Objective The hippocampal parallel excitatory pathways referred to as the trisynaptic pathway (entorhinal cortex → dentate gyrus → CA3 → CA1) and the monosynaptic pathways (entorhinal cortex → CA1, entorhinal cortex → CA3).The hippocampal function has been analyzed by deleting the NMDA receptor gene NR1 at each stage of the trisynaptic circuit with Cre-loxP recombination technology.Pattern completion, the process of retrieving entire memories from partial or degraded cues, was claimed to require NMDA receptors in the hippocampus.The CA3 subregion, along with the dentate gyrus and CA1, presents all the functional characteristics that could be essential for this cognitive activity.However, such notion was derived from a CA3-specific NMDA receptor knockout mice exhibited performance deficit during the recall of spatial reference memory under the partial-cue condition.While this result carried a significant caveat that the gene knockout applied in that study lacked inducible temporal manipulations, and as a result, the NMDA receptors were absent in all stages of memory processes.The multiplicity of NMDA receptors function in acquisition, consolidation, retrieval and retention has remained obscure, it is possible that the performance deficit observed in CA3-specific knockout mice under the partial-cue condition might have well reflected weak binding of various memory traces during the acquisition and consolidation stages due to the lack of the NMDA receptors during those stages.It is conceivable that weak memory traces formed during learning may not be obvious under the full-cue recall condition, but became detectable under the partial-cue recall condition.This pattern completion deficit during recall may lead to a false interpretation that the NMDA receptors in the hippocampus are required for associative memory recall.This study therefore aims to examine the role of the NMDA receptors in pattern completion via transient knockout of NMDA receptors limited to the memory retrieval stage.Methods We generated two independent lines of mice in which inducible knockout of the NR1 gene can be temporally restricted to the memory retrieval stage and spatially limited to either the hippocampal CA1 region (iCA1-KO) or the entire hippocampus regions (CA l, CA3, and dentate gyrus) and cortex (iFB-KO).In our experiments, the animals could acquire and consolidate memory normally in the presence of the NMDA receptors (before inducible knockout), but memory retrieval occurred in the absence of the NMDA receptors (after inducible knockout).We used the identical spatial reference memory protocol and subjected the iCA1-KO mice and iFB-KO mice to the hidden-platform version of the Morris water maze task to assess their ability to form a spatial reference memory.To confirm the spatial learning of this reference task, we subjected both groups of mice to a full-cue probe trial or a partial-cue probe trial on the day after the completion of the 10-day-training sessions.For further complement the genetic methods, we applied systemic injection of the NMDA receptor antagonist CPP in three groups of wild-type mice to examine the role of NMDA receptors restricted to the memory retrieval phase.Results We found that the two independent lines of inducible knockout mice during the specific phase of spatial memory retrieval (NMDA receptors are intact in the phases of memory acquisition and consolidation), exhibit normal recall of the target platform regardless of full-cue or partial-cue conditions.Not surprisingly, the pharmacological experiment of blocking NMDA receptors in wild-type mice render that, mice injected the CPP blockade represented impaired learning but not memory recall under the full-cue condition, normal retrieval ability under the partial-cue condition.Conclusion Our results revealed that the NMDA receptors in the hiprocampus and cortex are not crucial for spatial associated memory retrieval in a pattern completion task.