Add Destexhe 2001 Gfluct OU conductance based input source, tests, notebook showing conductances and distributions

[sanjayankur31]

No description provided.

[sanjayankur31]

Runs with jLEMS, but not with NEURON. When one runs the _nrn.py script, it gives:

]$ python LEMS_Gfluct_test_nrn.py 

    Starting simulation in NEURON of 3000ms generated from NeuroML2 model...

Population Pop0 contains 1 instance(s) of component: IF_curr_exp of type: IF_curr_exp
Processing 1 input lists
Finished processing 1 input lists
NEURON: syntax error
 near line 1
  { v_i_Currents_file.record(&syn_Pop0_0_0_noisyCurrentSource1_0.i) } 
                                                                ^
NEURON: syntax error
 near line 1
  { v_g_e_Conductance_file.record(&syn_Pop0_0_0_noisyCurrentSource1_0.g_e) } 
                                                                     ^
NEURON: syntax error
 near line 1
  { v_g_e1_Conductance_file.record(&syn_Pop0_0_0_noisyCurrentSource1_0.g_e1) } 
                                                                      ^
NEURON: syntax error
 near line 1
  { v_g_i_Conductance_file.record(&syn_Pop0_0_0_noisyCurrentSource1_0.g_i) } 
                                                                     ^
NEURON: syntax error
 near line 1
  { v_g_i1_Conductance_file.record(&syn_Pop0_0_0_noisyCurrentSource1_0.g_i1) } 
                                                                      ^
Setting up the network to simulate took 0.001689 seconds
Running a simulation of 3000.0ms (dt = 0.05ms; seed=54320)
Finished NEURON simulation in 0.072144 seconds (0.001202 mins)...
Saving results at t=3000.000000006531...
Saved data to: time.dat
Saved data to: i_gfluct.dat
Saved data to: g_gfluct.dat
Finished saving results in 0.110544 seconds
Done

So, the generation of the variable for the NEURON script doesn't quite work.

I'm not sure why this is, but based on what I read in jLEMS, this is treated like a synapse (the path has :), and so the NEURON variable is generated assuming its a synapse. I haven't seen any code that differentiates between a synaptic mechanism and an input mechanism either in jLEMS or in the Neuron writer. For the neuron script, to record conductances, one simply has to use Gfluct.g_e (i.e., <mechanism obj ref>.variable).

@pgleeson @borismarin : am I missing something here, or is recording from inputs in NEURON something we need to implement? (See also #6)

[sanjayankur31]

^Graphs from the jupyternotebok for convenience

^Graphs from the paper (Fig 3)

[sanjayankur31]

The distributions still do not match the paper figures---the E histogram should have a much higher peak frequency than the I one, but I can't get it to do this---perhaps they used different values? Not sure.