www.neuron.yale.edu

Hi All,

I wanna add a total noise ( synaptical interactions+tempreture+shot noise+...etc.) to neuron. I have done it in matlab as below:

Vm=Vm+deltat*Itot/Cm+sqrt(deltat*var_s)*randn;


like bold characters above. This is Gaussian noise. I can change variance of this noise by changing var_s and add a constant for the noise mean value.

But I couldnt achieved this in NEURON? Can we add a mechanism for this?
if anyone help, I will be happy

Thanks

It is possible to implement something like this
Vm=Vm+deltat*Itot/Cm+sqrt(deltat*var_s)*randn;
with NEURON, but this is far from "idiomatic." In NEURON, the membrane
potential of a biophysical model neuron is computed by numerical integration
of the cable equation. The implicit assumption is that each node has a finite
capacitance to ground, which implies that membrane potential can only be
changed by the accumulation of charge on local membrane capacitance.
Noise is typically introduced in the form of stochastic fluctuation of a current
or an ionic conductance. For an example of the latter, see Gfluct.mod
in this model published by Rudolph and Destexhe
http://senselab.med.yale.edu/ModelDB/Sh ... Gfluct.mod

To make membrane potential "jump" in the way you want, it will be necessary
to create a mechanism that delivers a current that is directly proportional
to the local membrane capacitance and inversely proportional to dt, the
integration time step. Such a mechanism will only be usable with fixed
time step integration.

Thanks for your detailed explanation. I will try it,