One of the most remarkable things about the brain
is its ability to change in response to sensory experience.
During development, sensory experience specifies how the neural circuits will wire
in order to be able to process information in a sensible manner
and throughout our lives, we learn through changing in the neural circuits
inside the brain.
The changes inside the brain are largely occurring at synapses,
so-called <i>synaptic plasticity</i> changes how synapses function,
and that's the topic of today's video.
Much of the synaptic plasticity that takes place in the brain
is so-called <i>post-synaptic plasticity</i>
where there are changes in the receptor composition
of the post-synaptic membrane
at excitatory glutamatergic synapses.
We've already discussed pre-synaptic plasticity,
and that's the dominant form of plasticity on the millisecond and second time scale
in the brain, but long-term forms of plasticity,
the types of changes in the brain that might underlie memories
that last for our lives, or at least for days,
that is thought to largely occur
through changes in the post-synaptic density.
And that's what we'll focus on today.
We'll think about changes that are largely driven by the NMDA receptor
that induces synaptic plasticity
and its changes in the AMPA receptor composition
that largely underlie the long-term functional differences at synapses.
Bliss and Lomo discovered Long Term Potentiation in the hippocampus.
They found that if they stimulated axonal fibers and recorded
synaptic potentials in the hippocampus, they could evoke a stable baseline
where each time they stimulated some fibers,
they would get excitatory post-synaptic potentials,
release of glutamate activating AMPA receptors, primarily.
They then made a LTP induction protocol.
LTP stands for Long Term Potentiation,
and they did this through high frequency stimulation
of the axons that they were stimulating,
so they would stimulate axons, say, at 100 Hz for one second.
That's a very typical protocol for inducing Long Term Potentiation.
So they did that for one second, maybe a few times.
You might want to repeat that 100Hz tetanic stimulation a few times,
and then when they returned to stimulating as they'd been doing here
during the baseline period, they found that the response amplitude
had dramatically increased, and that's the phenomenon of Long Term Potentiation.
That increase in the PSP amplitude can be very long-lasting.
It can last, basically, for as long as you're able to record.
In this case, we've plotted out two hours of recordings,
but if you are able to record for much longer periods of time,
these changes in synaptic efficacy
can last for many days when recorded in vivo.
So, this is a remarkable situation,
where a brief change in the activity of the neural circuits
can induce long-lasting and large changes
in the fixing of synaptic transmission.
How does that occur?
Experimentally, you can induce Long Term Potentiation
by a variety of ways.
One of the most common ones, the one I just pointed out,
is this high-frequency stimulation of many axons.
You might put, for example, field stimulation electrodes
where you might stimulate hundreds of axons at the same time.
You stimulate these at high frequencies, and during that stimulation period
where you stimulate every 10 milliseconds,
what's happening post-synaptically is that you get EPSP's coming in.
They summate, and often you get action potential firing
coming in on top of this.
So there's post-synaptic depolarization
and often, action potential firing of the post-synaptic neurons.
That's one of the most popular ways in which Long Term Potentiation
can be induced.
There are also other ways in which you can induce Long Term Potentiation
where you don't change the way that you stimulate the pre-synaptic cells,
and you just purely make post-synaptic changes.
One way in which you can do that is in your whole cell recording electrode.
If you're recording with a glass electrode the membrane potential of a cell,
you can simply inject current
and you can then, of course, depolarize the cell's membrane potential,
so during a baseline period,
the membrane potential of the cell might be sitting at, say, -70mV
You get EPSP's driven here that are largely driven by the AMPA receptors.
You can then depolarize the post-synaptic neuron.
That then unblocks the NMDA receptor.
You'll remember that there's a voltage dependent Mg block of the NMDA receptor.
If you now have depolarized the cell
and you gave the same stimulation to the pre-synaptic cell,
you can now induce long-lasting NMDA currents.
So, depolarizing the post-synaptic cell is another way
in which you can induce Long Term Potentiation.
And finally, you can also, again, it's a similar type of thing,
you are also injecting currents,
but here it's not a long-lasting depolarization,
but you can simply inject brief currents
to find action potentials in the post-synaptic neuron.
If we are recording an EPSP,
then we can make brief current injections in the post-synaptic cell
to depolarize and find action potentials here
during the actual EPSP that's occurring in the post-synaptic cell.
So there are two ways here in which we can depolarize a cell.
One is just simply just taking the membrane
potential to depolarized potentials,
and you may not meet any action potential firing.
Another way is to get brief current pulses,
find action potentials at the same time
that an EPSP is arriving in the post-synaptic neuron.
What all these different methodologies
for inducing Long Term Potentiation have in common
is that they induce activation of the NMDA receptor.
That's the key event for most forms of post-synaptic Long Term Potentiation.
We need to depolarize the cell, get the Mg out of the NMDA receptor,
and what that then allows is for calcium influx to occur.
You remember that AMPA receptors are not permeable to Ca,
but the NMDA receptor, in addition to its interesting voltage dependence,
also has Ca permeability so when we depolarize the cell
either through high-frequency stimulation,
injecting depolarizing current,
or firing action potentials in the post-synaptic cell
that unblocks the NMDA receptor.
Glutamate is still bound to the NMDA receptor,
and now, cytosolic Ca can increase in that post-synaptic area.
That, in turn, causes activation of protein kinases.
Many [cycling] molecules inside cells
are sensitive to the cytosolic Ca ion concentration
and at the post-synaptic density, there's a particularly high density
of a Ca dependent protein kinase
called *Calcium/calmodulin-dependent
protein kinase II* or CaMKII,
and activation of that molecule downstream of Ca increases
is thought to be the most important signal
that then, ultimately, leads to an increase in synaptic efficacy.
This is how we envisage the situation at the synapse.
In the baseline situation action potentials will propagate down the axon,
cause the release of neurotransmitter, glutamate
from the pre-synaptic specialization,
and that will largely activate AMPA receptors.
We imagine here that under these baseline conditions,
the membrane potential of the cell during rest
is maybe around -70 mV.
The NMDA receptor is blocked.
There's Mg that's sitting here at negative potentials,
and the released glutamate activates AMPA receptors,
and that then gives rise to an excitatory post-synaptic potential.
During the LTP induction protocol,
we're depolarizing this post-synaptic neuron.
Either because we're making high-frequency stimulation
of many different axons, and so, the whole post-synaptic neuron
is seeing a lot of glutamate across its entire dendritic operization,
and thus depolarized, or because experimentally
we've depolarized the membrane potential
to, say, zero mV, which is a typical manipulation to do.
That then gets the Mg out of the NMDA receptor
so the Mg is now gone.
The NMDA receptor can then contribute
when glutamate is bound to the NMDA receptor.
We're at these depolarized potentials
and now Ca can flow in though the NMDA receptor
and as mentioned, there's a high density
of Calcium/calmodulin-dependent protein kinase II
in this post-synaptic density region,
and this kinase adds phosphate groups to a variety of different targets.
One target is the AMPA receptor itself,
so it can add a phosphate group onto the AMPA receptor
and that seems to change the conductance of the AMPA receptor
and make it larger, and that in itself might contribute
to enhancing the amplitude of the post-synaptic potentials.
However, what turns out to be the more important contribution
to post-synaptic longterm plasticity
is the recruitment of additional AMPA receptors onto the post-synaptic area.
So AMPA receptors are actually floating around on the plasma membrane
at many locations across the dendritic operization,
and they're mobile.
They diffuse in the two-dimensional plane of the plasma membrane.
One of the things that seems to happen during Long Term Potentiation
is the ability of the post-synaptic density
to bind AMPA receptors and hold them
tightly locked in place here at the post-synaptic density.
So, one of the major features that CaMKII is thought to do
is to make more slots, as it were, binding partners here
where AMPA receptors can bind tightly at this post-synaptic area here
and you then simply recruit more AMPA receptors
into the post-synaptic area, so whereas before we might have had
ten AMPA receptors sitting in the post-synaptic density,
after Long Term Potentiation, we might now have 20 AMPA receptors
that have been driven in and bound here to the synapse
through the activation of these cycling cascades
downstream of Phosphorilation via CaMKII.
The net result is that we have more AMPA receptors
where we now return to the situation where we simply measure the potential,
so we've done the LTP induction protocol,
we depolarized here but only temporarily,
while we were inducing synaptic plasticity.
We then return to the same conditions as we had in our baseline recording period
but now the post-synaptic potentials are much larger
because we simply have more AMPA conductance
at the present, post-synaptically.
Now, if Long Term Potentiation were the only thing that went on in the brain,
synapses would always get larger and larger,
and that clearly isn't feasible, for all the synapses in the brain simply to grow.
So it turns out that, luckily, there's a converse process
called Long Term Depression that also takes place in the brain.
Through the process of Long Term Depression the baseline synaptic efficacy,
the post-synaptic potential amplitude can be made to grow smaller
through a so-called Long Term Depression induction protocol.
We start off with EPSP's that are relatively large in amplitude.
We do the Long Term Depression induction protocol,
and now, in a stable way across many hours or days,
we can decrease the amplitude of these excitatory post-synaptic potentials.
In order to induce Long Term Depression,
typically, many axons are stimulated at an intermediate frequency of, say, 1 Hz
so the baseline EPSP's might be stimulated with an interval of ten seconds,
so we record baseline EPSP's.
Then there's a period of stimulating at 1 Hz for many minutes,
maybe for five or ten minutes, there's a constant 1 Hz stimulation
of the same axons, and then we return to stimulating every ten seconds,
and then it turns out that the EPSP amplitude
has decreased in many different synapses across the brain.
So, that's one way in which we can induce Long Term Depression,
is through 1 Hz stimulation of many axons simultaneously.
Another way of doing it is without changing the inter-stimulus interval
of the pre-synaptic axons,
but again just making post-synaptic manipulations
where one can inject a weak depolarizing current
through the whole cell recording pipette.
So before, for Long Term Potentiation, we might have been depolarizing the cell
to 0 mV, and then strongly unblocking the NMDA receptor
In the induction of Long Term Depression,
you might depolarize the cell, say, to -50 mV.
A sort of an intermediate situation where one might expect some degree
of unblock of the NMDA receptors,
but nowhere near as big an unblock of the NMDA receptor
as you'd get at 0 mV.
So the NMDA receptors are presumably partially activated,
and indeed, many forms of Long Term Depression
require an NMDA receptor activation,
but it's at a smaller scale than for inducing Long Term Potentiation.
Downstream of NMDA receptor activation,
it's a little bit unclear what happens.
Ca rises, or at least cytosolic Ca is required post-synaptically,
but it's also possible that the NMDA receptor
has additional cycling capabilities
and it may be that protein-protein interactions between NMDA receptors
and other cyclying proteins in the post-synaptic density
are actually the ones that affect the expression of Long Term Depression.
It's also known that the activation of protein phosphatases
is essential for Long Term Depression,
and in particular, the protein phosphatase cancineurin
is essential for Long Term Depression.
Phosphatases do the opposite of kinases,
so a kinase adds a phosphate group to a protein,
a phophatase does exactly the opposite:
it'll take that phosphate away from the protein.
So in some respects, this process
looks like the opposite of Long Term Potentiation,
but in fact, there are interesting and important differences
between LTP and LTD.
They're not directly reversals of each other.
In terms of what happens after the NMDA receptors have been partially activated
and the protein phosphatases have been activated,
presumable, there's a dephosphorilation event that occurs,
and in particular it seems that, in part at least,
directly dephospohrilating AMPA receptors
appears to be a key aspect to Long Term Depression.
After that, the AMPA receptors are simply removed from the synapse at least in part.
And so, you start off with a situation where we might have many
AMPA receptors post-synaptically,
and after the Long Term Depression protocol,
depression, we end up with fewer receptors post-synaptically,
and so, we have less conductance,
and therefore, we get smaller post-synaptic potentials.
So that's the basis of Long Term Depression.
There's another form of synaptic plasticity
that probably relates closely to NMDA receptor-dependent Long Term Potentiation
and NMDA receptor dependent Long Term Depression that we've been talking about,
but this is a mechanism that has at least a more physiological bend to it
in the sense that it relies up`on action potential firing
and the precise timing of those action potentials,
so now we're not talking about long-lasting depolarizations
or high frequency stimulation manipulations that experimentally, of course,
can be carried out in a dish,
but we have no idea whether they actually take place normally
under physiological conditions inside the brain.
Spike timing-dependent plasticity relies on the precise temporal relationship
of an action potential in a pre-synaptic neuron
and an action potential in a post-synaptic neuron,
and of course, in the brain, we know
that both pre-synaptic and post-synaptic cells
fire action potentials,
and there was an interesting discovery by Henry Markram and Beth [inaudible]
that the precise timing of action potentials
in the pre-synaptic and post-synaptic neurons
can make a major impact upon the synaptic efficacy.
In particular, what they found was that if we imagine
a synaptically connected pair of neurons
where we have an action potential in a pre-synaptic neuron,
that then releases glutamate onto the post-synaptic cell,
causing, of course, a unitary excitatory post-synaptic potential.
They then began to inject currents into the post-synaptic cell
at the same time that there was also a synaptic input,
so we have the pre-synaptic action potential that releases glutamate
onto the post-synaptic cell as before
causing the post-synaptic potential,
but now, in addition to that EPSP there is also a current
that's being injected into the post-synaptic cell,
and that then makes it fire an action potential
riding on top of the post-synaptic potential.
When one then goes back and looks at what happens at the EPSP
when we just stimulate the pre-synaptic cell,
now the EPSP has grown.
It's important to note that typically this process, this pairing process
of pre-synaptic followed by post-synaptic action potential,
is typically repeated many times, maybe a hundred times,
in order to generate a robust increase in the efficacy of the connection
between the pre-synaptic and the post-synaptic neuron.
And so, we go from a small baseline EPSP to a large EPSP
after the Spike Timing Dependent Plasticity protocol has been run,
and in particular it's with a delay in the post-synaptic spike
relative to the pre-synaptic spike.
At these so-called positive delta-T intervals,
that causes an increase in the efficacy of the EPSP.
Remarkably, if you move the timing of the action potential
to be just ten ms before the pre-synaptic action potential,
then the plasticity reverses sign.
So here again we're looking at a unitary synaptic connection,
action potential in the pre-synaptic cell,
post-synaptic potential in the post-synaptic cell.
An action potential can then be evoked in the post-synaptic neuron
by injecting current into the post-synaptic cell
firing an action potential, and you can do that at a time
that precedes the action potential in the pre-synaptic cell,
so the pre-synaptic cell evokes an EPSP just like it did before,
but now that EPSP is preceded by an action potential here
during the spike timing-dependent plasticity induction period
where again you might pair a hundred times the action potential and the EPSP,
but now the timing is very different to before.
Here, the action potential was riding on top of the EPSP.
Now the action potential precedes the glutamatergic input onto the cell.
And what is found is that the EPSP now decreases in amplitude
and so we now have a decrease in the efficacy of the EPSP
when we've reversed the timing of the pre- and the post-synaptic cells,
so this is at so-called negative delta-T intervals
where the post-synaptic cell fires before the pre-synaptic cell.
By measuring what types of plasticity happen when you vary this time interval
between pre-synaptic and post-synaptic spikes,
you can build up a timing of the plasticity of the change
in EPSP amplitude, so that's what's plotted here on the Y axis
where positive values here indicate increased enhanced EPSP size
and the negative values are decreased amplitudes in the EPSP.
Here on the X axis we're looking at the difference in the timing
between the post-synaptic spike and the pre-synaptic spike
at positive times, that's when the post-synaptic cell fires
after the pre-synaptic cell, and that then gives rise
to enhanced synaptic plasticity.
When we have the timing the other way around,
the post-synaptic cell fires first.
That gives rise to decreases in the EPSP.
In this temporal relationship to synaptic plasticity,
the so-called spike timing-dependent plasticity (STDP)
has been now found at a large variety of synapses in the brain,
so it appears to be an interesting, and perhaps general, rule.
What's remarkable is really the short timescale here
where we're thinking about ten, twenty ms and maybe 50 to 100 ms on this side here
where on this millisecond time scale,
the order of the action potential firing makes a very big diference
to the types of plasticity that are involved.
To some extent, this makes sense,
so this EPSP didn't just come at a random time.
It came at a time when the neuron actually fired action potentials.
So one might then think that this input to the cell was useful
and something that should be kept and maybe even strengthened,
and that's what happens here in spike timing-dependent plasticity.
Apparently useful inputs that contribute to firing the post-synaptic cell
get strengthened.
In this scenario, the action potential occurred before the glutamate was released
from the pre-synaptic neuron.
Clearly, the firing of the post-synaptic cell
has nothing to do with what the pre-synaptic cell was doing.
This input from the pre-synaptic cell
might not be a particularly useful input for that cell.
It certainly doesn't contribute to this artificially flat spike
that's done under these experimental conditions.
So, the cell may think that this synapse
is not a particularly helpful one for the function of the neuron,
so it may be the reason for why the depression of that EPSP occurs
when the action potentials occur
in an uncorrelated way with the pre-synaptic neuron.
Changes in the post-synaptic receptor composition,
the number of AMPA receptors in particular
that are present at the post-synaptic density
are an important way in which the synapse strength can be regulated
between two different neurons.
It's the activity of these neurons, and in particular the depolarization
of the post-synaptic cell
concomitant with the release of glutamate from the pre-synaptic cell
that is typically involved in setting the plasticity
and the direction of the plasticity that occurs in these different paradigms.
Strong depolarization or action potentials riding directly on top of EPSP's coming in
strengthen that synaptic connection,
and the strengthening appears to occur, at least in part,
through the insertion and recruitment of AMPA receptors
that are diffusing on the plasma membrane
but then bind onto post-synaptic density proteins
and get locked onto the post-synaptic density
when they can contribute to driving post-synaptic potentials.
On the other hand, weak depolarization
or action potentials occurring at times that are different
than when the synaptic inputs arrive
tend to cause a depression of the synaptic inputs,
so there's a strengthening of what appear to be useful inputs,
and there's a weakening of inputs that appear to not contribute
to the post-synaptic firing.