Basic code to generate network of electrically coupled cerebellar Golgi cells with low frequency background inputs
Channel and synapse mechanisms are in Mechanisms
Cell descriptions are in Cells/Golgi
- /solinas includes the original Solinas morphology with soma, 3 dendrites and long axon compartment (see Solinas et al, 2007 or on ModelDB ). These are also the morphologies of GoC.cell.nml and GoC_2Pools.cell.nml
- /reduced includes 10 compartment reduced models, optimised by E. Piasini (see Piasini, 2015).
- An example of full reconstructed morphology is present here
More channel density sets are stored in Parameters that when inserted into the Solinas morphology, produce autonomous firing rates of 2-9 Hz and F-I slope of 14-25 Hz/nA.
To construct new GoC files using one of the morphologies and different parameter sets, use vary_channels.py
or vary_channels_2pools.py
- depending on whether you want a cell of class Cell or Cell2CaPools.
- simulate_one_cell.py generates a network with 1 cell (from specified GoC model). For example use, see example_gen_onegoc.ipynb
- generate_simple_network.py generates network of ~40 cells, with electrical connectivity and background inputs (low frequency Poisson spiketrains). For example use, see this nb
PythonUtils has function definitions for generating connectivity (electrical or chemical)
- Using reduced or full morphology with current channel densities causes model to fail (Vm goes to 80mV after around 500 ms -> which channel is unstable? Density adjustment for morphology?)
- ✅ Fix: NaT reaches very small time constants at spike -> use much smaller integration dt (0.001 ms or lower)
- Constructing network with a Population that has ComponentType from class Cell2CaPools -> LEMS file fails to be generated
- 🔹 Diagnosis: Cannot create events file using event port 'spike' - not supported for class Cell2CaPools?
- For some morphology?, synapses can only be inserted at location=0.5 along dendrite.
- 🔹 In that case, comment line 156 and uncomment 158 that create connectionWD instances for each background input and insert them into dendrites (in generate _simple_network.py )
-
python2.7
-
NeuroML and pyNeuroML python libraries
-
jvm
-
Neuron 7.3 or above (compile mod)