Hasaeam Cho Bio NanoStructure Lab of Prof MC Choi I will summary the paper Contents Introduction Bursting midbrain DA neurons beyond RPE signal Primate studies Rodent studies ID: 553005
Download Presentation The PPT/PDF document "BURSTING" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
BURSTING
Hasaeam
Cho
Bio-
NanoStructure
Lab
of Prof. MC Choi.Slide2
I will summary the paper …Slide3
Contents
Introduction
Bursting midbrain DA neurons – beyond RPE signal
Primate studies
Rodent studiesBeyond phenomenology – how bursts and pauses are generated in DA neurons
Afferent inputs controlling bursts in DA neuronsAfferent inputs controlling pausesIntrinsic conductances in DA neurons as gates for burst and pause controlIn vitro dynamic clamp approaches to channel function in DA bursting
Conclusions
References
1
2
3
4
5
6Slide4
STEP 1
INTRODUCTIONSlide5
DA midbrain neurons
Dopamine (DA) midbrain neurons
project to several striatal and cortical target areas
Are essentially involved in important brain functions such as
Action selectionMotor performance
MotivationReward-based learningWorking memoryCognitionSlide6
Reward Prediction Error RPE
The reward prediction error (RPE)
The difference between expected and actually delivered rewards
Prediction Error = actual reward – expected rewardSlide7
RPE
The reward prediction error (RPE)
The difference between expected and actually delivered rewards
It’s
Positive when the reward
> expectationPositive RPE is expressed as a phasic increase of firing above the tonic background rate.
It’s
Negative
when the reward
< expectation or not delivered at all (reward omission)
Negative RPE is expressed as a transient reduction of firing frequency below
background rate or even by a period of complete electrical silence (a pause)Slide8
RPE
Midbrain DA neurons
Short (<500
ms
) bursts of high-frequency
In vivo, occur time-locked (>50-300 ms after) to either unexpected reward delivery or, after learning, sensory cues that predict upcoming reward delivery within the next few seconds.Quantitative analysis revealed
the cue-related intra-burst firing frequency was associated with both
the expected reward amplitude and expected probability of delivery,
and the intra-burst frequency after reward delivery, RPE.
Schultz, 2007Slide9
Reinforcement-learning theory
The reinforcement
-learning
theory
Both negative and positive RPEs act as teaching signals.
Most likely
by changing synaptic
weights
of glutamatergic cortico-striatal synapses on the
most prominent target neurons of midbrain DA neuronsVia altering the occupancy and signaling of postsynaptic D1- and D2-type receptors.Slide10
Excitation, Inhibition, Disinhibition
Excitation
Excitatory postsynaptic potential (EPSP)
Causes depolarization
Inhibition
Inhibitory postsynaptic potential (IPSP)Causes hyperpolarizationDisinhibitionA temporary loss of inhibitionSlide11
STEP 2
BURSTING MIDBRAIN DA NEURONS – BEYOND RPE SIGNAL
Burst firing of DA neurons under diverse behavioral contexts in awake animals.Slide12
Primate studies
Burst
firing
was
induced by additional cues.As aversive, a blow of cold air to the eyes that triggers a protective blink response
In many previous studies have characterized dopamine neurons as a functionally homogeneous population.However, the largest population 40% of DA neurons did not
show phasic response to ACS
(ACS = air-puff-predictive conditioned stimulus).
But, displayed typical responses to reward-predicting cues.
An even smaller 10% responses to unexpected
airpuffsdorsolateral DA subpopulation showed large and relatively short (100
ms) phasic burst.Slide13
Primate studies
Matsumoto and
Hikosaka
, 2009Slide14
Primate studies
The midbrain DA population might be indeed relatively
uniform
in their responses to unexpected reward and reward-predicting cues,
But displays a topographically organized diversity in response to other salient events.
which might not directly instruct RPE-learning but initiate. E.g. orienting responses.Slide15
Primate studies
The recent wave of studies in awake primates
have significantly widened the functional context for burst firing
among different types of DA neurons as well as within the burst firing itself.
However, have remained descriptive and phenomenological
Because pharmacological or even optogenetic tools have not yet been used.Slide16
Rodent studies
SFB% : the percentage of spikes within a spike train
that were fired within bursts spikes fired in
bursts
The
degree of “burstiness” of DA neuronThe start (ISI < 80 ms) and stop conditions (ISI > 160
ms
) of burst firing, suggested by Grace and
Bunney
.With this criterion, bursting became quantifiable.Therefore, alternative burst detection methods have been introduced in recent years relate burst firing and pauses to the stochastic properties of the spike train
are independent of the absolute firing frequency.Also, DA recordings in awake rodents, either freely moving or head fixed are increased.Slide17
Rodent studies
In summary,
Cue- and reward-associated burst firing is present in some,
b
ut never all
in recent awake rodent studies even when ruled out by cell type-specific optogenetic tagging.Thus, functional diversity among DA VTA neurons appears to be the norm.Burst firing have also been recorded in behavioral settings not directly related to reward-driven classic or operant conditioning paradigms.
Recordings in awake rodents identified
a diverse phenomenology of burst firing associated with salient sensory cues, delivery of reward and the control of action sequences.
Ventral tegmental area; VTA, substantia
nigra
(pars compacta); SN(C)Slide18
STEP 3
BEYOND PHENOMENOLOGY – HOW BURSTS AND PAUSES ARE GENERATED IN DA NEURONS
Burst firing of DA neurons by synaptic excitation and GABAergic disinhibition.Slide19
Beyond phenomenology
The behavioral importance of bursts and pauses
Can only be appreciated from studies done in awake behaving animals
Most underlying mechanisms arise from studies
in anesthetized rodents
or in vitro brain slice studies.DA neurons exhibit spontaneous bursts, even in anesthetized animal.Much of studies of these spontaneous bursts.
Mechanisms in anesthetized preparations and in awake behaving animals
No guarantee that they are identical.
S
ome evidence that they might be very similar.Spontaneous bursts are indistinguishable
from reward related bursts in duration, firing frequency, and other structural features.The firing ratesOf single spike (
non-bursting), about 0-10 Hz.During bursts in vivo, rates up to 50 Hz.Slide20
Afferent inputs controlling bursts in DA neurons
Excitatory synaptic
input
in vivo
is necessary for bursting in DA neurons.
The synaptic inputs to midbrain DA neuronsPenduculopontine nucleus (PPN)/Lateral dorsal tegmentum (LDT)In vivo disinhibition of PPN increased the burst firing of putative DA neurons in the SN and VTA by about 50%.PPN inhibition reduced burst firing by about 50%
Subthalamic
nucleus (STN)
GABA-mediated disinhibition of the STN leads to increased burst firing in a subpopulation of DA neurons in an NMDA-sensitive manner.
In contrast to excitatory input, it has been estimated that
70% of all afferents to DA neurons are inhibitory
.Slide21
Zweifel
et al., 2009Slide22
Afferent inputs controlling pauses
in vivo
, IPSPs generated by GABAergic synaptic inputs are the obvious candidate afferents.
IPSPs are capable of delaying action potentials in a time-dependent way.
The powerful somatic oscillatory
currents drive the single-spike firing.The cell will fire again unless there is strong inhibition of long duration.Pauses are
only
possible with a synchronized increase in inhibition from significant GABAergic input or a decrease in tonic excitatory input.Slide23
Afferent inputs controlling pauses
Cohen et al., 2012Slide24
STEP 4
INTRINSIC CONDUCTANCES IN DA NEURONS AS GATES FOR BURST AND PAUSE CONTROL
Control of burst firing by distinct potassium channels in DA neurons.Slide25
Gates for burst and pause control
Two complementary approaches
used to study functional contribution of postsynaptic channels to DA bursting.
In vivo
extracellular
approachis to pharmacologically or molecularly modulate channel activityand then analyze the related changes in burst firing parameters.
I
n vitro
intracellular
brain slice approachprovides a high level of experimental controlis to induce bursts in DA neurons by stimulation of afferents
and then assess underlying channel mechanisms.Slide26
Gates for burst and pause control
ATP-sensitive potassium (K-ATP) channels
Necessary for bursting in vivo
for a medial subpopulation of SNC neurons.
In vitro
, selective opening of K-ATP channels in the presence of tonic NMDA receptor stimulation is sufficient to switch medial SNC neurons to a burst-firing mode.Enable
NMDA-mediated bursting of medial DA SN neuron
in vitro
and in vivo
.Slide27
Schiemann
et al., 2012Slide28
Schiemann
et al., 2012Slide29
Gates for burst and pause control
T
he bio-physical mechanisms is
not yet
clearhow K-ATP channel opening enables burst firing
which in vivo upstream mechanism controls K-ATP channel open probabilities in DA neurons Behavioral data suggest that K-ATP channel-mediated bursting in medial DA SN neurons is important for explorative behavior.Slide30
STEP 5
IN VITRO
DYNAMIC CLAMP APPROACHES TO CHANNEL FUNCTION IN DA BURSTING
Realistic burst firing dissected by dynamical clamp techniques.Slide31
Dynamic Clamp technique
A method that uses computer simulations to introduce virtual
conductances
into real neurons.
representing a hybrid between computational models and biological neurons.
can be used to study specific parameters of NMDA, AMPA, GABA, and other channel conductances in shaping bursts and pauses in DA neurons.Slide32
In vitro
dynamic clamp approaches to channel function
NMDA receptor
conductance
is capable of following the activity of DA neurons at bursting
rates.undergoes
block
, following each AP within a burst
.
helps to remove sodium channel inactivation.
allows full hyperpolarization during the repolarization following each AP within a burst.unblocks during the depolarizing phase prior to the next AP, increasing the speed of depolarization.Therefore, the voltage dependence of NMDA receptors allows dopaminergic neurons to fire at rates higher than those possible.N-methyl-D-aspartate (NMDA) Slide33
In vitro
dynamic clamp approaches to channel function
Deister
et al., 2009Slide34
In vitro
dynamic clamp approaches to channel function
Lobb
et al., 2010Slide35
In vitro
dynamic clamp approaches to channel function
Dynamic clamp experiments show that
Realistic burst
firings
the NMDA receptor conductance is capable of following the activity of DA neurons at bursting rates .Intrinsic cellular dynamics are an important part of bursting in the dopaminergic neuron.Future studiesMore complete burst mechanism by incorporating synaptic kinetics with the ion channel kinetics
how specific excitatory inputs may differentially affect the firing pattern of DA neurons.
more data regarding dendritic ion channel distributions.Slide36
In vivo study
In vivo studies show that
local
application of a variety of GAB
antagonists
by
pressure ejection onto the recorded neuron shifts the firing pattern of the DA neuron from a single-spiking mode to a bursting
one
tonic
GAB
receptor activation suppresses burst firing.However, it would need to overcome a substantially increased conductance and block.These limitations can be overcome by a mechanism whereby bursting is controlled by disinhibition, and pausing is controlled by phasic removal of excitation. Slide37
STEP 6
CONCLUSIONSlide38
Highlights
STEP 2
Burst firing of DA neurons under diverse behavioral contexts in awake animals.
STEP 3
Burst firing of DA neurons by synaptic excitation and GABAergic disinhibition.
STEP 4Control of burst firing by distinct potassium channels in DA neurons.STEP 5Realistic burst firing dissected by dynamic clamp techniques.Slide39
CONCLUSION
Transient changes of firing in midbrain DA neurons have received great attention.
Recent work broadened the behavioral framework of these changes both in non-human primates and rodents.
However, up to now most mechanistic studies are currently limited to
in vitro
preparations.Future in vivo studies are likely to advance mechanistic understanding.With the advent of optogenetic tagging of distinct DA neurons in the midbrain and the potential to selectively drive synaptic inputs,
However, intracellular
in vivo
recordings of DA neurons are needed to improve our mechanistic understanding of burst and pause firing of defined DA neurons in awake behaving animals.Slide40
REFERENCES
Figure 1
Generating bursts (and pauses) in the dopamine midbrain neurons, C. A.
Paladini
and J. Roeper
, Neuroscience (2014)Step 1 Introduction - Figure 1, 2, 3, 4, 5, 6, 7http://www.medizinische-fakultaet-hd.uni-heidelberg.de/Simon.102039.0.htmlPPTs of Brain Science FundamentalFigure 2 of Schultz, 2007a
Step 2 - Figure 7, 8
Figure 1, 2 of
Matsumoto and
Hikosaka, 2009Step 3 – Figure 9Figure 3 of
Cohen et al., 2012Step 4 – Figure 10, 11Figure 2, 3 of Schiemann et al.,
2012Figure 12http://blog.gametize.com/2014/03/7-smart-ways-to-improve-learning-with-gamification-part-2/boring/Step 5 The dynamic clamp – Figure 13, 14, 15, 16http://rtxi.org/docs/tutorials/2014/12/05/dynamic-clamp/Figure 1 of Dynamic clamp with StdpC software, Ildiko Kemenes et al., 2011, Nature ProtocolsFigure 6 of Diester et al., 2009Figure 1 of Lobb et al., 2010Slide41
TANK YOU!