Post-training systemic and intra-amygdala administration of the GABA-B agonist baclofen impairs retention

The effects of the GABA-B receptor agonist baclofen on memory storage were studied in two series of experiments. In the first series, CD-1 mice were trained in two aversively motivated tasks: a one-trial inhibitory avoidance task and a classical conditioning task (conditional emotional response). Immediate post-training ip administration of (+/-)baclofen (10 and 30 mg/kg) impaired retention of animals in both tasks. The effect was time-dependent: Retention was not affected by baclofen administered 120 min after training. In the second series of experiments, which used Sprague-Dawley rats, post-training intra-amygdala administration of baclofen impaired retention of an inhibitory avoidance response. These results support the view that the GABAergic system is involved in the modulation of memory storage and that the amygdaloid complex may be a critical site for effects of drugs affecting the GABAergic system.     

Benzodiazepine receptor ligand influences on acquisition: suggestion of an endogenous modulatory mechanism mediated by benzodiazepine receptors

In rats, pretraining ip administration of the central benzodiazepine receptor antagonist, flumazenil (5.0 mg/kg), or of the inverse agonist, n-butyl-beta-carboline-3-carboxylate (BCCB) (0.2 or 0.5 mg/kg), facilitated retention of a step-down inhibitory avoidance task; the central agonists, clonazepam and diazepam (0.4 or 1.0 mg/kg), had an opposite effect, and the peripheral agonist, 4'-chlordiazepam (1.25 or 6.25 mg/kg), was without effect. Pre- but not post-training flumazenil (2.0 mg/kg) blocked the effect of BCCB (0.5 mg/kg), clonazepam (1.0 mg/kg), or diazepam (1.0 mg/kg) given also pretraining. The post-training administration of all of these drugs had no effect on retention of the avoidance task. Flumazenil (5.0 mg/kg) and BCCB (0.5 mg/kg), given before training, enhanced retention test performance of habituation to a buzzer but not of habituation to an open field. In the three tasks studied, none of the drugs used had any appreciable effect on training session parameters. These results suggest that there is an endogenous mechanism mediated by benzodiazepine agonists, sensitive to inverse agonists, that normally down-regulates acquisition of certain behaviors; this mechanism becomes activated only when the tasks involve or occur with a certain degree of stress or anxiety (i.e., inhibitory avoidance or habituation to the buzzer) and not in less stressful or anxiogenic tasks (i.e., habituation to an open field).     

Glucose effects on memory: behavioral and pharmacological characteristics

Recent findings indicate that post-training glucose injections can modulate memory storage for inhibitory (passive) avoidance training. Experiment I extended these findings to determine whether glucose, like other memory modulating treatments, enhances memory storage when administered after training with low footshock and impairs memory storage after high footshock training. In Experiment I, male Sprague-Dawley rats were trained in a one-trial inhibitory avoidance task using either a brief footshock (0.5 mA, 0.7 s) or slightly more intense footshock kept on until escape (0.7 mA, mean escape latency = 3.4 s). Immediately after training, each rat received a subcutaneous injection of glucose (100 mg/kg). When tested for retention performance 24 h later, the glucose-injected animals exhibited enhanced retention performance for low footshock training and impaired retention for high footshock training. Experiment II determined whether pretreatment with adrenergic antagonists blocked the effects of glucose on memory. Pretreatment with the alpha- or beta-adrenergic receptor antagonists, phenoxybenzamine, or propranolol, respectively, had no effect on acquisition or retention in animals trained with the brief footshock and did not affect glucose facilitation of that memory. In animals trained to escape footshock, phenoxybenzamine did not attenuate the amnesia produced by glucose. Propranolol-pretreated animals had impaired retention whether or not they received post-training amnestic injections of glucose; glucose had no effect on retention in these amnestic animals. These findings add further support to the view that glucose release after training and treatment may represent a physiological response subsequent to epinephrine release in modulating memory storage processing.     

Simultaneous training on two hippocampus-dependent tasks facilitates acquisition of trace eyeblink conditioning

A common cellular alteration, reduced post-burst afterhyperpolarization (AHP) in CA1 neurons, is associated with acquisition of the hippocampus-dependent tasks trace eyeblink conditioning and the Morris water maze. As a similar increase in excitability is correlated with these two learning paradigms, we sought to determine the interactive behavioral effects of training animals on both tasks by using either a consecutive or simultaneous training design. In the consecutive design, animals were trained first on either the trace eyeblink conditioning task for six sessions, followed by training on the water maze task for six sessions, or vice versa. The simultaneous design consisted of six or 11 training days; animals received one session/day of both trace eyeblink conditioning and water maze training. Separate groups were used for consecutive and simultaneous training. Animals trained on both tasks simultaneously were significantly facilitated in their ability to acquire the trace eyeblink conditioning task; no effect of simultaneous training was seen on the water maze task. No effect was seen on acquisition for either task when using the consecutive training design. Taken together, these findings provide insight into how the hippocampus processes information when animals learn multiple hippocampus-dependent tasks.     

Differential roles of hippocampal metabotropic glutamate receptors 1 and 5 in inhibitory avoidance learning

Group I metabotropic glutamate receptors (mGlu1 and 5) have been implicated in synaptic plasticity and learning and memory. However, much of our understanding of how these receptors in different brain regions contribute to distinct memory stages in different learning tasks remains incomplete. The present study investigated the effects of the mGlu5 receptor antagonist, 2-methyl-6-(phenylethynyl)-pyridine (MPEP), and mGlu1 receptor antagonist, (S)-(+)-alpha-amino-4-carboxy-2-methylbenzene-acetic acid (LY 367385) in the dorsal hippocampus on the consolidation and extinction of memory for inhibitory avoidance learning. Male, Sprague-Dawley rats were trained in a single-trial step-down inhibitory avoidance task. MPEP, LY 367385 or saline were infused bilaterally into the CA1 region immediately after training or immediately after the first retention test which was given 24h after training. Rats receiving MPEP (1.5 or 5.0 microg/side) or LY 367385 (0.7 or 2.0 microg/side) infusion exhibited a dose-dependent decrease in retention when tested 24h later. MPEP was ineffective while LY 367385 significantly attenuated extinction when injected after the first retention test using an extinction procedure. These findings indicate a selective participation of hippocampal group I mGlu receptors in memory processing in this task.     

Distinct mechanisms underlying memory modulation after the first and the second session of two avoidance tasks

Rats were subjected to three consecutive sessions, one session per day, of either a step-down inhibitory avoidance task using a 60-Hz. 0.3-mA footshock, or a two-way active avoidance task using 20 presentations of a 5-s, 1-kHz tone and a 0.3-mA footshock. After either the first or the second training session animals received an intraperitoneal injection of ACTH (0.2 microgram/kg), epinephrine-HCl (5.0 micrograms/kg), or naloxone-HCl (0.8 mg/kg). All these treatments caused memory facilitation on both tasks when administered after the first training session. When administered after the second training session only ACTH and adrenaline were effective, on both tasks. As previous physiological and pharmacological reports point to the activation of the brain beta-endorphin system after the first, but not the second, session of a task, we propose that (a) memory facilitation by naloxone depends on the previous activation of the brain beta-endorphin system; and (b) memory facilitation due to ACTH or epinephrine does not depend on the opioid activity, so their effects are expressed after both the first and the second training sessions. It was also observed that the enhancement of performance in the second training session due to post-training facilitatory treatments carried over to the test session. These results suggest that some form of consolidation occurs both after the first and after the second training session.     

The "anxiety state" and its relation with rat models of memory and habituation.

Rats selected as "anxious", "nonanxious," or normal according to their behavior in an elevated plus maze were submitted to memory tasks and the densities of central benzodiazepine receptors in the amygdala and the hippocampus were studied. Anxious rats exibited better retention scores in the inhibitory avoidance task while nonanxious rats exibited worse retention scores in inhibitory and two-way active avoidance tasks compared to normal rats. No significant differences were detected in the retention scores for habituation to an open field. Nonanxious rats presented a lower benzodiazepine receptor density in the hippocampus but not in the amygdala compared to the other groups. These data suggest that the benzodiazepine receptors are involved in the effect of "anxiety" or emotional states on memory storage processes.     

Glucose modulation of memory storage processing

Epinephrine, derived from the adrenal medulla, enhances memory storage for several forms of learning. One physiological action of this hormone is to liberate hepatic glucose stores. This experiment tested the possibility that glucose could itself enhance memory. Rats were water deprived, pretrained to drink, pretrained to drink in the behavioral apparatus, and then trained in a one-trial inhibitory (passive) avoidance task. Immediately after the training footshock, the animals each received an injection of glucose (1.0-500 mg/kg). When tested for retention 24 h later, the animals which received 10 or 100 mg/kg doses of glucose exhibited enhanced retention performance; higher and lower doses had no significant effect on the memory tests. Also, glucose injections (100 mg/kg) delayed by 1 h after training had no effect on the retention tests. These findings suggest that the increase in plasma glucose levels subsequent to epinephrine injection may contribute to the effects of epinephrine on memory. In addition, the results suggest that peripheral glucose levels may exert important influences on memory storage.     

Exposure to novelty enhances retrieval of very remote memory in rats

Rats were trained in a one-trial step-down inhibitory avoidance task at the age of 3 months and tested for retention 1 day later, or 3, 6, 9, 12, 15, and 19 months later, i.e., when the animals were 3, 6, 9, 12, 15, 18, or 22 months of age. Retrieval performance declined with time and was undetectable in the last two age groups. Exposure to an unrelated novel environment (a square box lined with black plastic) 1 h before retention testing enhanced retrieval at all ages, regardless of the decline in the level of test session performance. The effect cannot be explained by an anxiogenic effect of the novelty box, or by an influence of novelty on locomotion or exploration, or by a nonspecific influence of exposure to novelty on step-down latency in the inhibitory avoidance apparatus.     

The genetic versus pharmacological invalidation of the cannabinoid CB(1) receptor results in differential effects on 'non-associative' memory and forebrain monoamine concentrations in mice

The endocannabinoid CB(1) receptor has been implicated in the inhibitory control of learning and memory. In the present experiment, we compared the behavioral response of CB(1) receptor knockout mice (CB(1)R(-/-)) with animals administered CB(1) receptor antagonist/inverse agonist SR141716A (rimonabant; 3 mg/kg IP, 30 min pre-trial) in terms of acquisition and retention of a habituation task and changes in cerebral monoamines. The results can be summarized as follows: (i) the acute and chronic invalidation of the CB(1) receptor resulted in an increase of behavioral habituation during the first exposure to an open field, indicative of enhanced acquisition of the task; (ii) CB(1)R(-/-) mice, but not rimonabant-treated animals, showed enhanced retention of the habituation task when re-tested 48 h and 1 week subsequent to the first exposure to the open field, respectively; (iii) the facilitation of retention of the habituation task in CB(1)R(-/-) mice was accompanied by a selective and site-specific increase in serotonin activity in hippocampus; and (iv) rimonabant-treated animals displayed 'antidepressant-like' neurochemical alterations of cerebral monoamines, that is, most parameters of monoaminergic activity were increased especially in dorsal striatum and hippocampus. Taken together, the present findings demonstrate that the genetic disruption of the CB(1) receptor gene can cause an improvement of behavioral habituation, which is considered to represent a form of 'non-associative' learning. Furthermore, our data support the assumption of a rimonabant-sensitive cannabinoid receptive site that is different from the 'classical' CB(1) receptor and which, under physiological conditions, might be involved in the inhibitory control of the acquisition but not retention of non-associative learning tasks.     

Two time windows of anisomycin-induced amnesia for inhibitory avoidance training in rats: protection from amnesia by pretraining but not pre-exposure to the task apparatus

We have studied the effect of training conditions on hippocampal protein synthesis-dependent processes in consolidation of the inhibitory avoidance task. Adult male Wistar rats were trained and tested in a step-down inhibitory avoidance task (0.4 mA foot shock, 24 hr training–test interval). Fifteen minutes before or 0, 3, or 6 hr after training, animals received a 0.8-μl intrahippocampal infusion of the protein-synthesis inhibitor anisomycin (80 μg) or vehicle (PBS, pH 7.4). The infusion of anisomycin impaired retention test performance in animals injected 15 min before and 3 hr after the training session, but not at 0 or 6 h post-training. Pretraining with a low foot shock intensity (0.2 mA) 24 hr before training, prevented the amnestic effect of anisomycin injected at 15 min before or 3 hr after training. However, simple pre-exposure to the inhibitory avoidance apparatus did not alter the amestic effects of anisomycin. The results suggest that hippocampal protein synthesis is critical in two periods, around the time of, and 3 hr after training. A prior weak training session, however, which does not itself alter step-down latencies, is sufficient to prevent the amnestic effect of anisomycin, suggesting that even if not behaviorally detectable, weak training must be sufficient to produce some lasting cellular expression of the experience.     

Stimulus selection in passive avoidance learning and retention: weanling, periadolescent, and young adult rats

Periadolescent rats exhibit a number of behavioral differences in comparison with younger or older animals. For instance, periadolescents tend to show enhanced acquisition of simple active avoidance tasks, but impaired acquisition of more complex appetitive and aversive discriminations. In this experiment, rats were trained on a simple passive avoidance task at one of three ages, as weanlings (25 days), periadolescents (35 days), or young adults (45 days). Training occurred in the presence of both a redundant discriminative stimulus and a specified, redundant contextual stimulus. The periadolescents did not differ from either younger or older rats in rate of learning the passive avoidance task. The retention performance of these animals was then tested following a change in either, neither, or both of the redundant cues. When a measure of performance that controls for baseline activity was used, it was observed that periadolescents were not disrupted by a change in the redundant discriminative stimulus, a cue change that clearly disrupted performance in 25- and 45-day-old animals, and tended to be more disrupted by the contextual change than younger or older rats. It is hypothesized that the alterations in performance exhibited by periadolescents may be related to an ontogenetic alteration in stimulus selection modulated by the catecholaminergic systems.     

Strength of scopolamine-induced amnesia as a function of time between training and testing

Variations in the strength of scopolamine-induced amnesia as a function of age of the habit were studied in Swiss Webster mice. Animals were trained in an active avoidance task to a criterion of 9/10 avoidances and immediately following training injected with scopolamine hydrochloride (1.0 mg/kg) or saline. Retention of the avoidance learning was evaluated by testing different groups of animals 1, 3, 7, 10, 14, and 28 days following training. The retention test consisted of five trials in which the CS but not the UCS was presented. Results indicated that saline-treated mice exhibited near-perfect retention up to 14 days post-training with forgetting beginning to be apparent at 28 days. Scopolamine treatment produced strong amnesia in animals tested 1 and 3 days post-training but normal retention in animals tested 7 and 10 days after learning. The amnesia abruptly reappeared at 14 days after which time it remained stable. The marked similarity of the scopolamine retention curve to changes in the strength of memory of discrimination learning in undertrained rats reported by Deutsch suggested that scopolamine resulted in the storage of a weak memory of the avoidance response. To explore this idea further we trained mice to a criterion (4/5) which would result in a weak avoidance response and tested different groups 1, 3, 10, 14, and 28 days following learning. Results showed that strength of the memory of avoidance learning increased up to 10 days and then decreased abruptly at 14 days thus replicating the general shape of the retention curve produced by injecting scopolamine following strong training. These data suggest that scopolamine disrupts processes essential for the formation of durable memories.     

Opposite effects of a single versus repeated doses of gabapentin on retention performance of an inhibitory avoidance response in mice

CF-1 male mice were trained in an inhibitory avoidance (IA) task. A single gabapentin (GBP) administration (50mg/kg, ip) immediately after training enhanced retention performance when mice were tested 8 days after training. On the contrary, when the same dose of the anticonvulsant drug was given twice a day for 7 days (repeated treatment), a significant impairment on retention performance 12h after the last injection of GBP was observed. When the retention test was delayed 7 days after the end of the repeated treatment, the retention performance was not significant different from the control group, whereas if the retention test was delayed 14 days, retention performance was higher than control group but similar to that observed when GBP was administered once immediately after training. The impairment on retention performance was correlated with a significant decrease in the high affinity choline uptake in the hippocampus at the end of the retention test. The pretest administration of the direct muscarinic cholinergic agonist oxotremorine (50 microg/kg, ip) reversed the impairment on retention performance. This reversion was prevented by the muscarinic cholinergic antagonist scopolamine (0.5 mg/kg, ip). Taken together, these results suggest that the impairment on retention performance of an IA task in mice induced by repeated administration of GBP affected memory retrieval but not memory consolidation and that this impairment may be attributable to a reduction on central cholinergic activity.     

Discrete and Contextual Cue Alterations Eliminate the Instrumental Appetitive-to-Aversive Transfer Impairment in Phenytoin-Treated Rats

We have shown previously that the antiepileptic phenytoin impairs transfer in an instrumental learning task (Banks et al., 1999). The present study examined the effects of contextual alterations on appetitive-to-aversive transfer performance of rats treated with either phenytoin or tang. Adult rats were tested in tone-signaled appetitive and aversive instrumental tasks, where the animal bar-pressed to obtain a food reward (sugar pellet) or to avoid shock. Rats were trained on the appetitive task for 31 days. Beginning on the twenty-first day, rats were gavaged with either phenytoin or tang twice daily. Animals were then transferred to aversive training, with the phenytoin or tang treatment continuing throughout the 25 testing days. For some animals, contextual changes were introduced as they shifted from appetitive to aversive training, while for other animals these changes were not made. Phenytoin-treated rats that were presented with changes in context as they transferred from the appetitive to the aversive task learned the avoidance response to levels substantially higher than drug-treated rats not presented with the contextual changes. These results indicate that phenytoin impairs avoidance learning following transfer from the appetitive task, and that this impairment can be eliminated by introducing changes in context at the point of transfer. In the tang-treated control subjects, on the other hand, there was no improvement in transfer learning performance associated with the changes in contextual cues. This pattern of results suggests that contextual encoding processes in rats being trained in an instrumental appetitive-to-aversive paradigm are dramatically affected by phenytoin.     

Discrete and contextual cue alterations eliminate the instrumental appetitive-to-aversive transfer impairment in phenytoin-treated rats

We have shown previously that the antiepileptic phenytoin impairs transfer in an instrumental learning task (Banks et al., 1999). The present study examined the effects of contextual alterations on appetitive-to-aversive transfer performance of rats treated with either phenytoin or tang. Adult rats were tested in tone-signaled appetitive and aversive instrumental tasks, where the animal bar-pressed to obtain a food reward (sugar pellet) or to avoid shock. Rats were trained on the appetitive task for 31 days. Beginning on the twenty-first day, rats were gavaged with either phenytoin or tang twice daily. Animals were then transferred to aversive training, with the phenytoin or tang treatment continuing throughout the 25 testing days. For some animals, contextual changes were introduced as they shifted from appetitive to aversive training, while for other animals these changes were not made. Phenytoin-treated rats that were presented with changes in context as they transferred from the appetitive to the aversive task learned the avoidance response to levels substantially higher than drug-treated rats not presented with the contextual changes. These results indicate that phenytoin impairs avoidance learning following transfer from the appetitive task, and that this impairment can be eliminated by introducing changes in context at the point of transfer. In the tang-treated control subjects, on the other hand, there was no improvement in transfer learning performance associated with the changes in contextual cues. This pattern of results suggests that contextual encoding processes in rats being trained in an instrumental appetitive-to-aversive paradigm are dramatically affected by phenytoin.     

Caudate nucleus and memory for egocentric localization

A large body of evidence suggests that the caudate nucleus (CN) plays a critical role in the processing of spatial localization information. Furthermore, evidence has begun to accumulate that the CN is involved in the processing of a very specific class of spatial cues, namely, egocentric cues (localization with reference to the organism). This is in contrast to allocentric localization, where an organism localizes on the basis of cues external to the organism. One would then predict that lesions to the CN should disrupt performance on any tasks that depend chiefly on egocentric spatial cues, while leaving performance on allocentric tasks intact. To test this prediction, two groups of rats were trained on two different egocentric memory tasks and two different allocentric memory tasks. Specifically, one group of rats was trained on an adjacent-arm (egocentric) and an 8-arm radial maze task (allocentric). A second group of rats was trained on a right-left discrimination (egocentric) and a place-learning task (allocentric). After training, both groups received bilateral lesions of the CN. Results showed that CN-lesioned animals were profoundly impaired on retention of the egocentric tasks. In sharp contrast to this, the same animals were not or were only transiently impaired or transiently affected on allocentric tasks. Sham-operated controls were either unimpaired or transiently affected on all tasks. These findings further support the idea that the CN plays a critical modulatory role in the processing of egocentric spatial and not allocentric spatial cues.     

Amnesia induced by 2-deoxygalactose in the day-old chick: lateralization of effects in two different one-trial learning tasks

2-Deoxy-D-galactose, an inhibitor of brain glycoprotein fucosylation, was injected intracranially (10 mumole dose in 10 microliters) into either the left or the right forebrain hemisphere of day-old chicks (Gallus domesticus). Bilateral injection of this dose of 2-deoxy-D-galactose is known to induce amnesia for several learning tasks including one-trial passive avoidance and sickness-induced learning. When a tritiated form of the drug was injected into one forebrain hemisphere only, a significantly large proportion of the dose remained in that hemisphere. Chicks were trained in two different one-trial learning tasks. The first was a passive avoidance task in which the chicks were allowed to peck at a green training stimulus (a small light-emitting diode, LED) coated in the bitter liquid, methylanthranilate, giving rise to a strong disgust response and consequent avoidance of the green stimulus. In the second paradigm the chicks were allowed to peck at a similarly colored dry stimulus but, 30 min later, were injected intraperitoneally with lithium chloride (0.1 ml of 1 M solution), causing a sickness-induced aversion for the green LED. 2-Deoxy-D-galactose caused amnesia for the passive avoidance task when injected before training into the right hemisphere but not the left. However, unilateral injection of the drug before training on the sickness-induced learning task did not cause amnesia. The results indicate that fucosylation of brain glycoproteins is required in the right hemisphere for learning the passive avoidance task but that memory for sickness-induced learning can be retained by either hemisphere.     

Longitudinal attentional engagement rescues mice from age-related cognitive declines and cognitive inflexibility

Learning, attentional, and perseverative deficits are characteristic of cognitive aging. In this study, genetically diverse CD-1 mice underwent longitudinal training in a task asserted to tax working memory capacity and its dependence on selective attention. Beginning at 3 mo of age, animals were trained for 12 d to perform in a dual radial-arm maze task that required the mice to remember and operate on two sets of overlapping guidance (spatial) cues. As previously reported, this training resulted in an immediate (at 4 mo of age) improvement in the animals' aggregate performance across a battery of five learning tasks. Subsequently, these animals received an additional 3 d of working memory training at 3-wk intervals for 15 mo (totaling 66 training sessions), and at 18 mo of age were assessed on a selective attention task, a second set of learning tasks, and variations of those tasks that required the animals to modify the previously learned response. Both attentional and learning abilities (on passive avoidance, active avoidance, and reinforced alternation tasks) were impaired in aged animals that had not received working memory training. Likewise, these aged animals exhibited consistent deficits when required to modify a previously instantiated learned response (in reinforced alternation, active avoidance, and spatial water maze). In contrast, these attentional, learning, and perseverative deficits were attenuated in aged animals that had undergone lifelong working memory exercise. These results suggest that general impairments of learning, attention, and cognitive flexibility may be mitigated by a cognitive exercise regimen that requires chronic attentional engagement.     

Calcium channel antagonists enhance retention of passive avoidance and maze learning in mice

Although a number of studies have shown that treatment with calcium channel antagonists (CCAs) can ameliorate impairments in learning and memory in aged animals, evidence for a general nootropic effect of CCAs in neurologically normal young adult animals is ambiguous. This study attempts to resolve some of this ambiguity by comparing the effects of several CCAs on retention of passive avoidance learning and acquisition and retention of appetitively motivated spatial discrimination learning in young adult mice. Animals were trained in a step through passive avoidance apparatus and, immediately after training, injected subcutaneously with different doses of nimodipine, nifedipine, amlodipine, flunarazine, diltiazem, or verapamil. Retention was tested 24 h after training. In the maze-learning task mice were treated with the same doses of the aforementioned CCAs immediately after a brief training session in a linear maze and retention was tested 24 h after training. The most effective dose of each agent in the maze-retention experiment was administered to additional groups of animals 1 h prior to training to determine the effects of CCAs on acquisition processes. The effects of central administration of CCAs were examined by intracerebroventricular injection of different doses of amlodipine immediately after passive avoidance training. Results showed (1) all peripherally administered drugs except verapamil facilitated retention of passive avoidance training in a dose-dependent manner, (2) all drugs dose dependently facilitated retention of linear maze learning, (3) all doses of the drugs (except verapamil) which facilitated maze retention also facilitated maze learning, and (4) central administration of the dihydropyridine amlodipine produced a dose-dependent facilitation of the retention of passive avoidance learning. These data indicate that drugs which block calcium channels can enhance retention of two different types of learning in mice.