The competitive NMDA antagonist AP5, but not the non-competitive antagonist MK801, induces a delay-related impairment in spatial working memory in rats.

Rats were trained to alternate responses on a discrete trial working memory task on a T-maze. In Experiment 1, the rats were then matched for choice accuracy and allocated to three treatment groups. These were: implantation of osmotic minipumps for intraventricular infusion of either (a) 15 mM D-2-amino-5-phosphonopentanoic acid (AP5) or (b) artificial cerebrospinal fluid (VEH); and an unoperated control group (UNOP). In Phase 1 we assessed alternation performance with a minimal delay between responses: the UNOP and VEH rats continued to choose accurately; the AP5 rats showed an impairment of choice accuracy, but recovered over days. In Phase 2 a 20-s delay between responses was enforced, and choice accuracy was assessed following injections either of saline or of Milacemide HCl (10 mg/kg). There was now a severe and enduring impairment of choice accuracy in the AP5 group, but Milacemide injections did not affect performance in any of the treatment groups. In Experiment 2 rats were trained in a similar way, and then given intraperitoneal injections of MK801 or of physiological saline in a within-subjects design and tested for T-maze performance with a minimal or a 20-s delay between responses. In the first Phase, MK801 was given 10-min before behavioural testing commenced; in the second Phase, it was given 28-40 min before behavioural testing commenced. The outcome depended critically on the time between drug injection and testing. There was a significant drug-induced impairment of choice accuracy in both Phases; but in Phase 1 there was no impairment in testing with a minimal retention interval and an impairment with a 20-s retention both retention intervals. We conclude that AP5, but not MK801, interferes with temporary memory storage in a delay-dependent manner.