The subfolders contain the model files used in the following paper:

Szoboszlay, Lorincz, Lanore, Vervaeke, Silver & Nusser: Functional
Properties of Dendritic Gap Junctions in Cerebellar Golgi Cells
2016, Neuron 90, 1-14

This model requires NEURON which is available for free from

The model can be autolaunched from modeldb if NEURON is installed and
the browser is properly configured. One of the gapCond.mod files need
to be compiled for your platform (see  Then
you can start mosinit.hoc or to select a simulation.  See below for
more information.

The "01_Soma-dendritic_recordings" folder contains the models which
used to estimate the specific axial resistivity (Ra) of cerebellar
Golgi cells (GoCs). Two sets of experiments are modeled, the only
difference between the two is the application of gap junction (GJ)
blocker mefloquine. These simulations are in the "02_Mefloquine"
folder, whereas the others are in the "01_Control" folder.

Figure 5 of the original paper shows the results of the mefloquine
simulations, which can be initiated by running "init_fitting.hoc"
file.  Please note that all the simulations contain the original
experimental data:
 - "Isoma" is the somatic membrane potential (evoked by somatic
   current injections)
 - "Vdend" is the dendritic membrane potential at the given dendritic
   compartment (evoked by the same somatic current injections).

The "01_Control" folder is essentially the same, the models can be
initiated as mentioned above. In case of 140311-C1, the simulation of
2 syncytia is from Figure S2. num3 is the syncytium with the lowest Ra
estimate, whereas num10 is the one with the highest. To recreate the
lower set of traces in figure S2A compile the modfiles (see in the folder

Golgi/01_Soma-dendritic_recordings/01 Control/140311-C1/syncytium/num10

and then start init_fitting.hoc in this folder.  After you click on
the "Error Value" button in the mutliple run fitter you should see:

Fig 2SA (lower)
"02_Somatic_paired_recordings" contain the simulation files used in
Figure 6, where we determined gap junctional conductance (Ggj).
Each of these simulations can be initiated - after compiling them - by
running "init_auto_fitting.hoc". The files
"Passive_GoC_membrane_kinetics_#.hoc" are all generated by
neuroConstruct (Gleeson, Steuber, Silver (2007). neuroConstruct: a
tool for modelling networks of neurons in 3D space. Neuron 54,

These files contain the randomly generated GJ locations used in the
paper, the different scenarios can be run directly by changing the
actual number at the end of the filename in the "init.hoc" file.

These simulation also contain experimental data, the "generator"
starting with "I" is always the one receiving the current injection,
whereas the one starting with "Att" is the one containing the
attenuated membrane response of the coupled GoC.

Any further questions related to the model should be addressed to:
Miklos Szoboszlay (,
R. Angus Silver ( or Zoltan Nusser (