Update sonata_tech.rst #5
Conversation
Defining the third cell class of synapses which is modulatory (In my case: cholinergic interneurons) which can be excitatory or inhibitory depends upon postsynaptic neuron. There are three types of cells in the striatum, excitatory, inhibitory, and neuromodulatory. In building the striatum circuit, we need to define a third class of cells that are neuromodulatory and whose synapse class depends upon their postsynaptic neurons.
mgeplf
requested review from
dkeller9,
ferdonline,
jamesgking,
matz-e,
MWolfR,
pramodk,
romani79 and
WeinaJi
| @@ -72,7 +72,7 @@ Group column represents the HDF group where the dataset is located under /<popul | |||
| /0 ``mtype`` utf8 Mandatory Defines the morphological type of the node. | |||
| /0 ``me_combo`` utf8 Optional Deprecated Replaced by dynamics_params. | |||
| Foreign key to me combo tsv file. | |||
| /0 ``synapse_class`` utf8 Mandatory Defines the synapse type of the node; whether the neuron is inhibitory or excitatory. "EXC" or "INH". | |||
| /0 ``synapse_class`` utf8 Mandatory Defines the synapse type of the node; whether the neuron is inhibitory or excitatory. "EXC" or "INH".(or "MOD" - Define a third class of synapse (other than excitatory and inhibitory class) which is neuromodulatory.) | |||
There was a problem hiding this comment.
| /0 ``synapse_class`` utf8 Mandatory Defines the synapse type of the node; whether the neuron is inhibitory or excitatory. "EXC" or "INH".(or "MOD" - Define a third class of synapse (other than excitatory and inhibitory class) which is neuromodulatory.) | |
| /0 ``synapse_class`` utf8 Mandatory Defines the synapse type of the node; whether the neuron is inhibitory, excitatory, or neuromodulatory. "EXC", "INH", or "MOD". |
For better wording…
There was a problem hiding this comment.
I think "MOD" is too general. Would suggest "Cholinergic." It will prevent confusion down the line once additional neuromodulatory types are added.
There was a problem hiding this comment.
@vignan21 I like your suggest. The synapse class remains MOD, but then we can have different synapse class for each neuromodulator. We can start with M1 for Acetylcholine. I think this is enough for now. We can always append new type/neuromodulator depending on our needs. Listing "all" the neuromodulators and their combinations may be too much. Let's start simple.
There was a problem hiding this comment.
what is the impact of the first, concretely?
My first reaction is that this seems redundant if we assume that for a given synapses the presynaptic terminal releases the neurotransmitter compatible with the postsynaptic receptors. This is a fair assumption and we avoid useless complications.
I would ask the other scientists if I am missing obvious exceptions for which we should specify presynaptic neurotransmitters.
There was a problem hiding this comment.
I think it depends a bit for which simulator aspects we are gonna use this. If e.g. this relates to which MOD files to load etc., it might not be redundant information. Let's say we have a presynaptic GABA release mechanism, but it gets picked up by GABA-A and GABA-B receptors, in that case it might be useful to separate neurotransmitter (or release mechanism?) from receptor. Of course, we're talking about the future, I'm not sure we have a use case right now.
There was a problem hiding this comment.
Apparently, I created a bit of mess here. The topic is about node file, while I was discussion more in terms of edge file. During the meeting, we commented that the column on Exc/Inh in the node file may be redundant if we have the postsynapse receptor information in the edge file.
There was a problem hiding this comment.
@wvangeit the redundancy is at the level of neurotransmitter release, but both receptor and model files are essential information as you pointed out. The proposal we had in the meeting is to store the information on receptor in the edge file, and the information on which mod file to use outside the edge file (e.g. recipe, Neurodamus).
There was a problem hiding this comment.
The proposal we had in the meeting is to store the information on receptor in the edge file, and the information on which mod file to use outside the edge file (e.g. recipe, Neurodamus).
I don't think I like this proposal. The format is a "circuit model definition", so which mod file to load should be defined in the circuit as we do with morphologies, or with the e-model to load. That is the information that is needed to the simulator.
Follow the convention for node files. This field should never be empty, but at the very least contain two values corresponding to mod files that have been previously used via `syn_type_id`. Follow up to #5. Recipe part to be defined in this PR still.
Defining the third class of synapses which is modulatory (In my case: cholinergic interneurons) which can be excitatory or inhibitory depends upon the postsynaptic neuron.
There are three types of cells in the striatum, excitatory, inhibitory, and neuromodulatory. In building the striatum circuit, we need to define a third class of cells that are neuromodulatory and whose synapse class depends upon their postsynaptic neurons.