The simulation, which is based on the MultiCell program by Matt Wilson,
closely follows the three neuron model of the tritonia CPG described in
Chapter 6 (by Peter Getting) of Methods in Neuronal Modeling, edited by
Koch and Segev.
The cells are all single-compartment models of a soma with synaptically
activated channels. (The one exception to this is the vsi neuron, which
also has a Hodgkin-Huxley K channel.) The channels are created with a
modification of the MultiCell "makechannel" function which has been changed
to calculate gmax from the nominal value and the two time constants.
float taudif = tau2 - tau1
float tpeak
float gmmm
if (taudif == 0.0)
gmmm = {gsyn / Vex}
else
tpeak = {tau2*tau1*log({tau2/tau1})/taudif}
gmmm = {gsyn*tau2*(exp({-tpeak/tau2})-exp({-tpeak/tau1}))/taudif}
end
"gsyn" is the nominal value as specified in Appendix 6B (p. 194). The
channel field "gmax" is set to "gmmm", and Vex = exp(1.0). At present, the
channel parameter menus allow one to set gmax, rather than gsyn. In this
simulation, the channels dsiK1 and vsiNa use values of gsyn which are
different from those specified in Appendix 6B. A private communication from
P. Getting to S. Ryckebusch indicated that an error in the original
simulation caused his published results to be slightly different from those
that would be obtained with the simulation as described in the chapter.
The gsyn values used here were chosen to reproduce the published simulation
results. Changing the parameters to the published values produces a slight
change in the timing of the bursts.