Administration of a non-competitive NMDA antagonist dizocilpine (MK-801) was proposed to be an animal model of psychosis. NMDA-receptor blockade is accompanied by increased locomotion, behavioral deficits, and other changes resembling psychotic symptoms. However, the role of NMDA-receptors in organizing brain representations is not understood yet. We tested the effect of NMDA-receptor blockade by systemic administration of dizocilpine at two different doses (0.1 or 0.2 mg/kg) in a recently designed Active Allothetic Place Avoidance (AAPA), a task which requires rats to separate spatial stimuli from two continuously dissociated subsets. The effect of dizocilpine on learning in the AAPA task was compared with its effect on acquisition of the reference memory version of the Morris Water Maze task. Both doses impaired performance in the Morris Water Maze task, whereas only the higher dose impaired performance in the AAPA task. The Morris Water Maze appears to be more sensitive to dizocilpine-induced behavioral deficit than the AAPA task. These findings support the notion that these two tasks are differentially dependent on the NMDA-receptor function.