Learning not to respond: Role of the hippocampus in withholding responses during omission training.

Autoshaping is a Pavlovian learning paradigm in which rats experience pairings of a CS and a US independently of their behavior. When the CS is a lever inserted into the test cage and the US is food delivered to an adjacent magazine, many rats acquire a lever-pressing response called 'sign-tracking' even though that response has no effect on the occurrence of either the CS or the US. Since these lever presses are always followed by the US, it has been suggested that sign-tracking could be due to unintended reinforcement of the response. To eliminate the possibility of such instrumental learning the omission schedule, in which a response to the CS cancels the US, was introduced. Previous research has shown that training rats on autoshaping and switching them to an omission schedule generally reduces but does not eliminate sign-tracking, suggesting that it may be due to both Pavlovian and instrumental learning. In the present study naive rats trained on an omission schedule sign-tracked less than a control group exposed to random, unpaired CS and US presentations, suggesting that they learned to withhold the lever press response because of the negative contingency between that response and the US. In a second experiment rats with dorsal hippocampus lesions sign-tracked more than sham-lesioned rats on omission schedules, suggesting that this case of learning not to respond is hippocampus-based. This conclusion is consistent with many previous findings on the inability of hippocampal rats to withhold or suppress responding, and with studies suggesting that one form of extinction of learned responses in normal rats is due to competition from hippocampus-based learning not to respond.

Parallel learning in an autoshaping paradigm.

In an autoshaping task, a single conditioned stimulus (CS; lever insertion) was repeatedly followed by the delivery of an unconditioned stimulus (US; food pellet into an adjacent food magazine) irrespective of the rats' behavior. After repeated training trials, some rats responded to the onset of the CS by approaching and pressing the lever (sign-trackers). Lesions of dorsolateral striatum almost completely eliminated responding to the lever CS while facilitating responding to the food magazine (US). Lesions of the dorsomedial striatum attenuated but did not eliminate responding to the lever CS. Lesions of the basolateral or central nucleus of the amygdala had no significant effects on sign-tracking, but combined lesions of the 2 structures impaired sign-tracking by significantly increasing latency to the first lever press without affecting the number of lever presses. Lesions of the dorsal hippocampus had no effect on any of the behavioral measures. The findings suggest that sign-tracking with a single lever insertion as the CS may consist of 2 separate behaviors learned in parallel: An amygdala-mediated conditioned orienting and approach response and a dorsal striatum-mediated instrumental response. (PsycINFO Database Record

Dissociation of memory systems: The story unfolds.

In this article we describe the ideas and circumstances that led to the experiment demonstrating a triple dissociation of memory systems. We then move on to discuss the results of 20 years of investigation of those ideas. First, evidence is described from animal studies consistent with the ideas that memory for different kinds of information is stored in different brain systems, and that the hippocampus, amygdala, and dorsal striatum are each central structures in one of the systems. We then focus on the 3 tasks used in the original triple dissociation: win-stay learning, conditioned cue preference, and win-shift learning. Each of these tasks is specific to behavior resulting from the type of information stored in one of the systems, but the use of other behavioral tests that are sensitive to the types of information stored in other systems has revealed that, in each case, other types of information are acquired in parallel. Next, evidence consistent with the idea that the outputs of the systems compete for control of behavior is discussed together with alternative forms of more direct interactions among the systems. Finally, some evidence that many of these ideas about multiple parallel memory systems may apply to humans is reviewed.

A triple dissociation of memory systems: Hippocampus, amygdala, and dorsal striatum.

This study investigated the respective roles of the hippocampus, the amygdala, and the dorsal striatum in learning and memory. A standard set of experimental conditions for studying the effects of lesions to the three brain areas using an 8-arm radial maze was used: a win-shift version, a conditioned cue preference (CCP) version, and a win-stay version. Damage to the hippocampal system impaired acquisition of the win-shift task but not the CCP or win-stay tasks. Damage to the lateral amygdala impaired acquisition of the CCP task but not the win-shift or win-stay tasks. Damage to the dorsal striatum impaired acquisition of the win-stay task but not the win-shift or CCP tasks. These results are consistent with the hypothesis that the mammalian brain may be capable of acquiring different kinds of information with different, more-or-less independent neural systems. A neural system that includes the hippocampus may acquire information about the relationships among stimuli and events. A neural system that includes the amygdala may mediate the rapid acquisition of behaviors based on biologically significant events with affective properties. A neural system that includes the dorsal striatum may mediate the formation of reinforced stimulus-response associations.

Memory enhancement produced by post-training exposure to sucrose-conditioned cues.

A number of aversive and appetitive unconditioned stimuli (such as shock and food) are known to produce memory enhancement when they occur during the post-training period. Post-training exposure to conditioned aversive stimuli has also been shown to enhance memory consolidation processes. The present study shows for the first time that post-training exposure to conditioned stimuli previously paired with consumption of a sucrose solution also enhances memory consolidation. Male Long Evans rats were trained on a one-session conditioned cue preference (CCP) task on a radial arm maze. Immediately or 2 hours after training, rats consumed a sucrose solution or were exposed to cues previously paired with consumption of sucrose or cues previously paired with water. Twenty-four hours later, the rats were tested for a CCP. Immediate, but not delayed, post-training consumption of sucrose enhanced memory for the CCP. Immediate, but not delayed, post-training exposure to cues previously paired with sucrose, but not with water, also enhanced CCP memory. The possibility that rewarding and aversive conditioned stimuli affect memory by a common physiological process is discussed.

Parallel processing of information about location in the amygdala, entorhinal cortex and hippocampus.

The conditioned cue preference paradigm was used to study how rats use extra-maze cues to discriminate between 2 adjacent arms on an 8-arm radial maze, a situation in which most of the same cues can be seen from both arms but only one arm contains food. Since the food-restricted rats eat while passively confined on the food-paired arm no responses are reinforced, so the discrimination is due to Pavlovian stimulus-reward (or outcome) learning. Consistent with other evidence that rats must move around in an environment to acquire a spatial map, we found that learning the adjacent arms CCP (ACCP) required a minimum amount of active exploration of the maze with no reinforcers present prior to passive pairing of the extra-maze cues with the food reinforcer, an instance of latent learning. Temporary inactivation of the hippocampus during the pre-exposure sessions had no effect on ACCP learning, confirming other evidence that the hippocampus is not involved in latent learning. A series of experiments indentified a circuit involving fimbria-fornix and dorsal entorhinal cortex as the neural basis of latent learning in this situation. In contrast, temporary inactivation of the entorhinal cortex or hippocampus during passive training or during testing blocked ACCP learning and expression, respectively, suggesting that these two structures co-operate in using spatial information to learn the location of food on the maze during passive pairing and to express this combined information during testing. In parallel with these processes we found that the amygdala processes information leading to an equal tendency to enter both adjacent arms (even though only one was paired with food) suggesting that the stimulus information available to this structure is not sufficiently precise to discriminate between the ambiguous cues visible from the adjacent arms. Expression of the ACCP in normal rats depends on hippocampus-based learning to avoid the unpaired arm which competes with the amygdala-based tendency to enter that arm. In contrast, there is cooperation between amygdala- and hippocampus-based tendencies to enter the food-paired arm. These independent forms of learning contribute to the rat's ability to discriminate among spatial locations using ambiguous extra-maze cues.

Effects of post-training heroin and d-amphetamine on consolidation of win-stay learning and fear conditioning.

It has been proposed that the reinforcing properties of drugs of abuse are due, in part, to their ability to enhance memory consolidation. To test this hypothesis, heroin (0.03-3 mg/kg, SC) and d-amphetamine (0.5-2 mg/kg, SC) were administered to male Sprague-Dawley rats immediately or 4 h after training on win-stay and fear conditioning tasks. On the win-stay, immediate post-training administration of lower doses of heroin and d-amphetamine enhanced acquisition, and probe tests further revealed that these drugs enhanced different aspects of learning. Higher doses had no effect or impaired performance, particularly when administered repeatedly. On fear conditioning, the memory-enhancing effects of immediate post-training administration of lower heroin and d-amphetamine doses were revealed only when a single tone-shock pairing procedure was employed. Therefore, under appropriate experimental conditions, mildly stimulatory doses of heroin and d-amphetamine enhanced the acquisition of tasks thought to involve different types of learning. These results support the hypothesis that one of the ways in which drugs of abuse such as opiates and psychomotor stimulants reinforce behavior is by enhancing memory consolidation processes.

Lesions of basolateral and central amygdala differentiate conditioned cue preference learning with and without unreinforced preexposure.

In the separated arms conditioned cue preference (CCP) task rats are trained by confining them in one arm of an eight-arm radial maze with food and in another arm on the opposite side of the maze with no food on alternate days. After two such trials, rats prefer the food-paired arm when allowed to move freely between the two arms, neither of which contains food. However, if the rats are preexposed to the maze by exploring it without food before training, no preference is observed and at least 4 training trials are required to produce a CCP, suggesting that unreinforced preexposure to the maze latently inhibits acquisition. If this interpretation is correct, preexposure should reduce the size of the preference acquired with both 2 and 4 training trials. In Experiment 1, this prediction was replicated for 2 training trials; however, with 4 training trials, eliminating preexposure also eliminated the CCP. A previous finding that basolateral amygdala lesions impair the CCP with preexposure and 4 training trials was replicated in Experiment 2, but similar lesions had no effect on the CCP in non-preexposed rats given 2 training trials. In contrast, lesions of the central nucleus impaired the 2 training trial CCP but had no effect on the 4 training trial CCP. This double dissociation suggests that the BLA-mediated 4 training trial CCP may be due to learning about the reward features of the maze space, while the central-nucleus-mediated 2 training trial CCP may be due to a conditioned approach response.

Ultrasonic vocalization ratios reflect the influence of motivational state and amygdala lesions on different types of taste avoidance learning.

Consumption of a sweet solution (the CS) and ultrasonic vocalizations (USVs) emitted by rats were recorded in a conditioned taste avoidance paradigm. The rats' affective states were inferred from a ratio of high to low-frequency ultrasonic calls, which have been associated with positive and negative affect, respectively. The interacting effects of deprivation state and lesions of the basolateral amygdala (BLA) on CS consumption and affective state were examined. Rats were trained during the light phase while either 23 h or 3h water deprived by exposing them to the CS and then injecting them with LiCl or saline. They were tested by re-exposing them to the CS while either 23 or 3h deprived. Sham-lesioned rats that received LiCl injections consumed significantly less of the CS and evidenced relatively negative affect (inferred from the USV ratio) compared to control rats that received saline injections, regardless of the deprivation state in which they were trained or tested. Rats with BLA lesions trained while 23 h deprived failed to exhibit either reduced consumption or negative affect, regardless of whether they were tested while deprived for 23 or 3h. Identical lesions had no effect on reduced consumption or on negative affect in rats trained while 3h deprived, regardless of whether they were tested while deprived for 3 or 23 h. The findings suggest that both reduced consumption and negative affect are the results of different learning processes in deprived (23 h) and nearly satiated (3h, during the light phase) rats. The amygdala-dependent negative affective shift observed in deprived rats may be due to an aversive Pavlovian conditioned response that acts to suppress drinking. The amygdala-independent negative affective response and reduced consumption in nearly satiated rats could be due to a form of latent learning of a stimulus-outcome association.

Temporary inactivation of the dorsal entorhinal cortex impairs acquisition and retrieval of spatial information.

We tested the effects of temporary inactivation of the dorsal entorhinal cortex on spatial discrimination using a conditioned cue preference (CCP) paradigm. The three phases of the procedure were: pre-exposure: unreinforced exploration of the center platform and two adjacent arms of an eight-arm radial maze; training: rats were confined to the ends of the two arms on alternate days - one arm always contained food and the other never contained food; testing: unreinforced exploration of the center platform and the two arms. Rats that received bilateral infusions of saline into the dorsal entorhinal cortex before the training trials or before the test trial spent significantly more time in the arm that previously contained food than in the arm that never contained food, demonstrating that they had acquired and were able to express information that discriminated between the two adjacent maze arms. In contrast, rats that received bilateral, intra-entorhinal infusions of muscimol, a gamma-aminobutyric acid(a) (GABA(a)) agonist, before either training or testing spent equal amounts of time in the two arms, indicating that they failed to acquire and were unable to express this information. Interactions between the entorhinal cortex and hippocampus in the acquisition and expression of the information required for this discrimination are discussed.

Some highlights of research on the effects of caudate nucleus lesions over the past 200 years.

This review describes experiments on the effects of caudate nucleus lesions on behavior in monkeys, cats and rats. Early work on monkeys and cats focused on the relationship of the caudate to the cortex in motor control, leading to the idea that the caudate serves to inhibit behaviors initiated by the cortex. However, investigation of this hypothesis with systematic behavioral testing in all three species did not support this idea; rather, these studies provided evidence that caudate lesions affect memory functions. Two main types of memory tasks were affected. One type involved reinforced stimulus-response (S-R) associations, the other involved spatial information, response-reinforcer contingencies, or working memory. Recent evidence, mainly from rats, suggests that the dorsolateral part of the caudoputamen is central to the processing and consolidation of memory for reinforced S-R associations, and that the more medial and anterior parts of the same structure are part of a neural circuit that (in some cases) also includes the hippocampus, and mediates relational information and certain forms of working memory. The possibility that the spatial distribution of the patch and matrix compartments within the caudoputamen underlies these regional differences is discussed.

Unreinforced spatial (latent) learning is mediated by a circuit that includes dorsal entorhinal cortex and fimbria fornix.

The relationship of the entorhinal cortex (EC) and fimbria-fornix (FF) in unreinforced spatial (latent) learning was studied using the conditioned-cue-preference task on an eight-arm radial maze. The maze was turned before every trial to eliminate the use of local cues. During three pre-exposure sessions, food-deprived rats explored the center platform and two adjacent arms of the maze. Since most of the same cues were visible from both arm locations, discriminating them required spatial learning. The rats were then alternately confined to the end of each arm over several days: one arm always contained food, the other was empty. Finally, the rats were allowed free access to both arms with no food present. Normal rats spent more time in their food-paired than in their unpaired arms showing that they learned to discriminate between the arm locations. Bilateral micro-injections of muscimol into the dorsal, but not into the ventral EC, given before the pre-exposure sessions only, impaired the discrimination. The discrimination was also impaired in rats with unilateral lesions of FF and contralateral injections of muscimol into the dorsal EC given before the pre-exposure sessions. Ipsilateral FF lesions and entorhinal inactivation had no effect. These results indicate that the acquisition of information during unreinforced exploration of a novel environment requires an intact circuit involving the dorsal EC and fimbria fornix. Together with previous reports, that this form of learning does not require a functional hippocampus, (Gaskin et al. (2005) Hippocampus 15:1085-1093) the findings also suggest that the acquisition of certain kinds of unreinforced information by this circuit is independent of the hippocampus.

Roles of learning and motivation in preference behavior: mediation by entorhinal cortex, dorsal and ventral hippocampus.

In the latent cue preference (LCP) task, water-deprived rats alternately drink a salt solution in one distinctive compartment of a conditioned cue preference (CCP) apparatus and water in the other compartment over 8 days (training trials). They are then given a choice between the two compartments with no solutions present (preference test). Previous findings showed that this training procedure results in two parallel forms of learning: conditioning to water-paired cues (a water-CCP) and latent learning of an association between salt and salt-paired compartment cues (a salt-LCP). Experiment 1 examined these two types of learning in isolation. Results showed that expression of the salt-LCP required salt deprivation during testing, but expression of the water-CCP did not require a deprivation state during testing. Other results showed that salt-LCP learning itself involves two distinct components: (1) the latent association among neutral cues in the salt-paired compartment, and (2) motivational information about salt deprivation during testing. Previous findings also demonstrated roles for the dorsal hippocampus (DH), ventral hippocampus (VH), and entorhinal cortex (EC) in salt-LCP learning. Experiment 2 examined the involvement of these structures during acquisition or expression of salt-LCP learning. Rats with cannulas aimed at DH, VH, or EC were given infusions of muscimol, either before exposure to the salt-paired, but not the water-paired, compartment during training or before the preference test. Inactivation of the DH or EC impaired both acquisition and expression of the association between salt and salt-paired compartment cues, while inactivation of the VH disrupted the influence of motivational information about salt deprivation required to express the salt-LCP. These results suggest unique roles for the EC-DH circuit and VH in salt-LCP learning, as well as a functional dissociation between the DH and VH.

Cooperation and competition between the dorsal hippocampus and lateral amygdala in spatial discrimination learning.

The conditioned cue preference (CCP) was used to study how rats discriminate between adjacent arms on a radial maze. Chai and White (Behav Neurosci 2004, 118:770-784) showed that an intact dorsal hippocampus is required to learn this discrimination and that an amygdala-based conditioned approach response that produces an equal tendency to enter both arms is simultaneously acquired. In the present experiments, rats were preexposed to the maze with no food and trained by alternately confining them at the ends of two adjacent arms, one that contained food and one that did not. When given a choice between these arms with no food present, the rats spent more time on their food-paired arms, suggesting they had learned to discriminate their locations. Temporary inactivation of the dorsal hippocampus with muscimol during confinement on the food-paired arm had no effect on the discrimination, but inactivation while on the no-food arm impaired it. This pattern of effects was reversed in rats with amygdala lesions (inactivation on the food-paired arm impaired, but inactivation on the no-food arm had no effect on the discrimination), showing that hippocampus-based and amygdala-based learning interact to influence the behavior of normal rats in this situation. The dorsal hippocampus learns about locations that contain food and about locations that do not contain food. The amygdala-based tendency to enter the food-paired arm cooperates with hippocampus-based foraging for food on the food-paired, but the amygdala-based tendency to enter the no-food arm competes with hippocampus-based learning about the absence of food on that arm.

Dorsal hippocampus function in learning and expressing a spatial discrimination.

Learning to discriminate between spatial locations defined by two adjacent arms of a radial maze in the conditioned cue preference paradigm requires two kinds of information: latent spatial learning when the rats explore the maze with no food available, and learning about food availability in two spatial locations when the rats are then confined in one arm with food and the other with no food. Previous research showed that a functional dorsal hippocampus is not required for latent learning. The present experiments show that it is required for learning about food availability, and during retrieval of both types of information.

Neural circuits mediating latent learning and conditioning for salt in the rat.

Male Long-Evans rats alternately drank a salt solution in one distinctive compartment of a conditioned cue preference (CCP) apparatus and water in a different compartment over 8 days (training trials) and were then given a choice between the two compartments with no solutions present (test trial). Rats that were water deprived during training, then salt+water deprived during testing, spent more time in their salt-paired compartments, a salt latent cue preference (LCP). Rats that were water-only deprived during training and testing spent more time in their water-paired compartments, a water CCP. Rats that were salt+water deprived during both training and testing spent more time in their salt-paired compartments, a salt CCP. Bilateral, pre-training lesions of the lateral amygdala impaired the water and salt CCPs but not the salt LCP, reflecting the role of the amygdala in Pavlovian conditioning. Lesions of the dorsal or ventral hippocampus impaired the salt LCP and the water and salt CCPs, possibly reflecting the role of the hippocampus in contextual learning. Lesions of the fimbria-fornix impaired the water and salt CCPs but not the salt LCP, while lesions of the entorhinal cortex impaired the salt LCP but not the CCPs. This suggests that the LCP depends on a circuit that includes dorsal and ventral hippocampus and entorhinal cortex, a major conduit of sensory information from the cortex. In contrast, the CCPs depend on the amygdala and a circuit that includes the hippocampus and fimbria-fornix, possibly as a conduit of motivational information from subcortical structures.

A latent cue preference based on sodium depletion in rats.

Three experiments show latent (or incidental) learning of salt-cue relationships using a conditioned cue-preference paradigm. Rats drank a salt solution while confined in one compartment and water in an adjacent, distinct compartment on alternate days. When given access to the two compartments with no solutions present, sodium-deprived rats preferred their salt-paired compartments; normal rats preferred their water-paired compartments. Reversing the deprivation states of the two groups reversed their preferences. These results show that rats can latently acquire associations between environmental cues and the taste of salt, and can use this information flexibly to guide behavior based upon internal cues produced by sodium deprivation.

Inactivation of the dorsal hippocampus does not affect learning during exploration of a novel environment.

The conditioned cue preference (CCP) task was used to study the ability of rats to discriminate between spatial locations. Food-deprived rats explored an eight-arm radial maze with no food present (pre-exposure). On subsequent days, they were alternately confined in one arm of the maze with food and in another arm with no food (training), followed by a preference test with no food present, to determine if they had learned to discriminate between the two arm locations. No injections were given during the two latter phases. With adjacent radial maze arms, rats given three 10-min pre-exposure sessions and four food-pairing trials exhibited a preference for their food-paired arms; rats not pre-exposed did not exhibit this preference. Rats pre-exposed 30 min after dorsal hippocampus injections of muscimol exhibited the preference. With widely separated maze arms, rats given two training trials with no pre-exposure exhibited a preference for the food-paired arm; rats that were given one pre-exposure session did not. Rats pre-exposed 30 min after dorsal hippocampus injections of muscimol did not exhibit the preference. The same intrahippocampal muscimol injections that failed to affect the influence of pre-exposure on CCP learning with both arm configurations impaired win-shift performance, a standard test of spatial learning. These findings suggest that a functional dorsal hippocampus is not required for the (incidental or latent) learning that occurs during unreinforced exploration of a novel environment. The information acquired during this activity subsequently produces a latent learning effect if it is used to discriminate between two ambiguous locations (adjacent arms) or a latent inhibition--like effect if it is used to discriminate between two unambiguous locations (separated maze arms).

How independent are parallel memory systems? A theoretical comment on Gibson and Shettleworth (2005).

Using only behavioral manipulations, B. M. Gibson and S. J. Shettleworth (2005) have shown that rats can simultaneously acquire both place and response information. Experiments in which these 2 kinds of information are presented sequentially (response followed by place), provide some evidence for blocking of place learning, suggesting an influence of information acquired by a response-learning system on the subsequent acquisition of information by a place-learning system. A detailed examination of the results in the context of information about the kinds of learning that take place on a maze suggests that the memory processes underlying the rats' behavior in this experiment were considerably more complex than is acknowledged by Gibson and Shettleworth. The discussion illustrates the importance of considering both behavioral and physiological information for understanding how learning and memory functions are organized in the brain.