MT4 Flashcards
cell assembly theory
- “cells that…”
- how does info get transferred?
- hebbian plasticity (recurrent connections + pattern completion)
“cells that fire together wire together”
Hebb’s idea was that info gets transferred in the form of the strength of synapses (neuroplasticity)
Hebbian plasticity: cell assemblies = associative memory
- initially CA3 neurons have weak connections
(CA3 gets inputs from LEC/MEC and can project to CA1 or loop back onto itself –> recurrent)
- something activates multiple neurons at the same time (pre/post syn) –> strengthens synapses (recurrent connections)
- the next time one of those neurons is activated, the others will be activated too (pattern completion –> can remain even without synapse –> memory playing out)
Long Term Potentiation (LTP)
- rabbit hippocampus study
- measurement
- procedure (3)
- result
Rabbit hippocampus study:
- if you do a train of excitation at dentate gyrus (short bursts) the synaptic connection becomes more efficient = bigger response to same input
- synaptic strength measured by size of EPSP (measured @ DG) –> dips down on graph bc electrode is extracellular
- low freq stim given as baseline
- high freq stim multiple times (similar to encoding experience) to presyn axons
- give low freq again and see response
result: saw that expt group EPSPs remain potentiated for hours/days –> response to low stim stays high (efficiency inc)
flashbulb memory (4)
- hypothesis
- mechanism
- highly vivid
- associated w strong emotion
- deja-vu like –> one aspect triggers whole memory (associative)
- like a snapshot (moment in time, little details included)
- feels very accurate but equally prone to mistakes as other memories
Hypothesis: emotionally valenced memory stored more diffusely instead of locally –> more avenues for retrieval, stronger connections/encoding
mechanism: weak stimulus (depol) E-LTP gets turned into L-LTP by borrowing PSPs from strong stimulus –> now linked = minor details associated with main aspect of memory
AMPAR - 3 steps
- type of synapse (what is NT?)
- synapse between what?
glutaminergic synapse: CA3 neuron terminal, CA1 neuron spine
- AP travelling down axon causes VG calcium channels open –> Ca++ enters
- Ca++ causes vesicles to merge with membrane and release NTs
- NTs bind to AMPAR (ligand-gated) and allow Na+ enters/K+ exits (non-selective) –> if there’s enough depol EPSPs summation creates AP
NMDAR
- type of receptor
- conditions (2)
- if sufficient, what happens? (regular stuff + 3)
- result
NMDAR = coincidence detector –> has block inside made of Mg++
- 1st condition: glu binds –> activates NMDAR, but Mg++ still blocking
- 2nd condition: Mg++ moves only when theres enough depol in post syn (enough NTs)
if sufficient: Na+ in, K+ out, influx of Ca++/calmodulin at postsyn NMDAR –> activates 3 signaling pathways
- CAMKII: causes phosphorylation of AMPAR –> increases conductance (lets in more current) = more depol
- NO synthase: creates NO –> gas can pass through membrane –> goes to presyn –> causes more NT release
- PKC (protein kinase C): summons more AMPAR to memb
Result: synapse bolstered –> inc EPSPs for each presynaptic AP
- happens in milliseconds (immed strengthening through bolstering of signal)
Fundamental properties of LTP
- synaptic transmission
- cooperativity
- associativity
- synapse specificity
synaptic transmission: single weak active input –> caused by minor changes (eg. perception) –> Mg++ no move = no LTP (even if glu binds, not enough depol to move Mg++) = no encoding
cooperativity: multiple weak active inputs = sufficient spatiotemporal summation = significant enough activity to initiate LTP (Mg++ block moves)
associativity: strong input can cause initiation of LTP in their neuron and also nearby neurons with weak inputs (small details take advantage of strong memory)
- if there is no activity, even strong input (lots of depol which would move Mg++) would not cause LTP since no NTs are being released
synapse specificity: LTP only occurs at active synapses –> prevents storage/association of unrelated info
early vs late LTP
- dependent on?
- reason for difference, what if you inhibit?
- Late LTP mechanism
early LTP: decays within hours (when there’s only one stimulus) –> 90 mins back to baseline (temporarily increased connection)
- NMDAR-dependent, occurs in milliseconds –> able to associate stimuli occurring ~1s apart
late LTP: requires more stimuli and persists for many hours in vitro, lasts much longer in vivo –> 300% increase in response
- PRP-dependent, lasts minutes to hours
Reason: protein synthesis –> epigenetic (experience-dependent) –> need gene to make protein
- when you inhibit protein synth with drug, early is not impacted since it takes advantage of local signalling, but for late LTP in inhibited rats (compared to controls), there is a decline
Late LTP mechanism:
- Ca++/calmodulin activates adenylyl cyclase, which makes cAMP
- cAMP causes CREB-dependent transcription –> causes upregulation of plasticity-related proteins (PRPs)
- PRPs are necessary for long term structural synapse strengthening + new synapse formation –> more long term than just NMDAR LTP (early)
synaptic tagging
- strong stimulus
- weak stimulus
strong stimulus: causes activated synapse to get tagged (with protein) and causes PRPs to be expressed –> PRPs targeted to synapse –> enables late LTP
weak stimulus: causes activated synapse to get tagged –> steals depol or PRPs from strong stim –> gets linked with strong stim –> not strong enough to cause gene transcription by itself
pattern separation and completion
pattern separation: dentate gyrus has a large number of cells (eg. granule cells) –> able to diverge sensory info –> separate memories that are similar but different –> represented as a population code (multiple inputs from MEC/LEC)
pattern completion: recurrent connections in CA3 link components of memory via LTP –> activation of a subset of neurons results in reactivation of complete memory (activates pyramidal cells in CA3)
inhibition
- lateral
- feedforward
- KO NMDAR
lateral inhibition: between dentate gyrus cells –> sharpens contrast –> pattern stands out, everything else inhibited (reduces activity of other neurons)
feedforward inhibition: activates pattern to complete memory, while also activating inhib cells that inactivate other patterns/neurons –> major role in selecting which CA3 neurons participate in the memory
KO NMDAR: in CA3/dentate = problems with pattern separation
Morris water maze
- control vs KO
- hidden platform task (function, set up, controls vs KOs and what that means)
- visible platform task
-/- = NMDAR knock out --> subunit of receptor KO'ed = nonfunctional = no LTP bc NMDAR is coincidence detector that starts LTP process \+/+ = control (WT)
Hidden platform water maze task = spatial orientation (test of necessity) –> multiple cues around the walls –> mouse can orient itself so it can find platform
- efficiency of controls increased (LTP is happening)
- KO group slightly slower to find platform/learn (did not completely prevent LTP) –> contributes to idea that LTP is part of learning (bc NMDAR is KO) –> does not fully prove tho
visible platform task: if you damage hippoc can it still find platform when water is clear (can see platform) –> confirming that movement is normal –> non-spatial learning
- both groups able to do it normally (LTP unneeded for normal navigation)
transfer test
- KO mice vs WT mice
- overall trend
- limitation
transfer test: remove platform and place mice in maze next day (after acquisition) –> mice should go to quadrant where platform was
- KO mice spent less time in quadrant + passed by old location of platform less than WT mice
- overall trend shows that they did spend more time in correct quadrant than other quads (learned something)
limitation: KO mice were very jumpy/easily startled –> may have something to do with amygdala –> emotional valence/fear conditioning affected due to KO of some gene in amygdala
Contextual fear conditioning - Dentate gyrus
- room A vs room B –> KO vs WT
- proof of concept
- conditioning
- DG is for? –> WT vs KO in room A vs room B
- CA3 unable to…
conclusion
DG-NMDAR KO mice vs WT –> assessing freezing behaviour
room A: shock through floor
room B: no shocks, identical floors, unique odours, roofs and lighting
- Proof of concept: KO mice did not show LTP at DG (inc in EPSPs) = successfully KOed gene
conditioning: rats will associate shock with room A –> exhibit freezing
- DG is important for pattern completion –> WT rats are able to distinguish room A from room B and don’t get scared in room B (discrimination ratio A/A+B high), but KO rats had harder time (still exhibited freezing behaviour in room B –> generalized fear of room A to room B)
- CA3 place cells unable to do rate remapping –> cannot modulate firing for diff contexts = cannot distinguish
conclusion: DG synapses (EC to DG) are important for context discrimination with limited experience (if many trials KO rats able to discriminate similar to WT)
CA3 NMDAR KO in Morris Water Maze
- what does it affect?
- full cue –> why? what region?
- partial cue
- no cue –> evidence for what?
NMDAR KO = no LTP in CA3 –> see how it affects pattern completion
Morris water maze: full cue vs partial cue vs no cue
- full cue: KO and WT same –> no pattern completion needed since all cues there
- CA1 is a major site for storage of spatial reference memory (can still store without CA3) - partial cue: big difference, WT able to do pattern completion, KO mice unable to (much lower recall)
- no cue: not much difference (both low) –> evidence for pattern completion (since controls are same as KO without pattern)
- KO may be slightly higher bc they’re already using some other brain area (not over-reliant on spatial cues)
cochlear implants
- tonotopic map
- sound processor
- electrode array
- learning curve
- tonotopic map: we know what frequencies activate basilar memb –> artificially activate hair cells
- receive signals (receiver electrodes) from sound processor (transmitting electrodes) and deliver to electrode array threaded into cochlea
electrode array = series of indiv electrodes that pass current at diff point along cochlea –> electric field causes depol of spiral ganglion cells (AN cells) which bypass hair cells and activates nerves
- learning curve: takes a while to learn what the sounds mean
optogenetics
- DNA expression
- promoters
1. channelrhodopsin
2. halorhodopsin - targeting –> how do we know where?
- all cells have same DNA, but different expression
- can target cells that we know are involved with behaviour using promoters –> drive specific expression of proteins that allow neurons to be manipulated/controlled –> proteins can control activity of cell
- channelrhodopsin: when ChR2 gets hit with 470 nm light –> AP when channel opens
- Halorhodopsin: activated by longer wavelengths (569nm) –> turns channel into chloride pump –> inibits cell –> stops AP (opposite of chr2)
targeting: making sure only cells we want get activated by using dental cement to glue a guide through head –> insert into brain = feeding wavelength to brain
- can figure out where activity is happening using EEG –> can reactivate memory by shining laser
