendrites different names?
Will there be a conflict of names, where there may be information
over-written?

No and no.

My main concern is that I cannot open the saved "states.dat" file.
So I cannnot view the contents of the saved binary file ?
Is there a way to view the contents of the file??

od -h
or
od -a
but the results will be meaningless--the file is not designed to be human-readable.

Why don't you just write a proc printstates so that
printstates(filename)
where filename is a string
iterates over all internal nodes of all sections and prints out the values of all states to
a file named filename?
Then you can
run a simulation
save states
printstates("control.txt")
exit NEURON
start NEURON
do a custom initialization that restores the states
printstates("test.txt")
then compare the contents of control.txt and test.txt

The other thing is the exchangers and the pumps? How can do I best deal with them ? ( I know this last question is rather broad, but I would appreciate
your advice)

Broader than you may think. What do you mean "deal with them"? Save & restore their
states? SaveState takes care of that. If you are still unsure, the best way to allay further
doubts is to perform the test described above.