Figure 7.

Model variables and their computational relations. (A) Category learning. (B) Direct pathway of top-down primed search. (C) Indirect pathway of top-down primed search. The dashed boxes correspond to the layers of the box diagram in Figure 1. Each layer has two neurons indicating the connections to the neighboring layers. Different types of connections correspond to excitatory, adaptive, or inhibitory effects between two layers. The letter inside each neuron refers to the variables or the constant values specified in the Appendix.