Learning by stimulation avoidance: A principle to control spiking neural networks dynamics
Fig 6
Trajectory of the network in the two-dimensional space of the firing rates of Output Zone A and Output Zone B.
Using LSA, we can steer the network in this space; by contrast, the “reinforcement only” experiment maintains the network balanced relatively to the two firing rates. (a) leads to the low external stimulation region but (b) does not. Statistical results for N = 20 networks.