Suppression of neurotoxic lesion-induced seizure activity: evidence for a permanent role for the hippocampus in contextual memory

Revisiting systems consolidation and the concept of consolidation

For more than 50 years, knowledge of memory processes has been based on the consolidation hypothesis, which postulates that new memories require time to become stabilized. Two forms of the consolidation model exist. The Cellular Consolidation concept is based upon retrograde amnesia induced by amnesic treatments, the severity of which decreases as the learning to treatment increases over minutes or hours. In contrast, The Systems Consolidation model is based on post-training hippocampal lesions, which produce more severe retrograde amnesia when induced after days than after weeks. Except for the temporal parameters, Cellular and Systems Consolidation show many similarities. Here we propose that Systems consolidation, much as Cellular Consolidation (see Gisquet- Verrier and Riccio, 2018), can be explained in terms of a form of state-dependency. Accordingly, lesions of the hippocampus induce a change in the internal state of the animal, which disrupts retrieval processes. But the effect of contextual change is known to decrease with the length of the retention intervals, consistent with time-dependent retrograde amnesia. We provide evidence supporting this new view.

Distributed learning episodes create a context fear memory outside the hippocampus that depends on perirhinal and anterior cingulate cortices

Damage to the hippocampus (HPC) typically causes retrograde amnesia for contextual fear conditioning. Repeating the conditioning over several sessions, however, can eliminate the retrograde amnesic effects. This form of reinstatement thus permits modifications to networks that can support context memory retrieval in the absence of the HPC. The present study aims to identify cortical regions that support the nonHPC context memory. Specifically, the contribution of the perirhinal cortex (PRH) and the anterior cingulate cortex (ACC) were examined because of their established importance to context memory. The findings show that context memories established through distributed reinstatement survive damage limited only to the HPC, PRH, or ACC. Combined lesions of the HPC and PRH, as well as the HPC and ACC, caused retrograde amnesia, suggesting that network modifications in the PRH and ACC enable context fear memories to become resistant to HPC damage.

Trajectory-modulated hippocampal neurons persist throughout memory-guided navigation

Trajectory-dependent splitter neurons in the hippocampus encode information about a rodent's prior trajectory during performance of a continuous alternation task. As such, they provide valuable information for supporting memory-guided behavior. Here, we employed single-photon calcium imaging in freely moving mice to investigate the emergence and fate of trajectory-dependent activity through learning and mastery of a continuous spatial alternation task. In agreement with others, the quality of trajectory-dependent information in hippocampal neurons correlated with task performance. We thus hypothesized that, due to their utility, splitter neurons would exhibit heightened stability. We find that splitter neurons were more likely to remain active and retained more consistent spatial information across multiple days than other neurons. Furthermore, we find that both splitter neurons and place cells emerged rapidly and maintained stable trajectory-dependent/spatial activity thereafter. Our results suggest that neurons with useful functional coding exhibit heightened stability to support memory guided behavior.

Has multiple trace theory been refuted?

Multiple trace theory (Nadel & Moscovitch, Current Opinion in Neurobiology, 1997, 7, 217-227) has proven to be one of the most novel and influential recent memory theories, and played an essential role in shifting perspective on systems-level memory consolidation. Here, we briefly review its impact and testable predictions and focus our discussion primarily on nonhuman animal experiments. Perhaps, the most often supported claim is that episodic memory tasks should exhibit comparable severity of retrograde amnesia (RA) for recent and remote memories after extensive damage to the hippocampus (HPC). By contrast, there appears to be little or no experimental support for other core predictions, such as temporally limited RA after extensive HPC damage in semantic memory tasks, temporally limited RA for episodic memories after partial HPC damage, or the existence of storage of multiple HPC traces with repeated reactivations. Despite these shortcomings, it continues to be a highly cited HPC memory theory.

Hippocampal Damage Causes Retrograde Amnesia and Slower Acquisition of a Cue-Place Discrimination in a Concurrent Cue-Place Water Task in Rats

Explanations of memory-guided navigation in rodents typically suggest that cue- and place-based navigations are independent aspects of behavior and neurobiology. The results of many experiments show that hippocampal damage causes both anterograde and retrograde amnesia (AA; RA) for place memory, but only RA for cue memory. In the present experiments, we used a concurrent cue-place water task (CWT) to study the effects of hippocampal damage before or after training on cue- and place-guided navigation, and how cue and place memory interact in damaged and control rats. We found that damaging the hippocampus before training caused a delay in the expression of cue-place navigation strategies relative to intact control animals; surprisingly, place navigation strategies emerged following pre-training hippocampal damage. With additional training, both control and damaged rats used local cues to navigate in the CWT. Damaged animals also show minor impairments in latency to navigate to the correct cue following a cue contingency reversal. By contrast to these anterograde effects, damage made after training causes RA for cue choice accuracy and latency to navigate to the correct cue. In addition, the extent of hippocampal damage predicted impairments in choice accuracy when lesions were made after training. These data extend previous work on the role of the hippocampus in cue and place memory-guided navigation, and show that the hippocampus plays an important role in both aspects of memory and navigation when present during the learning experience.

Hippocampal area CA1 and remote memory in rats

Hippocampal lesions often produce temporally graded retrograde amnesia (TGRA), whereby recent memory is impaired more than remote memory. This finding has provided support for the process of systems consolidation. However, temporally graded memory impairment has not been observed with the watermaze task, and the findings have been inconsistent with context fear conditioning. One possibility is that large hippocampal lesions indirectly disrupt (by retrograde degeneration) the function of areas that project to the hippocampus that are important for task performance or thought to be important for storing consolidated memories. We developed a discrete lesion targeting area CA1, the sole output of the hippocampus to neocortex, and tested the effects of this lesion on recent and remote memory in the watermaze task, in context fear conditioning, and in trace fear conditioning. In all three tasks, recent and remote memory were similarly impaired after CA1 lesions. We discuss factors that help to illuminate these findings and consider their relevance to systems consolidation.

Rat intra-hippocampal NMDA infusion induces cell-specific damage and changes in expression of NMDA and GABAA receptor subunits

Excessive stimulation of NMDA receptors with glutamate or other potent agonists such as NMDA leads to excitotoxicity and neural injury. In this study, we aimed to provide insight into an animal model of brain excitotoxic damage; single unilateral infusion of NMDA at mild dose into the hippocampal formation. NMDA infusion induced chronic, focal neurodegeneration in the proximity of the injection site. The lesion was accompanied by severe and progressive neuroinflammation and affected preferentially principal neurons while sparing GABAergic interneurons. Furthermore, the unilateral lesion did not cause significant impairment of spatial learning abilities. Finally, GluN1 and GluN2B subunits of NMDA receptor were significantly upregulated up to 3 days after the NMDA infusion, while GABAA α5 subunit was downregulated at 30 days after the lesion. Taken together, a single infusion of NMDA into the hippocampal formation represents an animal model of excitotoxicity-induced chronic neurodegeneration of principal neurons accompanied by severe neuroinflammation and subunit specific changes in NMDA and GABAA receptors.

Adult Hippocampal Neurogenesis, Fear Generalization, and Stress

The generalization of fear is an adaptive, behavioral, and physiological response to the likelihood of threat in the environment. In contrast, the overgeneralization of fear, a cardinal feature of posttraumatic stress disorder (PTSD), manifests as inappropriate, uncontrollable expression of fear in neutral and safe environments. Overgeneralization of fear stems from impaired discrimination of safe from aversive environments or discernment of unlikely threats from those that are highly probable. In addition, the time-dependent erosion of episodic details of traumatic memories might contribute to their generalization. Understanding the neural mechanisms underlying the overgeneralization of fear will guide development of novel therapeutic strategies to combat PTSD. Here, we conceptualize generalization of fear in terms of resolution of interference between similar memories. We propose a role for a fundamental encoding mechanism, pattern separation, in the dentate gyrus (DG)-CA3 circuit in resolving interference between ambiguous or uncertain threats and in preserving episodic content of remote aversive memories in hippocampal-cortical networks. We invoke cellular-, circuit-, and systems-based mechanisms by which adult-born dentate granule cells (DGCs) modulate pattern separation to influence resolution of interference and maintain precision of remote aversive memories. We discuss evidence for how these mechanisms are affected by stress, a risk factor for PTSD, to increase memory interference and decrease precision. Using this scaffold we ideate strategies to curb overgeneralization of fear in PTSD.

Memory consolidation

Conscious memory for a new experience is initially dependent on information stored in both the hippocampus and neocortex. Systems consolidation is the process by which the hippocampus guides the reorganization of the information stored in the neocortex such that it eventually becomes independent of the hippocampus. Early evidence for systems consolidation was provided by studies of retrograde amnesia, which found that damage to the hippocampus-impaired memories formed in the recent past, but typically spared memories formed in the more remote past. Systems consolidation has been found to occur for both episodic and semantic memories and for both spatial and nonspatial memories, although empirical inconsistencies and theoretical disagreements remain about these issues. Recent work has begun to characterize the neural mechanisms that underlie the dialogue between the hippocampus and neocortex (e.g., "neural replay," which occurs during sharp wave ripple activity). New work has also identified variables, such as the amount of preexisting knowledge, that affect the rate of consolidation. The increasing use of molecular genetic tools (e.g., optogenetics) can be expected to further improve understanding of the neural mechanisms underlying consolidation.

Inactivation of the anterior cingulate reveals enhanced reliance on cortical networks for remote spatial memory retrieval after sequential memory processing

One system consolidation model suggests that as time passes, ensembles of cortical neurons form strong connections to represent remote memories. In this model, the anterior cingulate cortex (ACC) serves as a cortical region that represents remote memories. However, there is debate as to whether remote spatial memories go through this systems consolidation process and come to rely on the ACC. The present experiment examined whether increasing the processing demand on the hippocampus, by sequential training on two spatial tasks, would more fully engage the ACC during retrieval of a remote spatial memory. In this scenario, inactivation of the ACC at a remote time point was hypothesized to produce a severe memory deficit if rats had been trained on two, sequential spatial tasks. Rats were trained on a water maze (WM) task only or a WM task followed by a radial arm maze task. A WM probe test was given recently or remotely to all rats. Prior to the probe test, rats received an injection of saline or muscimol into the ACC. A subtle deficit in probe performance was found at the remote time point in the group trained on only one spatial task and treated with muscimol. In the group trained on two spatial tasks and treated with muscimol, a subtle deficit in probe performance was noted at the recent time point and a substantial deficit in probe performance was observed at the remote time point. c-Fos labeling in the hippocampus revealed more labeling in the CA1 region in all remotely tested groups than recently tested groups. Findings suggest that spatial remote memories come to rely more fully on the ACC when hippocampal processing requirements are increased. Results also suggest continued involvement of the hippocampus in spatial memory retrieval along with a progressive strengthening of cortical connections as time progresses.

Relative contributions of CA3 and medial entorhinal cortex to memory in rats

The hippocampal CA1 field processes spatial information, but the relative importance of intra- vs. extra-hippocampal sources of input into CA1 for spatial behavior is unclear. To characterize the relative roles of these two sources of input, originating in the hippocampal field CA3 and in the medial entorhinal cortex (MEC), we studied effects of discrete neurotoxic lesions of CA3 or MEC on concurrent spatial and nonspatial navigation tasks, and on synaptic transmission in afferents to CA1. Lesions in CA3 or MEC regions that abolished CA3-CA1, or reduced MEC-CA1 synaptic transmission, respectively, impaired spatial navigation and unexpectedly interfered with cue response, suggesting that in certain conditions of training regimen, hippocampal activity may influence behavior otherwise supported by nonhippocampal neural networks. MEC lesions had milder and temporary behavioral effects, but also markedly amplified transmission in the CA3-CA1 pathway. Extensive behavioral training had a similar, but more modest effect on CA3-CA1 transmission. Thus, cortical input to the hippocampus modulates CA1 activity both directly and indirectly, through heterosynaptic interaction, to control information flow in the hippocampal loop. Following damage to hippocampal cortical input, the functional coupling of separate intra- and extra-hippocampal inputs to CA1 involved in normal learning may initiate processes that support recovery of behavioral function. Such a process may explain how CA3 lesions, which do not significantly modify the basic features of CA1 neural activity, nonetheless impair spatial recall, whereas lesions of EC input to CA1, which reduce the spatial selectivity of CA1 firing in foraging rats, have only mild effects on spatial navigation.

Retrograde and anterograde memory following selective damage to the dorsolateral entorhinal cortex

Anatomical and electrophysiological evidence suggest the dorsolateral entorhinal cortex (DLEC) is involved in processing spatial information, but there is currently no consensus on whether its functions are necessary for normal spatial learning and memory. The present study examined the effects of excitotoxic lesions of the DLEC on retrograde and anterograde memory on two tests of allocentric spatial learning: a hidden fixed-platform watermaze task, and a novelty-preference-based dry-maze test. Deficits were observed on both tests when training occurred prior to but not following n-methyl d-aspartate (NMDA) lesions of DLEC, suggesting retrograde memory impairment in the absence of anterograde impairments for the same information. The retrograde memory impairments were temporally-graded; rats that received DLEC lesions 1-3 days following training displayed deficits, while those that received lesions 7-10 days following training performed like a control group that received sham surgery. The deficits were not attenuated by co-infusion of tetrodotoxin, suggesting they are not due to disruption of neural processing in structures efferent to the DLEC, such as the hippocampus. The present findings provide evidence that the DLEC is involved in the consolidation of allocentric spatial information.

Recent memory for socially transmitted food preferences in rats does not depend on the hippocampus

The standard model of systems consolidation holds that the hippocampus (HPC) is involved only in the initial storage and retrieval of a memory. With time hippocampal-neocortical interactions slowly strengthen the neocortical memory, ultimately enabling retrieval of the memory without the HPC. Key support for this idea comes from experiments measuring memory recall in the socially-transmitted food preference (STFP) task in rats. HPC damage within a day or two of STFP learning can abolish recall, but similar damage five or more days after learning has no effect. We hypothesize that disruption of cellular consolidation outside the HPC could contribute to the amnesia with recent memories, perhaps playing a more important role than the loss of HPC. This view predicts that intraHPC infusion of Tetrodotoxin (TTX), which can block conduction of action potentials from the lesion sites, will block the retrograde amnesia in the STFP task. Here we confirm the previously reported retrograde amnesia with neurotoxic HPC damage within the first day after learning, but show that co-administration of TTX with the neurotoxin blocks the retrograde amnesia despite very extensive HPC damage. These results indicate that HPC damage disrupts cellular consolidation of the recent memory elsewhere; STFP memory may not ever depend on the HPC.

The impact of multiple memory formation on dendritic complexity in the hippocampus and anterior cingulate cortex assessed at recent and remote time points

Consolidation processes, involving synaptic and systems level changes, are suggested to stabilize memories once they are formed. At the synaptic level, dendritic structural changes are associated with long-term memory storage. At the systems level, memory storage dynamics between the hippocampus and anterior cingulate cortex (ACC) may be influenced by the number of sequentially encoded memories. The present experiment utilized Golgi-Cox staining and neuron reconstruction to examine recent and remote structural changes in the hippocampus and ACC following training on three different behavioral procedures. Rats were trained on one hippocampal-dependent task only (a water maze task), two hippocampal-dependent tasks (a water maze task followed by a radial arm maze task), or one hippocampal-dependent and one non-hippocampal-dependent task (a water maze task followed by an operant conditioning task). Rats were euthanized recently or remotely. Brains underwent Golgi-Cox processing and neurons were reconstructed using Neurolucida software (MicroBrightField, Williston, VT, USA). Rats trained on two hippocampal-dependent tasks displayed increased dendritic complexity compared to control rats, in neurons examined in both the ACC and hippocampus at recent and remote time points. Importantly, this behavioral group showed consistent, significant structural differences in the ACC compared to the control group at the recent time point. These findings suggest that taxing the demand placed upon the hippocampus, by training rats on two hippocampal-dependent tasks, engages synaptic and systems consolidation processes in the ACC at an accelerated rate for recent and remote storage of spatial memories.

Factors affecting graded and ungraded memory loss following hippocampal lesions

This review evaluates three current theories--Standard Consolidation (Squire & Wixted, 2011), Overshadowing (Sutherland, Sparks, & Lehmann, 2010), and Multiple Trace-Transformation (Winocur, Moscovitch, & Bontempi, 2010)--in terms of their ability to account for the role of the hippocampus in recent and remote memory in animals. Evidence, based on consistent findings from tests of spatial memory and memory for acquired food preferences, favours the transformation account, but this conclusion is undermined by inconsistent results from studies that measured contextual fear memory, probably the most commonly used test of hippocampal involvement in anterograde and retrograde memory. Resolution of this issue may depend on exercising greater control over critical factors (e.g., contextual environment, amount of pre-exposure to the conditioning chamber, the number and distribution of foot-shocks) that can affect the representation of the memory shortly after learning and over the long-term. Research strategies aimed at characterizing the neural basis of long-term consolidation/transformation, as well as other outstanding issues are discussed.

Remote context fear conditioning remains hippocampus-dependent irrespective of training protocol, training-surgery interval, lesion size, and lesion method

Systems consolidation involves the reorganization of brain circuits that support long-term memory. It is a prolonged process that can take days, weeks, or longer. An animal model of systems consolidation was established in the early 1990s and provided compelling support for the initial observations in humans, that hippocampal damage disproportionally impairs recent memory compared to remote memory. Context fear conditioning was the most frequently and successfully used task to study systems consolidation and demonstrate temporally graded retrograde amnesia. However, recent studies have failed to support these early findings of temporal gradients and instead reported that both recent and remote memories are equally impaired. Thus, the status of context fear conditioning as method to study the process of systems consolidation is at present uncertain. Accordingly, we evaluated classically conditioned fear memory in large groups of rats with hippocampal damage by manipulating several procedural variables including the training protocol, the training-surgery interval, the extent of hippocampal damage, and the method of damaging the hippocampus. The results indicate that hippocampal damage profoundly impairs context fear conditioning. These findings are unambiguous and independent of any particular procedural manipulation we evaluated. We suggest that the preponderance of currently available evidence indicates that context fear memory remains hippocampus-dependent indefinitely.

Role of adult neurogenesis in hippocampus-dependent memory, contextual fear extinction and remote contextual memory: new insights from ERK5 MAP kinase

Adult neurogenesis occurs in two discrete regions of the adult mammalian brain, the subgranular zone (SGZ) of the dentate gyrus (DG) and the subventricular zone (SVZ) along the lateral ventricles. Signaling mechanisms regulating adult neurogenesis in the SGZ are currently an active area of investigation. Adult-born neurons in the DG functionally integrate into the hippocampal circuitry and form functional synapses, suggesting a role for these neurons in hippocampus-dependent memory formation. Although results from earlier behavioral studies addressing this issue were inconsistent, recent advances in conditional gene targeting technology, viral injection and optogenetic approaches have provided convincing evidence supporting a role for adult-born neurons in the more challenging forms of hippocampus-dependent learning and memory. Here, we briefly summarize these recent studies with a focus on extra signal-regulated kinase (ERK) 5, a MAP kinase whose expression in the adult brain is restricted to the neurogenic regions including the SGZ and SVZ. We review evidence identifying ERK5 as a novel endogenous signaling pathway that regulates the pro-neural transcription factor Neurogenin 2, is activated by neurotrophins and is critical for adult neurogenesis. We discuss studies demonstrating that specific deletion of ERK5 in the adult neurogenic regions impairs several forms of hippocampus-dependent memory formation in mice. These include contextual fear memory extinction, the establishment and maintenance of remote contextual fear memory, and several other challenging forms of hippocampus-dependent memory formation including 48h memory for novel object recognition, contextual fear memory established by a weak foot shock, pattern separation, and reversal of spatial learning and memory. We also briefly discuss current evidence that increasing adult neurogenesis, by small molecules or genetic manipulation, improves memory formation and long-term memory.

The maintenance of established remote contextual fear memory requires ERK5 MAP kinase and ongoing adult neurogenesis in the hippocampus

Adult neurogenesis in the dentate gyrus of the hippocampal formation has been implicated in several forms of hippocampus-dependent memory. However, its role in the persistence of remote memory is unknown. Furthermore, whether the hippocampus plays a role in maintaining remote contextual memories is controversial. Here we used an inducible gene-specific approach for conditional deletion of erk5 in the adult neurogenic regions of the mouse brain to specifically impair adult neurogenesis. The erk5 gene was conditionally deleted under three different experimental conditions: prior to training for contextual fear, 6 days after training, or 5 weeks after training, We present evidence that remote memory was impaired under all three conditions. These data demonstrate that ongoing adult neurogenesis is required both for the initial establishment and the continued maintenance of remote contextual fear memory, even after the remote memory has transferred into extra-hippocampal regions of the brain 5 weeks after training.