Help creating a Basal Ganglia model

[lettsrj]

Hi,
I'm a beginner that doesn't know much about neuron. I have read the tutorials, and I have to do a coursework that requires me to create inhibitory and excitatory neurons. It's probably a really simple thing to do, but I don't know how.
A need to know the codes to type in the neuron program. Basically I have to:
Model the basal ganglia (BG). There are three principle synapses (dopaminergic, gabaergic and cholinergic) that are involved in the basal ganglia. These “synapses” really are anatomically distinct regions (nuclei) within the BG and your first task is to model these nuclei, incorporating the HH dynamics. The BG acts as an intermediary between the motor cortex, and corticospinal tract – with mediation via the thalamus. You should implement each of the 3 principle neuron classes (dopaminergic, cholinergic, and gabaergic) as either excitatory or inhibitory neurons. You do not have to generate the actual BG circuitry, but simulate each element as if you were going to create the circuit. The purpose is to create a model that could be used to investigate diseases of the BG such as Parkinson’s disease. Note that the principle cause of PD is a lack of dopaminergic activation within the circuit, yielding excessive gabaergic activation.

Please, can someone help me, I'm freaking out !!!

[ted]

Use the Network Builder's tools to define each cell class (type of cell) and create a toy network that contains one instance of each type of cell. Export from the Network Builder to a hoc file that contains reusable code that you can mine for cell class definitions and statements that create the synaptic connections. The Network Builder tutorial will help you get started--see the Documentation page http://www.neuron.yale.edu/neuron/docs

[lettsrj]

Thanks so much for your help, but I was told I needed to do all in coding. It's supposed to be simple. I already wrote the fallowing:
oc>tstop = 10
oc>ndend = 3
first instance of ndend
oc>create soma, axon, dend[ndend]
oc>access soma
oc>objectvar syn[ndend]
oc>proc init_cell() {
> oc>soma{
> oc>nseg = 1
> oc>diam = 100
> oc>L = 100
> oc>insert hh
> oc>}
> oc>axon {
> oc>nseg = 50
> oc>diam = 10
> oc>L = 5000
> oc>insert hh
> oc>}
> oc>for i = 0, ndend-1 dend {
> oc>nseg = 5
> oc>diam = 25
> oc>L = 500
> oc>insert pas
> oc>}
> oc>connect axon(1), soma (0)
> oc>for i = 0, ndend-1 connect dend (0), soma(1)
> oc>for i = 0, ndend-1 dend {
> oc>syn = new AlphaSynapse(0.8)
> oc>syn.tau = 0.1
> oc>syn.onset = 0
> oc>syn.gmax = 1
> oc>syn.e = 15
> oc>}
> oc>}
oc>init_cell()

But, I don't know how to make the cells excitatory or inhibitory.
I'm so sorry for asking you so many questions, but I'm a bit desperate.
Thanks again

Leticia

[ted]

I was told I needed to do all in coding. It's supposed to be simple.

It is. You can do it in an hour by combining NEURON's computer generated hoc code with a bit of your own. You can even do it in two or three hours--assuming you learn quickly by example and can also quickly and accurately decipher terse documentation--by emulating the example provided by Gillies and Sterratt (see their tutorial from the link on NEURON's documentation page). "Work hard or work smart."

I don't know how to make the cells excitatory or inhibitory.

A model cell is not inhibitory or excitatory. A synaptic mechanism is inhibitory or excitatory, depending on whether it moves membrane potential toward or away from firing threhsold.

Has your instructor told you to
0. Use a text editor to create hoc code. Save it to a file. Use NEURON to load the file, e.g. by double clicking on it (MSWin), drag & drop onto the nrngui icon (OS X), or typing
nrngui filename.hoc
at the system command prompt (OS X/UNIX/Linux).
1. Separate model specification from simulation control code or user interface code.
2. Create specifications of model cells by writing code that first specifies topology, then geometry, then biophysical properties, and, only after all that has been done, attaches synaptic mechanisms to the cell.
3. When specifying topology, always connect sections so that 0 ends are proximal and 1 ends are distal.
connect axon(1), soma (0)
is a good example of what NOT to do. Instead,
connect axon(0), soma (0)
Otherwise you will become confused about what is proximal and what is distal.
4. If the model cell is supposed to be an element in a network, use ExpSyn or Exp2Syn or some other mechanism that can be driven by events, not an AlphaSynapse.
5. Each instance of ExpSyn or Exp2Syn can have its own time constant(s) and equilibrium potential.
6. Construct networks by
----first specifying cell classes, and testing each cell class by creating an instance of it and driving its synaptic mechanisms with events delivered by NetCons
----then creating the desired number of instances of each cell class
----finally using NetCons to connect spike sources (presynaptic cells) to targets (synaptic mechanisms attached to instances of cell classes)

I'm so sorry for asking you so many questions, but I'm a bit desperate.

Don't be sorry. Ask your instructor some questions too. Like, where do I send my bill for doing his/her job?

[lettsrj]

Thank you so much.

My teacher didn't teach me anything properly. I will try with your instructions, but for me it's still very complicated. I follow the tutorials but I can only do what the tutorials do, I don't know how to do my on coding or anything.
I will try to read a bit more and see if I can understand your instructions better.
Thank you so much, really !
I wish I could send him the bill ! :)

Well, maybe falling a module at university won't be the worst thing !

Thanks again !

[ted]

Seriously, if time is limted and you really want to learn something, other than "programming is not a natural way for human beings to think," I strongly suggest that you follow the recommendations in my first reply. Work through the Network Builder tutorial. When you are finished with the "Renshaw cell" example, export that network to a hoc file and study the hoc file. Identify the templates that define the cell classes. Also identify the statements that connect spike sources to targets (synapse on the motoneuron, or Renshaw cell). Read through the code to understand how it works, referring to the Programmer's Reference as necessary.

Then use the GUI to create the three cell classes you need for your basal ganglia model, and make a toy network with one instance of each cell class connected in whatever wiring diagram you desire for the network. Export that to a hoc file and study the output, identifying the templates that define the cell classes, and the code that attaches cells to each other.

Finally, imagine how you might use "for loops" to create more than one instance of each cell class and set up the desired connectivity. Try it.