PSYC 3x14

Justin Harris
Week 8: Synaptic plasticity and associative learning

Week 9: Localisation of learning and memory;

Movement and motor control

Week 11: Basal ganglia; Neurogenerative diseases.

 Cellular Mechanisms of Learning:
Synaptic Plasticity

Learned changes in behaviour must correspond to neural changes.

• Ramon y Cajal: “plasticity” (changes) in synaptic connections

responsible for learning and memory.

• 50 years later, Konorski and Hebb described models of synaptic

plasticity.
 Hebbian synapse
Biologically significant events (USs) have hard-wired connections
controlling behaviour.
• Eg, neurons coding for food can directly excite neurons producing
salivation.

US US Output Behaviour
 Hebbian synapse
• Neurons for other events (eg CSs) form weak (ineffective) synapses
with neurons controlling that behaviour.

CS CS

US US Output Behaviour

Pavlov: these connections must be acquired through learning

Hebb & Konorski:
these “latent” connections must be strengthened through learning
 Hebbian synapse
Synaptic connection between CS and behavioural output is strengthened
when weak CS input arrives simultaneously with strong US input.
• “neurons that fire together, wire together”

CS now able to produce

CS CS response (CR).

US US Output Behaviour
Synaptic plasticity in hippocampus

Hippocampus proper
comprised of 3 regions:
CA1, CA2, and CA3
(Cornu Ammonis)

CA1
CA3

Dentate
gyrus
 Synaptic connections in the Hippocampus
Very hard to stimulate one fibre and measure from single neuron.
But organization of neural circuitry in hippocampus conveniently
segregates inputs and throughputs

pyramidal cells in CA1

granule cells pyramidal cells in CA3

perforant path (in dentate gyrus)
(from entorhinal cortex)
 Long-Term Potentiation (LTP)

LTP is a physiological example of synaptic plasticity.

- potential as a model for neural mechanisms of learning.

• Most demonstrations of LTP have been in the hippocampus,

usually in vitro, but also in vivo.
 Long-Term Potentiation (LTP)
Step 0:
Weak stimulation of Presynaptic input (eg, perforant path to hippocampus)
causes little or no activity in post-synaptic neurons (eg, in dendate gyrus
(DG) of hippocampus).

weak No
stimulation effect

CA3
dentate gyrus
perforant path (pyramidal cells)
(granule cells)
 Long-Term Potentiation (LTP)
Step 1:
Strong, high-frequency (eg 100Hz) stimulation of pre-synaptic input
causes long-lasting increase in sensitivity of post-synaptic neurons.

Strong, high-
frequency big
stimulation response

perforant path dentate gyrus

 Long-Term Potentiation (LTP)
Step 2:
Weak stimulation of the pre-synaptic input now produces action
potentials in the post-synaptic cells

weak moderate
stimulation response

perforant path dentate gyrus

 Long-Term Potentiation (LTP)
HFS HFS HFS
Neuronal response

Bliss & Lomo (1973). Journal of Physiology, 232, 331-356.

 LTP is “dose dependent”

• Weak high-frequency stimulation (HFS) can produce

short-lived potentiation (10 min), but long-lasting
potentiation (hours) achieved by strong HFS
 LTP is “dose dependent”

• Weak high-frequency stimulation (HFS) can produce

short-lived potentiation (10 min), but long-lasting
potentiation (hours) achieved by strong HFS

• HFS often as continuous volley, but can be patterned as

bursts at theta frequency (“Theta Burst Stimulation”, TBS):
eg, short bursts of 5 pulses in 50ms, repeated every 200ms.

50 ms 50 ms 50 ms
150 ms 150 ms

• Duration of LTP depends on number of TBSs.

 3 Hz
Long-Term Depression
(LTD)
• Plasticity is bi-directional:
Low-frequency stimulation can 10 Hz
reduce synaptic efficacy (LTD)

1 Hz
50 Hz

Dudek & Bear (1992). Proc Natl Acad Sci

 LTP has 3 properties to recommend it as a
model of learning and memory:
Persistence: Potentiation is enduring, sometimes lasting weeks.

Synaptic specificity:
Only stimulated pre-synaptic inputs show potentiation
ie, no increased sensitivity to other pre-synaptic inputs.

Associativity….
 Associativity of LTP
Can get LTP at pre-synaptic inputs weakly stimulated at the same
time as strong stimulation to separate (but converging) input.

Strong,
stimulation

Large
response
Weak, DG
stimulation
 3. Associative LTP
Strong
stimulation

Weak
stimulation
 3. Associativity of LTP
This property most resembles Hebb’s model for how associations are
acquired by nervous system.

US
Behavioural
Output response

CS
 Do LTP and learning share a common mechanism?

Correlations between LTP and learning:

• Age-related decline in learning correlates with age-related decline in
induction of LTP in hippocampus.
• Similar correlations between LTP and learning in mouse model of
Alzheimer’s Disease.
 Do LTP and learning share a common mechanism?

Is learning affected by saturation of LTP?

Bliss & Lomo (1973). Journal of Physiology, 232, 331-356.

• Evidence that saturation of LTP in hippocampus can prevent rats from

learning in a simple maze.
Do LTP and learning share a common mechanism?

Do learning and LTP share common neurochemistry?

Pharmacological interventions that prevent LTP (block NMDA receptors)

also disrupt learning.
• Conditioned Taste Aversion;
• Conditioned Fear;
• Conditioned Eyeblink;
• Maze learning.
 Neurochemical basis of LTP

LTP dependent on release of excitatory neurotransmitter glutamate.

1. Glutamate binding to AMPA receptors.

&
2. Glutamate binding to NMDA receptors.
 Neurochemical bases of LTP
Glu from pre-synaptic terminal binds to AMPA receptors on post-synaptic
neuron, causes immediate excitation (depolarisation) of post-synaptic
neuron (Fast excitatory post-synaptic potentials, EPSPs).

Na+

Post-
Glu synaptic
neuron
Pre-synatpic terminal
Fast EPSPs
 Neurochemical bases of LTP
Glu must also bind to NMDA receptors, opening Ca++ channels.

Ca++

Glu
 Intracellular cascade
increased number of
Strong, high- AMPA receptors =
frequency increased Na+ influx
stimulation
 Neurochemical bases of LTP
NMDA receptors have 2 special properties that underlie synaptic plasticity:

1. Admit Ca++ into the neuron (increase AMPA)

2. Ca++ channels on NMDA receptors are:
• ligand-gated – Glutamate (ligand)
&
• voltage-gated – post-synaptic neuron must be depolarised
Allows for
specificity
Strong,
high-frequency
Na+
stimulation
Glu

Na+
Ca++ Ca++
 Allows for
associativity
Strong,
high-frequency
Na+
stimulation
Glu

Na+
Ca++ Ca++
Weak
stimulation
 Intracellular Changes triggered during LTP
Stages that convert initial learning into long-term memory:

1) Generating the synaptic change

(creating the initial memory trace)

2) Stabilising changes

3) Consolidating changes

4) Maintaining changes (preventing forgetting)

Generating the synaptic change:
post-translational changes underlying LTP

• These rapid changes called “post-translational” because

they use existing proteins in neuron and do not require
synthesis of new proteins (requires translation from RNA).

• Transient (they revert back to previous state) unless other

intracellular processes are activated to stabilise the
changes

 May explain why recent memory traces can be

disrupted by head trauma (concussion causes amnesia for
events that occurred a few minutes before an accident.)
Generating the synaptic change:
Constitutive trafficking and recycling of receptors

Glu
Synapse Glu
Glu
Glu
AMPA NMDA

PSD PSD AMPA

Dendritic spine
(post-synaptic) endocytosis

endosome
Generating the synaptic change:
Constitutive trafficking and recycling of receptors

Glu
Synapse Glu
Glu
Glu
NMDA AMPA

PSD PSD
Ca++
PKs
Dendritic spine
(post-synaptic) endocytosis

endosome

Protein Kinases: PKs promote trafficking AMPA

back to post-synaptic density (PSD).
Generating the synaptic change:
Constitutive trafficking and recycling of receptors
Synapse

NMDA AMPA

Ca++

actin

endocytosis

endosome

Ca++ influx triggers enzymes ( ) to disassemble actin filaments

(cytoskeleton) that otherwise obstruct AMPA trafficking to PSD.
Stabilising the synaptic change:
rebuilding the cytoskeleton
Synapse

NMDA AMPA

Ca++
Polymerised actin
actin

endocytosis

endosome

Ca++ influx triggers other enzymes rebuild (polymerise) actin

cytoskeleton.
 Stabilising the synaptic change:
Cell adhesion molecules

Pre-synaptic Pre-synaptic
terminal terminal

Cadherin Cadherin
Synapse Synapse
monomer dimer

Ca++

post-synaptic post-synaptic
spine spine

Ca++-dependent cell adhesion molecules (neural cadherins)

form bridge between pre and post-synaptic membranes.
Ca++ influx through NMDA-R converts weakly-adhesive monomer
to strongly adhesive dimer, stabilising synapse.
 Consolidating synaptic change:
Translational processes (protein synthesis)

• Weak HFS can produce short-lived potentiation (10 min),

but long-lasting potentiation (hours) achieved by strong
HFS, or HFS at theta burst frequency.

• This LTP requires translational processes (protein synthesis);

drugs that block protein synthesis can prevent long-lasting
synaptic potentiation after strong HFS (but have no effect
on the initial short-lived potentiation).
 Consolidating synaptic change:
Local translation of mRNA (protein synthesis)

Dendritic Ca++
spine

mRNA

Nucleus
(DNA)

Dendritic spines contain machinery for local soma

synthesis of proteins to make AMPA receptors:
• Enduring LTP can be induced in CA1 dendrites that have been dissected
away from cell body (soma) that contains nucleus.
 Consolidating synaptic change:
Transcription of mRNA from nucleus

Voltage-dependent
Dendritic Ca++ Action
Ca++ channels
spine potential

mRNA Ca++

mRNA
Nucleus
(DNA)
• New mRNA must be transcribed from DNA soma
in nucleus to supplement mRNA in dendrites.
• Ca++ trigger processes that engage mRNA transcription in nucleus.
• Largely due to Ca ++ entering through voltage-dependent Ca ++ channels
at soma, triggered by action potentials passing from dendrites to axon.
 Consolidating synaptic change:
Spine growth
BDNF
Sustained synaptic activity leads to long-lasting
Dendritic polymerization of actin cytoskeleton
spine  enlarged dendritic spine.
Arc

Ca++
polymerised
actin
mRNA
Nucleus
(DNA)
• Depends on brain-derived neurotrophic soma
factor (BDNF) and transcription of mRNA.
• New cytoskeleton provides scaffold for routine trafficking of extra AMPA-Rs
• Enlarged spine more effective and more stable. May even lose plasticity.
(Small spines learn; large ones remember.)
 Maintaining synaptic change:
Self-perpetuation
GluA2 GluA1
Change in make-up of AMPA receptor:
Initially, AMPA receptors containing
GluA1 subunit.
Long-term up-regulation of synapse PK
involves replacing GluA1 with GluA2
subunit in AMPA receptors (captured
from the extra-synaptic pool). GluA2
GluA2 doesn’t support further potentiation.

A new type of PK that is self-activating (lacks inhibitory unit that turns it off)
 Overview
1. Generating a memory trace (~1 min)
Initial strengthening of synapse:
PKs that “traffic” local AMPA-Rs back to synapse.

2. Stabilising the memory trace (needs 10-15 min)

Breaking down and rebuilding cytoskeleton.
Strengthening bridge between pre and post-synaptic
membranes (CAMs).

3. Consolidating the memory trace (2-4 hr)

Synthesising proteins (from local RNA and newly transcribed
RNA from nucleus) for new AMPA receptors;
Building up cytoskeleton to promote spine growth.

4. Maintaining the memory trace (4+ hr)

Changing type of AMPA receptor (GluA1 replaced with GluA2)
Synthesis of PKs that remain active (self-activating).