Briefly, the Multiple-T task is a modified continuous T-maze consisting of a sequence of T-choices, with the final choice between two return arms (Schmitzer-Torbert and Redish, 2004a; Figure 2A). In any given 40-min session, only one return arm contained food reward, but the rewarded side, as well as the sequence of T's, was varied from session to session. If the correct return arm was chosen, two 45-mg food pellets were delivered electronically at each of two reward sites once the control software detected the rat crossing the “trigger line” for that site (solid lines in Figure 2A). Because the pellet dispensers were located some distance away from the track, pellets took about 1.5–2 s from the trigger time to arrive. Except for rare misses, rats were able to catch the pellets with their mouth as they appeared. As described before (Johnson and Redish, 2007; Schmitzer-Torbert and Redish, 2004a; van der Meer and Redish, 2009), rats were pre-trained to run laps on the task before surgery, and were physically prevented from running in the wrong direction by the experimenter. After being
