Figure 1.

A schematic drawing of the design of Experiment 1. (Left panel) The rats were trained on a dual‐solution T‐maze task in which the start arm and the correct arm were fixed throughout the training. (Middle panel) Immediately preceding the probe trial, rats were given 1‐hr free access to either the reinforcer pellets, for the devalued condition, or the maintenance diet for the non‐devalued condition. (Right panel) The probe trial was conducted in extinction, and the animals were released from the novel arm opposite to the start arm that had been used during training. Each animal received two probe trials, on separate days, one after prefeeding of the reinforcer (devalued probe trial) and the other after prefeeding of the maintenance diet (non‐devalued probe trial). The assignment of the start arm from the four possible arms (N, E, S, and W), the correct arm (left or right), and the order of probe trials (non‐devalued and devalued) were fully counterbalanced across animals