Inhibitory neurotransmitters/connections

Hello Nengo forums, I have a question about inhibitory neurotransmitters.
When you are creating inhibitory connections, where is the inhibition defined? I.e., does Nengo function such that you can define a presynaptic ensemble of neurons to project inhibitory activity (then any signal from the presynaptic ensemble will dampen the activity of the postsynaptic ensemble). Or does the inhibitory nature have to be defined within the connections (i.e., through “transform” kwarg), or even at the encoding of the signal (at the postsynaptic ensemble)?
I have an M1 model, and want to have projections from the BG-THA-M1 pathway such that M1 is under continuous inhibition. Eventually I want it such that M1 signals BG to release specific parts from inhibition in order to generate cortical patterns, however, for now I am just cruious how I can achieve inhibitory connectivity?

Any help much appreciated !!

Hello, to make an inhibitory connection, you would connect directly to the neurons of an Ensemble with negative values on the transform, e.g., something like

nengo.Connection(pre, ens.neurons, transform=-1 * np.ones((ens.n_neurons, pre_dimensions)))

where the -1 controls the strength of the inhibition (you may have to transpose the transform matrix i defined there, but i believe that’s correct)

1 Like