Hard-coded initial membrane potentials

[katduecker]

[adding this for documentation purposes]

state_init() in network_bulder.py hard codes the initial membrane potentials for each cell. https://github.com/jonescompneurolab/hnn-core/blob/master/hnn_core/network_builder.py#L611

This leads to issues for new models with different resting membrane potentials.

I think it was implemented in this way, instead of setting the initial values when creating the sections, based on the assumption that h.finitialize resets the starting membrane potential to -65 mV https://github.com/jonescompneurolab/hnn-core/blob/master/hnn_core/network_builder.py#L67

It turns out that h.finitialize does not reset the membrane potential to a default value, so the initial values can be implemented when creating the cell (I have tested this).

I cannot upload my .py files so I'll just briefly describe my current work around:

I pass a v_init dictionary to _get_dends in cells_default.py

  v_init = {'apical_1': -71.32,
            'apical_2': -69.08,
            'apical_tuft': -67.30,
            'apical_trunk': -72,
            'soma': -72.0,
            'basal_1': -72,
            'basal_2': -72,
            'basal_3': -72,
            'apical_oblique': -72}

  sections = _get_dends(p_all, cell_type='L5Pyr',
                        section_names=section_names, v_init=v_init)

add value v to the section class

 def __init__(self, L, diam, Ra, cm, v=-65, end_pts=None):

      self._L = L
      self._diam = diam
      self._Ra = Ra
      self._cm = cm
      self._v = v             # initial voltage for each section
      if end_pts is None:
          end_pts = list()
      self._end_pts = end_pts

      self.mechs = dict()
      self.syns = list()

      # For distance functionality
      self.nseg = _get_nseg(self.L)

and then I set v to the initial value in _set_biophysics in cell.py
setattr(sec, 'v', section.v)

I'm working on a new model in local copy of HNN, but let me know if and how soon we'd want this feature.