Issue with single cell simulations (firing with no input)

[Katherine Doxey]

Hello,

I've imported the NEURON version of the Layer 5b Pyramidal cell by Hay et al. (2011) from NeuroML-DB into NetPyNE (https://neuroml-db.org/model_info?model_id=NMLCL000073). When running a simulation with no input, i.e. no current injection, the cell is spiking. I’ve attached the voltage trace of the soma as well as my code. 

I’ve stepped through the NetPyNE code for createSimulateAnalyze(), but did not notice anything unexpected. The stims object returned from create() is empty.

So far I have tried modifying the initial voltage (-80 mV) and temperature (6.3 C) based on the original NEURON simulations (https://github.com/OpenSourceBrain/L5bPyrCellHayEtAl2011/tree/b75a81839a6f14e35f414ce554c75d97b1563931). 

[Katherine Doxey]

UPDATE:

I have simplified the model to contain just the soma and passive channels and ran simulations with no input in NetPyNE and NEURON. It appears that there is something in the NetPyNE simulation that is causing the cell to depolarize. I've attached the simplified cell hoc file, simulation scripts, and output voltage traces.

[Eugenio Urdapilleta]

Hi Katherine,

 I don't have the mechanism "pas_nml2", but it seems a simple leak. The line in the NetPyNE code:

netParams.cellParams['L5PC_hoc']['secs']['soma_0']['mechs']['pas_nml2']['conductance'] = -90

seems confuse as the variable "conductance" doesn't exist in your mechanism (as I infer from your hoc). Maybe

netParams.cellParams['L5PC_hoc']['secs']['soma_0']['mechs']['pas_nml2']['e_pas_nml2'] = -90

Let us know. Best,

Eugenio

[Katherine Doxey]

Hi Eugenio,

Thanks for the reply

In my NetPyNe code, the 'pas_nml2' mechanism object has the items 'gmax' and 'conductance'. There is no 'e_pas_nml2' object.

[Eugenio Urdapilleta]

Hi Katherine,

 can you provide us the .mod for the  "pas_nml2" mechanism so we can try to figure out what's going on? Thanks a lot,

Eugenio

[Katherine Doxey]

Thanks!

[Eugenio Urdapilleta]

Hi Katherine,

 the mod file is very preliminar. The parameters to be assigned (modifiable) from the exterior should be declared in the section "PARAMETER". So, please, move the variable "e" from ASSIGNED to PARAMETER. Then, NetPyNE (using a function from NEURON, which reads these PARAMETERS so they can be used/provided a value without knowing them a priori) would assign corresponding entries. If you make the above change, your mechanism would have the parameters: {gmax, e, conductance}. In your BREAKPOINT, the only thing that matters right now are:

gion = gmax * fopen -----> gmax

i = gion * (v - e)

So, you need "e" for sure, which wasn't provided to NetPyNE. In NEURON, it actually does in the hoc line "e_pas_nml2 = -90.0" because it is transmitted via the RANGE declaration. In NetPyNE this hoc line is discarted because NetPyNE doesn't know that "e" was a parameter as it wasn't declared so. Anything with no value goes normally to zero, so this reversal potential was set equal to 0 (see your NetPyNE evolution). 

With the above correction, you can change any of your parameters with:

netParams.cellParams['L5PC_hoc']['secs']['soma_0']['mechs']['pas_nml2']['e']

netParams.cellParams['L5PC_hoc']['secs']['soma_0']['mechs']['pas_nml2']['conductance'] ---> useless right now

netParams.cellParams['L5PC_hoc']['secs']['soma_0']['mechs']['pas_nml2']['gmax']

Hope it helps. Best,

[Katherine Doxey]

This fixed it!! Thank you so much for all your help.