Adding Synapses¶

In this tutorial, we will learn how to create and configure synaptic inputs in a neuronal model. We will create a population of “virtual” presynaptic neurons that will form synapses on our postsynaptic explicitely simulated neuron. Synapses within a population share the same kinetic properties but can have different activation times.

Creating a Population¶

First, let’s create a population of presynaptic neurons that will form 50 AMPA synapses on the apical dendrite of the postsynaptic neuron. We create the synapses and allocate them to the sections of the postsynaptic neuron. Each synapse in the population is assigned a random segment.

>>> segments = model.get_segments(group_names=['apical'])
>>> model.add_population(
...    name='excitatory',
...    segments=segments,
...    N=50,
...    syn_type='AMPA'
... )

We can now access the populations through the populations attribute.

>>> model.populations
{'excitatory': <Population(excitatory, N=50)>}

We can access the population properties for a given population though kinetic_params and input_params.

>>> model.populations['excitatory'].kinetic_params
{'gmax': 0.001, 'tau_rise': 0.1, 'tau_decay': 2.5, 'e': 0}

>>> model.populations['excitatory'].input_params
{'rate': 1, 'noise': 0, 'start': 100, 'end': 200, 'weight': 1, 'delay': 0, 'seed': None}

We can access individual synapses through the synapses attribute.

>>> model.populations['excitatory'].synapses
{
    (NeuronSection(idx=9), 0.018): [<Synapse(NeuronSection(idx=9)(0.018))>],
    (NeuronSection(idx=9), 0.361): [<Synapse(NeuronSection(idx=9)(0.361))>],
    (NeuronSection(idx=9), 0.535): [<Synapse(NeuronSection(idx=9)(0.535))>],
    (NeuronSection(idx=9), 0.545): [<Synapse(NeuronSection(idx=9)(0.545))>],
    ...
}

Alternatively, we can access the synapses through the flat_synapses attribute of the population, which is a flattened list of all synapses in the population.

We allocate the synapses to the sections of the postsynaptic neuron uniformly. A synapse is assigned a random section (syn.sec) and location (syn.loc) within the section.

>>> syn = model.populations['excitatory'].flat_synapses[0]
>>> syn.sec, syn.loc
(NeuronSection(idx=9), 0.018)

Each synapse has the following references to the NEURON objects:

_ref_syn - Reference to the synapse object
_ref_stim - Reference to the stimulus object (NetStim)
_ref_con - Reference to the connection object (NetCon)

>>> syn._ref_syn, syn._ref_stim, syn._ref_con
(AMPA[0], [VecStim[0], Vector[115]], NetCon[0])

Setting Activation Properties¶

We can now update the activation properties of the synapses of our choice, such as the rate, noise, start time, and end time.

>>> model.update_population_input_params(
...        pop_name='AMPA_0',
...        rate=30, # Hz
...        noise=1, # (1) between 0 and 1
...        start=100, # ms
...        end=900, # ms
...        weight=1, # (1)
...        seed=42 # random seed
...    )

The seed parameter controls the random number generator used to create the activation (spike) times for the presynaptic neurons in a virtual population. When you provide an integer seed, the procedure that draws the spike trains (Poisson or other stochastic processes controlled by the population input parameters) will be deterministic.

Warning

The seed only affects activation timing, not where synapses are placed on the postsynaptic morphology. Therefore, it ensures reproducibility of simulations only when reloading an exported stimulation protocol, where both synapse placement and activation times remain identical. A placement seed will be introduced in future releases.

Setting Kinetic Properties¶

Finally, we can update the kinetic properties of the synapses, such as the maximum conductance, rise time, decay time, and reversal potential.

>>> model.update_population_kinetic_params(
...        pop_name='AMPA_0',
...        gmax=0.1, # uS
...        tau_rise=0.1, # ms
...        tau_decay=2, # ms
...        e=0 # mV
...    )