Making whole cells out of Mcintyre myelinated axon model (2002) and testing short and long-term plasticity between them

Managing anatomically complex model cells with the CellBuilder. Importing morphometric data with NEURON's Import3D tool or Robert Cannon's CVAPP. Where to find detailed morphometric data.
Posts: 11
Joined: Sun Jun 10, 2018 6:51 am

Re: Making whole cells out of Mcintyre myelinated axon model (2002) and testing short and long-term plasticity between t

Post by moeskiski » Tue Jul 17, 2018 12:17 pm

Hello Ted,

Over the past couple of weeks, I've abandoned my previous model completely and have shifted my focus slightly (as I'll outline at the end of the post). Instead of working from the code first, I began designing my model from the GUI. Initially, I used neuromorpho models of pyramidal cells for the topography but that posed some complications when making networks and overall proved unnecessary for my purposes so I dropped it.

I then worked exclusively from the GUI to create a simplified pyramidal cell with a myelinated axon, a soma, and apical and basal dendrites. It was a working basic model that responded to stimulation using iClamp. After I realized it was okay I created a two-neuron network connected with ExpSyn. It worked. I then loaded about a dozen mods that I have been trying out ranging from synapses to channels and stimuli. That's when trouble arose.

I used different biophysics mods for the axons (fh or axnode) and somas (hh) and I assigned mods as I felt they were needed, like the extracellular mech. and xstim or xtra (I had previously, in other models, used the netstim/expsyn method you suggested and it was great - the issue I faced is that I don't know how to derive the power needed for stimulation in a real-life application from that). From that point onwards I kept getting a segmentation violation error. I didn't know how to get the extracellular mech working from the GUI so I tested the model with iclamp and that gave me the aforementioned error... iclamp and anything else I've attempted.

So I went back to the beginning and made a toy model. I thought adding all the mods at once might have introduced some incompatibility issues. It's working nicely without any added frills. I made a model with two neurons connected with ExpSyn.

Now, I checked out the zip folder you'd linked in the hot tips section on extracellular stimulation and the technical issue I'm facing is in using the GUI to stimulate the neuron with an extracellular source. It's showing in the biophysics section of cell builder and it's not a point process like the clamps so how do I bring it up with the coordinates like you did in the example in that zip folder? I tried to start from the initxstim hoc file in the zip folder and create my own model from there but it still gave me a segmentation violation if i tried to run extracellular stimulatin with anything but the cell that came with the demo. The other issue is in terms of how it works - I looked through the forum for clues as to how to measure net power transmitted to the target, which is still unclear.


Considering what I'm planning, the model will ultimately be one of two designs:
a. two myelinated pyramidal cells stimulated at the presynaptic axon
b. one fiber tract model and one postsynaptic cell with the stimulus giver at the fiber tract model. I have a question about this: The separate fiber tract model will have its own channel properties, and let us say I want to connect two neurons through it, is there a way to connect two axons so that they act as if they were a continuous single axon, as would be the case with this hypothetical presynaptic neuron?


What I'm currently doing:

1. I need to measure the power required in order to cause an EPSP or a spike in the postsynaptic cell by stimulating the presynaptic axon - this is important for my paper. Whatever I use has to be translatable to real-world application. Basically, I want to demonstrate that the power needed to cause EPSP wouldn't be enough to cause damage to the cell. I read up on the theory and experiments. Between iClamp, NetStim and extracellular I don't know which is best if any. I felt that xstim by Reilly looked promising since it was developed to test safety and efficacy, but he's looking at electrical stimulation and I'm looking at focused ultrasound (they both activate ion gated channels despite different mechanisms of action - would translating that to power as used in fUS be as direct as converting to mW/cm2?). If there's a way to have extracellular mimic the high spatial resolution of focused ultrasound by making it constricted to 1mm^2 or 1mm^3 and have its power evenly distributed then that would be exactly what I need.

2. Focused ultrasound can either stimulate neural tissue through mechanical means or suppress it through thermal means. I could not come up with a way to simulate those actions in NEURON. I want to have the presynaptic cell fire at a consistent rate before I introduce the modulatory component which would either increase or decrease activity. What steps do need to take and what method would be best to go about doing this?

3. To gauge plastic changes, both short and long-term. I'm trying out a group of different synapses for that. I do not know which would be best for modeling strict STDP as opposed to associative plasticity - the former meaning that the presynaptic soma wouldn't need to fire for plastic change to occur as long as the postsynaptic cell fires from a stimulus received at the axo-dendritic synapse, while the latter implies plasticity derived from the spiking of both cells. So far the Bi & Poo model seems to look at it that way (Do all the plasticity models focus on the activity in the synapse?)

4. Plotting. How do I produce paper-worthy plots on neuron? is there a way to export my results to matlab or anything esle where I could create beautiful plots?


Can I do everything I'm aiming for using only the GUI? What would I need to code directly?


Post Reply