Figure 4.

Evolution of weights in the D1, D2, and RP pathways, as well as the performance over a simulation of 15 blocks of 40 trials each. After eight blocks (black arrow), 16% (left column) or 66% (right column) of the dopaminergic neurons were rendered useless until the end of the simulation. In (A,D), the weights represented are the average weight to each action coding population in D1 and D2 from state 1. (B) and (E) correspond to the average weights for each state-action pairing striosomal sub-population to the dopaminergic neuron population. (C,F) display the moving average success ratio of the model over the simulation. Vertical dashed gray lines denote the start of a new block. In (A,B,D,E), the black line is the total average of the plotted weights. Color-code follows the one in Figure 3 and shaded areas represent standard deviations.