BackgroundSet-shifting tests represent a reliable paradigm to assess executive functions in humans and animals. In the rat, set-shifting in a cross-maze is a recognized method. In this test, rats must learn an egocentric rule to locate food reinforcement. Once acquired, a second rule, based on visual-cue strategy, allows the location of the food. Ability of rats to shift from the first to the second rule is considered to reflect cognitive flexibility. New methodThis study aimed at optimizing the most currently used set-shifting protocol in a cross-maze for standardized drug testing by modulating the parameters related to caloric restriction, reward preference, and by redefining the notion of turn bias and classification of errors sub-types, i.e. perseverative vs. regressive. The new protocol has then been used to assess rats treated by sub-chronic phencyclidine administration and investigate the potential reversal effect of tolcapone, a brain penetrant catechol-O-methyl transferase inhibitor. ResultsThe new procedure resulted in a decreased total duration and a re-definition of turn bias and error subtypes. Despite preferences for sweet rewards, caloric restriction had to be maintained to motivate animals. Overall, sub-chronic PCP-treated rats made mostly perseverative errors compared to controls and required more trials to shift between the two rules. Tolcapone partly reversed impairments observed in PCP-treated rats. ConclusionThe new protocol has improved the reliability of key parameters and has contributed to the decrease of the test duration. PCP-treated rats submitted to this protocol have been shown to have significant deficits that could be reversed by tolcapone.