Introduction
The phenomenon of memory is alike to layman, psychologist, and biochemist, but its intangible and nebulous nature makes it an entity that is difficult to describe in precise biochemical terms. Moreover, it is difficult to quantitate, and thereby experiments to lay the basis for a biological model of memory have been regarded by some as controversial. By and large, there is a general agreement that molecular events are involved in the storage of information in the nervous system. Experiments involving the use of drugs affecting learning and macromolecular changes with learning experiences (training etc.) have lent considerable support to this statement.
What the actual macromolecules which “code” for learning memory are still perhaps debatable. Most earlier evidence points to these macromolecules being either RNA or protein. There are two fundamental problems in memory expressed in biochemical parameters:
- are there any chemical changes in the brain that can be correlated with acquisition and storage of information;
- are these changes information-specific, i.e. do they represent an actual coding of neural information?
It is about to differentiate different kinds of memory, one which is short-term and the other is long-term. The processes responsible for short-term and long-term memory differ. This is suggested by the following experiments. If a lesion in the hippocampus region of an animal’s brain is introduced, the animal loses its memory for recently-acquired skills, sometimes as quickly as 1 h after the task has been learned. Memory for events before the lesion is normal.
A blow to the head may produce partial or complete amnesia for past events immediately preceding the trauma. This indicates that memory ordinarily forms after experience and consolidation of memory is nullified by the trauma. Similarly, animals are given electro-convulsive shock (ECS), soon after training for any task, showing impaired ability to perform. Subsequently, on the other hand, ECS given sometimes after training does not affect performance. The formation of long-term memory can be influenced by agents that have little effect once the memory has been consolidated. Those that inhibit long-term memory include general anesthesia (ether and pentobarbital) convulsions (induced either electrically or chemically, e.g. by Metrazol), hypoxia, RNase treatment, and many others.
The capability for change associated with learning is termed plasticity. This is an ability within the nervous system to change the behavioral pattern in response to changes in external and internal environments. This involves the synthesis of different proteins or spouting of new dendrites and strengthening of synaptic connections around neurons. The parts of the brain associated with memory in mammals include the association cortex of frontal, parietal, occipital, and temporal lobes; components of limbic system viz. hippocampus and amygdala, and part of the diencephalon.
Anatomical changes occur in neurons when they are stimulated like prolonged intense activity results in an increase in several presynaptic terminals and enlargement of synaptic end bulbs in postsynaptic neurons with growing age as they are used for a longer period. Just the opposite changes are seen when some neurons are underused, e.g. animals that have lost eyesight exhibit thinner visual area of the cerebral cortex. Believed or underlie some aspects of memory is a phenomenon called long-term potentiation (LTP), in which transmission at some synapses within the hippocampus is enhanced (potentiated) for a long period after a brief period of high-frequency stimulation.
Mammalian LTP in the amygdala in response to Pavlovian fear conditions
Maren (1999) compiled a set of studies on brain anatomy, in particular, the amygdala and its fear-conditioning neural circuit connectivity to other parts in the coronal section of the rat brain. The author revealed two prominent positions of the amygdala: the basolateral complex (BLA) comprising lateral, basolateral, and biomedical nuclei, and the central nucleus (CEA), comprising medial and lateral components.
Using electrodes, conditional stimulus like an alternating impulse of strong tone that elicits auditory fear conditions and aversive unconditional stimulus such as electric footshock was directed to the basolateral nucleus of the amygdala and electrical field potentials were recorded (Rogan, Stäubli & Ledoux, 1997). LTP is believed to mediate acoustic fear-associated memory. According to Maren, BLA is the focal region for convergence of several fear-related stimuli such as auditory, visual, and somatic (shock), etc, received from synaptic inputs of diverse sensory structures. The lesion in BLA would impair Pavlovian fear conditioning.
Intra-amygdaloid circulatory conveys this to CEA, where divergent projections to hippocampus, brainstem, and medulla mediate fear combating defensive responses. The lesion in CEA and lateral hippocampus also remarkably affects the acquisition of fear conditions. In this study, though BLA has been projected as being not only the main center for formation and storage of fear memories through LTP but also a center developing behavioral and associative firing to other parts. It was further revealed that LTP is associated with the triggering of N-methyl-D-aspartate (NMDA) receptors in BLA. To ascertain, NMDA receptor (NMDA-R) antagonist (6)-2-amino-5-phosphonovaleric acid (APV) was injected which completely reversed the fear adverting responses.
Surprisingly, some of the fear countering responses diminished while the others remained unchanged. The author had no explanation for this unexpected result of the partial APV perfusion effect. This review was well-timed to comprehend several scattered reports of divergent nature on the related subjects. The author has justified and narrowed down the conclusions to the amygdala as being the center of fear memory conditioning.
It would have been better to narrow down the electrode spacing between input and output brain centers within the amygdala itself rather than between hippocampus and amygdala. The interpretation that a single neuron in the BLA responding to auditory, visual, and somatic stimuli was not experimentally validated. It was concluded that lesions in BLA, CEA, and hippocampus regions can avert the fear memory responsive process was difficult to justify, given the fact that only BLA was made responsible for NMDA-R-mediated LTP acquisition. Most of the studies were based on in vivo electrophysiological tests and lesions in the amygdala may have bearing on the hippocampus and vice versa. Intracranial antagonist infusion may also not hamper just one neuronal function (APV on NMDA receptor blockade) but may also affect others.
Why APV perfusion reverted only some fear countering responses and not the others was not answered. It could well be that the hippocampus may participate in amygdala-independent LTP and memory storage, an issue not considered.
A more recent review published by Lynch (2004) gave an updated account of neuronal physiological and molecular changes attributed to LTP. Broadly, the author categorized memory in “explicit” means related to consciousness such as information of place, people things, etc., and “implicit” related to the nonconscious recall of tasks like motor skills. The author listed several brain areas, predominantly the hippocampus, taking part in the consolidation of the above forms of memory/learning. Amygdala was found responsible for emotional memory, particularly in response to Pavlovian fear conditioning, but its role in other emotional memory storage was suggested to be mediated through other brain areas.
Another interesting finding was that, while NMDA-R antagonists did disrupt behavioral fear conditions, its retrieval required activation of NMDA-R, including proteins synthesis and activation of transcription factor cAMP-responsive element-binding protein (CREB), and hence the withdrawal of these antagonists was not sufficient to bring back the fear-responsive conditions. Both hippocampus and entorhinal cortex receive direct projections from BLA and therefore it is imminent that the amygdala would exert a modulatory role on the hippocampus. The author has shown that BLA activation enhances LTP even in the distantly placed dentate gyrus. These findings explained the anomaly in conclusions arising in the paper of Maren (1999).
The strength of this paper is an elaborate explanation of biochemical and molecular events that lead to LTP. LTP in the hippocampus and other regions for different kinds of memory/learning consolidation exhibit characteristic changes in the overall biochemical setup of signal transduction in the pre-and postsynaptic cells. A common event as pointed out by the author is a postsynaptic increase in calcium from the opening of calcium channels, mediated through NMDA-R activation.
In a cascade of action calcium/calmodulin kinase II (CaMKII) is induced enabling postsynaptic calcium entry followed by CREB transcriptional factor hyperphosphorylation and increased expression of several genes like c–fos. The fundamental issue raised in this paper was what is the role of CaMKII in learning/memory? Here authors have demonstrated that glutamate mobilizing α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor (AMPA-R) activation by CaMKII was necessary for establishing LTP.
A new “silent synapse” theory was put forth, according to which even though NMDA-R is active in some synapses, lack of AMPA-R did not initiate LTP. AMPA-R building was manifold increased upon a training session though no change was seen in NMDA-R in these synapses, and consequently such synapses switched on for LTP. Enhances signal molecules CaMKII and AMPA proceed towards dendrites and play a role in neurotransmitter release. Moreover, the author claimed that CaMKII is associated with actin and increased postsynaptic density by cytoskeletal changes.
There is also the other signaling pathways viz. cAMP/Protein Kinase A (PKA), which associate pre-or postsynaptically with delayed sustenance of LTP. All these signal molecules converge to activate Mitogen-Activated Protein Kinase/Extracellular signal Regulated Kinase (MAPK/ERK). ERK activation modulates, either by self or by activating CREB protein, gene expression, and translation of further downstream signaling proteins, cytoskeleton proteins, synaptosomal proteins, and nuclear regulatory proteins. An overall consequence of this cascade of action is manifest at enhanced overall mRNA and protein syntheses.
The article was well-timed to sort out a lot of confusion from the previous work on the exact position of fear-related LTP and explained why only one brain part, the amygdala, can not solely consolidate LTP response. The article did suggest that creating a lesion in one part of the brain may affect the well-integrated adjacent centers of the brain, but it has not attempted to solve this issue in the in vivo research. Further, in the “silent synapse” theory no explanation was given as to why APMA-R is diminished only in selected neurons while NMDA-R is still plentiful. It could well be that in “silent synapse” cAMP/PKA-ERK is still functional and thereby they are also participating in LTP.
A clear rationale of NMDA-R and cAMP/MAPK/ERK-mediated signal transduction in postsynaptic neurons is a matter of extended work from this report. While the memory-selective brain parts are distinguished, it is apparent that overlapping reasons may superimpose the memory paradigm, and this issue also could have been addressed.
In continuation, work by Huang & Kandel (2007) projected a role of late amygdala LTP by cAMP/PKA pathway. In this paper duel effect of serotonin (5-hydroxytryptamine; HT), a neurotransmitter modulator, in the basolateral area of the amygdala was researched. In the normal course, this drug is known to affect neurotransmitters linked to behavior and anxiety. It induced a short-term depression of synaptic transmission, though at a longer duration the effect was just reversed.
There are several HT receptors in the amygdala and some of them also interact with cAMP/PKA mediated signaling. RS23597, an HT antagonist, suppressed some of the receptors, yet could not abolish the short-term acceleration of transmission and only brought down the long-term effect of HT. Authors envisaged two explanations: either blockade of HT receptors increased external HT concentration, and both HT and antagonist had a synergistic effect on HT receptors, or alternate HT receptors, not blocked by the antagonist, were still capable to take HT in. Once inside the cells, HT was reported to activate PKA/MAPK activation of CREB and induce long-term LTP facilitations. Authors have ruled out the role of the NMDA-R mediated pathway in HT’s action. Gene transcription inhibitors also blocked the late- rather than the early events of HT.
The question was asked whether the HT effect depends on the regulation of actin cytoskeleton expression and observations in this paper revealed that indeed this was the case. Broadly, the paper projected that the cAMP/PKA pathway functions independently of the NMDA-R pathway in the amygdala, and perhaps this also contributes to ERK-mediated signal transduction for LTP, an assumption already put forth in a previous paper.
However, the authors could not convincingly explain why the short- and long-term effects of HT on cAMP/PKA signaling differed. One possible reason could be that the short-term promotional action on LTP could be due to an effect on some alternate signaling pathway not deciphered till then. Another issue of concern was that, if transcriptional inhibitors suppressed the HT‘s downregulation of the cAMP/PKA/MAPK pathway then why the same explanation does not hold for NMDA-R mediated signaling.
Auditory information reaches the thalamus and from there transmitted to the lateral nucleus of the amygdala. Apergis-Schoute, Debiec, Doyere, LeDoux & Schafe (2005) carried out an investigation on the nature of signal transduction at the input level, i.e. in the MGm/PIN region of the thalamus. The authors have claimed that like the lateral amygdala (LA), ERK/MAPK is the primary pre-synaptic signal molecule in this region.
This means while postsynaptic changes are mainly on account of NMDA-R mediated trigger and subsequent cAMP/PKA/MAPK mediated sustenance of LTP, at presynaptic level (input region) only the later signaling mechanism is operational. And most likely calcium signaling that we see in post-synaptic cells may not exist here. Auditory inputs were given to the thalamus and the fear perception-based LTP was recorded by pharmacological means using an ERK antagonist, 1,4-diamino-2,3-dicyano-1,4-bis(o-aminophenylmercapto) butadiene (U1026). Collectively, the authors proposed that inhibition of ERK/MAPK inhibition just before training impaired the memory consolidation of auditory fear in the thalamus region. Further, inhibitors of mRNA and protein syntheses also exerted the same effect on long-term memory consolidation.
The authors further concluded that aside from LA which has been considered as a sole region for both input and output signaling for Pavlovian auditory fear memory consolidation, regions in the auditory thymus are also involved in LTP. The electrophysiological analysis done in this paper was quite standard but pharmacological testing within a small region of the brain had its problems. Let us consider the report of Maren (1999) which suggests that the input somatosensory circuits originate from LA and re-enter the thalamus. Hence in tissue sections, these postsynaptic neurons may be by mistake considered as the pre-synaptic ones. The antagonist’s action could well be on the output circuits and this aspect was not discussed to its best.
Most of the pharmacological experiments stated above were carried out in tissue slices (sections) and under in vitro conditions. These conditions are not suitable for immunological procedures. A different approach was to locate the signal molecules within the tissue as it appears under in vivo conditions. Using this approach, ERK/MAPK pathway for Pavlovian fear conditioning was stated to be anatomically restricted to regions of the lateral amygdala (LA) and blockade of ERK/MAPK activation had led to impairment of the fear sensation (Schafe, Swank, Rodrigues, Debiec & Doyère, 2008). In vivo analysis was carried out by the authors using Western blotting and immunohistochemical techniques.
High-frequency electrical stimulation (HFS) with the frequency adjusted to achieve LTP was introduced through electrodes implanted in the auditory thalamus. HFS induced the same pattern of ERK distribution in LA, particularly towards the ventral region, as was detected upon Pavlovian fear stimulation. Low-frequency stimulations for a prolonged period failed to induce any ERK staining and consequently, no LTP was imposed. One interesting finding by the authors in immunohistochemical analysis and different from slice (section) experiment was that while fear conditions and HFS induced identical biochemical changes in the amygdala, immunologically they are not synonymous processes.
The authors have not provided any explanation. By this method, it was possible to demarcate the LTP stimulated by conditions that otherwise impose identical biochemical events. However, there are also problems with immunological procedures; firstly tissue impregnation of the antibodies would be a vital limiting factor discriminating by the depth of the tissue, and secondly, between impulse and tissue processing the delay would cause diffusion of the signal proteins disturbing the immunological staining.
In order further to elucidate whether pathways other than those discussed above (NMDA-R or cAMP/ERK/MAPK consolidating the LTP at pre-and postsynaptic neurons), Oren, Nissen, Kullmann, Somogyi & Lamsa (2009) inhibited the AMPA-R mediated calcium intake in hippocampus CA1 area by using an antagonist, 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo[f]quinoxaline-2,3-dione (NBQX), which is supposed to trigger LTP initiated by NMDA-R downward signaling also known as “Hebbian LTP” in this context. Many interneurons in this area exhibited AMPA and calcium intake-independent LTP in this region of the brain.
Even when NMDA-R was blocked using APV, the LTP still prevailed. The entire studies in which several AMPA antagonists were used revealed that kainate receptors that are exclusively present in the hippocampus region get the induction under conditions when AMPA-R blockade takes place. Glutamate-dependent calcium uptake is possible using these alternate receptors somewhat like the way AMPA-R does. Further, these intersynaptic kainate receptors are in abundance in the hippocampus region. All this evidence points out the NMDA-R/APMA-R-independent prevalence of glutamate activated and calcium intake-induced LTP.
This work nonetheless also indicates that despite the similarity in somatodendritic shapes of these neurons, there exists vast diversity within interneurons falling in different categories. What are the categories and how can that be ascertained was not explained in the paper. The authors have used this explanation independent of cAMP/PKA/MAPK mediated downward signaling and quite likely this pathway may also be operative in conjunction with or in isolation to the kainate receptor-mediated pathway in the hippocampus CA1 area. The basic question is whether this signaling pathway is responsible for visuospatial memory consolidation for which the hippocampus is the principal site?
Spatial memory consolidation in food-storing birds has an analogy with LTP
Birds are the other species with extraordinary capabilities of memorizing many events like food caching and retrieving and “birdsong” training. It was assumed that the memory acquisition and storage in birds are similar to the mammals. Accordingly, some physiological and anatomical analogy between the memory-related centers of birds’ and mammals’ brains about LTP was established. Food-sorting birds remember the scattered locations of food caches and pick them.
In a pioneering review by Clayton (1998), it was revealed that food storing birds exhibit relatively larger hippocampus sizes and more neurons compared to those birds that do not, albeit the relative brain size remains unchanged in the two groups. This explanation may be exaggerated given the fact that the two groups of birds have different nutritional requirements and habitats which can affect hippocampus size. An analogy was drawn by the authors between mammalian and avian hippocampus.
While homologous regions were assumed to have similar memory-related functions like long-term spatial memory consolidation and immunological cross-reactivity with neuronal proteins, in reality, there was no experiment done to conclusively prove this assumption. Other than spatial cues, the food-storing birds also remember information like caching and retrieving the food. In an experiment, birds were given one category of food for a long period to adjust them to the food type.
Thereafter, another food type was supplied for a brief period. After a retrieval time without food, both food types were simultaneously presented. The birds displayed a tendency to sort only the first acclimatized food that they fed on initially. This experiment had a disadvantage of olfactory and visual cues which were operative while recognizing which food was to be chosen, and selecting a food need not necessarily be a memory-related attribute.
Shiflett, Tomaszycki, Rankin & DeVoogd (2004) executed successfully another investigation in which birds were trained in a few trials in indoor conditions to locate the holes in which food of choice was kept.
After a short or a long pause, they were again released in the chamber without having any visual or olfactory aid to determine how many times they pick the food from the same spots. As this is a spatial memory believed to be stored in the hippocampus, in the next set of experiments NMDA-R antagonists like D-2 amino-5-phosphonopentanoic acid (AP5) were infused precisely above the hippocampus formation region at three occasions: before the acquisition of memory, just following the training period and after the bird is accustomed to the training in several sessions of food sorting. When the blockade was imposed before training, long-term memory to seek the right food spots was impaired but short-term remembrance prevailed.
Post-training infusion had practically no effect on long-term memory. An infusion after a prolonged duration had a marked effect on the long-term memory of the second session of sorting food from different locations than what was kept in the first training session. In mammalian hippocampus formation, NMDA-R-regulated long-term spatial memory is assumed to be stored and the same mechanism and neural network seem to be operational in the experimental birds as well. However, these assumptions were not experimentally confirmed.
Clayton’s observation that hippocampal neural density is linked to food-storing habits (see above) was revisited by Hoshooley & Sherry (2007). Authors have compared chickadees and house sparrows in terms of their hippocampal neuron recruitment, testBromo bromo-dU immunohistochemistry, and expression of neuronal nuclei specific protein (NeuN), both signifying new neuron recruitment at hippocampus.
For experimentation in captive, ty the birds were captured in two seasons in a year, fall and spring. The authors claimed that in the case of chickadees and not house sparrows, neuron mass in the hyperpalliumapicale, a component of the hippocampus, increased food storing habits in the field. There were also seasonal changes in the hippocampal neural density, which corresponded to the extent of food caching habits at that particular time.
Chickadees caught in spring had more neurons than those captured in fall. The food-sorting activity in chickadees peaks in spring. In this study too, the impact of various environmental and nutritional conditions prevailing in the two seasons are expected to affect the conclusions as to whether the neural development is a memory-related attribute or it is simply a developmental process dependent on other factors. Appropriate controls such as comparison with permanently captive birds in the two seasons could have been relevant to include as an additional group.
Indeed the differences were found within chickadees populations wandering in the wilderness and those brought to captivity and then acclimatized to sort some desired food type. To resolve the likely effect of captivity vis-à-vis wilderness, LaDage, Roth, Fox, & Pravosudov (2009) executed an interesting set of experiments as follows: the wild-caught birds were grouped in two batches, experienced, in which case the birds navigated freely and sorted food and were also trained for sorting the food in a closed chamber, and the deprived birds, which remained in captivity and were served food before allowing to enter a closed chamber in which food pans were placed but there was no food.
The experienced and deprived groups were compared with the wild birds caught for the experiments and not kept in captivity at all. Authors experienced that wild chickadees had larger hippocampal volume than both the captive experienced and deprived individuals. Moreover, there was not much difference in hippocampal neural density and telencephalon volume between the two captive groups or between these groups and the wild group. Interestingly, neither the experienced nor the deprived captive individuals exhibited any significant difference in hippocampus and telencephalon volume or neutral density.
The authors did not produce any convincing explanation for this unexpected result. It is quite likely that just having kept in captivity brings about a kind of stress imposed on the normal routine of chickadees, and this condition is sufficient a factor to reduce hippocampus volume, even though the neuron density and telencephalon volume remained unchanged. This could well be that decrease in hippocampus volume had nothing to do with memory-related exercises as perceived by the authors. The time taken for chickadees to be trained perhaps was not adequate to exert any learning associated changes in hippocampus size and this experiment needed further validation.
Conclusions
In this critical review, long-term memory acquisition and storage in mammals and birds have been compared. Pavlovian fear was chosen as emotional memory for rodents and food sorting as spatial memory for food caching birds. Briefly, the LTP and fear perception and response were focused mainly in the amygdale, but some centers in the hippocampus region were also found to be equally responsible. The pre-and postsynaptic signal transduction, induced by NMDA-R/AMPA-R mediated pathways and/or cAMP/PKA pathway induced CREB transcriptional factor to up-regulate several genes related to neuronal cytoskeleton development.
At sensory input centers like the thalamus, the ERK/MAPK downstream signaling was responsible for sending the fear messages. In the hippocampus, the kainate receptor-mediated pathway seems to be additionally operational for establishing LTP. Much less information is available in food storing birds, though NMDA-R mediated signal transduction for visual/spatial cues is expected to be functional in long-term memory consolidation of food locations whether this can be equated with mammalian LTP is not certain. Further research would be necessary to make a homology between mammalian and avian brain memory centers in terms of their anatomy and physiology.
References
Apergis-Schoute, A.M., Debiec, J., Doyere, V., LeDoux, J.E., & Schafe, G.E. (2005). Auditory Fear Conditioning and Long-Term Potentiation in the Lateral Amygdala Require ERK/MAP Kinase Signaling in the Auditory Thalamus: A Role for Presynaptic Plasticity in the Fear System. Journal of Neuroscience, 25 (24), 5730 –5739.
Clayton, N.S. (1998). Memory and the hippocampus in food-storing birds: A comparative approach. Neuropharmacology, 37 (4-5), 441-452.
Hoshooley, J.S. & Sherry, D.F. (2007). Greater Hippocampal Neuronal Recruitment in Food-Storing Than in Non-Food-Storing Birds. Developmental Neurobiology, 67, 406–414.
Huang, Y-Y., & Kandel, E.R. (2007). 5-Hydroxytryptamine Induces a Protein Kinase A/Mitogen-Activated Protein Kinase-Mediated and Macromolecular Synthesis Dependent Late Phase of Long-Term Potentiation in the Amygdala. The Journal of Neuroscience, 27 (12), 3111–3119.
LaDage, L.D., Roth, T.C. Fox, R.A. & Pravosudov, V.V. (2009). Effects of Captivity and Memory-Based Experiences on the Hippocampus in Mountain Chickadees. Behavioral Neuroscience, 123 (2), 284–291.
Lynch, M.A. (2004). Long-Term Potentiation and Memory. Physiological Reviews, 84, 87–136.
Maren, S. (1999). Long-term potentiation in the amygdala: a mechanism for emotional learning and memory. Trends in Neuroscience, 22, 561–567.
Oren, I., Nissen, W., Kullmann, D.M., Somogyi, P., & Lamsa, K.P. (2009). Role of Ionotropic Glutamate Receptors in Long-Term Potentiation in Rat Hippocampal CA1 Oriens-Lacunosum Molecular Interneurons. The Journal of Neuroscience, 29 (4), 939 –950.
Rogan, M.T., Stäubli, U.V., & Ledoux, J.E. (1997). Fear conditioning induces associative long-term potentiation in the amygdala. Nature, 390, 604-607.
Schafe, G.E., Swank, M.W., Rodrigues, S.M., Debiec, J., & Doyère, V. (2008). Phosphorylation of ERK/MAP kinase is required for long-term potentiation in anatomically restricted regions of the lateral amygdala in vivo. Learning & Memory, 15, 55–62.
Shiflett, M.W., Tomaszycki, M.L., Rankin, A.Z., & DeVoogd, T.J. (2004). Long-Term Memory for Spatial Locations in a Food-Storing Bird (Poecile atricapilla) Requires Activation of NMDA Receptors in the Hippocampal Formation during Learning. Behavioral Neuroscience, 118 (1), 121–130.