Support retrieving the current value of state variables without recording

[apdavison]

I propose to extend the get() method of Population, PopulationView and Assembly to allow returning the current value of state variables, such as the membrane potential.

[ruthvik92]

Hello, Does the current version has this feature?

[apdavison]

No, this has not yet been implemented.

[ruthvik92]

Any idea about when this feature will be released? May be I could try looking into this if this is not a juggernaut task.

[apdavison]

There is no particular timeline for implementing this. I would be happy for you to look into it.

A few suggestions for the implementation:

• in the definition for common.Population.get(), you will see that the parameters are retrieved through a private method _get_parameter_dict(), which then calls _get_parameters(). The latter method is implemented separately for each simulator, i.e. as neuron.Population._get_parameters(), nest.Population._get_parameters(), etc.
• to retrieve state variables through Population.get(), I suggest first creating a method _get_state_variables() for each simulator Population class.
• implementing _get_state_variables() should be quite simple
  • for NEURON, something like [cell._cell.v for cell in population.all_cells] will give you the values of v for a whole population
  • for NEST, the equivalent is nest.GetStatus(p.all_cells.tolist(), 'V_m')
  • for Brian, I don't remember off the top of my head how to access the state variables, but I think it's quite simple
• you will have to take care of translating variable names (e.g. 'v' --> 'V_m' for NEST) and units (e.g. pA --> nA for currents in NEST). The information needed for this is mostly in the celltype definitions in standardmodels.cells or in nest.recording, etc.

[ruthvik92]

I wrote a few lines of code as you suggested, I tested it with Neuron and I could read the membrane potentials. I haven't tried it with Brian or Nest. Thank you for your valuable insights.