[NEST Users] Re: Adding weight to synaptic currents

Dear Atiyeh, Thanks for the clarifications. It looks like the S() function is only used to make sure that positive weights stay positive, and negative weights stay negative, i.e. the weights never change sign. Am I right to assume that you are not using any synaptic plasticity rules in your network (i.e. you only use the static_synapse)? You could ensure that this condition holds by construction of your network. For example, if you make a connection to this neuron with a positive w_ji, then S(w_ji) is nonzero but S(-w_ji) is zero, so there is an excitatory postsynaptic response but no inhibitory postsynaptic response. If you make a connection to this neuron with a negative w_ji, then the opposite holds, so you get only the inhibitory contribution and not the excitatory one. So S() might be better implemented outside of the NESTML model; you can ensure the same result during network construction, when you are making calls to nest.Connect(). For your second question, yes, that sounds correct. For your third question, do you mean for the simulation to run faster? Are you already running multi-threaded? See e.g. setting nest.local_num_threads to a value larger than one; detailed documentation for running NEST using multiprocessing can be found at https://nest-simulator.readthedocs.io/en/latest/hpc/parallel_computing.html. I hope this helps, please let us know how you fare! With kind regards, Charl On Mon, Nov 27, 2023, at 16:06, atiye nejadebrahim wrote: