record synaptic currents

[coschroeder]

at the moment it is not possible to record that

[huangziwei]

synaptic currents actually got recorded (only) in network level, as {synapse_name}_current in states.

    net0.cell(0).branch(0).loc(0.5).stimulate(current, verbose=False)

    # record the membrane potential of the pre- and post-synaptic compartments
    net0.cell(0).branch(0).loc(0.5).record(verbose=False)
    net0.cell(1).branch(0).loc(0.5).record(verbose=False)

    net0.cell(1).branch(0).loc(0.5).record("AMPA_R", verbose=False)
    net0.cell(1).branch(0).loc(0.5).record("AMPA_C", verbose=False)

    net0.record("AMPA_current", verbose=False)

maybe we should update the synapses to the same convention used in channel.

[ntolley]

Is there any interest in adding functionality for recording extracellular fields (since they're calculated directly from these currents)?

I'm planning to start playing around with jaxley to simulate LFP/EEG, happy to try and contribute if there's interest in a dedicated API

[michaeldeistler]

Yes, it would definitely be great to have this! Let me know if you need any input or would like to discuss, I would be happy to have a quick chat over zoom!

Michael

Aside: my guess is that achieving this will require to have access to x, y, z of postsynaptic compartments. Once #464 is merged, those can be accessed by doing:

net._update_nodes_with_xyz()
net.copy_node_property_to_edges("x", "post")
net.edges

[ntolley]

Working on this a bit more, I think it might be useful to add a column to net.recordings that indicates which compartment the recordings corresponds to. Perhaps adding the glocal_comp_index as a column here:

What do you think? This can be handled with some bookkeeping in the code, but adding it as a column would help merging the dataframes for downstream analyses

[ntolley]

I'm actually realizing that the index on this dataframe does seem to correspond to global_comp_index! At least the index values repeat when adding recordings for multiple current types...

[jnsbck]

Indeed, the index of whatever you are recording is written to rec_index, see

jaxley/jaxley/modules/base.py

Lines 1479 to 1481 in 32068e9

	in_view = self._nodes_in_view if state in comp_states else self._edges_in_view
	new_recs = pd.DataFrame(in_view, columns=["rec_index"])

. This can be the edge or comp index depending on whether the state is part of a synapse or channel.