Encoding versus retrieval of spatial memory: double dissociation between the dentate gyrus and the perforant path inputs into CA3 in the dorsal hippocampus

The hippocampus is an essential neural structure for spatial memory. Computational models suggest that the CA3 subregion of the hippocampus plays an essential role in encoding and retrieval of spatial memory. The perforant path (PPCA3) and dentate gyrus (DG)-mediated mossy fibers (MFs) compose major afferent inputs into CA3. A possible functional dissociation between these afferent inputs was attempted using a simple navigation test (i.e., the modified Hebb-Williams maze). Behavioral testing was combined with electrolytic lesions of PPCA3 or neurotoxic lesions of the DG, to eliminate each afferent input into CA3. Lesions in either afferent input into CA3 affected learning of an effective navigational path on the maze. The contributions of the two CA3 afferent inputs, however, were different regarding encoding and retrieval of memory measured based on indices operationally defined for the behavioral paradigm (i.e., encoding, the number of errors reduced within a day; retrieval, the number of errors reduced between days). The DG-lesioned animals exhibited deficits regarding the encoding index, but not the retrieval index, whereas the PPCA3-lesioned rats displayed deficits regarding the retrieval index, but not the encoding index. The results suggest that the two major afferent inputs of CA3 may contribute differentially to encoding and retrieval of spatial memory.     

Role of dCA3 efferents via the fimbria in the acquisition of a delay nonmatch to place task

Dorsal CA3, but not dorsal CA1, lesioned rats are impaired in the acquisition of a delay nonmatch to place task. In this study, dorsal CA3 efferent fibers in the fimbria were transected; while taking care to spare afferent fibers from the medial septum. Neurophysiological, anatomical tracing, and histochemical data suggest that the transection was selective to dorsal CA3 efferent fibers and spared afferents from the medial septum. Rats showed a deficit for acquisition, but not for performance once learned. One possible explanation is that a small change to the cholinergic inputs to dCA3 caused by a decrease in dorsal CA3 efferent signals reaching the medial septum may impair new learning but not performance of a task once learned.     

Validation of a single-day Morris Water Maze procedure used to assess cognitive deficits associated with brain damage

This experiment was designed to validate a single-day Morris Water Maze procedure used to assess cognitive functioning in rats. Separate groups of randomly assigned rats received either bilateral or unilateral fimbria fornix transections, bilateral or unilateral cortical ablations, or a sham surgical control procedure. Subjects were tested 7 days postoperatively with a modified version of the Morris Water Maze procedure that requires only a single day of training. The results indicated that bilateral fimbria fornix transections severely disrupted acquisition. Unilateral fimbria fornix transections and bilateral and unilateral cortical lesions disrupted acquisition less severely but impaired subsequent test performance. In general, unilateral lesions of both types produced less severe deficits than bilateral lesions. The practical and analytical advantages of the single-day procedure are discussed.     

Hebb-Williams performance and scopolamine challenge in rats with partial immunotoxic hippocampal cholinergic deafferentation

Recent studies suggested that the cholinergic innervation of the hippocampus is not crucial for spatial learning, but it might be important for other forms of learning. This study assessed the effects of partial immunotoxic cholinergic lesions in the medial septum and concurrent scopolamine challenge in a complex learning task, the Hebb-Williams maze. Long-Evans rats were given intraseptal injections of 192 IgG-saporin (SAPO). Rats injected with phosphate-buffered saline (PBS) served as controls. Starting 25 days after surgery, behavioural performance was assessed in the Hebb-Williams maze test without prior or after injection of scopolamine (0.17 or 0.5 mg/kg, i.p.). In SAPO rats, histochemical analysis showed a 40-45% decrease in the density of hippocampal AChE staining. The number of ChAT-positive cell bodies in the medial septum was also significantly decreased (-56%) and there was a non-significant reduction of the number of parvalbumine-positive neurons. The behavioural results demonstrated that the lesions induced small but significant learning deficits. At 0.17 mg/kg, scopolamine produced more impairments in SAPO rats than in PBS-injected rats, suggesting an additive effect between the partial lesion and the drug. These observations indicate that the Hebb-Williams test may be more sensitive to alterations of septohippocampal cholinergic function, than radial- or water-maze tasks. They also show that subtle learning deficits can be detected after partial lesions of the cholinergic septohippocampal pathways. Finally, the data from the scopolamine challenge are in keeping with clinical results showing higher sensitivity to muscarinic blockade in aged subjects in whom weaker cholinergic functions can be presumed.     

Selective lesions of the fimbria and the fornix in the rat: differential effects on CA1 and dentate theta

The effects of electrolytic lesions of the dorsal fornix and the dorsomedial fimbria on the CA1 and the dentate theta rhythms (theta s) recorded from the dorsal hippocampal formation were investigated in the ether-anesthetized rat. The results showed that (i) fornix lesions mainly affected CA1 theta, (ii) fimbrial lesions mainly affected dentate theta, and (iii) combined fornix-fimbria lesions suppressed both CA1 and dentate theta s. When considered in connection with other observations, these data suggest that the septal projections pacing the CA1 theta may course essentially within the dorsomedial fornix whereas those pacing the dentate theta may pass essentially within the dorsomedial fimbria. Moreover, our data provide new support for the hypothesis that at least two septohippocampal neural systems are anatomically and functionally independent and capable of controlling the theta activity of the dorsal hippocampal formation of the anesthetized rat.     

The effects of knife cuts of hippocampal pathways on epileptic activity in the seizure-sensitive gerbil

Previous studies have shown morphological differences in the hippocampal formation of seizure-sensitive gerbils as compared to seizure-resistant gerbils. To determine the significance of these differences, lesions were made of hippocampal afferents and efferents. Seizure-sensitive gerbils with bilateral knife cuts of the perforant path, including those with bilateral fornix lesions, showed no seizure activity following surgery. However, bilateral transections of the fimbria of the fornix, unilateral lesions of the perforant path and sham surgeries had no significant effect on seizure activity. The termination of seizure activity with bilateral lesions of the perforant path suggests that this pathway, as opposed to the fornix, is required for motor seizures in this strain of gerbils.     

Effects of discrete kainic acid-induced hippocampal lesions on spatial and contextual learning and memory in rats

Substantial information is available concerning the influence of global hippocampal lesions on spatial learning and memory, however the contributions of discrete subregions within the hippocampus to these functions is less well understood. The present investigation utilized kainic acid to bilaterally lesion specific areas of the rat hippocampus. These animals were subsequently tested on a spatial orientation task using a circular water maze, and on an associative/contextual task using passive avoidance conditioning. The results indicate that both the dorsal CA1 and the ventral CA3 subregions play important roles in learning. Specifically, CA1 lesions produced a deficit in the acquisition of the water maze task and a significant memory impairment on the passive avoidance task. CA3 lesions also caused learning deficits in the acquisition of the water maze task, and produced even greater impairments in performance on the passive avoidance task. We conclude that CA1 and CA3 hippocampal subregions each play significant roles in the overall integration of information concerning spatial and associative learning.     

神经元型一氧化氮合酶在学习记忆过程中的变化和作用

目的　探讨神经元型一氧化氮合酶 (neuronalnitricoxidesynthase,nNOS)及一氧化氮 (nitricoxide ,NO)在学习记忆机制中的相关作用。方法　采用免疫组化方法观察Y迷宫空间辨别学习训练后大鼠不同脑区nNOS表达变化 ,并探讨特异性nNOS抑制剂 7 nitroindozal(7 NI)、钙拮抗剂尼莫通 (nimotop)腹腔注射对大鼠学习获得和记忆再现能力的影响。结果　学习训练后海马各亚区nNOS样神经元数量及染色强度明显增加 ,而皮层和纹状体区则无显著变化 ;7 NI以剂量依赖方式损伤大鼠的学习获得能力 ,但不影响记忆再现 ,尼莫通则对这两种能力均有破坏。结论　提示学习记忆过程可能伴有nNOS合成及活性增加 ,nNOS/NO在学习获得阶段具有重要作用。     

The role of the direct perforant path input to the CA1 subregion of the dorsal hippocampus in memory retention and retrieval

Subregional analyses of the hippocampus have suggested a selective role for the CA1 subregion in intermediate/long-term spatial memory and consolidation, but not short-term acquisition or encoding processes. It remains unclear how the direct cortical projection to CA1 via the perforant path (pp) contributes to these CA1-dependent processes. It has been suggested that dopamine selectively modulates the pp projection to CA1 while having little to no effect on the Schaffer collateral (SC) projection to CA1. This series of behavioral and electrophysiological experiments takes advantage of this pharmacological dissociation to demonstrate that the direct pp inputs to CA1 are critical in CA1-dependent intermediate-term retention and retrieval function. Here we demonstrate that local infusion of the nonselective dopamine agonist, apomorphine (10, 15 microg), into the CA1 subregion of awake animals produces impairments in between-day retention and retrieval, sparing within-day encoding of a modified Hebb-Williams maze and contextual conditioning of fear. In contrast, apomorphine produces no deficits when infused into the CA3 subregion. To complement the behavioral analyses, electrophysiological data was collected. In anesthetized animals, local infusion of the same doses of apomorphine significantly modifies evoked responses in the distal dendrites of CA1 following angular bundle stimulation, but produces no significant effects in the more proximal dendritic layer following stimulation of the SC. These results support a modulatory role for dopamine in the EC-CA1, but not CA3-CA1 circuitry, and suggest the possibility of a more fundamental role for EC-CA1 synaptic transmission in terms of intermediate-term, but not short-term spatial memory.     

Visual discrimination learning impairments produced by combined transections of the anterior temporal stem, amygdala and fornix in marmoset monkeys

Marmoset monkeys (Callithrix jacchus) with bilateral transections of the anterior temporal stem, amygdala and fornix were unable to relearn a 2-choice object discrimination first learnt prior to surgery, and were very severely impaired at relearning a concurrent object discrimination task which they had learnt and relearnt prior to surgery, indicating that they had a dense retrograde amnesia. They also had difficulty learning new visual object discriminations but were only mildly impaired on spatial learning. When tested on new learning of concurrent discriminations 8 to 10 weeks after surgery, three operated monkeys were unable to reach criterion in 400 trials while the remaining two operated monkeys performed within the normal range. The operated monkeys were subsequently shown to be impaired on acquisition of shape discriminations using black objects. These anterograde effects suggest that the impairment runs mainly in the domain of visual analysis. The monkeys also exhibited many of the features of the Klüver-Bucy syndrome. Histological analysis indicated that in addition to cutting some of the subcortical temporal lobe efferent pathways, the surgical procedures had cut the cholinergic afferents to the temporal neocortex, entorhinal cortex, and hippocampus. In a second experiment we found that treatment with the cholinergic agonist pilocarpine, which is effective in monkeys with specific cholinergic lesions, was unable to remediate the lesion-induced impairments. This suggests that transection of the non-cholinergic afferents, or the temporal lobe subcortical efferents, contributed to the behavioural syndrome and the learning and retention deficits seen in these monkeys.     

Cholinergic innervation of the primate hippocampal formation: II. Effects of fimbria/fornix transection

The distribution of choline acetyltransferase (ChAT)-immunoreactive and acetylcholines-terase (AChE)-positive fibers and terminals was analyzed in the hippocampal formation of macaque monkeys subjected to transection of the fimbria/fornix. Cases with either unilateral or bilateral transections were prepared, with post transection survival times ranging from 2 weeks to 1.5 years. The fimbria/fornix transection resulted in a dramatic decrease in the number of cholinergic fibers in most regions of the hippocampal formation. Some hippocampal regions. however, showed relatively greater sparing of ChAT- or AChE-positive fibers. In practically all regions of the hippocampal formation, residual AChE-positive fibers were more abundant than ChAT-immunoreactive fibers. In animals with unilateral lesions, the distribution patterns and density of AChE and ChAT staining on the side contralateral to the lesion were generally similar to those of sections from unlesioned control brains. The largest decreases in the densities of positive fibers were observed in the dentate gyrus, CA3 and CA2 fields of the hippocampus, subiculum, parasubiculum, and medial and caudal parts of the entorhinal cortex. Fibers were relatively better preserved in the rostral or uncal portion of the hippocampus and dentate gyrus and in the rostral portion of the entorhinal cortex. The presubiculum demonstrated remarkable sparing that contrasted with the almost complete loss of fibers in the parasubiculum. Interestingly, animals killed approximately 1.5 years after the fornix transection showed essentially the same pattern of fiber loss as the cases with shorter survival periods. This indicates that the residual ChAT-immunoreactive fibers, many of which reach the hippocampal formation through a ventral cholinergic pathway, are not capable of reinnervating the denervated portions of the hippocampal formation. This appears to distinguish the monkey from the rat, for which substantial sprouting and reinnervation of cholinergic fibers have been reported after similar lesions. © 1996 Wiley-Liss, Inc.     

大鼠纹状体边缘区和海马空间学习功能比较的研究

目的　 研究纹状体边缘区和海马在空间学习功能方面有无区别。方法　 将大鼠穹隆海马伞切断及向边缘区注射海人藻酸 ,用Morris水迷宫测试两实验组和相应对照组的空间学习能力的变化。结果　 穹隆海马伞组和边缘区注射海人藻酸组的平均逃避潜伏期与对照组相比均有延长 ,而且延长幅度大致相同。穿环数均有所下降 ,幅度也大致相同。 结论　 纹状体边缘区同海马一样 ,与大鼠的空间学习功能有着密切关系。     

Selective fimbria lesions impair acquisition of working and reference memory of rats in a complex spatial discrimination task

It has been reported that transections of the fimbria-fornix or lesions of the hippocampus selectively impair spatial working memory. Disruptive effects of these lesions on reference memory performance, however, have also been reported. We studied the effects of selective fimbria lesions on the acquisition of a complex spatial discrimination in the cone field. The cone field task is a place learning task that permits the simultaneous assessment of working and reference memory performance. Reproducible bilateral stereotaxic lesions were made by knife-cuts parallel to the midline. Sham lesions consisted of similar knife-cuts that were restricted to the overlying tissue. The rats were randomly started from 1 of 4 positions in order to prevent the development of a fixed food search pattern. On both memory components, fimbria-lesioned rats made about twice as many errors as the sham-lesioned and intact subjects, even after extensive training. Transection of the fimbria caused pronounced cholinergic denervation, predominantly at the more ventral part of the hippocampus, as indicated by reduced acetylcholinesterase histochemistry. Our results suggest a major role of the cholinergic innervation of the ventral hippocampus in spatial discrimination.     

GABAergic septohippocampal neurons are not necessary for spatial memory.

The medial septum/vertical limb of the diagonal band of Broca (MSDB) provides a major input to the hippocampus and is important for spatial memory. Both cholinergic and GABAergic MSDB neurons project to the hippocampus, and nonselective lesions of the MSDB or transections of the septohippocampal pathway impair spatial memory. However, selective lesions of cholinergic MSDB neurons using 192-IgG saporin (SAP) do not impair or only mildly impair spatial memory. Previously, intraseptal kainic acid was found to reduce levels of glutamic acid decarboxylase, a marker of GABAergic neurons, but not to alter the levels of choline acetyltransferase, a marker of cholinergic neurons. The present study further characterized the effects of kainic acid on GABAergic MSDB neurons and examined the effects of intraseptal kainic acid on spatial memory. Saline, kainic acid, SAP, or the combination of kainic acid and SAP was administered into the MSDB of rats. Spatial memory was assessed in an eight-arm radial maze and a water maze. Kainic acid destroyed GABAergic septohippocampal neurons, but spared cholinergic neurons. SAP eliminated MSDB cholinergic neurons, sparing noncholinergic neurons. Coadministration of kainic acid and SAP destroyed GABAergic and cholinergic MSDB neurons. Acquisition of the radial maze task and performance on this task with 4-h delays were unimpaired by intraseptal kainic acid or SAP, but were impaired by coadministration of kainic acid and SAP. Acquisition of the water maze task was unaffected by intraseptal kainic acid, delayed slightly by SAP, and impaired severely by coadministration of kainic acid and SAP. These results provide evidence that kainic acid at appropriate concentrations effectively destroys GABAergic septohippocampal neurons, while sparing cholinergic MSDB neurons. Furthermore, lesions of the GABAergic septohippocampal neurons do not impair spatial memory. While lesions of cholinergic MSDB neurons may mildly impair spatial memory, the combined lesion of GABAergic and cholinergic septohippocampal neurons resulted in a memory impairment that was greater than that observed after a selective lesion to either population. Thus, damage of GABAergic or cholinergic MSDB neurons, which together comprise the majority of the septohippocampal pathway, cannot totally account for the spatial memory impairment that is observed after nonselective lesions of the MSDB.     

Conditional associative learning and the hippocampal system

Rats with lesions of the fornix, the dorsal hippocampus, or a control operation were trained on a spatial-visual conditional associative learning task in which they had to learn to associate particular locations with specific visual stimuli. Animals with damage of the fornix were able to learn the task at a rate comparable to that of the control animals, but the performance of the hippocampal rats was significantly impaired in comparison with that of both the control and the fornix groups. In a second experiment, lesions to the fornix or the dorsal hippocampus significantly impaired performance on a spatial working memory task, the eight-arm radial maze. These findings suggest that the interaction between the hippocampus and subcortical structures via the fornix may be critical only for certain types of spatial learning and memory. Hippocampus 1998;8:131–137. © 1998 Wiley-Liss, Inc.     

NGF Deprivation of Adult Rat Brain Results in Cholinergic Hypofunction and Selective Impairments in Spatial Learning

Cholinergic hypofunction has often been correlated with a variety of behavioural impairments. In the present study, adult Wistar rats were intraventricularly infused with antibodies to nerve growth factor (anti-NGF) to examine the effects on cholinergic neurons of the basal forebrain, and on behavioural performance. Immunocytochemical techniques indicated that chronically infused anti-NGF penetrates into the basal forebrain, cortex, striatum, corpus callosum and hippocampus, confirming previous findings after a single injection. Treatment with anti-NGF for 1 or 2 weeks resulted in a significant decrease of 27-33% in density of choline acetyltransferase immunostaining of the cholinergic cell bodies in the medial septum and vertical diagonal band, and a 26% reduction in choline acetyltransferase enzyme activity in the septal area. An array of spatial learning Morris water maze tasks was used to distinguish between acquisition skills and the flexible use of learned information in novel tests. Rats subjected to the spatial learning paradigm received anti-NGF infusion for 2 weeks prior to and for another 2 weeks during the behavioural testing. The anti-NGF-treated animals were found to be no different from those receiving control serum in the Morris water maze acquisition task, either in the latency to find the platform or in the time spent searching in the training quadrant when the platform was removed. However, in consecutive extinction trials, anti-NGF rats continued to search in the empty training quadrant, suggesting the occurrence of perseveration; control rats expanded their search over other areas of the pool. This inflexibility of the anti-NGF rats was also evident from their difficulty in learning to find a relocated platform in the reversal task. Finally, the anti-NGF-treated animals demonstrated hyperactivity in the open field, resembling in this respect the behaviour of animals after septal and fimbria fornix lesions, and during pharmacological cholinergic blockade. While these data support a role for NGF in cholinergic function and spatial learning behaviour, they indicate that deficits in the latter, e.g. those demonstrated in impaired aged rats, may not be attributable selectively to deficits in the function of NGF-sensitive cholinergic neurons of the basal forebrain.     

Both the dorsal hippocampus and the dorsolateral striatum are needed for rat navigation in the Morris water maze

The multiple memory systems theory proposes that the hippocampus and the dorsolateral striatum are the core structures of the spatial/relational and stimulus-response (S-R) memory systems, respectively. This theory is supported by double dissociation studies showing that the spatial and cue (S-R) versions of the Morris water maze are impaired by lesions in the dorsal hippocampus and dorsal striatum, respectively. In the present study we further investigated whether adult male Wistar rats bearing double and bilateral electrolytic lesions in the dorsal hippocampus and dorsolateral striatum were as impaired as rats bearing single lesions in just one of these structures in learning both versions of the water maze. Such a prediction, based on the multiple memory systems theory, was not confirmed. Compared to the controls, the animals with double lesions exhibited no improvement at all in the spatial version and learned the cued version very slowly. These results suggest that, instead of independent systems competing for holding control over navigational behaviour, the hippocampus and dorsal striatum both play critical roles in navigation based on spatial or cue-based strategies.     

Transection of CA3 does not affect memory performance in rats

Longitudinal hippocampal pathways are needed for seizure synchronization, and there is evidence that their transection may abolish seizures. However, the effect of such transection on memory is unknown. In this study, we investigated the effect of transverse CA3 transections on memory function in Sprague-Dawley rats. With a stereotactic knife, a single CA3 transection was made unilaterally (n=5) or bilaterally (n=5). Sham surgery was done in another group (n=4). Morris water maze and novel object recognition tests were started 18 days later and revealed no significant differences between transected animals and controls. Cresyl-violet brain staining confirmed the locations of transections in the CA3 region. We conclude that normal performances in Morris water maze and novel object recognition tests do not appear to require intact transmission throughout the whole length of CA3, supporting the hypothesis that CA3 transections may be used in temporal lobe epilepsy to interrupt seizure circuitry while preserving memory.     

Transplantation of neurospheres after granule cell lesions in rats: cognitive improvements despite no long-term immunodetection of grafted cells

EGF-responsive C17 murine-derived neural stem cells (neurospheres) were grafted into the dentate gyrus of adult male rats after dentate granule cells lesions produced by colchicine injections. Behavioural performance was evaluated over two post-grafting periods, using tests sensitive to hippocampal dysfunctions. The first period began 1 month after grafting and testing conducted in the water maze and the radial maze distinguished working- and reference-memory performance. The second period began 9 months after grafting and learning performance was also evaluated in a Hebb-Williams maze, in addition to both other tests. The lesions induced lasting deficits in all tests. During the first period, the grafts had no effect in either test. Conversely, during the second period, grafted rats showed a weak improvement in the water maze and a significant increase of reference memory performance in the radial maze. In the Hebb-Williams maze, performance of grafted rats was close to normal.Strengthening the idea that dentate gyrus granule cells play an important role in the acquisition of new (perhaps more configural than only spatial) information, our results, moreover, suggest that neurosphere grafts may foster recovery after damage to point-to-point connection systems in the adult brain.