Memory Flashcards
Whitlock
2006, evidence learning induced LTP in vivo
− Used inhibitory avoidance (IA) paradigm, which has been shown to evoke stable memory trace in a single trial
− IA mimicked the effects of HFS in CA1
− Immediate NMDAR-dependent increase in phosphorylation of GluR1
− Delivery of SluR1 and GluR2 (but not NR1) to the synaptoneurosome biochemical fraction
− Increase in the slope of the evoked fEPSP
− IA-induced increases in fEPSP partially occluded subsequent LTP by HFS
− Evidence LTP and learning induce change by common mechanism
− (But note – this is association, doesn’t show causality)
O’Keefe and Nadel
1978, argued for the Hc as a cognitive map
− Perhaps most influential and enduring of the theories. Suggested that primary role of Hc was to encode spatial information and to form spatial memories
O’Keefe and Dostrovsky
1971, discovered place cells
− Recorded from rat Hc (CA1) PCs whilst rat explored spatial environment
− Response of cells varied as a function of rat’s spatial location: place cells (type of PC) only fire when animal occupies a specific location within its environment
− Different place cells have different place fields (locations where they fire), so are thought to provide a cognitive map for the animal
− No apparent topography to pattern of place cells
Morris (1)
1982, lesion evidence for Hc in spatial memory
− Hc lesions impair spatial memory performance in morris watermaze task in mice
− Lesioned mice do not learn or get faster, spend equal time in each quadrant
Morris (2)
1986, NMDAR antag (D-AP5) imparis performance on spatial watermaze task
− rats (controls and D-AP5) trained in watermaze task for 4 days
− controls learned to escape relatively rapidly
− D-AP5 rats learned more slowly, but their impairment was not striking
− Transfer test: take platform away, track movement for 60seconds
− Controls spend most time in trained quadrant, least in opposite
− D-AP5 showed little bias towards training quadrant, mostly equal
− Further training
− D-AP5 stabilised their escape latencies at much higher level, taking longer, more indirect routes
− Visual discrimination showed no difference between groups (using visually different platforms)
− So not likely to be a secondary sensorimotor/motivational impairment
− Confirmed D-APV blocked LTP in vivo, but not basal synaptic transmission
− Implies blocking NMDAR-LTP learning impairment
Pastalkova
2006, ZIP infusion produced loss of 1-day old spatial information
− Injected a cell-permeable PKM(zeta) inhibitor into the rat Hc
− It blocked LTP maintenance in vivo, and produced persistent loss of 1-day old memories
− The inhibitor could block LTP and impair Hc-dependent memory even if administered days after acquisition
− Argued mechanism maintaining LTP also sustains spatial memory
Keith and Ruddy
1990, argued that D-AP5 animals were impaired in sensory motor domains, as they often fell back off the platform
Bannerman (1)
1995, found that spatial pretraining ameliorates the AP5 effects on learning
− Carried out above experiment and got same results
− Pretrained rats in a watermaze task in different lab, then implanted minipumps and trained rats in second watermaze task
− AP5 rats were still slower, but searched predominantly in the trained quadrant during transfer test
− Showed that Hc was still necessary for spatial learning after pretraining (lesioning showed deficits)
− AP5-induced deficit still seen, however, is animal is pretrained using a non-spatial task
− So, NMDARs may not be required for encoding spatial representations of a specific environment
Tsien
1996,
− Using Cre/loxP, created mouse strain with CA1-specific deletion of the GluN1 subunit in PCs. Mice grow into adulthood without obvious abnormalities
− Adult mice lack NMDAR-mediated currents and LTP at CA1 synapses (shown in slice experiments)
− Spatial watermaze task: KO mice were deficient in learning (longer escape latency, didn’t prefer trained quadrant in transfer task)
− Non-spatial landmark task: KO mice learned slightly slower, but reached same level of optimal performance as controls
− Argue KO mice display a selective and significant deficit in spatial memory
Fukaya
2003, argued that the KO was not Hc specific
− At 2 months of age, immunohistochemistry showed mice have reduced GluN1 levels in cortex and neocortex
Bannerman (2)
2012, found that KO differentially effected different spatial tasks
− Used a Hc-specific GluN1 KO mouse line (NMDARs knocked out in PCs in CA1 and granule cells in DG)
− Some Cre expression in olfactory bulb granule cells, and in fewer L2 piriform cortex neurons, but only minimal NMDAR cortex ablation
− Showed a lack of LTP in CA1
− 1) Watermaze task
− KO not impaired: acquired spatial reference memory as well as controls, spent MORE time in trained quadrant than controls
− Impaired in spatial reversal testing (move platform to opposite side)
− 2) Radial maze task (comparable to land watermaze task)
− Mice must discriminate between arms containing food/no food
− KO impaired at acquiring spatial reference memory (makes more choice errors)
− Seems Hc NMDARs play a role in aspects of spatial memory performance, but not necessarily in the formation of associative, long-term memories
Bannerman (3)
2014, argued that blocking NMDAR results in a behavioural inhibition deficit
− Watermaze task with naïve KO/control mice
− A) non-spatial discrimination: 2 visually distinct beacons (1 with platform), position moved between trials
− KO not impaired
− B) Spatial discrimination: 2 visually identical beacon (1 with platform), mouse had to use allocentric spatial location to discriminate
− KO significantly impaired (made more choice errors)
− BUT, showed no deficit in transfer test
− KO spent as much time in training quadrant as controls- they knew as much about the spatial location of the platform
− Number of errors varied systematically depending on where the trails started form
− Increased choice errors when started closer to decoy beacon
− No deficits when always started equidistant
− So, seems mice able to learn spatial location (agreeing with earlier findings)
− Seems not a spatial learning deficit, but rather reflects an inability to use the spatial cue to behaviourally inhibit very strong conditioned response to swim towards closest beacon
Hoffman
2002, could still induce LTP by theta-burst pairing protocol in CA1-GluN1 KO mice (mimics theta waves generated in exploratory behaviour)
Kohl
2011, found that not all synapses in Hc are ‘as’ plastic
− Used optogenetic tools for cell type- and hemisphere-specific recruitment of CA3 axons
− Found that inputs from left CA3 onto CA1 PCs more able to produce t-LTP than inputs from the right CA3
− The hemispheric asymmetry in plasticity could be explained by differential GluN2B expression at left/right CA3 targeted synapses
− NMDAR with GluN2B favour LTP induction
− Behavioural consequences of this remain to be investigated
Bannerman (review)
2014,
− Argue Hc NMDARs are not required to encode associative spatial memories, but are important for inhibiting responding to ambiguous or uncertain cues
− might be important when must disambiguate between overlapping/competing memories
− Would explain results above:
− Spatial task with identical cues ambiguity impairment
− Pre-training with then new allocentric cues no ambiguity
− Spatial reversal task, unchanged allocentric cues ambiguity
− Perhaps Hc is important for memory, but is not site of encoding
− May be more for memory retrieval, particularly of retrieving correct memory in situation of ambiguous/overlapping memories
− Also argue that Hc plays a key role in anxiety
− Anxiety can be described as uncertainty arising from concurrently available response choices or goals
− Earlier idea can be extended – Hc acts as a comparator when a mismatch occurs between the perceived state of the world and expectation based on memory
− dHc receives input of highly processed polymodel sensory information from cortical areas, vHc is more closely linked to subcortical structures like AmG and HPA axis
− dHc damage affect spatial learning
− vHc damage reduces anxiety in social interaction, elevated zero-maze and neophobia tests, doesn’t effect spatial memory
