Structural learning and the hippocampus

The neural substrates of higher-order conditioning: A review

Sensory preconditioned and second-order conditioned responding are each well-documented. The former occurs in subjects (typically rats) exposed to pairings of two relatively neutral stimuli, S2 and S1, and then to pairings of S1 and a motivationally significant event [an unconditioned stimulus (US)]; the latter occurs when the order of these experiences is reversed with rats being exposed to S1-US pairings and then to S2-S1 pairings. In both cases, rats respond when tested with S2 in a manner appropriate to the affective nature of the US, e.g., approach when the US is appetitive and withdrawal when it is aversive. This paper reviews the neural substrates of sensory preconditioning and second-order conditioning. It identifies commonalities and differences in the substrates of these so-called higher-order conditioning protocols and discusses these commonalities/differences in relation to what is learned. In so doing, the review highlights ways in which these types of conditioning enhance our understanding of how the brain encodes and retrieves different types of information to generate appropriate behavior.

A role for the medial temporal lobes in category learning

Despite much research, the role of the medial temporal lobes (MTL) in category learning is unclear. Two unstructured categorization experiments explored conditions that might recruit MTL category learning and memory systems—namely, whether the stimulus display includes one or two stimuli, and whether category membership depends on configural properties of the stimulus features. The results supported three conclusions. First, in agreement with prior research, learning with single stimulus displays depended on striatal-mediated procedural learning. Second, and most important, learning with pair displays was mediated by MTL declarative memory systems. Third, the use of stimuli in which category membership depends on configural properties of the stimulus features made MTL learning slightly more likely. Overall, the results suggested that the MTL are most likely to mediate learning when the participant must decide which of two configural stimuli belongs to a selected category.

Exploring the neurocognitive basis of episodic recollection in autism

Increasing evidence indicates that the subjective experience of recollection is diminished in autism spectrum disorder (ASD) compared to neurotypical individuals. The neurocognitive basis of this difference in how past events are re-experienced has been debated and various theoretical accounts have been proposed to date. Although each existing theory may capture particular features of memory in ASD, recent research questions whether any of these explanations are alone sufficient or indeed fully supported. This review first briefly considers the cognitive neuroscience of how episodic recollection operates in the neurotypical population, informing predictions about the encoding and retrieval mechanisms that might function atypically in ASD. We then review existing research on recollection in ASD, which has often not distinguished between different theoretical explanations. Recent evidence suggests a distinct difficulty engaging recollective retrieval processes, specifically the ability to consciously reconstruct and monitor a past experience, which is likely underpinned by altered functional interactions between neurocognitive systems rather than brain region-specific or process-specific dysfunction. This integrative approach serves to highlight how memory research in ASD may enhance our understanding of memory processes and networks in the typical brain. We make suggestions for future research that are important for further specifying the neurocognitive basis of episodic recollection in ASD and linking such difficulties to social developmental and educational outcomes.

Belief state representation in the dopamine system

Learning to predict future outcomes is critical for driving appropriate behaviors. Reinforcement learning (RL) models have successfully accounted for such learning, relying on reward prediction errors (RPEs) signaled by midbrain dopamine neurons. It has been proposed that when sensory data provide only ambiguous information about which state an animal is in, it can predict reward based on a set of probabilities assigned to hypothetical states (called the belief state). Here we examine how dopamine RPEs and subsequent learning are regulated under state uncertainty. Mice are first trained in a task with two potential states defined by different reward amounts. During testing, intermediate-sized rewards are given in rare trials. Dopamine activity is a non-monotonic function of reward size, consistent with RL models operating on belief states. Furthermore, the magnitude of dopamine responses quantitatively predicts changes in behavior. These results establish the critical role of state inference in RL.

Unraveling the Role of the Hippocampus in Reversal Learning

Research in reversal learning has mainly focused on the functional role of dopamine and striatal structures in driving behavior on the basis of classic reinforcement learning mechanisms. However, recent evidence indicates that, beyond classic reinforcement learning adaptations, individuals may also learn the inherent task structure and anticipate the occurrence of reversals. A candidate structure to support such task representation is the hippocampus, which might create a flexible representation of the environment that can be adaptively applied to goal-directed behavior. To investigate the functional role of the hippocampus in the implementation of anticipatory strategies in reversal learning, we first studied, in 20 healthy individuals (11 women), whether the gray matter anatomy and volume of the hippocampus were related to anticipatory strategies in a reversal learning task. Second, we tested 20 refractory temporal lobe epileptic patients (11 women) with unilateral hippocampal sclerosis, who served as a hippocampal lesion model. Our results indicate that healthy participants were able to learn the task structure and use it to guide their behavior and optimize their performance. Participants' ability to adopt anticipatory strategies correlated with the gray matter volume of the hippocampus. In contrast, hippocampal patients were unable to grasp the higher-order structure of the task with the same success than controls. Present results indicate that the hippocampus is necessary to respond in an appropriately flexible manner to high-order environments, and disruptions in this structure can render behavior habitual and inflexible.SIGNIFICANCE STATEMENT Understanding the neural substrates involved in reversal learning has provoked a great deal of interest in the last years. Studies with nonhuman primates have shown that, through repetition, individuals are able to anticipate the occurrence of reversals and, thus, adjust their behavior accordingly. The present investigation is devoted to know the role of the hippocampus in such strategies. Importantly, our findings evidence that the hippocampus is necessary to anticipate the occurrence of reversals, and disruptions in this structure can render behavior habitual and inflexible.

The computational nature of memory modification

Retrieving a memory can modify its influence on subsequent behavior. We develop a computational theory of memory modification, according to which modification of a memory trace occurs through classical associative learning, but which memory trace is eligible for modification depends on a structure learning mechanism that discovers the units of association by segmenting the stream of experience into statistically distinct clusters (latent causes). New memories are formed when the structure learning mechanism infers that a new latent cause underlies current sensory observations. By the same token, old memories are modified when old and new sensory observations are inferred to have been generated by the same latent cause. We derive this framework from probabilistic principles, and present a computational implementation. Simulations demonstrate that our model can reproduce the major experimental findings from studies of memory modification in the Pavlovian conditioning literature.

DOI:http://dx.doi.org/10.7554/eLife.23763.001

Our memories contain our expectations about the world that we can retrieve to make predictions about the future. For example, most people would expect a chocolate bar to taste good, because they have previously learned to associate chocolate with pleasure. When a surprising event occurs, such as tasting an unpalatable chocolate bar, the brain therefore faces a dilemma. Should it update the existing memory and overwrite the association between chocolate and pleasure? Or should it create an additional memory? In the latter case, the brain would form a new association between chocolate and displeasure that competes with, but does not overwrite, the original one between chocolate and pleasure.

Previous studies have shown that surprising events tend to create new memories unless the existing memory is briefly reactivated before the surprising event occurs. In other words, retrieving old memories makes them more malleable. Gershman et al. have now developed a computational model for how the brain decides whether to update an old memory or create a new one. The idea at the heart of the model is that the brain will attempt to infer what caused the surprising event. The reason the chocolate bar tastes unpalatable, for example, might be because it was old and had spoiled. Every time the brain infers a new possible cause for a surprising event, it will create an additional memory to store this new set of expectations. In the future we will know that spoiled chocolate bars taste bad.

However, if the brain cannot infer a new cause for the surprising event – because, for example, there appears to be nothing unusual about the unpalatable chocolate bar – it will instead opt to update the existing memory. The next time we buy a chocolate bar, we will have slightly lower expectations about how good it will taste. The dilemma of whether to update an existing memory or create a new one thus boils down to the question: is the surprising event the consequence of a new cause or an old one? This theory implies that retrieving a memory nudges the brain to infer that its associated cause is once again active and, since this is an old cause, it means that the memory will be eligible for updating.

Many experiments have been performed on the topic of modifying memories, but this is the first computational model that offers a unifying explanation for the results. The next step is to work out how to apply the model, which is phrased in abstract terms, to networks of neurons that are more biologically realistic.

DOI:http://dx.doi.org/10.7554/eLife.23763.002

The role of the hippocampus in generalizing configural relationships

The hippocampus has been implicated in integrating information across separate events in support of mnemonic generalizations. These generalizations may be underpinned by processes at both encoding (linking similar information across events) and retrieval (“on‐the‐fly” generalization). However, the relative contribution of the hippocampus to encoding‐ and retrieval‐based generalizations is poorly understood. Using fMRI in humans, we investigated the hippocampal role in gradually learning a set of spatial discriminations and subsequently generalizing them in an acquired equivalence task. We found a highly significant correlation between individuals’ performance on a generalization test and hippocampal activity during the test, providing evidence that hippocampal processes support on‐the‐fly generalizations at retrieval. Within the same hippocampal region there was also a correlation between activity during the final stage of learning (when all associations had been learnt but no generalization was required) and subsequent generalization performance. We suggest that the hippocampus spontaneously retrieves prior events that share overlapping features with the current event. This process may also support the creation of generalized representations during encoding. These findings are supportive of the view that the hippocampus contributes to both encoding‐ and retrieval‐based generalization via the same basic mechanism; retrieval of similar events sharing common features. © 2016 The Authors Hippocampus Published by Wiley Periodicals, Inc.

Contextual fear conditioning in humans using feature-identical contexts

Contextual fear conditioning studies in animals and humans found an involvement of the hippocampus and amygdala during fear learning. To exclude a focus on elements of the context we employed a paradigm, which uses two feature-identical contexts that only differ in the arrangement of the features and requires configural processing. We employed functional magnetic resonance imaging to determine the role of the hippocampus and neocortical areas during the acquisition of contextual fear in humans. For contextual fear acquisition, we paired one context (CS+) with an aversive electrical stimulus, whereas the other (CS-) was never followed by aversive stimulation. Blood oxygen level dependent activation to the CS+ was present in the insula, inferior frontal gyrus, inferior parietal lobule, superior medial gyrus and caudate nucleus. Furthermore, the amygdala and hippocampus were involved in a time-dependent manner. Psychophysiological interaction analyses revealed functional connectivity of a more posterior hippocampal seed region with the anterior hippocampus, posterior cingulate cortex and superior parietal lobule. The anterior hippocampus was functionally coupled with the amygdala and postcentral gyrus. This study complements previous findings in contextual fear conditioning in humans and provides a paradigm which might be useful for studying patients with hippocampal impairment.

Early exposure to volatile anesthetics impairs long-term associative learning and recognition memory

BACKGROUND

Anesthetic exposure early in life affects neural development and long-term cognitive function, but our understanding of the types of memory that are altered is incomplete. Specific cognitive tests in rodents that isolate different memory processes provide a useful approach for gaining insight into this issue.

METHODS

Postnatal day 7 (P7) rats were exposed to either desflurane or isoflurane at 1 Minimum Alveolar Concentration for 4 h. Acute neuronal death was assessed 12 h later in the thalamus, CA1-3 regions of hippocampus, and dentate gyrus. In separate behavioral experiments, beginning at P48, subjects were evaluated in a series of object recognition tests relying on associative learning, as well as social recognition.

RESULTS

Exposure to either anesthetic led to a significant increase in neuroapoptosis in each brain region. The extent of neuronal death did not differ between groups. Subjects were unaffected in simple tasks of novel object and object-location recognition. However, anesthetized animals from both groups were impaired in allocentric object-location memory and a more complex task requiring subjects to associate an object with its location and contextual setting. Isoflurane exposure led to additional impairment in object-context association and social memory.

CONCLUSION

Isoflurane and desflurane exposure during development result in deficits in tasks relying on associative learning and recognition memory. Isoflurane may potentially cause worse impairment than desflurane.

Isoflurane exposure in newborn rats induces long-term cognitive dysfunction in males but not females

Volatile anesthetics are used widely for achieving a state of unconsciousness, yet these agents are incompletely understood in their mechanisms of action and effects on neural development. There is mounting evidence that children exposed to anesthetic agents sustain lasting effects on learning and memory. The explanation for these behavioral changes remains elusive, although acute neuronal death after anesthesia is commonly believed to be a principal cause. Rodent models have shown that isoflurane exposure in newborns induces acute neuroapoptosis and long-term cognitive impairment. However, the assessment of predisposing factors is lacking. We investigated the role of sex by delivering isoflurane to postnatal day (P)7 male and female Sprague Dawley rats for 4 h. Brain cell death was assessed 12 h later using FluoroJade C staining in the thalamus, CA1-3 regions of hippocampus, and dentate gyrus. Behavior was assessed separately using a series of object recognition tasks and a test of social memory beginning at P38. We found that isoflurane exposure significantly increased neuronal death in each brain region with no difference between sexes. Behavioral outcome was also equivalent in simple novel object recognition. However, only males were impaired in the recognition of objects in different locations and contexts. Males also exhibited deficient social memory while females were intact. The profound behavioral impairment in males relative to females, in spite of comparable cell death, suggests that males are more susceptible to long-term cognitive effects and this outcome may not be exclusively attributed to neuronal death.

Effects of the C57BL/6 strain background on tauopathy progression in the rTg4510 mouse model

Background

Cross-breeding of transgenic mice is commonly used to assess gene-gene interactions, particularly in the context of disease. Strain background changes can influence the phenotype of mouse models and can confound crossbreeding studies. We sought to determine if changing the strain background of a commonly used mouse model of tauopathy (rTg4510) would significantly impact the originally reported phenotype. On the original F1 FVB/N x 129S6 background, rTg4510 mice present with progressive cognitive decline, increased insoluble tau, robust tau pathology and age-dependent neurodegeneration. One of the most common strains in mouse modeling is C57BL/6. We and others have previously reported that this strain background alters the phenotypes of various models, including the JNPL3 model of tauopathy. To determine if the phenotype of rTg4510 mice was similarly affected by the introduction of the C57BL/6 background, we compared rTg4510 mice on the original F1 FVB/N x 129S6 background to rTg4510 mice on an F1 FVB/N x C57BL/6NTac (B6/NTac) background, herein termed rTg4510_B6.

Results

Despite a small, but significant increase in soluble human tau levels, young rTg4510_B6 mice had equivalent levels of tau phosphorylation, aggregation and cognitive impairments as age-matched rTg4510 mice. At 6.5 months of age, rTg4510_B6 mice displayed hyperphosphorylated insoluble tau and robust cortical tau neuropathology that was equivalent to age-matched rTg4510 mice; however, 10.5-month-old rTg4510_B6 mice had greater amounts of phospho-tau in the cortex and hippocampus when compared to age-matched rTg4510 mice. Non-transgenic (NT) littermates of rTg4510_B6 (NT_B6) mice also had greater amounts of cortical and hippocampal phospho-tau at 10.5 months of age when compared to NT littermates of rTg4510 mice. Additionally, older rTg4510_B6 mice had gross forebrain neurodegeneration that was equivalent to age-matched rTg4510 mice.

Conclusions

Overall, our data shows that introduction of the C57BL/6 strain into the rTg4510 mouse background modestly alters the tau pathology that was originally reported in rTg4510 on the F1 FVB/129 background. In contrast, behavioral and neurodegenerative outcomes were not altered. These studies support the use of the rTg4510 mouse model on a partial C57BL/6 strain background without losing fidelity of the phenotype and suggest that the C57BL/6 background does not inherently protect against tauopathy.

Physical exercise during pregnancy improves object recognition memory in adult offspring

Exercising during pregnancy has been shown to improve spatial learning and short-term memory, as well as increase brain-derived neurotrophic factor mRNA levels and hippocampal cell survival in juvenile offspring. However, it remains unknown if these effects endure into adulthood. In addition, few studies have considered how maternal exercise can impact cognitive functions that do not rely on the hippocampus. To address these issues, the present study tested the effects of maternal exercise during pregnancy on object recognition memory, which relies on the perirhinal cortex (PER), in adult offspring. Pregnant rats were given access to a running wheel throughout gestation and the adult male offspring were subsequently tested in an object recognition memory task at three different time points, each spaced 2-weeks apart, beginning at 60 days of age. At each time point, offspring from exercising mothers were able to successfully discriminate between novel and familiar objects in that they spent more time exploring the novel object than the familiar object. The offspring of non-exercising mothers were not able to successfully discriminate between objects and spent an equal amount of time with both objects. A subset of rats was euthanized 1h after the final object recognition test to assess c-FOS expression in the PER. The offspring of exercising mothers had more c-FOS expression in the PER than the offspring of non-exercising mothers. By comparison, c-FOS levels in the adjacent auditory cortex did not differ between groups. These results indicate that maternal exercise during pregnancy can improve object recognition memory in adult male offspring and increase c-FOS expression in the PER; suggesting that exercise during the gestational period may enhance brain function of the offspring.

Association rules for rat spatial learning: the importance of the hippocampus for binding item identity with item location

Three cohorts of rats with extensive hippocampal lesions received multiple tests to examine the relationships between particular forms of associative learning and an influential account of hippocampal function (the cognitive map hypothesis). Hippocampal lesions spared both the ability to discriminate two different digging media and to discriminate two different room locations in a go/no-go task when each location was approached from a single direction. Hippocampal lesions had, however, differential effects on a more complex task (biconditional discrimination) where the correct response was signaled by the presence or absence of specific cues. For all biconditional tasks, digging in one medium (A) was rewarded in the presence of cue C, while digging in medium B was rewarded in the presences of cue D. Such biconditional tasks are “configural” as no individual cue or element predicts the solution (AC+, AD−, BD+, and BC−). When proximal context cues signaled the correct digging choice, biconditional learning was seemingly unaffected by hippocampal lesions. Severe deficits occurred, however, when the correct digging choice was signaled by distal room cues. Also, impaired was the ability to discriminate two locations when each location was approached from two directions. A task demand that predicted those tasks impaired by hippocampal damage was the need to combine specific cues with their relative spatial positions (“structural learning”). This ability makes it possible to distinguish the same cues set in different spatial arrays. Thus, the hippocampus appears necessary for configural discriminations involving structure, discriminations that potentially underlie the creation of cognitive maps.

The operation of pattern separation and pattern completion processes associated with different attributes or domains of memory

Pattern separation and pattern completion processes are central to how the brain processes information in an efficient manner. Research into these processes is escalating and deficient pattern separation is being implicated in a wide array of genetic disorders as well as in neurocognitive aging. Despite the quantity of research, there remains a controversy as to precisely which behavioral paradigms should be used to best tap into pattern separation and pattern completion processes, as well as to what constitute legitimate outcome measures reflecting impairments in pattern separation and pattern completion. This review will discuss a theory based on multiple memory systems that provides a framework upon which behavioral tasks can be designed and their results interpreted. Furthermore, this review will discuss the nature of pattern separation and pattern completion and extend these processes outside the hippocampus and across all domains of information processing. After these discussions, an optimal strategy for designing behavioral paradigms to evaluate pattern separation and pattern completion processes will be provided.

The hippocampus and visual perception

In this review, we will discuss the idea that the hippocampus may be involved in both memory and perception, contrary to theories that posit functional and neuroanatomical segregation of these processes. This suggestion is based on a number of recent neuropsychological and functional neuroimaging studies that have demonstrated that the hippocampus is involved in the visual discrimination of complex spatial scene stimuli. We argue that these findings cannot be explained by long-term memory or working memory processing or, in the case of patient findings, dysfunction beyond the medial temporal lobe (MTL). Instead, these studies point toward a role for the hippocampus in higher-order spatial perception. We suggest that the hippocampus processes complex conjunctions of spatial features, and that it may be more appropriate to consider the representations for which this structure is critical, rather than the cognitive processes that it mediates.

Pavlovian biconditional discrimination learning in the C57BL/6J mouse

Four experiments using mice examined acquisition of Pavlovian biconditional discriminations in which two stimulus compounds were paired with food (AX+ and BY+) and two were not (AY- and BX-). Temporally asynchronous compounds were generated by using contextual stimuli (Experiment 1) and 15-s discrete visual cues (Experiments 2A, 2B and 3) to disambiguate when embedded noise or tone stimuli would be paired with food. When food pellets followed both reinforced compounds, successful acquisition was obtained in Experiment 1 but not in Experiments 2A and 2B even though the order of trials was modeled after that used in Experiment 1. However, when differential outcomes followed the reinforced compounds in Experiment 3, acquisition was obtained with discrete cue stimulus compounds. The implications of these results for modulatory models of conditional discrimination learning in animals are discussed.

Passive spatial perception facilitates the expression of persistent hippocampal long-term depression

Learning-facilitated plasticity describes the ability of hippocampal synapses to respond with synaptic plasticity when weak afferent stimulation is coupled with a spatial learning event. Qualitative differences appear to influence whether long-term potentiation or long-term depression (LTD) are facilitated by spatial learning. At many hippocampal synapses, LTD is facilitated when rats actively explore a novel spatial context. We investigated whether learning-facilitated plasticity is expressed when an unconstrained but stationary rat observes a computer-generated spatial environment. Visual fields were separated. Novel object configurations were presented to one field; familiar constellations were presented to the other field. LTD was facilitated in the CA1 region of the hemisphere to which novel object constellations were presented. Familiar constellations had no effect. LTD facilitation was prevented by treatment with the protein translation inhibitor, anisomycin. LTD in the dentate gyrus was not facilitated by novel object constellations, suggesting that effects are not common to all hippocampal subfields. These data support a unique association of LTD in the CA1 region with learning about spatial context and indicate that rats can passively perceive space.

Physical exercise during adolescence versus adulthood: differential effects on object recognition memory and brain-derived neurotrophic factor levels

It is well established that physical exercise can enhance hippocampal-dependent forms of learning and memory in laboratory animals, commensurate with increases in hippocampal neural plasticity (brain-derived neurotrophic factor [BDNF] mRNA/protein, neurogenesis, long-term potentiation [LTP]). However, very little is known about the effects of exercise on other, non-spatial forms of learning and memory. In addition, there has been little investigation of the duration of the effects of exercise on behavior or plasticity. Likewise, few studies have compared the effects of exercising during adulthood versus adolescence. This is particularly important since exercise may capitalize on the peak of neural plasticity observed during adolescence, resulting in a different pattern of behavioral and neurobiological effects. The present study addressed these gaps in the literature by comparing the effects of 4 weeks of voluntary exercise (wheel running) during adulthood or adolescence on novel object recognition and BDNF levels in the perirhinal cortex (PER) and hippocampus (HP). Exercising during adulthood improved object recognition memory when rats were tested immediately after 4 weeks of exercise, an effect that was accompanied by increased BDNF levels in PER and HP. When rats were tested again 2 weeks after exercise ended, the effects of exercise on recognition memory and BDNF levels were no longer present. Exercising during adolescence had a very different pattern of effects. First, both exercising and non-exercising rats could discriminate between novel and familiar objects immediately after the exercise regimen ended; furthermore there was no group difference in BDNF levels. Two or four weeks later, however, rats that had previously exercised as adolescents could still discriminate between novel and familiar objects, while non-exercising rats could not. Moreover, the formerly exercising rats exhibited higher levels of BDNF in PER compared to HP, while the reverse was true in the non-exercising rats. These findings reveal a novel interaction between exercise, development, and medial temporal lobe memory systems.

Novel age-dependent learning deficits in a mouse model of Alzheimer's disease: implications for translational research

Computational modeling predicts that the hippocampus plays an important role in the ability to apply previously learned information to novel problems and situations (referred to as the ability to generalize information or simply as 'transfer learning'). These predictions have been tested in humans using a computer-based task on which individuals with hippocampal damage are able to learn a series of complex discriminations with two stimulus features (shape and color), but are impaired in their ability to transfer this information to newly configured problems in which one of the features is altered. This deficit occurs despite the fact that the feature predictive of the reward (the relevant information) is not changed. The goal of the current study was to develop a mouse analog of transfer learning and to determine if this new task was sensitive to pathological changes in a mouse model of AD. We describe a task in which mice were able to learn a series of concurrent discriminations that contained two stimulus features (odor and digging media) and could transfer this learned information to new problems in which the irrelevant feature in each discrimination pair was altered. Moreover, we report age-dependent deficits specific to transfer learning in APP+PS1 mice relative to non-transgenic littermates. The robust impairment in transfer learning may be more sensitive to AD-like pathology than traditional cognitive assessments in that no deficits were observed in the APP+PS1 mice on the widely used Morris water maze task. These data describe a novel and sensitive paradigm to evaluate mnemonic decline in AD mouse models that has unique translational advantages over standard species-specific cognitive assessments (e.g., water maze for rodent and delayed paragraph recall for humans).

Challenging the classical distinction between long-term and short-term memory: reconsidering the role of the hippocampus

The hippocampus and surrounding medial temporal lobe structures have long been held to be critical for long-term declarative memory, but not for short-term or working memory. In fact, the notion that patients with selective and bilateral medial temporal lobe lesions have intact short-term memory has been a key argument to support the classical distinction between long- and short-term memory. However, recent behavioral, neuroimaging and electrophysiological data collected in humans have begun to challenge this classical distinction. Converging evidence now suggests that the ability to maintain the configural relationships of visual information in working memory for periods as short as a few seconds critically depends on the hippocampus. In functional terms, the hippocampus may be necessary for coordinating short-term maintenance when it relies on distributed cortical representations of objects, locations and their conjunctions. These findings indicate a need for modifying the current diagnostic work-up of patients with hippocampal lesions and the neuropsychological criteria for hippocampal dysfunction, which are currently centered upon the theory that hippocampal lesions will primarily affect long-term memory.

Lesions of the rat perirhinal cortex spare the acquisition of a complex configural visual discrimination yet impair object recognition

Rats with perirhinal cortex lesions were sequentially trained in a rectangular water tank on a series of 3 visual discriminations, each between mirror-imaged stimuli. When these same discriminations were tested concurrently, the rats were forced to use a configural strategy to solve the problems effectively. There was no evidence that lesions of the perirhinal cortex disrupted the ability to learn the concurrent configural discrimination task, which required the rats to learn the precise combination of stimulus identity with stimulus placement (“structural” learning). The same rats with perirhinal cortex lesions were also unimpaired on a test of spatial working memory (reinforced T maze alternation), although they were markedly impaired on a new test of spontaneous object recognition. For the recognition test, rats received multiple trials within a single session in which on every trial, they were allowed to explore 2 objects, 1 familiar, the other novel. On the basis of their differential exploration times, rats with perirhinal cortex lesions showed very poor discrimination of the novel objects, thereby confirming the effectiveness of the surgery. The discovery that bilateral lesions of the perirhinal cortex can leave configural (structural) learning seemingly unaffected points to a need to refine those models of perirhinal cortex function that emphasize its role in representing conjunctions of stimulus features.

Hippocampal-anterior thalamic pathways for memory: uncovering a network of direct and indirect actions

This review charts recent advances from a variety of disciplines that create a new perspective on why the multiple hippocampal–anterior thalamic interconnections are together vital for human episodic memory and rodent event memory. Evidence has emerged for the existence of a series of parallel temporal–diencephalic pathways that function in a reciprocal manner, both directly and indirectly, between the hippocampal formation and the anterior thalamic nuclei. These extended pathways also involve the mammillary bodies, the retrosplenial cortex and parts of the prefrontal cortex. Recent neuropsychological findings reveal the disproportionate importance of these hippocampal–anterior thalamic systems for recollective rather than familiarity-based recognition, while anatomical studies highlight the precise manner in which information streams are kept separate but can also converge at key points within these pathways. These latter findings are developed further by electrophysiological stimulation studies showing how the properties of the direct hippocampal–anterior thalamic projections are often opposed by the indirect hippocampal projections via the mammillary bodies to the thalamus. Just as these hippocampal–anterior thalamic interactions reflect an interdependent system, so it is also the case that pathology in one of the component sites within this system can induce dysfunctional changes to distal sites both directly and indirectly across the system. Such distal effects challenge more traditional views of neuropathology as they reveal how extensive covert pathology might accompany localised overt pathology, and so impair memory.

Going beyond LTM in the MTL: a synthesis of neuropsychological and neuroimaging findings on the role of the medial temporal lobe in memory and perception

Studies in rats and non-human primates suggest that medial temporal lobe (MTL) structures play a role in perceptual processing, with the hippocampus necessary for spatial discrimination, and the perirhinal cortex for object discrimination. Until recently, there was little convergent evidence for analogous functional specialisation in humans, or for a role of the MTL in processes beyond long-term memory. A recent series of novel human neuropsychological studies, however, in which paradigms from the animal literature were adapted and extended, have revealed findings remarkably similar to those seen in rats and monkeys. These experiments have demonstrated differential effects of distinct stimulus categories on performance in tasks for which there was no explicit requirement to remember information across trials. There is also accruing complementary evidence from functional neuroimaging that MTL structures show differential patterns of activation for scenes and objects, even on simple visual discrimination tasks. This article reviews some of these key studies and discusses the implications of these new findings for existing accounts of memory. A non-modular view of memory is proposed in which memory and perception depend upon the same anatomically distributed representations (emergent memory account). The limitations and criticisms of this theory are discussed and a number of outstanding questions proposed, including key predictions that can be tested by future studies.

Hippocampus-dependent and -independent theta-networks of active maintenance

Recent studies in humans and animals raise the possibility that actively maintaining a detailed memory of a scene within working memory may require the hippocampus, a brain structure better known for its role in long-term memory. We show that the hippocampus is behaviorally and functionally critical for configural-relational (CR) maintenance by orchestrating the synchrony of occipital and temporal brain regions in the theta-frequency range. Using magnetoencephalography in healthy adults and patients with bilateral hippocampal sclerosis, we distinguish this hippocampus-dependent theta-network from one that is independent of the hippocampus and used for non-CR scene maintenance. This non-CR theta-network involved frontal and parietal brain regions. We also show that the functional and topographical dissociation between these two networks cannot be accounted for by perceptual difficulty or the amount of information to be maintained ("load"). Also, we confirm in healthy adults that active maintenance of the CR arrangement of objects within a scene is impaired by task-interference during the delay in a manner akin to working-memory maintenance processes. Together, these findings demand reconsideration of the classical functional-anatomical distinctions between long- and short-term memory.