how to connect 2 neurons

Where are they and what do they have?

Moderator: tom_morse

Post Reply
apoorva bedekar
Posts: 11
Joined: Wed Jan 26, 2011 11:57 pm

how to connect 2 neurons

Post by apoorva bedekar »

Code: Select all

begintemplate single_CA1

public soma , axon , apicular_dentrite , nclist


create soma , axon , apicular_dentrite 
objectvar nclist
access soma

proc init ()  {
create soma , axon , apicular_dentrite
nclist = new List()


soma {
/* topology*/
nseg = 1
diam = 30
L = 30
/* biophysics*/
Ra =100
cm = 1
insert hh

}

axon {

/* topology*/
nseg = 3
diam = 1
L = 1000
/* biophysics*/

Ra = 100
cm = 1
insert hh

}


apicular_dentrite {
/* topology*/
nseg = 23
diam = 1
L = 600
/* biophysics*/
Ra = 100
cm = 1
insert pas
g_pas = 0.002
e_pas = -65


}
connect axon(0) , soma(0)
connect apicular_dentrite(0) , soma(1)
}
endtemplate single_CA1


/*creating 2 identical neurons*/

no_of_cells = 2
objectvar cells[no_of_cells]


for i = 0 , no_of_cells-1   {
cells[i] = new single_CA1()
   }
   

 
   
/*inserting current in both somas*/ 


objectvar current_injection[no_of_cells]


for i = 0, no_of_cells-1 cells[i].soma {
    current_injection[i] = new IClamp(0.5)
    current_injection[i].del = 100
    current_injection[i].dur = 100
    current_injection[i].amp = 0.45
}


/*placing synapses */
objectvar syn[no_of_cells]


for i=0 , no_of_cells-1 {
access cells[i].apicular_dentrite 
syn[i] = new Exp2Syn(0.5)
syn[i].tau1 = 0.5
syn[i].tau2 = 1.0
syn[i].e = 0


}


/*connecting two neurons*/


access cells[1].soma
cells[0].nclist.append(new NetCon (current_injection[1], syn[0] , -20 , 1 , 1 ))


access cells[0].soma
cells[1].nclist.append(new NetCon (current_injection[0], syn[1] , -20 , 1, 1))






[code]
we are trying to model a bi directionally coupled ca1 neuron , we have written a code for 2 neuron morphology but the output shows a single neuron...i would like to know :
1.) How to view if a synapse has been created ?
2.) This program does not give an error , it gives an output showing 1 1 1 in neuron 6.1 window. What does this mean ?
3.) When i viewed the output in shape plot window , i got as single neuron as output
ted
Site Admin
Posts: 6350
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: how to connect 2 neurons

Post by ted »

apoorva bedekar wrote:we have written a code for 2 neuron morphology but the output shows a single neuron
What "output" do you mean?
i would like to know :
1.) How to view if a synapse has been created ?
For model networks that are created by executing hoc code, there is no graphical tool that will show the cells, the locations of the synaptic mechanisms that are attached to them, or the NetCons that connect spike sources to synaptic mechanisms. If you are looking for GUI tools for setting up model networks, you might be interested in the Network Builder--for a link to tutorials, see http://www.neuron.yale.edu/neuron/docs. The Network Builder and its related tools are very useful for prototyping small nets, and then exporting hoc code that can be mined for templates and other code that can be reused when you are ready to write a network model specification entirely in hoc. An advantage of using the Network Builder in this way is that it spares the programmer the effort of having to write a lot of administrative code; properly written templates seem to be full of that stuff, and having to write it is a tedious task that is prone to human error.
2.) This program does not give an error , it gives an output showing 1 1 1 in neuron 6.1 window. What does this mean ?
Those are probably "status codes." Many hoc statements generate status codes that can be useful for debugging purposes, but most of the time these numbers aren't interesting. If a statement is generating a status code and you want to prevent that code from being printed to standard output, wrap the statement inside a pair of curly brackets { like this }. To get your model to do anything, you'll have to execute
run()
after all the model setup code has been executed.
3.) When i viewed the output in shape plot window , i got as single neuron as output
Ah, this is what you meant by "output" earlier. This is entirely a cosmetic issue. You might have multiple cells, but they are being drawn on top of each other. It doesn't hurt the operation of the network, it just makes the space plot look like there is only one cell.

If you want to be sure that you have multiple cells, at the oc> prompt type the command
topology()
and you should see a printout that shows, in rather crude "teletype" form, the branched architecture of each cell.

If you want the model cells to be drawn at different locations in a space plot, you have to control the locations of their root nodes (the 0 ends of their root sections--the "root section" is the section that has no parent section). If you examine the hoc code generated by the Network Builder from a toy network model, you will see that the template for a biophysical model cell may contain a public procedure that can be used to control the xyz coordinates of the root node.
apoorva bedekar
Posts: 11
Joined: Wed Jan 26, 2011 11:57 pm

connection of two neurons within a time window of 10ms

Post by apoorva bedekar »

I have connected two neurons in the following configuration.

Code: Select all

      --------O---<>-                ["---------" = target axon ;  "O" = target soma ;  "----<>-"  = target dendrite with "<>" synapse]  
    
      -<>---O--------                ["-<>---" = source dendrite with "<>" synapse ;   "O" = source soma ;  "------" = source axon ]

      left          right
On the right side source axon is connected to target dendrite with synapse & conversely on the left target axon is connected to source dendrite with synapse

One netstim input is given to each of the somas.

I am trying to implement a paper by Markram et al (1997). The aim is that the EPSP value in the target soma increases when netstim input in source soma is given at say 0 ms and the netstim input at the target soma is given at any time in the window of 0 - 10 ms. However if the input at target soma is given after 10 ms then there is no increase in the EPSP recorded at the target soma.
ted
Site Admin
Posts: 6350
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: how to connect 2 neurons

Post by ted »

I'm not entirely sure about your diagram, don't know which of Markram's 1997 papers you're referring to, and don't know if you are asking a question or just describing what you are trying to accomplish.

Do you mean that your network has
--two cells
--one cell is called source and the other called target
--each cell has an axon and a dendrite
and
--each cell's axon makes a synaptic attachment onto the dendrite of the other cell
?
One netstim input is given to each of the somas.
Well, you can't really attach a NetStim to a section. You can attach it to a synaptic mechanism that is attached to a section. But you didn't say that either soma had a synaptic mechanism attached to it.

Do you mean to say that a current clamp (IClamp) is attached to each soma so you can inject a depolarizing pulse into either (or both) cells?
apoorva bedekar
Posts: 11
Joined: Wed Jan 26, 2011 11:57 pm

Re: how to connect 2 neurons

Post by apoorva bedekar »

[quote]

[The diagram is only a schematic representation as I could not paste the shape plot of my model.

It has two neurons, one of the two neurons is named source neuron while the other is named target neuron.

Each neuron has three sections namely dendrite, soma and axon .

The title of the paper is ,' Regulation of Synaptic Efficacy by Coincidence of Postsynaptic Action Potentials and Excitatory Postsynaptic Potentials ' by Henry Markram, Joachim Lubke, Michael Frotscher, Bert Sakmann.

My sincere apologies for ambiguous description. Initially I am describing the model which is then followed by the aim.

The soma of both the neurons have Exp2Syn point processes attached. Thus the NetStim input is given to both the Exp2Syn point processes.

Firstly I want to know that how can I modify the mod file of Exp2Syn such that it behaves like an alpha synapse, is it true that if tau1=tau2 in the mod file of Exp2Syn then i will get the required modification.

Secondly I want to know how do I count the time elapsed after injection of current in the soma of the source neuron so as to co-ordinate the current injection in the soma of the target neuron

Thirdly do I have to specify a condition such as T2 should be greater than or equal to T1+10ms in mod file of the target neuron(Exp2Syn in the soma of the target neuron )

where: T2= time at which current is injected in the target neuron
T1= time at which current is injected in the source neuron /quote]
ted
Site Admin
Posts: 6350
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: how to connect 2 neurons

Post by ted »

apoorva bedekar wrote:The soma of both the neurons have Exp2Syn point processes attached.
Why the soma? In the experiments by Markram et al., which were all about the effect of backpropagating spikes on synaptic strenght, the synases were onto dendrites.
how can I modify the mod file of Exp2Syn such that it behaves like an alpha synapse, is it true that if tau1=tau2 in the mod file of Exp2Syn then i will get the required modification.
Second question first: to quote the Programmer's Reference, Exp2Syn
produces a synaptic current with alpha function like conductance (if tau1/tau2 is appoximately 1)
but if you have any doubts it's easy enough to verify. Just record the time course of the synaptic conductance (Exp2Syn.g), and compare those numerical values with an alpha function.
how do I count the time elapsed after injection of current in the soma of the source neuron so as to co-ordinate the current injection in the soma of the target neuron
Why "count the time elapsed"? You're using IClamps to inject current into the somas, aren't you? Read the Programmer's Reference documentation of IClamp to discover how to control when current starts.
Thirdly do I have to specify a condition such as T2 should be greater than or equal to T1+10ms in mod file of the target neuron(Exp2Syn in the soma of the target neuron )
Don't tinker with the NMODL source code for Exp2Syn. If you're looking for an implementation of a synaptic mechanism that shows spike timing dependent plasticity, see this: http://www.neuron.yale.edu/neuron/stati ... 23_06.html
apoorva bedekar
Posts: 11
Joined: Wed Jan 26, 2011 11:57 pm

Re: how to connect 2 neurons

Post by apoorva bedekar »

I have injected current as per your suggestion in the soma using IClamp. Now i am looking for modl files for the NMDA-Ampa and channels to be placed in the dendrite of the neuron where will i find those. I am trying to best replicate CA1 neuron so I also want to know which all channels I need to add.

Also now there is a glitch is it true that the paper that I am trying to implement will not be implemented in Neuron 6.1. The title of the paper is Regulation of Synaptic Efficacy by Coincidence of EPSPs and APs by Henry Markram et al. Please please advice as soon as possible. I was told that the code will have to incorporate change of synaptic weights when an coincidence event occurs.

Is there a way of knowing when the dentrite of the postsynaptic neuron recieves input and when does the backpropogated action potential reach that same point. Can I have flags to mark these events and then change the weight parameter of the netcon object.
apoorva bedekar
Posts: 11
Joined: Wed Jan 26, 2011 11:57 pm

Re: how to connect 2 neurons

Post by apoorva bedekar »

[quote][ also I am trying to observe EPSP in the dendrite of the post synaptic or target neuron but I either see no voltage or an Action potential is generated. Why does this happen /quote]
ted
Site Admin
Posts: 6350
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: how to connect 2 neurons

Post by ted »

apoorva bedekar wrote:Now i am looking for modl files for the NMDA-Ampa and channels to be placed in the dendrite of the neuron where will i find those. I am trying to best replicate CA1 neuron so I also want to know which all channels I need to add.
Sounds like the "kitchen sink" approach to me. What to include in a model should be dictated by the hypothesis that the model is supposed to address, tempered by experience and judgment. To quote Larry Abbott:

"Identifying the minimum set of features needed to account for a particular phenomenon and describing these accurately enough to do the job is a key component of model building. Anything more than this minimum set makes the model harder to understand and more difficult to evaluate." (Theoretical neuroscience rising. Neuron 60:489, 2008).

Code: Select all

Also now there is a glitch is it true that the paper that  I am trying to implement will not be implemented in Neuron 6.1.
How is NEURON 6.1 relevant? The current standard distribution of NEURON is 7.1.
Is there a way of knowing when the dentrite of the postsynaptic neuron recieves input and when does the backpropogated action potential reach that same point. Can I have flags to mark these events and then change the weight parameter of the netcon object.
Sounds like what the synaptic mechanism that I mentioned in my previous post does. Doesn't sound like you have tried it.
ted
Site Admin
Posts: 6350
Joined: Wed May 18, 2005 4:50 pm
Location: Yale University School of Medicine
Contact:

Re: how to connect 2 neurons

Post by ted »

apoorva bedekar wrote:I am trying to observe EPSP in the dendrite of the post synaptic or target neuron but I either see no voltage or an Action potential is generated. Why does this happen
Insufficient information provided--I can't even begin to guess.
Post Reply