Membrane potential dynamics of GABAergic neurons in barrel cortex of behaving mice

Gentet et al. (Petersen lab) Neuron 65: 422 (2010)

Mac HooksJournal Club #5

MethodsGABAergic input suppresses AP firing in L2/3 pyramids

All whiskers except C2 trimmed

Awake head restrained GAD67-GFP knockin mouse; habituated to restraint

Location of C2 barrel via intrinsic optical imaging

In vivo 2p imaging of GFP+ GABAergic neurons

Whole cell recording (K-gluconate; Alexa 594; biocytin)

Recordings L2/3 (about 180-220µm deep)

Sac after recording for anatomy

GABAergic input suppresses AP firing in L2/3 pyramids

Slow, large amplitude Vm fluctuations

Note AP frequency shiftQuantification of AP frequency shift

Mode shift to “bursting” (2 or more APs) in gabazine

“Quiet wakefulness”

This shows that GABAergic transmission is capable of modulating firing, but not which circuits are engaged (note later that AP firing in excitatory is not affected by shift from quiet to active wakefulness)

GAD67-GFP knockin mouse Tamamaki et al. J Comp Neurol (2003) 467:60

Two lines:GAD67-GFPGAD67-GFP(Δneo): mated to CAG Cre mouse to eliminate loxP-flanked PGK-neo cassette (in case it affects expression)

(Almost) all GFP+ neurons are GABAergicFigure S1 from Petersen (left)

Tamamaki (below ,neocortex) )gives GFP+ with GAD67 (156/196) and no GFP+/GAD- neurons

GAD67-GFP knockin mouse:Interneuron Subtypes (Motor cortex)

Characterization of Interneuron Subtypes in Frontal/Motor Cortex

All GFP+ cells are NeuN+ (about 19.5% of cortical neurons)

L1/2

L2/3

L2/3

L5/6

GAD67-GFP knockin mouse:Interneuron Subtypes (S1 cortex)

Figure S5

Tamamaki et al.

marked for comparison

Whole-cell recordings of GABAergic neuronsClassifying 3 cell types

GFP+ GFP+

Ok. Should we ask for a more fine division of interneuron population?

AP half width AP frequencyResting VmInput resistance

Classifying 3 cell types, Quantified

Whole-cell recordings of GABAergic neurons

Fig S3 also adds more data ….

Behavioral modulation of Vm dynamics in GABAergic neurons

Examples

Tracking whisker angle (video; 20s limit)

Vm (aligned)Expanded to show subthreshold

Quantify:AP frequency

Vm

Variance of Vm

1-5Hz area (V dot Hz) of Vm fourier transform

Divide time into (Q) and (W)

Behavioral modulation of Vm dynamics in GABAergic neurons

3 cell types, quantified

Behavioral modulation of Vm dynamics in GABAergic neurons

Fast (subthreshold) membrane potential oscillations phase locked to whisking

Whisker angle

Average Vm(aligned to peak of protraction)

Peak amplitude of fast (subthreshold) membrane potential oscillations, plotted at the time in the whisking cycle where it

occursNO DIFFERENCES BETWEEN CELL

TYPES

Correlated activity of excitatory and inhibitory neurons during (Q) quiet wakefulness

Two excitatory neurons(Dual in vivo patch: distance averaged 140±19 μm; in a 300 μm diameter barrel, this could put one in

the middle and the other at the edge)

Cross correlogram

Essentially examining subthreshold correlation

Does whisker trimming affect correlation across cortex?

Correlated activity of excitatory and inhibitory neurons during (Q) quiet wakefulness

Highly synchronous slow oscillations

Interneurons then can’t drive the hyperpolarized

phase in pyramids …

Behavioral modulation of correlated activity

Behavioral modulation of correlated activity

Why not subdivided for FS/non-FS?

Large brief specific events

Action potentials in excitatory neurons are driven by large, brief, cell-specific depolarization

(e.g. events in excitatory neurons are not correlated)

Shuffled events

Histogram of relative spike times

(10ms bins)

Data

Essentially examining suprathreshold correlation

Action potentials in inhibitory neurons are driven by broadly synchronized depolarization

Synchronized depolarizations

Summary of Vm prior to APs

Spike triggered averages for …

Spikes triggered in excitatory neurons include events not present in …

…excitatory, or… FS interneuron

Whereas these events are more similar (e.g. broad depolarizations)

Vm slope in 20ms prior to spike initiation

Is there any way this could be some artifact of the cell type?

A Model for Behavioral Modulation of Vm Dynamics

But …

(3) L3→L3 connections are among the strongest excitatory local

connections … wouldn’t we expect the most synchrony in excitatory

neurons of L3 and L4? Or does the sparseness of connectivity explain?

(1) Is there any evidence for the shift in excitatory input from FS to Non-FS interneurons during activity?

Could we look for this in pathways (POm?) that might be more excited during active whisking?

(2) How does this model account for the shared broad depolarizations of FS and Non-FS interneurons? (is the “shared input” from L3 pyramids or elsewhere)

Feedforward inhibition engaged by thalamocortical inputCruikshank et al. (2010) Neuron 65:230

Connors’ data for VB feedforward to L4 is similar to what Petersen proposes for L3:

-general principle of feedforward from layer to layer?

-would POm input be more similar to ‘active’ wakefulness, while VB is more similar to ‘quiet’

wakefulness?