Behavioral Neuroscience 1996, Vol. 110, No. 5,1040-1048 Copyright 19% by the American Psychological Association, Inc. 0735-7044/96/S3.00 Early Acquisition, but Not Retention, of the Classically Conditioned Eyeblink Response Is Af-Methyl-D-Aspartate (NMDA) Receptor Dependent Richard J. Servatius and Tracey J. Shors Princeton University Af-methyl-D-aspartate (NMDA) antagonists impair performance in some tasks, but whether they impair learning directly or through effects on sensorimotor performance remains controversial. Rats administered a competitive NMDA antagonist, CGP-39551, 24 hr before training could not acquire a classically conditioned eyeblink response. The associative deficit remained evident during training with a high-intensity conditioned stimulus, even though sensory reactivity was unaffected. The antagonist did not alter retention and thus did not affect motor performance of the task. These results extend and confirm studies that implicate NMDA-receptor activation in the acquisition of classically conditioned associations and specifically in tasks not dependent on the hippocampus for learning itself. Moreover, they substantiate recent claims that NMDA receptor activation (and by association, long-term potentiation) may be involved in early processes of procedural memory formation. Considerable evidence implicates glutamatergic neurotrans- mission, specifically activation of the Af-methyl-D-aspartate (NMDA) type of glutamate receptor, in the biochemical processes that underlie learning and memory. The NMDA receptor is unique among the glutamate receptors in that stimulation of the channel ionophore requires both glutamate binding and depolarization of the postsynaptic cell (Collin- gridge, Kehl, & McLennan, 1983; Dingledine, 1983). Further- more, there is an optimal time window during which these two events must occur. The dual regulation of the NMDA receptor by the occurrence of two temporally contiguous events is reminiscent of the conditions that occur during basic associa- tive learning. Reinforcing a connection to learning, NMDA- receptor activation is necessary for the induction of some forms of long-term potentiation (LTP, Collingridge et al., 1983). LTP is a long-lasting increase in synaptic efficacy in response to high-frequency stimulation of afferent fibers (Bliss & Lomo, 1973). First reported in the hippocampus, LTP possesses a number of cellular characteristics (e.g., cooperativ- ity, persistence, and synapse specificity) that are shared by or proposed for memory formation. Thus, LTP has been widely promoted as a neural model of learning and memory in the mammalian brain (Lynch & Baudry, 1984). Alternatively, we have proposed that the induction of LTP by means of activa- tion of NMDA receptors increases attention to environmental cues before and during the early stages of learning (Shors & Dryver, 1992; Shors & Matzel, in press; Shors & Servatius, Richard J. Servatius and Tracey J. Shors, Department of Psychol- ogy, Princeton University. This work was supported by grants from the Whitehall Foundation, McDonnell-Pew Program in Cognitive Neuroscience, the Office of Naval Research (N00014-92-J-1897), and the National Science Foun- dation (IBN-9511027). Correspondence concerning this article should be addressed to Tracey J. Shors, Department of Psychology, Green Hall, Princeton University, Princeton, New Jersey 08544-1010. 1995). In this model, LTP would not constitute a mechanism of memory formation, per se, but rather its induction could influence the rate of acquisition during early memory forma- tion. Despite some correlation to learning, it has been difficult to establish exactly how NMDA-dependent LTP might contrib- ute to the learning process. One obstacle has been the nonspecific, sensory, and motor abnormalities that accompany administration of NMDA-receptor antagonists (Keith & Rudy, 1990). For example, Morris and colleagues reported deficits in spatial learning in the water maze after administration of APS, a competitive antagonist (Butcher, Hamberger, & Morris, 1991; Davis, Butcher, & Morris, 1992; Morris, 1989; Morris, Anderson, Lynch, & Baudry, 1986). Two recent studies, how- ever, reported that the impairment only occurred during the initial learning of a spatial maze and that pretraining on a visual or nonvisual maze ameliorates the deficit in response to new spatial cues (Bannerman, Good, Butcher, Ramsay, & Morris, 1995; Saucier & Cain, 1995). These results indicate that NMDA receptor activation is not necessary for learning the spatial location of the cues, but rather for learning the procedures of the task (i.e., how external cues can be used to navigate the maze or maze navigation itself). Generally, procedural or skill-like memories like these are distinguished from the more declarative memories associated with spatial learning in that they are slow to acquire and are not necessarily dependent on an intact hippocampal formation. In addition to potential effects on procedural memory formation, some results suggest that the effects of NMDA antagonists on learning are, at least in part, attributable to deficits in sensory awareness (Hargreaves & Cain, 1992; Keith & Rudy, 1990; Morris, 1990). As an example, Staubli, Thibault, DiLorenzo, and Lynch (1989) reported that competitive NMDA antagonists impaired olfactory discrimination learning. How- ever, the effect only occurred when the odors were delivered at a less than optimal intensity. These results suggest that the 1040 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.