In contrast to the long range
GABAergic projection neurons
that we considered in the last video,
many GABAergic neurons in the brain
don't have long range axonal projections,
they have local axons that maybe stretch
some hundreds of microns and they're often termed
GABAergic interneurons for that reason.
They regulate the activity
of the local microcircuits in which they are imbedded.
The inhibitory neurons of the neural cortex are such local interneurons
and in general, they don't have long-range axonal projections.
The GABAergic neurons of the neural cortex are a small fraction
of the total number of neurons in the neural cortex.
85% if the neurons in the neural cortex are excitatory glutamatergic neurons
and that leaves about 15% of the cells
as GABAergic local inhibitory neurons.
Despite the fact that there's not so many of them,
they show a striking diversity in almost any feature
that you care to analyze.
The morphology of the cells is different,
their electric physiological properties are different,
their synaptic connectivity,
and perhaps, most importantly in terms of characterizing the cell types,
their molecular expression patterns are distinct and diverse.
In this video we're going to consider four different types
of GABAergic neocortical neurons.
We'll think about three types of GABAergic neurons
that are defined through their expression of different marker genes,
so we'll discuss GABAergic neurons that express parvalbumin;
we'll discuss others that express somatostatin;
and others that express vasoactive intestinal peptide.
These are non-overlapping expression patterns,
at least largely,
so they're different classes of neurons that don't overlap with each other.
Then finally, we'll talk about neuroglialform cells
that at this time don't have a good molecular marker for them.
We'll begin by talking about parvalbumin expressing GABAergic neurons.
In fact, we've already discussed these briefly in previous lectures,
when we were thinking about the action potential waveform.
The parvalbumin expressing cells
have fast-spiking intrinsic electric physiological properties
so if you have one of these cells on your recording electrode in a brain slice
and you inject some depolarizing current into that cell,
it'll depolarize and fire action potentials,
and if you inject a lot of current into it,
it'll fire action potentials at a very high rate,
some hundreds of Hertz are the maximum firing rates
of the fast-spiking GABAergic neurons.
If we zoom in on each individual action potential,
we'll see that they have fast waveforms,
and the A-P half width is something like 300 microseconds.
That contrasts with the glutamatergic excitatory pyramidal cells
where if you inject depolarizing current into them,
they certainly fire action potentials,
but the maximum firing rates are much lower
than the fast-spiking parvalbumin-expressing GABAergic neurons,
about an order of magnitude lower.
Each individual action potential in the glutamatergic cells
has a broad waveform lasting more than a millisecond.
The axonal output of the fast-spiking parvalbumin-expressing GABAergic cells
comes in two different flavors,
and there are two different types of PV, parvalbumin, expressing cells.
The major one, the so-called parvalbumin expressing basket cell,
sends its axon close to soma, and often it'll decorate the soma
with a basket-shape around the soma
causing a fast hyperpolarizing IPSP
mediated by GABA-A here at the soma,
or on proximal dendrites.
The other type of parvalbumin-expressing cell
specifically sends its axon to the axon initial segment
of the excitatory pyramidal cells.
These axial, axonic cells
or these chandelier cells, as they're also called,
then target their inhibition to a very specific place of the pyramidal cell,
the axon initial segment,
where action potentials are typically thought to initiate.
So, these cells specifically inhibit that area
and it may be that they leave dendritic integration
in their main dendrites and the soma of the cell largely untouched,
and they specifically edit whether the cell is able
to output an action potential or not.
A very interesting type of cell.
The parvalbumin-expressing GABAergic neurons,
especially this type of soma and proximal dendrite targeting cell,
form the largest family of the GABAergic neurons
in the neural cortex.
We can study the connectivity of these fast-spiking parvalbumin-expressing
GABAergic neurons in brain slices,
where we can record from different neurons simultaneously,
and we can then initiate action potentials in different cells
and study their synaptic connectivity.
This brain slice is from a specific type of transgenic mouse
that expresses green fluorescent protein in all the GABAergic neurons.
So we have green GABAergic neurons, and excitatory pyramidal cells
which form the majority of the cells here, are invisible,
except when they're being recorded from
where we have the red fluorescence from the recording electrode
that fills the soma and dendritic operizations.
And so, cell one and cell two are excitatory pyramidal cells,
and you might be able to make out their dendritic operizations here.
If we inject depolarizing current into these two cells,
they fire action potentials with broad waveforms
and at slow firing rates, typical of excitatory pyramidal cells.
Recording electrode three is sitting on a cell here that's turned yellow.
That's because it's a superposition of the red
from the recording electrode, and the green from the green
fluorescent protein of the mouse indicating that this is a GABAergic neuron
and upon injecting current into it,
we see that it's a parvalbumin and fast-spiking GABAergic neuron.
And so, we can study the synaptic connectivity
between these three different cells
by firing action potentials in each cell in turn.
If we evoke an action potential in cell one,
and we measure what happens in cell two and cell three,
we see that there's an excitatory post-synaptic potential
in our fast-spiking GABAergic neuron,
so these are synaptically connected.
Not shown is that there's no impact on cell two,
so cell one connects to cell three, but not to cell two.
We can find an action potential in cell two,
and we see that also causes an excitatory post-synaptic potential in cell three.
And so, both cell one and cell two are excitatory pyramidal cells,
provide input into the fast-spiking parvalbumin GABAergic neuron.
This type of connectivity is entirely typical.
We saw in last week's lectures that excitatory neurons in the neural cortex
connect to each other with something like a 10% connection probability.
On average, excitatory neurons and fast-binding GABAergic neurons
connect with something like 50% probability,
so the connectivity of excitatory neurons
is much higher onto these parvalbumin-expressing GABAergic neurons
than it is onto other excitatory neurons.
We can also evoke action potentials across any parvalbumin expressing neuron
and see what happens in these inhibitory cells,
and that's what we do here.
An action potential is fired in cell three,
a fast-spiking GABAergic neuron,
and that causes an inhibitory post-synaptic potential in cell one.
You can do the same here, fire an action potential in cell three,
and that also causes an IPSP in cell two.
So cell three inhibits both cell two and cell one,
and so there are bidirectional synaptic connections here.
In order to see these IPSP's
we have to depolarize the post-synaptic cell
and so here we're not at the resting membrane potential,
but we're depolarized to -50 mV, and same here.
We have to depolarize the cell to -50mV,
and that's because of course, the chloride reversal potential of the GABA-A receptor
is sitting at around -70 mV,
so at resting membrane potential of cell one or cell two,
you don't see that IPSP.
It's just a conductance, and the hyperpolarization
is only visible if you depolarize the cell.
This type of connectivity where we have excitatory neurons
that connect to other excitatory neurons with relatively low rates
and very strong connectivity to the inhibitory GABAergic neuron,
here we have about 50% connectivity,
and equally the inhibitory connectivity which is also strong,
we said this is a strong inhibitory loop.
If there's more excitation and excitatory neurons
that's rapidly going to be damped by this very strong inhibition,
and in fact, the whole system here, at least in layer two / three,
our primary somatosensory cortex, is strongly over-damped,
so it's actually very difficult to get polysynaptic explosive excitation
because of this very strong feedback inhibition loop
from the local parvalbumin-expressing fast-spiking GABAergic neurons.
The second type of GABAergic neuron that I'd like to consider
are the so-called somatostatin-expressing GABAergic neurons.
These neurons are particularly interesting
because many of them send their axonal arborizations
to a very specific region of the excitatory pyramidal cells.
They target distal dendritic regions, and in particular,
as we've seen in previous lectures,
many of the excitatory pyramidal neurons have extensive arborizations here,
right at the top of the brain.
So this is the outside of the brain and this the surface of the brain here
and there's sort of a layer of the brain called layer one
that has GABAergic neurons in it, but very few excitatory cells,
mainly just synapses,
and this is where the pyramidal cells extend their dendritic tuft region here
and that is a region that seems to be specifically targeted
by the inhibition of the somatostatin-expressing neurons.
There are very interesting inputs that arrive here in layer one.
For example, if we're thinking about the primary somatosensory cortex
of the mouse, then signals from the motor cortex
arrive here specifically in layer one.
So it may be that the somatostatin cells are specifically involved in gating
the input of motor information into the primary somatosensory cortex.
A very interesting role of this distal dendritic innovation
that somatostatin cells might drive.
Another fascinating feature of the somatostatin cells
is that they receive facilitating excitatory synaptic input
from the nearby pyramidal cells.
So, if this is a pyramidal cell that's somatically connected
to a somatostatin cell, if we fire a single action potential,
then the EPSP recorded here in the somatostatin cell
is relatively small and variable in amplitude,
but if we fire repetitive action potentials
at high frequency, say at 50 Hertz as in this example,
then the EPSP amplitude grows
and it becomes much larger than it was on the first action potential.
So this is typical of pre-synaptic facilitation:
an increase in the neurotransmitter release probability
that occurs at the pre-synaptic end of the specialization,
and this facilitation is very specific onto the somatostatin cells.
The same pre-synaptic neuron targeting a parvalbumin-expressing cell
will not show facilitation.
Rather, it will show something constant,
or perhaps even something that's depressing.
So there's something very specific about the synapses made here
onto somatostatin cells as opposed to parvalbumin cells,
even from the same axon of the same pre-synaptic neuron,
that give it remarkably different properties
in terms of what's sensed post-synaptically.
So there's something interesting occurring here
in the somatostatin cells where they seem to be integrating
bursts of action potentials from the pyramidal cells
over longer periods of time
when they can become strongly depolarized
in response to trains of action potentials
and that is unique to somatostatin cells.
There's no other cells in the neural cortex
that have such prominent synaptic facilitation.
The VIP-expressing GABAergic neurons are also extremely interesting.
VIP stands for vasoactive intestinal peptide,
and we don't know what the functional role of VIP is
in the neural cortex, but we can use it as a useful marker
for our cell type in the brain.
In the neural cortex, the VIP cells seem to have a specific role
in inhibiting other GABAergic neurons
so they release GABA here, they inhibit cells,
and the cells that they inhibit are not the excitatory cells,
but there are somatostatin cells,
and a little bit less, the parvalbumin cells.
And so, the VIP cells, rather than causing overall inhibition
in the neocortical microcircuit, actually cause excitation,
and they do that via disinhibition.
The last type of neocortical GABAergic neuron that we'll consider
is the so-called neurogliaform cell.
These are small, compact cells which have an extremely dense
axonal arborization close to the soma.
What's specifically interesting about the neurogliaform cells
is that along their axons they form, of course butons
and they release GABA,
but they don't have an obvious post-synaptic partner.
So these cells seem to release GABA into the extracellular space,
and they may then regulate the extracellular GABA concentrations,
and they can signal via GABA-B receptors
causing slow inhibitory post-synaptic potentials.
In this video we've seen that there are quite different types
of GABAergic neurons in the neural cortex
and they appear to serve quite different functions.
Parvalbumin expressing GABAergic neurons can fire at extremely high rates,
are highly connected to the excitatory neurons,
and probably mediate some form of fast feedback inhibition.
The somatostatin cells, on the other hand,
target distal dendritic elements,
and they receive facilitating input from the nearby excitatory neurons
and they presumably respond to quite different patterns
of activity in the neural cortex and furthermore, their inhibition
is of a rather different element in the circuit
than the parvalbumin cells.
The parvalbumin cells inhibit proximal areas close to the soma.
The soma itself acts on initial segment proximal dendrites, fast inhibition,
whereas the somatostatin cells are inhibiting a slower form of inhibition
at the distal dendrites.
We also talked about VIP expressing cells
that have an interesting role in disinhibition.
They inhibit other inhibitory neurons.
And finally, we mentioned neurogliaform cells
that might be involved in controlling the extracellular space
GABAergic concentrations, and mediate slower IPSP's
via the GABA-B receptor.
So there's a whole diversity of GABAergic neurons
and a challenge for the future is now to assemble these different cell types,
these different mechanisms,
and see how that functions in the living brain during behavior.