MK-801 can facilitate passive avoidance memory when retention is not present in control animals, and can fail to facilitate when it is present.

The results confirm that NMDA receptor blockade can result in improved retention performance of mice in step-down passive avoidance. A series of behavioural variations and analyses revealed that memory in the task depended mainly on the appearance of the grid that had been associated with shock, rather than on the execution of an instrumental response or on the spatial locus of the punishment. When the grid was made invisible during retest, retention was never found. However, MK-801 did not facilitate retention based on the appearance of the grid. In contrast, conditions were found in which, even though control animals showed no learning, MK 801 given after the learning trial facilitated retention. Thus, the dominant mode of learning and memory in the step-down task is insensitive to the drug, whereas the drug raises a weaker or alternative mode to above threshold levels.