Cascade models of synaptically stored memories

Storing memories of ongoing, everyday experiences requires a high degree of plasticity, but retaining these memories demands protection against changes induced by further activity and experience. Models in which memories are stored through switch-like transitions in synaptic efficacy are good at storing but bad at retaining memories if these transitions are likely, and they are poor at storage but good at retention if they are unlikely. We construct and study a model in which each synapse has a cascade of states with different levels of plasticity, connected by metaplastic transitions. This cascade model combines high levels of memory storage with long retention times and significantly outperforms alternative models. As a result, we suggest that memory storage requires synapses with multiple states exhibiting dynamics over a wide range of timescales, and we suggest experimental tests of this hypothesis.     

Efficient Partitioning of Memory Systems and Its Importance for Memory Consolidation

Long-term memories are likely stored in the synaptic weights of neuronal networks in the brain. The storage capacity of such networks depends on the degree of plasticity of their synapses. Highly plastic synapses allow for strong memories, but these are quickly overwritten. On the other hand, less labile synapses result in long-lasting but weak memories. Here we show that the trade-off between memory strength and memory lifetime can be overcome by partitioning the memory system into multiple regions characterized by different levels of synaptic plasticity and transferring memory information from the more to less plastic region. The improvement in memory lifetime is proportional to the number of memory regions, and the initial memory strength can be orders of magnitude larger than in a non-partitioned memory system. This model provides a fundamental computational reason for memory consolidation processes at the systems level.     

Molecular constraints on synaptic tagging and maintenance of long-term potentiation: a predictive model

Protein synthesis-dependent, late long-term potentiation (LTP) and depression (LTD) at glutamatergic hippocampal synapses are well characterized examples of long-term synaptic plasticity. Persistent increased activity of protein kinase M ζ (PKMζ) is thought essential for maintaining LTP. Additional spatial and temporal features that govern LTP and LTD induction are embodied in the synaptic tagging and capture (STC) and cross capture hypotheses. Only synapses that have been "tagged" by a stimulus sufficient for LTP and learning can "capture" PKMζ. A model was developed to simulate the dynamics of key molecules required for LTP and LTD. The model concisely represents relationships between tagging, capture, LTD, and LTP maintenance. The model successfully simulated LTP maintained by persistent synaptic PKMζ, STC, LTD, and cross capture, and makes testable predictions concerning the dynamics of PKMζ. The maintenance of LTP, and consequently of at least some forms of long-term memory, is predicted to require continual positive feedback in which PKMζ enhances its own synthesis only at potentiated synapses. This feedback underlies bistability in the activity of PKMζ. Second, cross capture requires the induction of LTD to induce dendritic PKMζ synthesis, although this may require tagging of a nearby synapse for LTP. The model also simulates the effects of PKMζ inhibition, and makes additional predictions for the dynamics of CaM kinases. Experiments testing the above predictions would significantly advance the understanding of memory maintenance.     

Memory Maintenance in Synapses with Calcium-Based Plasticity in the Presence of Background Activity

Most models of learning and memory assume that memories are maintained in neuronal circuits by persistent synaptic modifications induced by specific patterns of pre- and postsynaptic activity. For this scenario to be viable, synaptic modifications must survive the ubiquitous ongoing activity present in neural circuits in vivo. In this paper, we investigate the time scales of memory maintenance in a calcium-based synaptic plasticity model that has been shown recently to be able to fit different experimental data-sets from hippocampal and neocortical preparations. We find that in the presence of background activity on the order of 1 Hz parameters that fit pyramidal layer 5 neocortical data lead to a very fast decay of synaptic efficacy, with time scales of minutes. We then identify two ways in which this memory time scale can be extended: (i) the extracellular calcium concentration in the experiments used to fit the model are larger than estimated concentrations in vivo. Lowering extracellular calcium concentration to in vivo levels leads to an increase in memory time scales of several orders of magnitude; (ii) adding a bistability mechanism so that each synapse has two stable states at sufficiently low background activity leads to a further boost in memory time scale, since memory decay is no longer described by an exponential decay from an initial state, but by an escape from a potential well. We argue that both features are expected to be present in synapses in vivo. These results are obtained first in a single synapse connecting two independent Poisson neurons, and then in simulations of a large network of excitatory and inhibitory integrate-and-fire neurons. Our results emphasise the need for studying plasticity at physiological extracellular calcium concentration, and highlight the role of synaptic bi- or multistability in the stability of learned synaptic structures.     

Mechanisms of long-term memory

Long-term potentiation (LTP) of the hippocampus provides an excellent model on which to build hypotheses for the laying down of memories in the cerebral cortex. After repetitive activation, the primary happening seems to be the increase in transmitter sensitivity brought about by the increased Ca2+ in the recipient neurons. There may be secondary presynaptic changes. The extreme duration of the LTP may require structural changes of the synaptic spines. The hippocampus plays an essential role in the laying down of cognitive memories, the pathway to the frontal lobe being via the MD thalamus. The thalamo-cortical fibres activate stellate cells whose axons make climbing fibre-like ramifications up the apical dendrites of the pyramidal cells. On the Marr hypothesis repetitive conjunction of synaptic activation by these climbing fibres with synaptic activation by horizontal fibres on the apical dendrites produces prolonged potentiation of the horizontal fibre synapses, which is the neural basis of memory. Presumably this is a consequence of the raised Ca2+ of the apical dendrites, acting as it does on the hippocampal LTP. It will be considered how this elemental unit for cerebral memory can be developed into the varieties of cerebral memories that have been located by study of the regional cerebral blood flow during their retrieval. These sites for memory are in the frontal lobe, usually in the superior prefrontal area.     

Temporal Sensitivity of Protein Kinase A Activation in Late-Phase Long Term Potentiation

Protein kinases play critical roles in learning and memory and in long term potentiation (LTP), a form of synaptic plasticity. The induction of late-phase LTP (L-LTP) in the CA1 region of the hippocampus requires several kinases, including CaMKII and PKA, which are activated by calcium-dependent signaling processes and other intracellular signaling pathways. The requirement for PKA is limited to L-LTP induced using spaced stimuli, but not massed stimuli. To investigate this temporal sensitivity of PKA, a computational biochemical model of L-LTP induction in CA1 pyramidal neurons was developed. The model describes the interactions of calcium and cAMP signaling pathways and is based on published biochemical measurements of two key synaptic signaling molecules, PKA and CaMKII. The model is stimulated using four 100 Hz tetani separated by 3 sec (massed) or 300 sec (spaced), identical to experimental L-LTP induction protocols. Simulations show that spaced stimulation activates more PKA than massed stimulation, and makes a key experimental prediction, that L-LTP is PKA-dependent for intervals larger than 60 sec. Experimental measurements of L-LTP demonstrate that intervals of 80 sec, but not 40 sec, produce PKA-dependent L-LTP, thereby confirming the model prediction. Examination of CaMKII reveals that its temporal sensitivity is opposite that of PKA, suggesting that PKA is required after spaced stimulation to compensate for a decrease in CaMKII. In addition to explaining the temporal sensitivity of PKA, these simulations suggest that the use of several kinases for memory storage allows each to respond optimally to different temporal patterns.     

Dynamics of memory representations in networks with novelty-facilitated synaptic plasticity

The ability to associate some stimuli while differentiating between others is an essential characteristic of biological memory. Theoretical models identify memories as attractors of neural network activity, with learning based on Hebb-like synaptic modifications. Our analysis shows that when network inputs are correlated, this mechanism results in overassociations, even up to several memories "merging" into one. To counteract this tendency, we introduce a learning mechanism that involves novelty-facilitated modifications, accentuating synaptic changes proportionally to the difference between network input and stored memories. This mechanism introduces a dependency of synaptic modifications on previously acquired memories, enabling a wide spectrum of memory associations, ranging from absolute discrimination to complete merging. The model predicts that memory representations should be sensitive to learning order, consistent with recent psychophysical studies of face recognition and electrophysiological experiments on hippocampal place cells. The proposed mechanism is compatible with a recent biological model of novelty-facilitated learning in hippocampal circuitry.     

A fresh look at the role of CaMKII in hippocampal synaptic plasticity and memory

Advances in molecular, genetic, and cell biological techniques have allowed neuroscientists to delve into the cellular machinery of learning and memory. The calcium and calmodulin-dependent kinase type II (CaMKII) is one of the best candidates for being a molecular component of the learning and memory machinery in the mammalian brain. It is present in abundance at synapses and its enzymatic properties and responsiveness to intracellular Ca(2+) fit a model whereby Ca(2+) currents activate the kinase and lead to changes in synaptic efficacy. Indeed, such plastic properties of synapses are thought to be important for memory formation. Genetic analysis of the alpha isoform of CaMKII in mice support the hypothesis that CaMKII signaling is required to initiate the formation of new spatial memories in the hippocampus. CaMKII is also required for the correct induction of long-term potentiation (LTP) in the hippocampus, consistent with the widely held belief that LTP is a mechanism for learning and memory. Recent cell biological, genetic, and physiological analyses suggest that one of the cellular explanations for LTP and CaMKII function might be the trafficking of AMPA-type receptors to synapses in response to neural activity.     

Structural Synaptic Plasticity Has High Memory Capacity and Can Explain Graded Amnesia,Catastrophic Forgetting,and the Spacing Effect

Although already William James and, more explicitly, Donald Hebb''s theory of cell assemblies have suggested that activity-dependent rewiring of neuronal networks is the substrate of learning and memory, over the last six decades most theoretical work on memory has focused on plasticity of existing synapses in prewired networks. Research in the last decade has emphasized that structural modification of synaptic connectivity is common in the adult brain and tightly correlated with learning and memory. Here we present a parsimonious computational model for learning by structural plasticity. The basic modeling units are “potential synapses” defined as locations in the network where synapses can potentially grow to connect two neurons. This model generalizes well-known previous models for associative learning based on weight plasticity. Therefore, existing theory can be applied to analyze how many memories and how much information structural plasticity can store in a synapse. Surprisingly, we find that structural plasticity largely outperforms weight plasticity and can achieve a much higher storage capacity per synapse. The effect of structural plasticity on the structure of sparsely connected networks is quite intuitive: Structural plasticity increases the “effectual network connectivity”, that is, the network wiring that specifically supports storage and recall of the memories. Further, this model of structural plasticity produces gradients of effectual connectivity in the course of learning, thereby explaining various cognitive phenomena including graded amnesia, catastrophic forgetting, and the spacing effect.     

Soft-bound Synaptic Plasticity Increases Storage Capacity

Accurate models of synaptic plasticity are essential to understand the adaptive properties of the nervous system and for realistic models of learning and memory. Experiments have shown that synaptic plasticity depends not only on pre- and post-synaptic activity patterns, but also on the strength of the connection itself. Namely, weaker synapses are more easily strengthened than already strong ones. This so called soft-bound plasticity automatically constrains the synaptic strengths. It is known that this has important consequences for the dynamics of plasticity and the synaptic weight distribution, but its impact on information storage is unknown. In this modeling study we introduce an information theoretic framework to analyse memory storage in an online learning setting. We show that soft-bound plasticity increases a variety of performance criteria by about 18% over hard-bound plasticity, and likely maximizes the storage capacity of synapses.     

Interrogating structural plasticity among synaptic engrams

Our daily experiences and learnings are stored in the form of memories. These experiences trigger synaptic plasticity and persistent structural and functional changes in neuronal synapses. Recently, cellular studies of memory storage and engrams have emerged over the last decade. Engram cells reflect interconnected neurons via modified synapses. However, we were unable to observe the structural changes arising from synaptic plasticity in the past, because it was not possible to distinguish the synapses between engram cells. To overcome this barrier, dual-eGRASP (enhanced green fluorescent protein reconstitution across synaptic partners) technology can label specific synapses among multiple synaptic ensembles. Selective labeling of engram synapses elucidated their role by observing the structural changes in synapses according to the memory state. Dual-eGRASP extends cellular level engram studies to introduce the era of synaptic level studies. Here, we review this concept and possible applications of the dual-eGRASP, including recent studies that provided visual evidence of structural plasticity at the engram synapse.     

Do thin spines learn to be mushroom spines that remember?

Dendritic spines are the primary site of excitatory input on most principal neurons. Long-lasting changes in synaptic activity are accompanied by alterations in spine shape, size and number. The responsiveness of thin spines to increases and decreases in synaptic activity has led to the suggestion that they are 'learning spines', whereas the stability of mushroom spines suggests that they are 'memory spines'. Synaptic enhancement leads to an enlargement of thin spines into mushroom spines and the mobilization of subcellular resources to potentiated synapses. Thin spines also concentrate biochemical signals such as Ca(2+), providing the synaptic specificity required for learning. Determining the mechanisms that regulate spine morphology is essential for understanding the cellular changes that underlie learning and memory.     

Increasing numbers of synaptic puncta during late-phase LTP: N-cadherin is synthesized, recruited to synaptic sites, and required for potentiation

It is an open question whether new synapses form during hippocampal LTP. Here, we show that late-phase LTP (L-LTP) is associated with a significant increase in numbers of synaptic puncta identified by synaptophysin and N-cadherin, an adhesion protein involved in synapse formation during development. During potentiation, protein levels of N-cadherin are significantly elevated and N-cadherin dimerization is enhanced. The increases in synaptic number and N-cadherin levels are dependent on cAMP-dependent protein kinase (PKA) and protein synthesis, both of which are also required for L-LTP. Blocking N-cadherin adhesion prevents the induction of L-LTP, but not the early-phase of LTP (E-LTP). Our data suggest that N-cadherin is synthesized during the induction of L-LTP and recruited to newly forming synapses. N-cadherin may play a critical role in L-LTP by holding nascent pre-and postsynaptic membranes in apposition, enabling incipient synapses to acquire function and contribute to potentiation.