Figure 4.

Reinforcement learning circuit of the CogEM model (Grossberg, 1971, 1975; Grossberg and Seidman, 2006). (A) Processing stages of invariant object category, object-value category, and drive representation (value category) representations. CS, conditioned stimuli; S, sensory representations; and M, motor representations. (B) Conditioned reinforcer learning enables sensory events to activate emotional reactions at drive representations. Incentive motivational learning enables emotions to generate a motivational set that biases the system to process information consistent with that emotion. Motor learning allows sensory and cognitive representations to generate actions. (C) Anatomical interpretations of the processing stages. [Adapted from Grossberg and Seidman (2006), Figures 4, 5, with permission].