[Question] Multiple receptors for a single neuron

[alexfanqi]

Please:

• Check for duplicate requests.
• Describe your goal, and if possible provide a code snippet with a motivating example.

Hi,

I would like to model a neuron that has a fixed number of receptors with same synaptic dynamics model, but different parameters. These receptors are AlignPost. I considered splitting out connection matrix for each receptor, or alternatively having dummy neuron groups. But they seem less convenient. So right now, I flatten out receptors and post neurons, and use something like this,

class CUBAMultiReceptor(bp.dyn.CUBA):
    def __init__(
        self,
        n_receptors: int = 1,
        name: Optional[str] = None,
        scaling: Optional[bm.Scaling] = None,
    ):
        super().__init__(name=name, scaling=scaling)
        self.n_receptors = n_receptors

    def update(self, conductance, potential=None):
        conductance = super().update(conductance, potential)
        return bm.ein_reduce(
            conductance,
            "... (neuron receptor) -> ... neuron",
            receptor=self.n_receptors,
            reduction="sum",
        )

This will require STDP_Song2000 to be modified accordingly to propagate post spikes. I wonder if there are better solutions?

[alexfanqi]

Maybe another question, are there ways to let two Projections use the same instance of SynOut and SynDyn? ie two current=comm(pre) within different Projections go into same SynDyn instance.