Memory impairment induced by an interfering task is reverted by pre-frontal cortex lesions: a possible role for an inhibitory process in memory suppression in mice.

Interference theory refers to the idea that forgetting occurs because the recall of certain items interferes with the recall of other items. Recently, it has been proposed that interference is due to an inhibitory control mechanism, triggered by competing memories, that ultimately causes forgetting [Anderson MC (2003) Rethinking interference theory: Executive control and the mechanisms of forgetting. J Mem Lang 49:415-4453]. In the present research we study the interference process by submitting CD1 mice to two different hippocampal-dependent tasks: a place object recognition task (PORT) and a step-through inhibitory avoidance task (IA). Our results show a mutual interference between PORT and IA. To elucidate the possible neural mechanism underlying the interference process, we submit hippocampus- and prefrontal cortex-lesioned mice to PORT immediately before IA training. Results from these experiments show that prefrontal cortex lesions completely revert the impairing effect exerted by PORT administration on IA memory, while hippocampus lesions, that as expected impair memory for both PORT and IA, increase this effect. Altogether our results suggest that interference-induced forgetting is driven by an inhibitory control mechanism through activation of hippocampus-prefrontal cortex circuitry. The hippocampus seems to be crucial for storing information related to both behavioral tasks. Competition between memories triggers the inhibitory control mechanism, by activating prefrontal cortex, and induces memory suppression.

A role for ERK2 in reconsolidation of fear memories in mice.

Recent studies have shown that consolidated fear memories, when reactivated, return to a labile state that requires a new protein synthesis for reconsolidation. Post-retrieval infusion of an inhibitor of protein synthesis blocks memory reconsolidation processes. In a previous research, the role of MAPKs in memory consolidation has been shown in emotional tasks, such as passive and active avoidance. In particular, mice knockout for ERK1 had a better performance in comparison to wild type mice in both passive and active avoidance tasks. In the present study, in order to investigate the involvement of MAPKs in memory reconsolidation processes we administered immediately after retrieval, different doses of SL327 (an inhibitor of MEK, a kinase that activates both ERK1 and ERK2) both in C57BL/6 (C57) mice and ERK1 mutant mice tested in a fear conditioning task. Systemic administration of SL327 dose-dependently reduced the memory reconsolidation of fear memories in C57 mice. Moreover, SL327 administration impaired memory reconsolidation also in ERK1 mutant mice. Altogether, these results clearly indicate a central role for ERK2 protein in memory reconsolidation processes in mice.

Symmetry and binding in visuo-spatial working memory.

Three experiments study the impact of symmetry on a sequential block tapping immediate memory task in human subjects. Experiment 1 shows an advantage from vertical symmetry over non-symmetrical sequences, while finding no effect of horizontal or diagonal symmetry. Experiment 2 tests the possible role of verbal labeling by means of a secondary task that prevents this by articulatory suppression. No evidence of verbalization was observed. A third study examines the effects of a concurrent executive load, finding an overall impairment, that did not differ between symmetrical and asymmetric patterns, suggesting that the effect of symmetry reflects automatic rather than executive processes. Implications for the episodic buffer component of working memory are discussed.

Essential role for TrkB receptors in hippocampus-mediated learning.

Brain-derived neurotrophic factor (BDNF) and its receptor TrkB regulate both short-term synaptic functions and long-term potentiation (LTP) of brain synapses, raising the possibility that BDNF/TrkB may be involved in cognitive functions. We have generated conditionally gene targeted mice in which the knockout of the trkB gene is restricted to the forebrain and occurs only during postnatal development. Adult mutant mice show increasingly impaired learning behavior or inappropriate coping responses when facing complex and/or stressful learning paradigms but succeed in simple passive avoidance learning. Homozygous mutants show impaired LTP at CA1 hippocampal synapses. Interestingly, heterozygotes show a partial but substantial reduction of LTP but appear behaviorally normal. Thus, CA1 LTP may need to be reduced below a certain threshold before behavioral defects become apparent.

What do comparative studies of inbred mice add to current investigations on the neural basis of spatial behaviors?

In the present article a number of comparative lesion studies in two inbred strains of mice (C57BL/6 and DBA/2) with different levels of radial maze performance are reviewed. The effects of lesions in several brain areas on maze learning were investigated, thus revealing strain differences in the neural circuitry subserving spatial cognition. Results showed that the hippocampus and parietal cortex appear to be involved in the control of radial maze learning in both C57 and DBA mice, although in a strain-dependent fashion. Lesions in other structures such as the medial frontal cortex and the amygdala only affected spatial learning in the C57 strain. Lastly, the results showed some improvement in radial maze performance in DBA mice with nucleus accumbens lesions. The data highlight the variability in the neural mechanisms subserving well-differentiated levels of spatial performance. The contribution of inbred mice to our general understanding of the neural basis of spatial cognition is discussed.

A role for the Ras signalling pathway in synaptic transmission and long-term memory.

Members of the Ras subfamily of small guanine-nucleotide-binding proteins are essential for controlling normal and malignant cell proliferation as well as cell differentiation. The neuronal-specific guanine-nucleotide-exchange factor, Ras-GRF/CDC25Mm, induces Ras signalling in response to Ca2+ influx and activation of G-protein-coupled receptors in vitro, suggesting that it plays a role in neurotransmission and plasticity in vivo. Here we report that mice lacking Ras-GRF are impaired in the process of memory consolidation, as revealed by emotional conditioning tasks that require the function of the amygdala; learning and short-term memory are intact. Electrophysiological measurements in the basolateral amygdala reveal that long-term plasticity is abnormal in mutant mice. In contrast, Ras-GRF mutants do not reveal major deficits in spatial learning tasks such as the Morris water maze, a test that requires hippocampal function. Consistent with apparently normal hippocampal functions, Ras-GRF mutants show normal NMDA (N-methyl-D-aspartate) receptor-dependent long-term potentiation in this structure. These results implicate Ras-GRF signalling via the Ras/MAP kinase pathway in synaptic events leading to formation of long-term memories.

The differences shown by C57BL/6 and DBA/2 inbred mice in detecting spatial novelty are subserved by a different hippocampal and parietal cortex interplay.

Inbred C57BL/6 (C57) and DBA/2 (DBA) mice with hippocampus, posterior parietal cortex or sham lesions were placed in an open-field containing five objects and their reactivity to the displacement (spatial novelty) or the substitution (object novelty) of some of these objects was examined. C57 mice reacted to spatial novelty by exploring more the displaced than the non-displaced objects while DBA mice did not show any consistent reaction. In the highly reactive C57 strain, the peak of exploratory responses directed towards the displaced objects was completely abolished by hippocampal and posterior parietal cortex lesions. In the non-reactive DBA strain, hippocampal lesions induced an aspecific decreased interest towards the two categories of objects while posterior parietal cortex lesions did not produce any behavioral modification. The high reactivity of C57 mice to spatial change appears to be subserved by the conjunctive participation of the hippocampus and the posterior parietal cortex. Conversely, the deficit shown by DBA mice in that situation seems to be related to: (i) a poorly functional hippocampus; and (ii) the non-involvement of the posterior parietal cortes. The present data suggest that the participation of the posterior parietal cortes to the detection of spatial novelty may depend on the degree of functionality of the hippocampus.

Radial maze performance and open-field behaviours in aged C57BL/6 mice: further evidence for preserved cognitive abilities during senescence.

C57BL/6 mice, aged 2 or 24 months, were tested in a radial maze and observed for an 8-min period, repeated on 3 consecutive days, in an open-field situation with a novel object. In the eight-arm maze, the number of unrepeated path choices made by old mice does not significantly increase with training, whereas it does in young mice. Older animals also take significantly longer to solve the task but the two age groups do not differ with respect to how many paths they run before making the first error or in the strategies used to solve the task. In the open-field situation, the two age groups differ with regard to grooming and rearing behaviour, while in the novelty situation, older animal show a higher level of locomotor activity, perform less freezing, and interact more with the novel object. Habituation curves for all parameters, except grooming in the open field, do not differ between the two groups, thus indicating that this form of nonassociative learning does not vary substantially with increasing age. Results are discussed in terms of preserved cognitive abilities during senescence in that strain.

Learning in inbred mice: strain-specific abilities across three radial maze problems.

Mice belonging to the C57BL/6, DBA/2 (DBA), and C3H/He (C3H) strains were compared in three different eight-arm radial maze tasks requiring various degrees of spatial and nonspatial information processing. The results show that, on the standard radial maze task, C57 performed better than DBA, which, in turn, performed better than C3H. Fewer differences in the four-baited arm task and no difference in the cued version task were found between C57 and DBA, while C3H still performed more poorly. The high performance shown by C57 mice in all problems seems to be related to their ability to build up maze-running patterns based upon an optimal proportion of 45 degrees angle turns, according to the demand of the situation. The cognitive and discriminative mechanisms involved in the solving of each task, the sensorial characteristics of the three strains, and the limits of an approach based upon neuroanatomical-behavioral correlations are discussed.

Mechanical deafferentation of basal forebrain-cortical pathways and neurotoxic lesions of the nucleus basalis magnocellularis: comparative effect on spatial learning and cortical acetylcholine release in vivo.

Rats were assigned to one of the following treatments: bilateral cut of basal forebrain-cortical fibers (DEAFF), ibotenic (IBO) or quisqualic (QUIS) acid lesions of the NBM and sham operations (SHAM). They were trained to perform a radial eight-arm maze task with all the paths or only four paths baited. Cortical cholinergic release measured by microdialysis in vivo and choline acetyltransferase activity were also assessed in the four lesion conditions. The results show that, in the full baited maze task, only the DEAFF group showed a severe spatial learning impairment. In the four-baited path task, the DEAFF group was still more impaired than the other groups but a performance deficit also emerged in rats with IBO lesions. Neurochemical data indicated that cortical choline acetyltransferase activity was reduced by 25% after IBO lesions, by 52% after DEAFF and by 46% after Quis lesions. However, cortical cholinergic release, which dropped in the same fashion after DEAFF or QUIS lesions, was unaffected by IBO lesions. Thus, in spite of the distinctive patterns of behaviour exhibited by the three lesioned groups, no correlation between cortical cholinergic deficiencies and spatial learning impairment was found. The similar behavioural effects produced by DEAFF and fornix sections suggests that, among the basal forebrain-cortical pathways, descending fibers projecting onto the septo-hippocampal system could exert a strong control on spatial learning performance.

Genotype-dependent involvement of limbic areas in spatial learning and postlesion recovery.

Male C57BL/6 (C57) and DBA/2 (DBA) mice with hippocampal, amygdaloid, or sham lesions were tested in a radial eight-arm maze 1 or 4 weeks after surgery. The results show that the effect of the lesions varied according to the performance level of the strain considered. In the high-learner C57 strain, the two lesions impaired acquisition at both postlesion intervals. Conversely, in the low-learner DBA strain, only hippocampal lesions impaired acquisition 1 week but not 4 weeks after lesioning. It is hypothesized that if more limbic areas are involved in controlling spatial learning in C57 mice, these structures could be processing distinct but complementary memory attributes, thus contributing to a high baseline performance. This, however, also entails an increased sensitivity of C57 performance to brain damage with reduced possibilities of long-term recovery.

Dose-dependent effect of GM1 ganglioside during development on inhibitory avoidance behaviour in mice: influence of the period of administration.

Groups of C57BL/6 mice were injected intraperitoneally with GM1 monosialoganglioside at different ages during development and subsequently tested for the retention of an inhibitory avoidance task 24 h after training. Results show improvements in inhibitory avoidance retention according to the age of the animals, the doses of GM1 used and the length of treatment. The effective doses ranged from 20 mg/kg for all age groups after 7 days treatment to 280 mg/kg for 6- and 7-week old animals after pre-trial treatment. Six- and 7-week-old mice are more sensitive to GM1 treatment than 5-week-old animals and, with decreasing lengths of treatment, increasing doses of GM1 are needed to improve the performance of the animals. These findings show that short treatment durations can be effective in improving inhibitory avoidance retention as long as the doses of GM1 administered are increased and that animals are more sensitive to the treatment when they are 6 or 7 weeks of age than when they are 5 weeks old.

Modifications of open field and novelty behaviours by hippocampal and amygdaloid lesions in two inbred strains of mice: Lack of strain × lesion interactions.

C57BL/6 (C57) and DBA/2 (DBA) mice with dorsal hippocampus, central amygdala and sham lesions were observed for an eight-minute period, repeated on three consecutive days, in an open field situation with a novel object. Strain-dependent differences emerged when comparing sham lesioned mice; higher rearing and grooming scores as well as more defecation boli were found in DBA. In both strains, hippocampal lesions enhanced locomotor activity in the open field while amygdaloid lesions increased the number of contacts with the novel object. No significant lesion × strain interaction was recorded for any of the behaviours analyzed in the open field of the novelty situation. Significant strain × day, lesion × day and strain × lesion × day interactions were found thus indicating that previously described strain or lesion main effects were more pronounced on a particular day of testing. Despite the behavioural differences shown by sham lesioned C57 and DBA mice, the effects of the lesions differ in intensity but generally go in the same direction whatever the strain considered. Thus, while a strain-dependent selective involvement of subcortical areas in associative spatial learning has been previously reported, this does not seem to be the case for non-associative forms of learning.

Spatial learning in two inbred strains of mice: genotype-dependent effect of amygdaloid and hippocampal lesions.

Spatial learning performance and maze-running strategies were estimated in two inbred strains of mice, C57BL/6 and DBA/2, submitted to an 8-arm radial maze task. Subsequently the genotype-dependent effect of hippocampus and amygdala on the mastering of this task was examined as a function of the different acquisition model provided by each strain. The results firstly show that unoperated C57BL/6 mice reach a higher level of performance and develop a stronger preference for adjacent arms - 45 degrees angle - turns than unoperated DBA/2 mice. In the high learner C57BL/6 strain, both hippocampal and amygdaloid lesions impair performance and modify maze-running strategies. With practice, however, the difference between amygdala-lesioned mice and controls disappears while that between hippocampus-lesioned mice and controls persists. Conversely, in the low learner DBA/2 strain, hippocampal lesions have a negative effect on a single parameter of performance, while amygdaloid lesions only affect maze-running strategies. Taken together, these results confirm the specific control exerted by the hippocampus on spatial learning. Moreover, they suggest that the amygdala can parallel the role of the hippocampus as far as the baseline level of performance of the strain considered is high.

Open field behaviours and spatial learning performance in C57BL/6 mice: early stage effects of chronic GM1 ganglioside administration.

One month intact C57BL/6 mice were treated with GM1 ganglioside for 3 consecutive weeks. At 2 months of age, treated and control mice were observed in the open-field situation and tested for spatial learning in a radial eight-arm maze. The results showed that, in the open-field, treated mice displayed less freezing but more rearing behavior than control animals. In the radial maze, GM1-treated mice made more correct path choices before the first error within each trial than control mice. However, this improvement was limited to the first stage of training. These results suggest an early stimulating effect of the GM1 ganglioside treatment which could facilitate adaptive reactions to new situations.

Early experience and reinforcer quality in delayed flavour-food learning in the rat.

The ability of hungry rats to associate flavours with the consequences of ingesting glucose solutions was studied in three experiments. Experiment 1 used a procedure in which on some days one flavour, e.g. cinnamon, was presented and followed after 20 min by 20% glucose, while on other days a second flavour, e.g. wintergreen, was presented, but not followed by any event. During this training, subjects who received quinine-tainted glucose increased their consumption of the predictive flavour relative to groups given no quinine, but quinine tainting did not affect conditioned preference for the predictive flavour in choice tests. With the aim of discovering whether prior experience of a variety of foods improves ability to learn new flavour-calorie associations, Experiment 2 and used a similar procedure to compare subjects raised on a varied diet ("supermarket" rats) with controls previously given only chow. Contrary to expectation, the supermarket rats showed some impairment both on this delay task and, in Experiment 3, on one using a "mixtures" procedure. This involved presenting a mixture of one cue flavour with glucose-quinine on some days and a mixture of a second cue with an equally palatable saccharin solution on other days. Acquisition was particularly rapid in control subjects, reaching asymptote after only two flavour-glucose pairings. It was concluded that neither a decrease in palatability, as in Experiment 1, nor prior experience with a range of foods, as in Experiments 2 and 3, improve a rat's ability to associate a new flavour with calories.