Julian R. Keith

Why NMDA-receptor-dependent long-term potentiation may not be a mechanism of learning and memory: Reappraisal of the NMDA-receptor blockade strategy

The hypothesis that the processes mediating the induction of long-term potentiation (LTP) in the hippocampus also subserve hippocampally mediated learning and memory has recently been examined using an NMDA-receptor antagonist (AP5) to block the induction of LTP during training on learning tasks (Morris, Anderson, Lynch, & Baudry, 1986; Staubli, Thibault, DiLorenzo, & Lynch, 1989). The authors claim that their results support the NMDA-receptor-dependent LTP/memory hypothesis. Our reappraisal of the empirical support for this claim, however, suggests that it is premature, and that some aspects of the data can be used to argue just the opposite: that NMDA-receptor-dependent synaptic changes are not critically involved in the learning and memory processes that were under investigation. We support our position by arguing that (1) these studies did not convincingly rule out sensorimotor interpretations of the behavioral impairments that were observed, and (2) in all tasks studied, the AP5-treated animals displayed substantial evidence of learning, a result that contradicts the NMDA-receptor-dependent LTP/memory hypothesis. We also consider data from studies in which the noncompetitive NMDA antagonist MK-801 was used, and we conclude that the results obtained with this agent cannot be used to support the LTP/memory hypothesis, because MK-801 does not block NMDA-receptor-dependent LTP in vivo in doses that permit animals to behave.

Latent place learning in a novel environment and the influences of prior training in rats

Latent place learning in a novel environment was studied in rats, using a variation on the Morris (1981) water task. Groups of rats received 30 trials over 3 days in a training environment in which they learned to swim to an escape platform, were permitted to swim around the maze with no platform, or were placed on the platform by the experimenter. In a subsequent test, groups that had learned to swim to the escape platform during initial training were quicker to locate the platform in a novel environment after viewing the room from the location of the escape platform than were similarly pretrained subjects that were not allowed to view the room from the platform. Groups that were not trained to swim to the platform but received training on other procedural aspects of the task in the first environment did not benefit from standing on the escape platform in the novel environment. These results show that animals that are familiar with the requirements of the water maze task can generate accurate trajectories through a novel area based on information obtained while viewing the environment from the goal location.

Retrograde amnesia for visual memories after hippocampal damage in rats

It is generally believed that the hippocampus is not required for simple discrimination learning. However, a small number of studies have shown that hippocampus damage impairs retention of a previously learned visual discrimination task. We propose that, although simple discrimination learning may proceed in the absence of the hippocampus, it plays an important role in this type of learning when it is intact. In order to test the role of the hippocampus in simple discrimination learning, we performed a series of experiments utilizing a two-choice picture discrimination task. Our experiments confirm that rats readily learn simple two-choice picture discriminations after hippocampus damage. However, if such discriminations are first learned while the hippocampus is intact, subsequent hippocampus damage causes severe retrograde amnesia for the discriminations. Furthermore, retrograde amnesia for simple picture discriminations was equally severe when the interval between training and damage was 1 d or 60 d; remote picture memories are not spared. Similarly, the rule or schema underlying a recently or remotely acquired picture discrimination learning set was lost after hippocampus damage. The severity of retrograde amnesia for simple picture discriminations is negatively correlated with the volume of spared hippocampus tissue. Thus, the hippocampus plays an essential role in long-term memories supporting simple picture discriminations.

Fluoxetine and the dentate gyrus: memory, recovery of function, and electrophysiology

Chronic fluoxetine increases neurogenesis in the dentate gyrus (DG). In view of the widespread clinical use of fluoxetine and the well-established role of the DG in memory, surprisingly few studies have examined the effects of fluoxetine on memory and hippocampal electrophysiology. Additionally, few studies have evaluated the potential for fluoxetine to promote recovery of function after DG damage. Therefore, we studied the effects of long-term administration of fluoxetine on both spatial-reference memory and working memory, recovery of function after intrahippocampal colchicine infusions, which can destroy 50–70% of DG granule cells, and electrophysiological responses in the DG to perforant path stimulation in freely moving rats. Chronic fluoxetine did not affect matching-to-place or reference-memory performance in intact rats in the Morris water-maze task. Surprisingly, in rats with DG damage, recovery of function on both tasks was adversely affected by chronic fluoxetine. Finally, unlike an earlier study that reported fluoxetine-induced increases in hippocampal population spike amplitudes and excitatory postsynaptic potential slopes in urethane-anesthetized rats, electrophysiological measures in DG of freely moving rats were not affected by chronic fluoxetine treatment.

Acquisition in the Morris swim task is impaired by a benzodiazepine but not an NMDA antagonist: A new procedure for distinguishing acquisition and performance effects

The problem of learning versus performance continues to challenge researchers interested in drug effects on place learning and memory. To address this problem, the present study adapted the repeated acquisition/performance (RAP) procedure to the Morris swim task. The procedure involved training subjects to swim to a hidden platform that was always in the same location in one pool and in a new location each day in another pool. The advantages of this procedure are that acquisitions of new place responses and performances of previously learned place responses can be directly compared in individual subjects within single sessions, and dose–response functions can be determined for each individual subject. Experiment 1 demonstrated the ability of rats to learn new platform locations in one pool while maintaining stable performance on a previously learned place response in another pool within individual sessions. Experiment 2 used the RAP adaptation of the Morris swim task to study effects of chlordiazepoxide (CDZ), a benzodiazepine, and dizocilpine (DZP; a noncompetitive N-methyl-d-aspartate antagonist), on place learning and performance. Both drugs impaired performance in a dose-dependent manner. More interestingly, a moderate dose (10 mg/kg) of CDZ caused a small but reliable acquisition impairment without affecting performance of the previously learned place response. Doses of DZP that were substantially greater than doses previously shown to disrupt induction of NMDA-dependent long-term potentiation in the hippocampus did not impair acquisition in this procedure. Acquisition was affected only at DZP doses that also impaired the performance measure.

Persistent increases in the pool of doublecortin-expressing neurons in the hippocampus following spatial navigation training.

In addition to its role in neuronal migration during embryonic development, doublecortin (DCX) plays a role in hippocampal neurogenesis across the lifespan. Hippocampal neurons exhibit a high degree of synaptic plasticity while they are in the DCX phase. While previous studies have reported that behavioral training on hippocampus-dependent tasks can enhance neuron survival, little was known about the stage of development of those neurons and, particularly, whether a large pool of the surviving new neurons remains in the DCX phase for a prolonged period after training. Here we report that spatial navigation training increases the pool of neurons that are in the DCX phase 4 weeks after training ended. Thus, the stock of DCX-expressing neurons in the hippocampus is affected by whether a hippocampus-dependent task has been encountered during the preceding few weeks.

Repeated spatial acquisition: effects of NMDA antagonists and morphine.

Effects of morphine and 2 N-methyl-D-aspartate (NMDA) receptor antagonists, phencyclidine and LY235959, were studied using a within-subject, repeated-acquisition/performance procedure adapted to the Morris Swim Task. In the performance component, subjects swam to a hidden platform that was always in the same location in the pool. In the acquisition component, the platform was moved to a different place for each session. Baseline training produced rapid and direct swims to the platform in the performance component and steep within-session learning curves in the acquisition component. All 3 compounds increased swim distances, escape latencies, and slowed swim speed in a dose-dependent manner, but only morphine consistently produced selective impairments on acquisition. NMDA antagonists generally affected acquisition only at doses that also disrupted performance, although phencyclidine produced selective effects in some animals. These outcomes were different than those from studies of response chains in primates, suggesting that task and species variables may be important determinants of drug effects on acquisition.

Deficiencies in the Detection of Cognitive Deficits

Numerous studies purport to show that cardiopulmonary bypass (CPB) surgery is associated with persistent postoperative cognitive decline. In J. R. Keith et al. (2002), the authors argued that reports of post-CPB cognitive declines have often been quantified using data analysis methods that were based on tenuous assumptions and overlooked problems associated with familywise Type I errors. Four peers who are recognized for their expertise in neuropsychological outcomes research evaluated the arguments developed in the J. R. Keith et al. article, critiqued the study presented in that article, and offered suggestions for how to investigate whether cognitive decline occurs reliably after CPB. In this reply article, the authors respond to the open-peer commentaries made regarding the J. R. Keith et al. study.

Chlordiazepoxide interactions with scopolamine and dizocilpine: novel cooperative and antagonistic effects on spatial learning.

The authors investigated the effects on spatial behavior of coadministrations of a benzodiazepine, chlordiazepoxide (CDP), with a noncompetitive N-methyl-d-aspartate receptor antagonist (NMDAR), dizocilpine (DZP), and a muscarinic cholinergic receptor antagonist, scopolamine (SCP). Rats solved the Morris swim task in 2 settings; 1 in which a hidden escape platform was always in the same location (performance) and a 2nd in which the platform had been moved to a different location (acquisition) for repeated daily sessions. CDP (3.0 mg/kg) administered alone did not disrupt escape latencies or swim path accuracies. SCP and DZP each impaired acquisition and performance in a dose-dependent manner. CDP coadministered with 0.3 mg/kg SCP impaired escape only in the acquisition setting and when coadministered with 1.0 mg/kg SCP selectively exacerbated the escape impairment in the acquisition setting. CDP ameliorated deleterious effects of DZP in both settings.

Effects of ethanol and naltrexone on aggressive display in the siamese fighting fish, Betta splendens.

The present study used the aggressive display of Betta splendens in response to a mirror as an index of the effects of ethanol and the opiate antagonist, naltrexone. Naltrexone produces an opiate receptor blockade and thus provided a test of the hypothesis that ethanol effects on aggression are mediated by the opioid system. Eighty fish were randomly assigned to one of eight groups in a 4 X 2 factorial design with Ethanol (0, 0.25, 0.50, and 0.75 g%) and Naltrexone (0 and 5 mg/l) as the main factors. The 0.75 g% dose of ethanol reliably suppressed aggressive display as measured by number of gill show responses, but lower doses had no effect or tended to increase aggressive display. At the same time, all doses of ethanol increased arousal in the fish as measured by airgulping. Naltrexone alone did not have effects on aggression or arousal, and did not interact with ethanol on either measure. Thus the results did not support the ethanol-opioid common-link hypothesis.