Experimental and subject to dramatic changes!

I make no claim of correctness

I am attempting to model Mitral cell myelinated axons. My model was originally based on Shen…Hines 1999, but I have adapted the basic parts of that model to Python using an object oriented style to simplify experimental control and data analysis.

I used a number of channels and techniques described very well by Zbili et al. 2020 (which were based on channels from Hu et al. 2009)

The base class is called MitralCell, and is constructed with a unique ID and the number of nodes to construct. This class is not yet well documented (and likely never will be), so refer to the code for details.

This this notebook contains explanations of some of the parameters and reasoning behind parts of the model.

Ion channels were downloaded from https://senselab.med.yale.edu/ModelDB/showmodel?model=263053#tabs-2, and compiled with this Makefile in the nmodl directory. The channels must be manually loaded at the top of the controller script (see: https://www.neuron.yale.edu/phpBB/viewtopic.php?f=2&p=18501&sid=8dff1ac9d4bbc4a34fb50d42d8b55fa4#p18501 or the example notebook).

This was put together as a proof of concept for a class, and will be developed further elsewhere. As such, it is not yet pip-installable. I used basic sys.path manipulations from my notebooks to import and use MitralCell and actionpotential.py. See the notebooks for the within-script path modifications necessary to use them.

These experiments were run on Mac OS in a python 3.8 virtual environment with NEURON accessible via the PATH and PYTHONPATH. For instructions on how to configure python virtual environments to use NEURON on Mac OS, see my post: https://nickgeorge.net/programming/NEURON-neuronal-modeling-software-and-python-with-virtual-environments/

For more on the basic implementation (including myelin sheaths, ion channel distributions, and nodes of Ranvier), see this notebook.

For examples of experiments, please see the final presentation here.

I have written some helper AP analysis functions, (so far) just a set of functions for AP amplitude and to measure FWHM for AP shape. Those functions are in actionpotential.py. I have also written tests for many of these functions which can be run with pytest. I am also working on tests for the main MitralCell class, to ensure that experimental changes are accurately reflected in the model as I think they are. I have not implemented those yet, but it will involve using cell.ais.psection() and verifying the output is as expected after changing.