Performing calculations involving values from multiple Sections and mod files

[Robert Claar]

Hello,

I currently have created a framework built on NEURON to run simulations to investigate Islet of Langerhans cells with variable numbers of different cells/Sections under different conditions (e.g., glucose levels, 3D islet architectures, etc.). Each cell type has a corresponding mod file and each of these mod files contain POINTERs for signaling molecules that affect the different cell types. Since I can use the model to simulate an arbitrary number of cells, I created a rule that the signaling molecule concentrations would be set by pointing to the concentration in the closest cell of the appropriate type (e.g., if A cells are affected by a signaling molecule X that is released by B cells, my code finds the closest B cell to said A cell and points to that B cells X concentration). This is too simplistic for the kinds of hypotheses I hope to test using the model. Using my example with A and B cells, what I would like to do is set the concentration of X in this A cell at each time step equal to a distance-weighted average the concentrations of X produced by all the B cells in the Islet. Currently I am creating a mod file that will hold the distances between a given cell and the other cells of different types, and will also contain POINTERs to all the concentrations of the other signaling molecules so that I can perform these calculations and then use that calculated concentration. However, this is going to involve tons of POINTERs and be kind of hacky. I'll also need to look into how/if I can do matrix math in mod files.

I previously had my model setup to use the rxd module but had some issues with boundary conditions that turned into too much of a time sink for me so I want to try the approach above as what I see as a next-best option. Are there any NEURON tools for performing calculations using STATE variables from multiple Sections and mod files at each timestep?

Any time/consideration on this is greatly appreciated.