Memory And Plasticity Flashcards
Introduction
The significance of synaptic plasticity in learning and memory is a central issue in neuroscience.
A wealth of evidence implicates the hippocampus in the acquisition of episodic memory, most notably patient HM who after removal of large portions of the hippocampi suffered from anterograde and retrograde amnesia.
The pivotal 1973 study by Bliss and Lomo who used in vivo rabbit recordings to demonstrate that a tetanic stimulus triggers a long-lasting increase in EPSP amplitude in dentate gyrus neurons, first identified LTP which has become the dominant model of plasticity and learning.
The presumptive causality between plasticity and memory was formalised by Morris and postulates that plasticity is both necessary and sufficient for memory acquisition.
Although a wealth of data suggests necessity, current data fails to support the notion of sufficiency.
Nevertheless, a shift from mechanistic investigations of plasticity in single neurons towards probing how neural networks encode memory could validate the notion of sufficiency.
By first evaluating the significance of hippocampal LTP, we will also consider additional forms of plasticity and their associations with memory in the cerebellum and amygdala.
(Paragraph 5)
Neural circuits for memory/memory recall
Although LTP could learn about the correlations between synaptic inputs, it’s not immediately obvious how this information could be recalled and so plasticity must be considered in the context of circuits.
The architecture of the hippocampal CA3 circuits has inspired computational memory models since these circuits are well adapted for the rapid storage and retrieval of memories.
The dense recurrent excitatory connections between CA3 pyramidal cells (PCs) are thought to serve as an attractor network in which associative memories can be stored and recalled.
The network ‘attracts’ incoming activity to a stored firing pattern, where activation of some of the neurons in the network by elements of the memory reactivates the whole activity pattern by pattern completion.
To explore whether NMDAR-dependant plasticity in CA3 PCs is required for hippocampal learning and memory recall, Tonegawa et al (2002) used mice with GRIN1 KO restricted to CA3 PCs.
They demonstrated impaired spatial memory retrieval in the water maze task when these mice were presented with a fraction of the original cues, underscoring the significance of CA3 PCs in associative recall.
However, all synaptic inputs onto CA3 PCs are affected by neutralization of NMDARs and so we cannot elucidate for which pathway the removal of plasticity causes behavioural deficits.
Thus causality between plasticity in the CA3-attractor network and memory is yet to be established.
(Paragraph 5)
Neural circuits/memory recall
Part 1 - CA3 attractor network
Although LTP could learn about the correlations between synaptic inputs, it’s not immediately obvious how this information could be recalled and so plasticity must be considered in the context of circuits.
The architecture of the hippocampal CA3 circuits has inspired computational memory models since these circuits are well adapted for the rapid storage and retrieval of memories.
The dense recurrent excitatory connections between CA3 pyramidal cells (PCs) are thought to serve as an attractor network in which associative memories can be stored and recalled.
The network ‘attracts’ incoming activity to a stored firing pattern, where activation of some of the neurons in the network by elements of the memory reactivates the whole activity pattern by pattern completion.
(Paragraph 5)
Neural circuits/memory recall
Part 2 - Tonegawa et al (2002)
To explore whether NMDAR-dependant plasticity in CA3 PCs is required for hippocampal learning and memory recall, Tonegawa et al (2002) used mice with GRIN1 KO restricted to CA3 PCs.
They demonstrated impaired spatial memory retrieval in the water maze task when these mice were presented with a fraction of the original cues, underscoring the significance of CA3 PCs in associative recall.
However, all synaptic inputs onto CA3 PCs are affected by neutralization of NMDARs and so we cannot elucidate for which pathway the removal of plasticity causes behavioural deficits.
Thus causality between plasticity in the CA3-attractor network and memory is yet to be established.
(Paragraph 6)
To demonstrate the sufficiency of plasticity in memory
To demonstrate the sufficiency of plasticity in memory, we could examine the idea that neural representation of memory is encoded in the network containing synapses which were modified during the acquisition of that memory, then if selective inactivation of these neurons results in memory loss it would demonstrate sufficiency.
Kitamura et al (2017) used Intermediate Early Gene expression combined with Tet transcriptional control to interrogate neocortical and subcortical circuits for memory consolidation.
The authors found that prefrontal memory engrams are generated immediately whilst the memory is acquired, thus redefining the role of the hippocampus in rapidly generating the cortical engram cells during learning as well as their functional maturation.
Using a similar approach, future experiments could exploit the transient and activity dependant expression of IEGs to use their promoters to drive the expression of proteins which alter membrane excitability (e.g. ChR2 and the use of optogenetics), in order to reversibly activate those hippocampal cell assemblies which represent the memory acquired during a spatial memory task.
(Paragraph 6)
To demonstrate the sufficiency of plasticity in memory
Part 1
To demonstrate the sufficiency of plasticity in memory, we could examine the idea that neural representation of memory is encoded in the network containing synapses which were modified during the acquisition of that memory, then if selective inactivation of these neurons results in memory loss it would demonstrate sufficiency.
(Paragraph 6)
To demonstrate the sufficiency of plasticity in memory
Part 2 - Kitamura et al (2017)
Kitamura et al (2017) used Intermediate Early Gene expression combined with Tet transcriptional control to interrogate neocortical and subcortical circuits for memory consolidation. The authors found that prefrontal memory engrams are generated immediately whilst the memory is acquired, thus redefining the role of the hippocampus in rapidly generating the cortical engram cells during learning as well as their functional maturation.
(Paragraph 6)
To demonstrate the sufficiency of plasticity in memory
Part 3 - future experiments
Using a similar approach, future experiments could exploit the transient and activity dependant expression of IEGs to use their promoters to drive the expression of proteins which alter membrane excitability (e.g. ChR2 and the use of optogenetics), in order to reversibly activate those hippocampal cell assemblies which represent the memory acquired during a spatial memory task.
(Paragraph 7)
The ultimate paradigm illustrating sufficiency
The ultimate paradigm illustrating sufficiency would use synaptic plasticity to build a memory.
For instance a hippocampus-dependant memory could be synthesised using a training procedure and then erased before being re-installed by exploiting the knowledge gained about the synaptic changes underlying the original learning episode.
A large multi-electrode array could be used to monitor plasticity between hippocampal cell pairs implicated in encoding the new memory during training on the water maze task, where cross correlation of APs from cell pairs before and after training allows us to identify which pairs are encoding the memory.
Subsequent depotentiation of those identified synapses by immediate post-training application of low-frequency stimuli would remove the memory (resultant impaired performance on the water maze task).
STDP could then be used to re-tune synapses back to the memory state, thus allowing selective reversal of synaptic changes implicated in the memory acquisition with subsequent reinstallation of the memory.
Conclusion
Investigations into the neural basis of memory have been dominated by the dissection of the molecular and cellular basis of plasticity.
However in order to prove sufficiency, we require tools which emphasise the significance of neural networks implicated in learning.
Whilst circuit-specific memory erasure demonstrates that plasticity is necessary for memory acquisition, techniques which permit re-installation, at the network level, of silenced memories would establish that plasticity is also sufficient for memory acquisition.
Although this has been an intangible goal, recent advances in recording synaptic activation in vivo allude to significant progress;
Macpherson et al (2015) described X-GRASP, where activity-dependant GRASP (GFP reconstitution across synaptic partners) uses synaptobrevin as a carrier for the GFP fragment in order to retrospectively label active synapses in vivo.
Nonetheless given the sophisticated nature of the field requiring a myriad of multidisciplinary techniques, progress and the ultimate achievement of proving sufficiency certainly demands a collaborative approach.
Conclusion
Part 1 - from single neurons to networks, plasticity and sufficiency
Investigations into the neural basis of memory have been dominated by the dissection of the molecular and cellular basis of plasticity.
However in order to prove sufficiency, we require tools which emphasise the significance of neural networks implicated in learning.
Whilst circuit-specific memory erasure demonstrates that plasticity is necessary for memory acquisition, techniques which permit re-installation, at the network level, of silenced memories would establish that plasticity is also sufficient for memory acquisition.
Conclusion
Part 2 - Macpherson et al (2015)
Although this has been an intangible goal, recent advances in recording synaptic activation in vivo allude to significant progress.
Macpherson et al (2015) described X-GRASP, where activity-dependant GRASP (GFP reconstitution across synaptic partners) uses synaptobrevin as a carrier for the GFP fragment in order to retrospectively label active synapses in vivo.
Nonetheless given the sophisticated nature of the field requiring a myriad of multidisciplinary techniques, progress and the ultimate achievement of proving sufficiency certainly demands a collaborative approach.
