Network Description

The following XML file specifies a very simple network of cells, comprising only one instance of a SAC, located at the origin (0, 0, 0).

<?xml version="1.0" encoding="utf-8"?>
<network xmlns="RSME.simulation_parameters">
  <network_description> TEST </network_description>
  <populations>
    <population name="SAC" cell_id="0">
      <instances size="1">
        <instance id="0">
          <location x="0" y="0" z="0" />
        </instance>
      </instances>
    </population>
  </populations>
</network>

In the following XML, we are specifying an overlayed grid of SACs:

<?xml version="1.0" encoding="utf-8"?>
<network xmlns="RSME.simulation_parameters">
  <network_description> TEST </network_description>
  <populations>
    <population name="SACs" cell_id="0">
        <layered_2d_grid>
            <grid_arrangement>
                <rectangular_location>
                    <corner x="0" y="0" z="0" />
                    <cells_number x="3" y="3" />
                </rectangular_location>
                <spacing x="125" y="125" />
                <synapse type="TBD" />
            </grid_arrangement>
            <grid_arrangement>
                <rectangular_location>
                    <corner x="62" y="67" z="0" />
                    <cells_number x="2" y="2" />
                </rectangular_location>
                <spacing x="125" y="125" />
                <synapse type="TBD" />
            </grid_arrangement>
        </layered_2d_grid>    
    </population>
  </populations>
</network>

In this XML file, we define a layered 2D grid, which comprises two grid arrangements. Each grid is defined with a location (x,y,z of its corner) and the grid size (x, y). For each grid, you can specify the spacing between the morphologies (x, y distances).

Note:

  • We use the SAC morphology defined in the previous layer, identified as "0".

  • In RSME we can define several networks or populations. Here we identify the population ID as "0". We will use this ID in the visual stimulation layer, enabling the specification of particular stimuli per population.

In the following XML code, we define another population, containing one direction-selective ganglion cell:

<population name="RGCs" cell_id="1">
      <instances size="1">
        <instance id="0">
          <location x="0" y="0" z="0" />
        </instance>
      </instances>
</population>

Now, we have two populations of cells. In the network XML, we can now define a projection - connectivity scheme between two populations. In the following example, we define a projection from a source population (SAC) to a target population (DSGC). We will define a synapse, placed in accordance with the cosine similarity tule (see RSME paper for an elaborated description). We will define a synaptic connection following a planar (x, y) distance criteria while aligning the two populations one against each other.

<projections>
    <projection source="0" target="1" name="SACs-RGCs">
      <synapse>
        <preffered_direction type="cosine_similarity" x="1" y="0" />
      </synapse>
      <connectivity_pattern>
        <x_y_intersection>
          <x_y_alignment x="135" y="100" />
        </x_y_intersection>
      </connectivity_pattern>
    </projection>
  </projections>

Last updated