www.neuron.yale.edu

Posted: **Sun Oct 31, 2010 5:25 am**

I run a model of a complicated reconstructed cell using nrn7.1 version.
I insert only nseg, e_pas, g_pas, cm, global_ra and celsius and I run the simulation with a voltage graph.

This is a simple simulation, but it results in a crazy voltage graph from the first millisecond:

.
Changing to other global_ra and celsius parameters causes another very odd voltage graph:

.

What do you think is wrong?

Posted: **Sun Oct 31, 2010 5:28 am**

I hope it will let me insert the URL of the images this time:

http://www.flickr.com/photos/55323445@N04/5130708303
http://www.flickr.com/photos/55323445@N04/5130715545

Posted: **Sun Oct 31, 2010 9:56 am**

After working on this strange phenomenon for 3 days, I found the bug!
One of the many dendrites was actually composed of a couple of the same XYZ location points:

Code: Select all

apic[64] {pt3dclear()
   pt3dadd(-31.1, 253.26, -34.24, 2.07)
   pt3dadd(-31.1, 253.26, -34.24, 2.07)
}

This probably caused Neuron to think this is a circle, and caused an endless loop or something...
Now everything looks normal :)

Posted: **Sun Oct 31, 2010 12:12 pm**

The graphs are very strange and ugly.

The morphology error would produce a section with length 0. I don't know if that is sufficient to account for the results you got; I would expect that NEURON knows how to deal with such model specification errors. Wildly erroneous results can be a symptom of serious problems in one or more mod files. If you would like me to examine your mod files to see if such problems exist, zip them up and email them to ted dot carnevale at yale dot edu

www.neuron.yale.edu

Odd voltage results

Odd voltage results

Re: Odd voltage results

Update

Re: Odd voltage results