Welcome to Arborize’s documentation!#

Arborize lets you describe your model definitions, import schematics from multiple sources and build equivalent multicompartmental models on the supported backends (Arbor and NEURON).

Contents:

Defining a model#

Model definitions consist of cable_types and synapse_types. Cable types define the cable and ionic properties of the types of compartments in the model, such as the soma, axon and dendrites, and which biophysical mechanisms should be present. Synapse types define the properties of the synapses that can be inserted.

{
   "cable_types": {...},
   "synapse_types": {...}
}

You then pass this dictionary style definition to the define_model() function:

from arborize import define_model

definition = define_model({
   "cable_types": {...},
   "synapse_types": {...},
})

A definition can then be combined with a schematic by a builder to create instances of your model.

Cable types#

Each cable type has a name. A cable type definition consists of cable, ions, mechanisms, and synapses. During the schematic consutrction multiple cable types can be applied to a single piece of cable in the model.

{
   "cable_types": {
      "soma": {
         "cable": {...},
         "ions": {...},
         "mechanisms": {...},
         "synapses": {...}
      }
   }
}

Cable properties#

Ra: Axial resistivity (ohm/cm)
cm: Membrane capacitance (uF/cm^2)

{
   "cable": {
      "Ra": 0.34,
      "cm": 1,
   }
}

Ion properties#

The ions block can contain any key, representing the ion name, conventionally lowercase, and each can set the following properties:

e: Reversal potential

{
   "ions": {
      "h": {"rev_pot": 0},
      "ca": {"rev_pot": 30},
   }
}

Mechanisms#

Mechanisms are defined by their mechanism ID, which is either a string name, or a tuple of up to 3 strings: name, variant, and package; and a set of parameters.

{
   "mechanisms": {
      "Kv1": {
         "gbar": 1.4
      },
      ("Kv1", "burst"): {
         "gbar": 1.4
      }
   }
}

Hint

Arborize uses Glia to manage the mechanisms for the NEURON and Arbor backends. In order to use mechanisms you need to add them to your Glia library.

Synapses#

A synapse definition is defined by its name, mechanism ID and parameter set. If you do not specify a mechanism ID, the name is used instead.

{
   "synapses": {
      "AMPA": {
         "gmax": 3200
      },
      "depressing": {
         "mechanism": ("AMPA", "TM"),
         "parameters": {
            "gmax": 1300,
            "U": 0.60
         }
      },
      "facilitating": {
         "mechanism": ("AMPA", "TM"),
         "parameters": {
            "gmax": 10000,
            "U": 0.12,
            "tau_facil": 1.1
         }
      },
   }
}

Synapse types#

The previous section described how to add synapses to a cable type, but you can also define a synapse type available on the entire model by adding them to the synapse_types field.

{
   "cable_types": {...},
   "synapse_types": {
      "AMPA": {
         "gmax": 3200
      }
   }
}

Drawing a schematic#

Schematics can come from any source, and Arborize supports 2 sources out of the box:

BSB schematics from BSB morphologies and parameters.
File schematics load morphologies from file using MorphIO.

from arborize import file_schematic, define_model, neuron_build

definition = define_model({
   "cable_types": {...},
   "synapse_types": {...},
})
schematic = file_schematic("my_cell.swc", definition)
n_cells = 100
cells = [neuron_build(schematic) for i in range(n_cells)]