I apologize if this is a dumb question, or asked before. I didn’t see any other questions about this, so it is probably obvious. However I would have assumed that a positive stimulus as input to a neuron would result in action potential, but the nengo simulations show the opposite?
Hello! I suspect what you’re running into is the use of encoders in Nengo. When you make a standard Nengo connection between a stimulus Node and an Ensemble of neurons, the input signal will be dot multiplied with a set of ‘encoders’ for each neurons in the ensemble. The result is what’s fed into the neuron model. Whether or not a neuron responds to the input will depend on how well the input signal aligns with the encoders of the neuron (plus the scaling from the gain term).
To connect directly to the neurons, you’d use “nengo.Connection(input, ens.neurons)”, and then you should see the behaviour you’re expecting!