`IF_cond_exp` but with `tau_syn_NMDA` and `tau_syn_AMPA` instead of a single `tau_syn_E`

[Daniel Bliss]

Hi all,

What's the easiest way to get all the functionality of `IF_cond_exp`, but with separate time constants for NMDA and AMPA transmission?  Rather than specify a single `tau_syn_E`, I'd like to specify separate values for NMDA and AMPA current.

Thanks,

Dan

[Andrew Davison]

Currently there is no such model in the PyNN standard library. It should be fairly straightforward to add it though. Do I understand that both synaptic conductances should be "instantaneous rise, exponential decay", as for `IF_cond_exp`, just with separate time constants? Or should the NMDA model be more complex, e.g. finite rise time, magnesium block, etc.? 

Cheers,

Andrew

[Daniel Bliss]

Right, the NMDA model is more complex due to its voltage dependence, as you say, but I wasn't sure whether that complexity should be defined when the synapses are defined, rather than at the level of the cell type.

Many thanks,

Dan

[Daniel Bliss]

I've decided to write .mod files for my synaptic models.  The AMPA one was easy enough, but I'm struggling to implement my NMDA model in NMODL.  Here's the model:

s' = - 1 / tau_s * s + alpha_s * x * (1 - s)

x' = -1 / tau_x * x + sum_i(delta(t - t_i)),

where s is the fraction of open channels, x is an intermediate gating variable, t_i are the presynaptic spike times, tau_s = 100 ms (decay time), tau_x = 2 ms (rise time), and alpha_s = 0.5 kHz.

What's confusing me is how to get access inside the .mod file to whether there are any other coincident spikes.  NET_RECEIVE accomplishes this, I think, but I don't see how I can use NET_RECEIVE as a term in the equation for x', which is what is needed.

This isn't a PyNN issue, really, but I figured I'd share this to see whether anyone has any ideas.  My goal is to have the ultimate model be run using PyNN.

-- Dan

[Daniel Bliss]

Left out a crucial part of the model:

I_NMDA = g_NMDA * s * (V_m - V_E) / (1 + [Mg^2+] * exp(-0.062 * V_m) / 3.57),

where [Mg^2+] = 1.0 mM.

[Brian Hu]

Did you happen on a solution to this problem? I have also wanted to add NMDA channels to my PyNN models, but haven't been able to do so successfully. Best,

brian

[Daniel Bliss]

Yeah, I ended up implementing one NMDA model for each synapse, so that I didn't have to have each model worry about spikes from other presynaptic neurons.  Having that many models caused a substantial memory load, however, so I split my NMDA model into presynaptic and postsynaptic parts.  This enables me to have one model per neuron rather than one per synapse.

Best,

Dan

--
You received this message because you are subscribed to the Google Groups "Neural Ensemble" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neuralensembl...@googlegroups.com.
To post to this group, send email to neurale...@googlegroups.com.
Visit this group at http://groups.google.com/group/neuralensemble.
For more options, visit https://groups.google.com/d/optout.

[Brian Hu]

Sorry- I didn't see that you responded to my post. Would you be willing to share the code for your implementation? Thanks,

brian