Differences between Jansen-Rit Model and Coupling

[DK]

Hi, I have a couple of questions regarding the JR model implementation:

• The TVB implementation says that nu_max "Determines the maximum firing rate of the neural population [s^-1]." As far as I see, nu_max is equivalent to e0 in the 1995 JR paper, right? Which means that it determines the half-maximum firing rate.
• Why is the default v0 = 5.52 mV and not 6 mV as in the original 1995 JR paper?
• I see that two variables c and src enter the y4 equation. src seems to be set to 0, while c seems to be the input from other regions. Looking into the SigmoidalJansenRit coupling function, I understood that this function should take regions’ EPSPs+IPSPs, transforms them through the sigmoidal function into firing rate, multiplies by the connection strength and sends it to a local JR mass. The code indexes state variables 0 and 1, shouldn’t it use 1 and 2 to represent EPSPs and IPSPs? Why is the default v0 / midpoint = 6 mV here and not 5.52 mV as in the model?

Thanks a bunch for your help in advance!

[TVB Users]

hi,

The TVB implementation of JR 1995 represents a single column while the paper formulates a pair of columns; more specifically the synaptic model in JR 1995 requires a number of ODEs which increase with the square of the number of columns.  In the interest of computational simplicity, TVB implements an approximation via the SigmoidalJansenRit that avoids the per-connection ODEs.  The deviations in parameters are likely due to this modification and were chosen to ensure the model is numerically stable by default. 

cheers,

Marmaduke