University of Southern California Dissertations and Theses

Imaging neuromodulator dynamics in somatosensory cortex

(USC Thesis Other)

Imaging neuromodulator dynamics in somatosensory cortex

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Imaging neuromodulator dynamics in somatosensory cortex

By
Jing Zou

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)

August 2022

ii

Acknowledgements
I would first like to thank my mentor Samuel Andrew Hires, without whom this
dissertation would be impossible. I really appreciate that you gave me the opportunity to
keep studying neuromodulators even though it is not your expertise. Thanks for being
willing to jump out of your comfort zone and explore the neuromodulator field with me. I
feel very lucky being able to study what I am passionate about. Also thank you for
patiently teaching me coding. When I joined your lab, I had zero coding experience. It
was not easy at the beginning when I started to do some data analysis. You
demonstrated how you did analysis in front of me, which really expedited my learning
process. You were also very supportive for everyone in the lab to have a good work life
balance. And that is one of the major reasons after 6 years in graduate school I still
have a lot of passion for science, even more passionate compared to when I first started
my PhD. Thank you for making my PhD journey so pleasant.
I would also like to thank two professors I did rotation with, Dion Dickman and
Mara Mather. I enjoyed my rotation a lot, during which I got to meet so many good
friends and also learned about different research fields. I am also grateful for my
previous and current committee members, Thai Truong, Mara Mather, David McKemy,
and Donald Arnold. Thank all of you for making time to be on my committee and giving
feedback on my research.
Thanks to all the lab members in Hires lab, Jinho Kim, Andrew Erskine, Jonathan
Cheung, Samson King, Phillip Maire, Tanisha London, B. Isaac Cohen, Chao Wang,
and Stefanie Walker. It is truly a great pleasure to work with all of you. I couldn’t ask for
iii

better labmates. I am very lucky to not only have you guys as my labmates, but also as
my friends. I had so much fun with y’all in and outside of the lab.
Thanks to all the friends I met along the way. There are too many to list here.
You guys are all in my heart. I am truly grateful for meeting every one of you. You guys
showed me so much love. My PhD life wouldn’t be this fun without you guys.
Thanks to my two housemates, Xiling Li and Ahyun Jung. Thank you both for
taking care of me. I know sometimes I can be really childish. Thanks for bearing with
me. You guys are like the older sister that I wish I had. Also thank you for introducing
your cute pets Leo, Neo, Popcorn and Oreo to my life.
Lastly, I want to thank my parents. Thank you both for being so supportive,
financially and mentally. Without you, I wouldn’t be able to travel around the world,
watch so many sports games and concerts during my PhD time. Thank you for always
being so understanding, always respecting my choice. Love you always.

iv

Abstract

Neuromodulators (acetylcholine, norepinephrine, dopamine and serotonin) play
important roles in a wide range of behaviors. And the malfunction of these four
neuromodulatory systems are found related to multiple brain disorders. The importance
of these neuromodulators are largely appreciated, but the temporal and spatial release
dynamics of these neuromodulators in the primary somatosensory cortex (S1) are still
left largely unrevealed. With the recently developed GPCR based genetically-encoded
neuromodulator indicators, we are able to monitor multiple neuromodulators release
dynamics in S1 in awake behaving animals. This thesis reviewed latest studies on
cholinergic, noradrenergic, dopaminergic and serotonergic systems, as well as the
methods being developed to monitor the neuromodulator release in the brain. Besides,
two projects are included in this thesis. One focused on the characterization of multiple
neuromodulator indicators in vivo performance under two photon microscopy. The other
one characterized the acetylcholine release dynamics during tactile-based associative
learning in detail. These studies provide insight into the different roles that these
different cortical neuromodulator releases play during associative learning.
v

Table of contents

Acknowledgements ....................................................................................................................... ii
Abstract ........................................................................................................................................ iv
List of Tables ................................................................................................................................ 1
List of Figures ............................................................................................................................... 2
Chapter 1: Introduction of acetylcholine/ norepinephrine/ dopamine/ serotonin systems in the
brain .............................................................................................................................................. 5
Acetylcholine ............................................................................................................................. 5
Norepinephrine ....................................................................................................................... 10
Dopamine ................................................................................................................................ 15
Serotonin ................................................................................................................................. 22
Chapter 2: Summary of methods to detect neuromodulators in vivo .......................................... 26
Chapter 3: Characterize multiple genetically encoded neuromodulator indicators in vivo
performance under two photon microscopy ............................................................................... 32
Summary ................................................................................................................................. 32
Results .................................................................................................................................... 33
Discussion ............................................................................................................................... 38
Chapter 4: Monitor acetylcholine dynamics in somatosensory cortex during tactile guided
associative learning .................................................................................................................... 40
Summary ................................................................................................................................. 40
Introduction ............................................................................................................................. 41
Results .................................................................................................................................... 44
Discussion ............................................................................................................................... 58
Materials and Methods ............................................................................................................ 62
Supplemental Figures ............................................................................................................. 69
Chapter 5: Concluding Remarks ................................................................................................ 73
References ................................................................................................................................. 77
1
List of Tables
Chapter 4
Table 1: Key resources table. 62

2
List of Figures
Chapter 1
Figure 1.1: Cholinergic projection neurons in the brain. 6
Figure 1.2: Sensory modality specific topographic organization of BF cholinergic
neurons. 6
Figure 1.3: BF cholinergic neurons respond to both positive and negative reinforcers. 9
Figure 1.4: BF cholinergic neurons activity and acetylcholine release in BLA aligned to
first lick. 10
Figure 1.5: LC-NE neuron projections in the brain. 12
Figure 1.6: LC has different sub-cell types. 12
Figure 1.7: LC-NE neurons showed suppressed activities during food retrieval. 15
Figure 1.8: LC-NE axon terminals respond to different sensory stimuli. 15
Figure 1.9: 4 dopaminergic pathways in the brain. 17
Figure 1.10: Midbrain dopaminergic neurons encode RPE. 19
Figure 1.11: SNc dopaminergic neurons respond differently to the aversive stimuli
based on their projection. 19
Figure 1.12: VTA dopaminergic neurons show heterogeneous response to reward and
aversive stimuli based on their projection at nucleus accumbens. 20
Figure 1.13: Midbrain dopaminergic neuron axon terminals at striatum have
heterogeneous response. 20
Figure 1.14: Serotonergic system in the brain. 23
Figure 1.15: DR serotonergic neurons showed increased response to sucrose (middle)
and food (right) intake. 24
3
Figure 1.16: DR serotonergic neurons showed ramped up activities during reward
waiting time and a phasic response to reward delivery. 25
Chapter 2
Figure 2.1: Microdialysis and FSCV diagram. 28
Figure 2.2: Genetically-encoded glutamate indicator. 29
Figure 2.3: GPCR-based neuromodulator indicators. 31
Figure 2.4: Kinetics for different neuromodulator indicators. 32
Chapter 3
Figure 3.1: Different GRAB indicator variants were expressed in the mouse
somatosensory cortex. 34
Figure 3.2: GRAB sensor can be stably expressed in mouse somatosensory cortex
across one month period with some extent of expression pattern change. 34
Figure 3.3: GRAB-ACh3.0 is photostable under two photon microscopy imaging. 35
Figure 3.4: Whisker-guided object localization task behavioral paradigm. 36
Figure 3.5: GRAB-ACh2.0 and GRAB-ACh3.0 comparison. 37
Figure 3.6: Different GRAB variants fluorescence change sorted by trial types (Blue: Hit,
Black: Miss, Red: Correct Rejection, Green: False Alarm) under two photon microscopy
while animals were performing the tactile-guided object localization task. 38
Chapter 4
Figure 4.1: Learning of whisker-guided object location discrimination and associated
motor actions. 46
Figure 4.2: Acetylcholine release in S1 varies with trial outcome. 47
4
Figure 4.3: Whisking drives acetylcholine release in S1. 50
Figure 4.4: Licking strongly drives acetylcholine release in S1. 52
Figure 4.5: Reward delivery does not drive acetylcholine release in S1. 54
Figure 4.6: Learning selectively potentiates acetylcholine release to choice-signalling
licks in S1. 57
SFigure 4.1: Pole out cue induced whisking in all trial types. 69
SFigure 4.2: Technical details of targeting and imaging. 70
SFigure 4.3: Whisking drives acetylcholine release in S1. 71
SFigure 4.4: Inter-lick interval distribution comparison between early and expert
sessions. 72

5
Chapter 1: Introduction of acetylcholine/ norepinephrine/
dopamine/ serotonin systems in the brain

Acetylcholine

Acetylcholine plays an important role in our central nervous system (CNS) and
peripheral nervous system (PNS). It functions as both a fast point-to-point
neurotransmitter and a modulatory neuromodulator. In the CNS, acetylcholine is
believed to majorly serve as a neuromodulator, which modifies the excitability of
neurons. The primary sources for acetylcholine in the brain are mainly from local
cholinergic interneurons and cholinergic projection neurons. Cholinergic projection
neurons are largely located in the basal forebrain (BF) complex (including vertical limbs
of the diagonal band of Broca (VDB), horizontal limbs of the diagonal band of Broca
(HDB), nucleus basalis magnocellularis (NB), and medial septum (MS)) and
pedunculopontine tegmental nucleus (PPT) / laterodorsal pontine tegmentum (LTD) in
the brainstem (Newman et al., 2012, Li et al., 2018, Figure 1.1). More recent studies
revealed that media habenula (MHb) also has cholinergic projection neurons (Ren et al.,
2011, Li et al., 2018). Among these cholinergic projection neuron hubs, BF has been
through extensive studies. Numerous studies have shown that BF cholinergic neurons
send projections to a broad range of cortical areas, hippocampus, entorhinal cortex,
basolateral amygdala (BLA) and so on (Mesulam et al., 1983, Zaborszky et al., 2015a,
Bloem et al., 2014, Kim et al., 2016, Li et la., 2018). Although BF cholinergic neurons
6
send broad projections over the brain, multiple lines of evidence showed that the BF
complex has topographic organization based on the projection target (Bloem et al.,
2014, Zaborszky et al., 2015b, Kim et al., 2016). For example, BF cholinergic neurons
show sensory modality specific topography, with cholinergic neurons in HDB preferably
projecting to primary visual cortex (V1), in anterior NB projecting to primary
somatosensory cortex (S1), in posterior NB projecting to primary auditory cortex (A1)
(Kim et al., 2016, Figure 1.2).

Figure 1.1 Cholinergic projection neurons in the brain (Figure from Newman et al., 2012).

7
Figure 1.2 Sensory modality specific topographic organization of BF cholinergic neurons
(Figure from Kim et al., 2016).

Apart from the large body of anatomical studies on the cholinergic system in the
brain, vast research on the function of the cholinergic system have been conducted.
Acetylcholine has been shown to influence diverse cognitive processes that span
different timescales, including neural plasticity (Kilgard and Merzenich, 1998, Froemke
et al., 2013, Jiang et al., 2016), reinforcement learning (Parikh et al., 2007, Chubykin et
al., 2013, Liu et al., 2015, Crouse et al., 2020), selective attention (Donoghue and
Carroll, 1987, Tremblay et al., 1990, Hars et al., 1993, Goard et al., 2009), arousal
(Parikh et al., 2007, Zhang et al., 2011, Reimer et al., 2016), sleep-wake cycles (Xu et
al., 2015), food intake (Mineur et al., 2011), and memory (Croxson et al., 2011, Bang
and Brown, 2009). The ways to understand the cholinergic system’s functions can be
roughly categorized to 4 routes: 1) Anatomy studies, comparing the cholinergic neurons
and cholinergic receptors between patients and normal people; 2) Pharmacology
studies, seeing the effects of different cholinergic drugs; 3) neural recording, directly
monitoring cholinergic neurons activities; 4) manipulation studies, exploring the
consequences of changing cholinergic neurons’ activities. Through anatomy studies, it
has been found that there is severe loss of cholinergic neurons in BF in Alzheimer’s
patients (Giacobini, 2003, Zaborszky et al., 2008). And a wide decrease of nicotinic
acetylcholine receptors (nAChR) in brain regions like cerebral cortex and striatum is
seen in Parkinson’s disease patients (Lange et al., 1993, Quick et al., 2004). These
anatomy studies imply the important role acetylcholine plays in cognitive performance.
Alternatively, numerous pharmacological studies also support the cognitive function that
acetylcholine plays. Nonspecific metabotropic acetylcholine receptor (mAChR)
8
antagonist scopolamine can induce delirium (Crow and Grove-White, 1973), impair
attention (Broks et al., 1988) and impair new memories encoding (Aigner et al., 1991).

With the developments of genetic tools to target cholinergic neurons, neural
recording tools with high spatial and temporal resolution, and the neural manipulation
tools, researchers are able to explore more detailed roles that acetylcholine plays in
various processes. For instance, accumulating studies have revealed BF cholinergic
neuron’s role in reinforcement learning. As described in work from Adam Kepecs group,
opto-tagging identified BF cholinergic neurons respond to both positive and negative
reinforcers (Hangya et al., 2015, Figure 1.3A, Figure 1.3B). Their followup study further
characterized BF cholinergic neuron activities during reinforcement learning, through
electrophysiologically recording opto-tagging identified cholinergic neurons, GCaMP
recording in BF cholinergic neurons, and directly monitoring acetylcholine release at
basolateral amygdala (BLA) (Sturgill et al., 2020). And they proposed that BF derived
acetylcholine encodes valence-free reinforcement prediction error. However, one of the
caveats of these studies is that they didn’t fully dissociate cue response from cue-
induced movement response, and reward response from licking response. Works from
Adam Kepecs group reported that there is no licking induced BF cholinergic neuron
activities, because the BF cholinergic neuron activities ramping up precedes the first lick
(Sturgill et al., 2020, Figure 1.4A). But it is possible that BF cholinergic neurons encode
first lick initiation, and this hypothesis is supported by the recent study from Kay Tye
group. They saw BF cholinergic activities and acetylcholine release in BLA in licking
trials even in the absence of actual reward (Kimchi et al., 2022, Figure 1.4B, Figure
9
1.4C). Since reward is always correlated with licking, quantitative behavioral analysis is
needed to help us dissociate true reward induced BF cholinergic neuron activities from
licking induced activities. In summary, with current novel neural monitoring tools, which
cholinergic neurons fire at what times during reinforcement learning can be identified.
But when drawing conclusions, more attention needs to be paid to animal behaviors,
which we can acquire from monitoring their body and facial movements. And the new
neural manipulation tools allow us to investigate the causal role that cholinergic neurons
play in different processes.

Figure 1.3 BF cholinergic neurons respond to both positive and negative reinforcers.
A) NB cholinergic neurons respond to water reward. B) NB cholinergic neurons respond to
punishment. (Figures from Hangya et al., 2015)

10

Figure 1.4 BF cholinergic neurons activity and acetylcholine release in BLA aligned to first lick.
A) BF cholinergic neurons activities in an odor reward association task (left) and BF cholinergic
neurons activities aligned to first lick in the trial (right) (Figure from Sturgill et al., 2020). B) BF
cholinergic activities aligned to first lick in absence of reward delivery. C) Acetylcholine release
in BLA aligned to first lick in absence of reward delivery (Figure from Kimchi et al., 2022).

Norepinephrine

Norepinephrine released in the brain is mostly from locus coeruleus
noradrenaline (LC-NE) neurons located in the brainstem. Despite the relatively small
number (~1500 neurons in rodents), LC-NE neurons send broad output to virtually all
brain regions (Berridge and Waterhouse, 2003, Sara, 2009, Schwarz et al., 2015,
Figure 1.5). Unlike BF cholinergic neurons, LC-NE neurons don’t show sensory
modality specific topographic organization (Kim et al., 2016). Along with reported
11
synchronous firing properties (Ishimatsu and Williams, 1996), LC-NE neurons are
thought to be homogenous. However, the homogeneity of LC-NE neurons has always
been challenged. Although LC solely contains noradrenergic neurons, LC-NE neurons
show different co-expression with other genetic markers like neuropeptides (Holets et
al., 1988). LC-NE neurons co-express with galanin and neuropeptide Y (NPY) show
some spatial segregation (Schwarz and Luo, 2015, Figure 1.6a). But whether these two
groups of LC-NE neurons have different functions is still under investigation. Apart from
co-expression with other genetic markers, LC-NE neurons also show some extent of
spatial segregation according to the cell morphologies. Two types of cells are observed,
with multipolar cells being larger preferably in the ventral LC and smaller fusiform cells
preferably in the dorsal LC (Swanson, 1976, Figure 1.6b). Again whether these LC-NE
neurons with different morphologies are functionally distinct needs to be further studied.
Some optogenetic studies give us some hints that dorsal and ventral LC-NE neurons
might play different roles. Anthony Pickering group reported that photoactivating LC
evoked antinociceptive or pronociceptive in rats. And through post-hoc characterization,
they found the animals with more expression of channelrhodopsin2 (ChR2) at ventral
LC show antinociceptive response while animals with more ChR2 expression at dorsal
LC show pronociceptive response. (Hickey et al., 2014). Inspired by this, future
research can optogenetically manipulate genetically different LC-NE neurons to explore
the potential different functions they play.
12

Figure 1.5 LC-NE neuron projections in the brain (Figure from Sara, 2009).

Figure 1.6 LC has different sub-cell types (Figure from Schwarz and Luo, 2015).
A) LC cell types defined by different genetic markers. B) LC cell types defined by cell
morphology

Despite whether LC-NE neurons function in a homogeneous or heterogeneous
manner is still under debate, myriads of studies have shown that LC-NE neurons are
involved in diverse behaviors, arousal (Aston-Jones and Bloom., 1981, Carter et al.,
2010), sleep/wake states (Hobson et al., 1975, Carter et al., 2010), stress response
(McCall et al., 2015), feeding behavior (Margules et al., 1972, Sciolino et al., 2021,
Yang et al., 2021), sensory gating (Manunta et al., 2004), and perceptual learning
(Martins and Froemke., 2015, Glennon et al., 2019). The major ways to unravel the
function of the LC-NE system are the same as studying the cholinergic system. Through
anatomy studies, it was shown in animals (Leslie et al., 1985) and humans (Lohr and
13
Jeste, 1988, Marcyniuk et al., 1989, Dahl et al., 2019) that the integrity of the LC-NE
system is linked to memory performance in aging. Postmortems of Alzheimer’s patients
show around 50% LC cell loss in the rostral LC (Matthews et al., 2002). Various
pharmacology studies also support the essential role that the LC-NE system plays in the
brain. For example, β-adrenergic receptors blockers impair exercise-induced learning
and memory improvement (Ebrahimi et al., 2010); in nonhuman primates, α2-adrenergic
receptor agonist clonidine alleviates the cognitive deficits in aged animals (Arnsten and
Goldman-Rakic, 1985). Both anatomy and pharmacology studies imply LC-NE’s role in
cognitive function and memory.

Generation of the norepinephrine neuron marker dopamine β-hydroxylase (Dbh)
cre lines allows the researchers to monitor noradrenergic neuron’s activities under
different behaviors more specifically. Recently, the LC-NE system’s role in feeding was
further revealed. Through expressing GCaMP in LC Dbh+ cells, Patricia Jensen group
observed an enhanced LC-NE activity during food approaching and suppressed LC-NE
activity during feeding, and this food-induced response was attenuated in sated animals
(Sciolino et al., 2021, Figure 1.7). The paper didn’t mention whether this LC-NE activity
pattern also holds true for water restricted mice. But during a conference, the first author
mentioned they saw the same LC-NE activities for mice that were under water
restriction. So very likely the LC-NE activity during the food intake process is a reflection
of brain state change induced by food consumption. Apart from directly recording
noradrenergic neurons in LC, monitoring the LC-NE axonal terminal activities is another
way to see how LC-NE modulates its target brain regions. As shown in the study from
14
Michael London group, LC-NE axon terminals at the barrel cortex respond to different
sensory modality stimuli, albeit to different extent (Deitcher et al., 2019, Figure 1.8a).
They also recorded the LC-NE axon terminal at the visual cortex, seeing response to
different sensory modality stimuli (Figure 1.8b). These results suggest that the sensory
cortex projecting LC-NE neurons might function homogeneously. Future experiments
can further explore whether cortex projecting LC-NE axon terminal activities are
homogenous. Using the mesoscope to monitor LC-NE axon terminal activities or
norepinephrine release across cortical regions can help us understand whether LC-NE
acts on cortex homogeneously or heterogeneously. Recording LC-NE neural activity
shows us the correlation between LC-NE and different behaviors. To see the causal role
of the LC-NE system, manipulation of LC-NE neurons are needed. Through
chemogenetically inhibiting LC-NE neurons, stress induced anxiety-like behavior is
ameliorated while optogenetically increasing LC-NE tonic activities promotes anxiety-
like behavior (McCall et al., 2015). Optogenetically inhibiting LC-NE neurons reduces
wakefulness, while activating LC-NE neurons induce sleep-to-wake transitions (Carter
et al., 2010). In summary, all these manipulation experiments prove a causal role of the
LC-NE system in different behaviors. Subsequent studies can manipulate genetically
defined subpopulations of LC-NE neurons or specific regions projecting LC-NE, so we
can have a better understanding of a specific LC-NE circuit’s function.
15

Figure 1.7 LC-NE neurons showed suppressed activities during food retrieval (Figure from
Sciolino et al., 2021).

Figure 1.8 LC-NE axon terminals respond to different sensory stimuli (Figure from Deitcher et
al., 2019).
A) LC-NE axon terminals at the barrel cortex respond differently to stimuli with different sensory
modalities. B) LC-NE axon terminals at the visual cortex respond differently to stimuli with
different sensory modalities.

Dopamine

Dopaminergic neurons are primarily located in two midbrain regions, ventral
tegmental area (VTA) and substantia nigra pars compacta (SNc). Based on the
16
projection of the dopaminergic neurons, there are 4 main pathways, 1) nigrostriatal,
dopaminergic neurons in SNc project to caudate nuclei and putamen of the striatum; 2)
mesocortical, dopaminergic neurons in VTA project to cortical areas, especially frontal
cortex, 3) mesolimbic, dopaminergic neurons in VTA project to limbic structures; 4)
tuberoinfundibular, dopaminergic neurons in arcuate nucleus in hypothalamus projects
to the media eminence (Scarr et al., 2013, Figure 1.9). The input/ output to both VTA
midbrain dopaminergic neurons (Watabe-Uchida et al., 2012, Beier et al., 2015) and
SNc dopaminergic neurons (Watabe-Uchida et al., 2012, Lerner et al., 2015) have been
carefully characterized. Based on projection onto different parts of striatum, two
subpopulations of SNc dopaminergic neurons have been identified, dorsolateral
striatum (DLS) projecting neurons and dorsomedial striatum (DMS) projecting neurons
(Lerner et al., 2015). DLS projecting SNc dopaminergic neurons and DMS projecting
SNc dopaminergic neurons show spatial segregation within SNc, with more DLS
projecting neurons being seen in lateral SNc while more DMS projecting neurons being
seen in medial SNc. Although DLS projecting and DMS projecting dopaminergic
neurons receive similar input from a lot of brain regions, a significant difference was
seen in DMS and DLS. This indicates a preferential reciprocal connectivity between
SNc dopaminergic neurons and dorsal striatum. For VTA dopaminergic neurons, they
receive inhibitory, excitatory and modulatory inputs from various cell types in multiple
sources (Beier et al., 2015). And VTA dopaminergic neurons receive distinct inputs
based on their outputs. From these midbrain dopaminergic neurons input/ output
studies, more circuit-dependent heterogeneous activities of midbrain dopaminergic
neurons have been revealed.
17

Figure 1.9 4 dopaminergic pathways in the brain (Figure from Scarr et al., 2013).

Extensive studies have been done on midbrain dopaminergic neurons.
Foundational studies from Schultz and colleagues demonstrate dopaminergic neurons
encoding reward prediction error (RPE). Before associative learning, midbrain
dopaminergic neurons respond to reward; after associative learning, dopaminergic
neurons’ response to reward diminished and shifted to the conditional stimuli (CS) that
predicted the reward; when the reward was omitted, dopaminergic neurons still fired to
the CS and had depressed activities at the expected reward time (Schultz et al., 1997,
Figure 1.10). And they proposed that the midbrain dopaminergic neurons play a
homogeneous role as encoding RPE. However, accumulating evidence has indicated
that midbrain dopaminergic neurons encode more than RPE and show heterogeneous
activities depending on specific circuits. From the previously mentioned anatomy study
18
on SNc dopaminergic neurons, DMS projection dopaminergic neurons and DLS
projecting dopaminergic neurons not only innervate distincts part of striatum, but also
respond differently to aversive stimuli (Lerner et al., 2015). DMS projecting SNc neurons
showed a decreased response (Figure 1.11a), while DLS projecting SNc dopaminergic
neurons showed an increased response at the time of aversive stimuli (Figure 1.11b).
VTA dopaminergic neurons also show a diverse response according to where they
project to (de Jong et al., 2019). VTA dopaminergic neuron terminals at ventral nucleus
accumbens medial shell (vNAcMed) responded to unexpected aversive stimuli and the
sensory cue that predicted the aversive stimuli, while showing largely absent response
to reward predictive cue (Figure 1.12). VTA dopaminergic neuron terminals at nucleus
accumbens (NAc) lateral shell are mostly respond to reward and reward predictive cue
(Figure 1.12). Apart from playing an important role in reinforcement learning, midbrain
dopaminergic neurons were also found to be associated with locomotion (Barter et al.,
2015, Howe and Dombeck, 2016, da Silva et al., 2018). Midbrain dopaminergic neuron
axon terminals at dorsal striatum fired before acceleration onsets at locomotion initiation
(Howe and Dombeck, 2016, Figure 1.13a). There are heterogeneous responses in
midbrain dopaminergic axon terminals at dorsal striatum, some responding to
unexpected reward (Figure 1.13b) while some responding to the locomotion onsets
(Figure 1.13c). In alignment with playing a role in movement, midbrain dopaminergic
neurons are also thought of as encoding motivation (Phillips et al., 2003, Howe et al.,
2013., Hamid et al., 2016, Berke, 2018, Mohebi et al., 2019). With the capability of
recording large populations of neurons in midbrain with single cell resolution,
dopaminergic neurons are found playing even more roles. Through recording ~300 VTA
19
dopaminergic neurons when animals were performing a complex decision-making task
in a virtual-reality (VR) environment, dopaminergic neurons were found encoding
reward predictive cues, erward, reward history, movement kinematics, behavioral
choices and spatial position (Engelhard et al., 2019).

Figure 1.10 Midbrain dopaminergic neurons encode RPE (Figure from Schultz et al., 1997).

20
Figure 1.11 SNc dopaminergic neurons respond differently to the aversive stimuli based on
their projection (Figure from Lerner et al., 2015)
A) DMS-projecting SNc dopaminergic neurons showed decreased response to the aversive
stimuli. B) DLS-projecting SNc dopaminergic neurons showed increased response to the
aversive stimuli.

Figure 1.12 VTA dopaminergic neurons show heterogeneous response to reward and aversive
stimuli based on their projection at nucleus accumbens (Figure from de Jong et al., 2019).

Figure 1.13 Midbrain dopaminergic neuron axon terminals at striatum have heterogeneous
response (Figure from Howe and Dombeck, 2016).
A) Midbrain dopaminergic neuron axon terminals at striatum show increased response to
locomotion initiation on average. B) Sample dopaminergic axon at striatum responded to reward
(yellow) and sample dopaminergic axon at striatum didn’t respond to reward (magenta). C)
21
Sample dopaminergic axon at striatum responded to locomotion initiation (magenta) and sample
dopaminergic axon at striatum didn’t respond to locomotion initiation (yellow).

Apart from recording dopaminergic neuron’s activities, manipulation, lesion and
pharmacological studies also imply the multi-function that midbrain dopaminergic
neuron plays and its potential role in brain disorders like Parkinson’s disease (PD),
addiction, depression and schizophrenia. Transgenic mice that lack tyrosine
hydroxylase in dopaminergic neurons are hypoactive and aphagic, which can be
rescued by daily treatment with L-dopa (Palmiter, 2008). Optogenetically stimulating the
dopaminergic neuron axon terminals at dorsal striatum initiated locomotion and
controlled acceleration frequency (Howe and Dombeck, 2016). Besides, it has also
been shown that optogenetic activation of dopaminergic neurons was sufficient to cause
long-lasting increases in cue-induced reward seeking behavior (Steinberg et al., 2013).
From clinical studies, one major pathological feature of PD is the loss of midbrain
dopaminergic neurons (Dauer and Przedborski, 2003). And individuals with tumors in
limbic areas were likely to be diagnosed with schizophrenia (Malamud, 1967). Through
human in vivo imaging, a robust increase in striatal dopamine synthesis and release in
psychosis was seen (McCutcheon et al., 2018). All studies mentioned so far have
indicated that midbrain dopaminergic neurons show heterogeneous activities and
function more than just signaling RPE.

22
Serotonin

The serotonergic system broadly innervates almost all parts of the brain. More
than 60% of the serotonin-producing neurons are located in the dorsal and median
raphe nuclei in the brainstem (Figure 1.14). In the mouse brain, around 35% of the
serotonergic neurons are located in the dorsal raphe (DR) and they heavily innervate
the forebrain (Ishimura et al., 1988). DR serotonergic neurons show stereotyped
locations based on their projections, with subcortical projecting neurons being localized
more in dorsal DR and cortical regions projecting neurons preferably localized in ventral
DR (Ren et al., 2018). Apart from the heterogeneity in their projection patterns, DR
serotonergic neurons also show molecular heterogeneity. Part of the DR serotonergic
neurons co-express vesicular glutamate transporter vglut3 (Gras et al., 2002, Ren et al.,
2018). And it has been shown that the DR serotonergic neurons co-release glutamate
(Liu et al., 2014). A more detailed study demonstrated that vglut3 expressed DR
serotonergic neurons preferentially projected cortical regions/ olfactory bulb and largely
located in the ventral DR (Ren et al., 2018). RNA-seq study of DR serotonergic neurons
also showed that they can further be divided into different subtypes according to the
differential gene expression (Okaty et al., 2015). These molecular and projection
heterogeneity imply that DR serotonergic plays a heterogeneous role.
23

Figure 1.14 Serotonergic system in the brain (Figure from Scarr et al., 2013 ).

And indeed DR serotonergic neurons have been reported in playing a wide range
of functions, induction of anxiety-like behaviors (Teissier et al., 2015, Marcinkiewcz et
al., 2016) or antidepressant-like behavioral responses (Urban et al., 2016), locomotion
inhibition (Correia et al., 2017), encoding reward (Liu et al., 2014, Cohen et al., 2015, Li
et al., 2016, Ren et al., 2018), encoding punishment (Cohen et al., 2015, Ren et al.,
2018), promoting waiting for reward (Miyazaki et al., 2011, Miyazaki et al., 2014,
Fonseca et al., 2015). Through direct recording in DR, serotonergic neurons are found
responding to sucrose and food intake (Li et al., 2016, Figure 1.15). Not only do DR
serotonergic neurons have phasic response to reward intake, they also show ramped
up activities during the reward expectation period (Li et al., 2016, Figure 1.16).
Manipulation experiments further support DR serotonergic neuron’s role in patience for
24
reward. Optogenetically activating DR serotonergic neurons promotes animals’ waiting
for delayed reward (Miyazaki et al., 2014, Fonseca et al., 2015). Other manipulation
experiments revealed more functions that DR serotonergic neurons are involved in.
Optogenetically activating bed nucleus of the stria terminalis (BNST) projecting DR
serotonergic neurons increases anxiety-like behaviors (Marcinkiewcz et al., 2016).
Phasic optogenetic activation of DR serotonergic neurons can increase pupil size, and
this effect is linked to the level of uncertainty (Cazettes et al., 2021). Few seconds
activation of DR serotonergic neurons led to animals moving slowly when they were not
occupied with other activities, while repeatedly activating DR serotonergic neurons over
weeks led to animals moving more quickly (Correia et al., 2017). These manipulation
experiments suggest that the function DR serotonergic neurons play depends on their
projection target, firing mode,and more specific genetic identities.

Figure 1.15 DR serotonergic neurons showed increased response to sucrose (middle) and food
(right) intake (Figure from Li et al., 2016).
25

Figure 1.16 DR serotonergic neurons showed ramped up activities during reward waiting time
and a phasic response to reward delivery (Figure from Li et al., 2016).

Among all the functions that DR serotonergic neurons play, its role in multiple
neuropsychiatric diseases has drawn enormous attention, due to its therapeutic
potential. For example, selective serotonin reuptake inhibitors (SSRI) is one of the most
widely-used antidepressants. Since the classical psychedelics lysergic acid
diethylamide (LSD) and psilocybin are agonists for 5-HT2A receptor, recently
therapeutic use of psychedelics in treating a range of psychiatric disorders has drawn
increasing attention. Although these are still pilot studies, with the improved
understanding of the serotonergic system, more detailed structural information of
serotonin receptors and better molecular engineering techniques, therapeutic
psychedelic drugs with less off-target effect will be available in the not far future.

26
Chapter 2: Summary of methods to detect neuromodulators
in vivo

To understand the roles that different neuromodulators play in the brain, it is
important to know when and where they are released. Unlike classical excitatory
neurotransmitter glutamate and inhibitory neurotransmitter gamma-aminobutyric acid
(GABA), neuromodulators in the brain generally will not directly cause action potential
through directly activating ligand-gated ionotropic receptors. Thus to monitor the
neuromodulators release, various methods have been developed in the past.
Traditionally, people have used microdialysis coupled with high-performance liquid
chromatography (HPLC) to detect neuromodulators release (Schultz and Kennedy,
2008), like acetylcholine (Marrosu et al., 1995), serotonin (Peñalva et al., 2003) and so
on. Through a dialysis probe, chemical samples in a target region were collected into
the probe. Then the dialysate collected will be analyzed using HPLC. A diagram for
microdialysis combined with HPLC is shown in Figure 2.1A. Although this method has
high specificity to the neuromodulator that is being detected, the low temporal resolution
prevents it from detecting neuromodulators that work in second and sub-second time
scales. Moreover, this method has low spatial resolution, not to mention being capable
of detecting the neuromodulator release in specific cell types. The intrinsic invasiveness
of this method makes it hard to monitor neuromodulator release longitudinally. To
monitor the neuromodulator release in higher temporal resolution, people used fast-
scan cyclic voltammetry (FSCV), which enabled researchers to capture the
neuromodulator release in sub-second. Generally carbon-fiber microelectrodes are
27
implanted to detect the voltage change during the neurotransmitter redox reactions
(Venton and Cao, 2020). A diagram for FSCV to measure neuromodulator release is
shown in Figure 2.1B. However, FSCV lacks high specificity to different neuomulators.
For example, the discriminability between dopamine and norepinephrine is poor
(Robinson et al., 2003). Thus FSCV is not ideal to detect the neuromodulator release in
the brain regions that have multiple neuromodulator release. Like microdialysis, FSCV
is also invasive and doesn’t have cell type specificity. Thus it is also unsuitable for
monitoring neuromodulator release in a cell-type specific manner across multiple
imaging sessions.

To reach cell-type specific chronic recording, researchers were dedicated to
developing genetically-encoded neuromodulator indicators. And much progress has
been made in the past decade. These indicators generally can be divided into two
groups, according to the neurotransmitter binding protein being adopted. One of the
proteins being chosen is bacterial periplasmic-binding protein (PBP). CFP-YFP FRET
pair (SuperGluSnFR (Hires et al., 2008), Figure 2.2A) or circularly permuted
GFP(cpGFP) (iGluSnFR (Marvin et al., 2013), Figure 2.2B) were linked to glutamate-
binding PBP from Escherichia coli (E. Coli). Taking advantage of the successfully
developed iGluSnFR, structure and machine learning guided mutagenesis enables the
expansion to the development of other neurotransmitter indicators, like GABA indicator
iGABASnFR (Marvin et al., 2019), acetylcholine indicator iAChSnFR (Borden et al.,
2020), serotonin indicator iSeroSnFR (Unger et al., 2020). These microbial PBP based
neurotransmitter indicators make it possible to spy on the neuromodulators release onto
28
specific cell types. Although with great maximum fluorescence change in vitro, these
PBP based indicators all have relatively low affinity, whose Kd for their corresponding
neuromodulator is at micromolar range. iSeroSnFR’s Kd for serotonin is around 400μM,
iAChSnFR’s Kd for acetylcholine is ~4μM, and iGABASnFR’s Kd for GABA is ~9μM.
Thus they suffer from detecting endogenous neuromodulator release, which is in the
hundred nanomolar range, with high signal-to-noise ratio (SNR).

Figure 2.1 Microdialysis and FSCV diagram (Figure from Wang et al., 2018).
A) Diagram for using microdialysis to detect neuromodulators. B) Diagram for using FSCV to
detect neuromodulators.
29

Figure 2.2 Genetically-encoded glutamate indicator.
A) Diagram for SuperGluSnFR (Figure from Hires et al., 2008). B) Diagram for iGluSnFR
(Figure from Marvin et al., 2013).

Alternatively, another strategy is to use the natural neuromodulator binding
proteins G-protein coupled receptors (GPCRs), which have higher affinity for the
neuromodulators. In the past decade, multiple GPCR-based neuromodulator indicators
have been developed. According to extensive structure studies, GPCRs will go through
a conformational change upon ligand binding (Rusmussen et al., 2007, Kruse et al.,
2013). Thus multiple groups utilized conformational change sensitive fluorescent protein
pair FRET to detect this conformational change (Vilardaga et al., 2003, Nakanishi et al.,
2006, Maier-Peuschel et al., 2010). Generally, one fluorescent protein was fused at the
third intracellular loop (ICL3), while the other fluorescent protein was fused at the C
terminal of the GPCR (Figure 2.3A). When the ligand binds to the GPCRs, the
conformational change the receptor goes through will bring the two fluorescent proteins
closer. Consequently, the GPCR activation can be read out as a FRET signal. However,
these FRET based neuromodulator indicators show poor cell membrane trafficking and
30
very small SNR, which prevents these indicators from in vivo application. In order to get
a higher SNR, multiple groups switched to monitor the amplified downstream signaling
induced by GPCR activation. For example, cell-based neurotransmitter fluorescent
engineered reporters (CNiFERS) (Nguyen et al., 2009, Muller et al., 2014) are cultured
cells that stably expressed engineered GPCRs that coupled to Gq protein and
genetically encoded calcium indicator TN-XXL (Figure 2.3B). When corresponding
ligands bind to the GPCR, there will be an increase of cytosolic calcium through
Gq/inositol (IP3) second messenger passway. And this cytosolic calcium change can be
robustly captured by the calcium indicator TN-XXL. CNiFERs strategies have been
successfully applied to develop indicators for acetylcholine (Nguyen et al., 2009),
norepinephrine and dopamine (Muller et al., 2014). Notwithstanding the high specificity
for individual neuromodulators, the perplexing procedure to implant HEK cells into the
brain render researchers’ reluctance to use it, and its low temporal resolution (in second
scale) hinders the detection of neuromodulator release in sub-second scale. Similarly
another method TANGO assay also uses GPCR as the neuromodulator detection
protein. But unlike CNIFERs taking advantage of the G protein pathway, TANGO assay
monitors the beta-arrestin binding to GPCR (Barnea et al., 2008). A transcription
activator tTA was fused to the C terminal of the GPCR with a cleavage site for a specific
protease from tobacco etch virus (TEV) in between. Meanwhile the TEV protease was
fused to beta-arrestin. When the ligand binds to the GPCR, the beta-arrestin will bind to
the GPCR, which brings the TEV protease in proximity to the TEV cleavage site
between the tTA and GPCR C terminal. Thus the tTA will be released from the GPCR
and traveled to the nucleus to induce the expression of a reporter gene (Figure 2.3C).
31
Unfortunately, TANGO assay has even slower temporal resolution than CNiFERs (in
hour scale), which makes it even more difficult to track sub second neuromodulator
release. To get higher temporal resolution GPCR-based indicators, two groups fused
cpGFP at the ICL of GPCR to directly detect the conformational change of GPCR upon
ligand binding.Through multiple rounds of screening, linker optimization, and semi-
rational mutagenesis, these two groups succeeded in developing GPCR based
indicators for multiple neuromodulators (dopamine (Patriarchi et al., 2018, Sun et al.,
2018, Sun et al., 2020), norepinephrine (Feng et al., 2019), acetylcholine (Jing et al.,
2018, Jing et al., 2020), serotonin (Wan et al., 2021), ATP (Wu et al., 2022), and
endocannabinoid (Dong et al., 2021) with sub second temporal resolution (Figure 2.4A-
F). Because of the simple principle, the strategy can be easily extended and applied to
develop a wide range of neuromodulator indicators. Apart from developing indicators for
different neuromodulators, it is also important to expand the color spectrum of the
indicators to make them compatible with GFP based calcium indicators or blue light
activated opsins. Previously, a red version of iGluSnFR was developed (Wu et al.,
2018). More red GPCR-based neuromodulator indicators are currently under
development.

Figure 2.3 GPCR-based neuromodulator indicators.
A) Diagram for FRET based GPCR activation indicator (Figure from Vilardaga et al.,2003). B)
Diagram for CNiFER (Figure from Nguyen et al., 2009). C) Diagram for TANGO (Figure from
Barnea et al., 2008).
32

Figure 2.4 Kinetics for different neuromodulator indicators.
A) Dopamine indicators on kinetics (Figure from Sun et al., 2020). B) Norepinephrine indicators
on kinetics (Figure from Feng et al., 2019). C) Acetylcholine indicator on kinetics (Figure from
Jing et al., 2020). D) Serotonin indicator on and off kinetics (Figure from Wan et al., 2021). E)
ATP indicator on kinetics (Figure from Wu et al., 2022). F) Endocannabinoid indicator on and off
kinetics (Figure from Dong et al., 2021).

Chapter 3: Characterize multiple genetically encoded
neuromodulator indicators in vivo performance under two
photon microscopy

Summary

As mentioned in Chapter 2, multiple genetically-encoded neuromodulator
indicators have been developed. Even though all the indicators showed good
33
expression and detectable fluorescence change in cultured cells in response to their
corresponding ligands, these in vitro experiments can’t guarantee the indicators’ ability
to detect endogenous neuromodulator release in vivo. Here we characterized the
expression of multiple GPCR-activation based (GRAB) neuromodulator indicator
variants for acetylcholine, norepinephrine, serotonin, and dopamine in the primary
somatosensory cortex (S1) and their in vivo performance under two photon microscopy
when the animals were performing a whisker-guided object localization task. We
provided these indicators in vivo performance feedback to our collaborator and helped
them further optimize these indicators.

Results

We firstly tested whether we can express different GRAB sensor variants in vivo.
We separately injected the AAV viruses for acetylcholine indicators (GRAB-ACh2.0,
GRAB-ACh3.0, GRAB-ACh3.8), norepinephrine indicators (GRAB-NE2m, GRAB-NE2h,
GRAB-AC-NE1h), dopamine indicators (GRAB-DA2.0, GRAB-DA3m), and serotonin
indicators (GRAB-5HT3.0) into mice somatosensory cortex supragranular layers. We
saw that all of these GRAB indicators can be expressed in the mouse somatosensory
cortex (Figure 3.1) and the expression can be mostly stable at least across one month
period, with some extent of expression pattern change likely due to cell death (Figure
3.2). Then we test the photostability of GRAB sensors during ~one hour continuous
imaging under two photon microscopy. We saw an initial rapid fluorescence decrease
34
upon illumination, but the fluorescence got stable after ~100 seconds, being stable for
more than an hour (Figure 3.3).

Figure 3.1 Different GRAB indicator variants were expressed in the mouse somatosensory
cortex.

Figure 3.2 GRAB sensor can be stably expressed in mouse somatosensory cortex across one
month period with some extent of expression pattern change.
35

Figure 3.3 GRAB-ACh3.0 is photostable under two photon microscopy imaging.

To check these GRAB sensors’ ability to capture the endogenous
neuromodulator release in S1, we imaged mice with the expression of GRAB sensor
when the mouse was performing a whisker-guided object localization task. The mice
were trimmed down to a single whisker, and they used this whisker to discriminate
between two pole positions. After pole onset, the mice had one second sampling period
to explore the pole position. And then they have one second answer period to show
their decision. The posterior pole position (go position) is associated with water reward,
and the mice need to lick during the one second answer period to collect reward. When
the pole is at the anterior position (no-go position), the mice need to hold their lick
during the one second answer period. Failure to hold their lick leads to timeout with
various duration depending on which learning stage the mice are at. Based on mice’s
behaviors and pole position, there are 4 trial outcomes, hit, miss, false alarm, and
correct rejection (Figure 3.4). When the pole is at go position, licking during the answer
period leads to hit trials while absence of licking during the answer period leads to miss
trials. When the pole is at no-go position, licking during the answer period leads to false
36
alarm trials, while correctly holding the lick during the answer period leads to correct
rejection trials. We trained the mice with the pole at no-go position either out of reach or
reachable for the whisker to touch. We compared the in vivo performance of two GRAB-
ACh variants GRAB-ACh2.0, and GRAB-ACh3.0, which is reported to have significant
increase of signal to noise ratio (SNR) in cultured cell over GRAB-ACh2.0 (Jing et al.,
2020, Figure 3.5a). Consistent with the in vitro results, we also saw a larger
fluorescence change in response to the same stimulus when the mice were performing
the object localization task (Jing et al., 2020, Figure 3.5b). This reassures that
optimization of the GRAB sensors in vitro can be applied in vivo.

Figure 3.4 Whisker-guided object localization task behavioral paradigm.
37

Figure 3.5 GRAB-ACh2.0 and GRAB-ACh3.0 comparison (Figure from Jing et al., 2020).
A) GRAB-ACh2.0 and GRAB-ACh3.0 performance comparison in HEK cells. B) GRAB-ACh2.0
and GRAB-ACh3.0 performance comparison in awake behaving animals under two photon
microscopy.

Similarly, we evaluate other GRAB sensor variants in vivo performance. We
plotted the average fluorescence trace aligned to the pole onset time sorted by trial type
for different GRAB variants. A single session for GRAB-ACh3.0, GRAB-ACh3.8, GRAB-
NE2m, GRAB-NE2h, GRAB-AC-NE1h, GRAB-DA1.0, GRAB-DA3m, and GRAB-5HT3.0
are shown in Figure 3.6. We found that different GRAB sensor variants for the same
neuromodulator showed different in vivo performance. GRAB-ACh3.8 showed a slower
off rate compared to GRAB-ACh3.0, with longer time going back to baseline
fluorescence. GRAB-AC-NE1h showed larger SNR compared to GRAB-NE2m and
GRAB-NE2h. GRAB-DA3m has larger fluorescence change compared to GRAB-DA1.0.
And we also noticed that GRAB sensors for different neuromodulators exhibited
different release patterns. For instance, there is a larger fluorescence change difference
between hit trials and false alarm trials for serotonin and norepinephrine compared to
38
acetylcholine and dopamine. These different release patterns of different
neuromodulators during mice performing the same task implies that the same stimulus
might trigger different neuromodulator release in S1 and different neuromodulators
might play different roles in this whisker-guided object localization task.

Figure 3.6 Different GRAB variants fluorescence change sorted by trial types (Blue: Hit, Black:
Miss, Red: Correct Rejection, Green: False Alarm) under two photon microscopy while animals
were performing the tactile-guided object localization task. All traces are aligned to pole onset
time.

Discussion

Through imaging different GRAB sensor variants under two photon microscopy
while mice were performing a whisker-guided object localization task, we were able to
compare the in vivo performance of different GRAB variants for the same
neuromodulator. These results gave us insight of choosing which GRAB variant to use
for monitoring endogenous neuromodulator release. However, our results can only
suggest the better variant to use in S1. We should always keep in mind that in different
39
brain regions, the local neuromodulator concentration varies. Thus whether the best
GRAB variant to detect neuromodulator release in S1 can be applied in other brain
regions still needs to be tested. We also found that different variants showed different
kinetics in vivo. Thus we should also keep in mind that for the variant with slower
kinetics, we might fail to capture the local neuromodulator concentration change that
happens in a faster time scale.

By imaging multiple neuromodulator GRAB sensors when animals were
performing the same task, we were able to compare the release dynamic difference of
different neuromodulators. With more detailed data analysis being done, these results
can give us a hint of the different functions of these cortical neuromodulators release
during associative learning. With this rich dataset, we should at least be able to answer
the following two questions: 1) Do the same behavioral event and sensory stimulus
trigger different neuromodulator release? 2) Does learning shape different
neuromodulators release dynamics differently? In the next chapter, we use S1
acetylcholine release dynamics dataset as an example to demonstrate the detailed
analysis we did to uncover the function of cortical neuromodulator release in associative
learning.

40
Chapter 4: Monitor acetylcholine dynamics in somatosensory
cortex during tactile guided associative learning

Summary

Numerous cognitive functions including attention and learning are influenced by
the dynamic patterns of acetylcholine release across the brain. How acetylcholine
mediates these functions in cortex remains unclear, as the relationship between cortical
acetylcholine and behavioral events has not been precisely measured across task
learning. To dissect this relationship, we quantified motor behavior and sub-second
acetylcholine dynamics in primary somatosensory cortex during acquisition and
performance of a tactile-guided object localization task. We found that acetylcholine
dynamics were directly attributable to whisker motion and licking, rather than sensory
cues or reward delivery. As task performance improved across training, acetylcholine
release associated with the first lick in a trial became dramatically and specifically
potentiated, paralleling the emergence of a choice-signaling basis for this motor action.
These results show that acetylcholine dynamics in sensory cortex are driven by directed
motor actions to gather information and act upon it.

41
Introduction

Acetylcholine is a major neuromodulator in the brain that influences diverse
cognitive functions that span timescales, including arousal (Parikh et al., 2007, Zhang et
al., 2011, Kim et al., 2016, Reimer et al., 2016) selective attention (Donoghue and
Carroll, 1987, Tremblay et al., 1990, Hars et al., 1993, Gourd et al., 2009),
reinforcement learning (Parikh et al., 2007, Chubykin et al., 2013, Liu et al., 2015,
Crouse et al., 2020, Sturgill et al., 2020) and neural plasticity (Kilgard et al., 1998,
Froemke et al., 2013, Jiang et al., 2016, Guo et al., 2019). Many of these functions are
mediated through acetylcholine’s influence on cortical circuits (Hasselmo et al., 2011,
Lee et al., 2012). Cholinergic projections to cortex are complex, arising from multiple
basal forebrain (BF) nuclei that contain neuronal subgroups with distinct projection
specificity and arbor distributions within and across projection areas (Kim et al., 2016,
Zaborszky et al., 2015, Li et al., 2018). Individual nuclei also show significant
differences in the behavioral events which correlate with their activity patterns (Kim et
al., 2016, Robert et al., 2021). Direct observation of the spatiotemporal dynamics of
acetylcholine in cortex can bypass this organizational complexity and provide insight to
how these convergent cholinergic inputs influence cortically regulated cognitive
functions.

Numerous lines of evidence demonstrate that the cortical release of acetylcholine
regulates arousal and attention (Hasselmo et al., 2011, Lee et al., 2012). Increased
cortical acetylcholine levels are associated with (Inglis et al., 1995) and required for
42
induction of cortical desynchronization during active sensing (Eggermann et al., 2014).
Acetylcholine release causes layer-specific modulation of responses in primary
somatosensory cortex (S1) to whisker stimulation (Donoghue and Carroll, 1987,
Chaves-Coira et al., 2018), enhances sensory evoked responses in A1 7 and V1
(Herrero et al., 2008, Pinto et al., 2013), and suppresses spontaneous activity in S1
during whisker movement (Eggermann et al., 2014). These effects improve stimulus
discriminability and provide a mechanism for selective attention.
Attention is crucial for learning and performing tasks. Acetylcholine’s role in these
processes is becoming better appreciated through accumulating studies identifying
which cholinergic neurons fire at what times during task acquisition and execution.
Cholinergic neurons in BF respond (Crouse et al., 2020) and cortical acetylcholine
transients are evoked to reinforcement-predictive sensory cues (Parikh et al., 2007).
The extent to which association learning sculpts these responses varies across reports,
with such stimuli driving increasing amounts of cholinergic activity in BF (tones and
punishment (Guo et al., 2019), odors and reward (Sturgill et al., 2020)) and in nucleus
basalis (NB) to basolateral amygdala (BLA) projections (tones and reward (Crouse et
al., 2020)) across training, contrasting with the finding that reward-predictive tones show
stable acetylcholine release in BF with learning (Robert et al., 2021)21. Cholinergic
neurons throughout BF also strongly respond to negative reinforcement (Sturgill et al.,
2020, Guo et al., 2019, Robert et al., 2021, Hangya et al., 2015, Harrison et al., 2016),
but are inconsistently reported to respond to positive reinforcers like reward. In
primates, 70% of all BF neurons were significantly modulated in a choice period, but
only 25% in a reward period (Richardson et al., 1990). Cell-type specific recordings in
43
rodents found cholinergic neurons within BF do respond to positive reinforcers, scaled
by reinforcement surprise (Hangya et al., 2015) and encode valence-free reinforcement
error (Sturgill et al., 2020). However, in rat medial prefrontal cortex (mPFC),
acetylcholine levels increased for reward-predictive cues, but not for reward delivery
(Parikh et al., 2007).

A challenge in interpreting studies of cholinergic activity is that both
reinforcement-predictive cues and reinforcing events are tightly correlated with orofacial
and body movements (Clarke et al., 1971, Bindra et al., 1967). These movements are
also associated with transient increases in cortical acetylcholine levels (Reimer et al.,
2016, Jing et al., 2020, Lohani et al., 2020). Of particular interest to reward related
mechanisms, licking is reported to drive cholinergic activity, though reports vary from
strong acetylcholine release at lick bout onset (Harrison et al., 2016, Kimchi et al., 2022)
and offset (Robert et al., 2021), to weak release to licks (Sturgill et al., 2020) to none at
all (Hangya et al., 2015). This variability may arise from experimental differences in task
conditions and sensory modalities or may reflect a compartmentalization of
acetylcholine functions based on nucleus, cell-type, and projection targets. Regardless,
since whisking and licking patterns are strongly influenced by reward expectation and
delivery (Clarke et al., 1971, Bollu et al., 2021), dissociating sensory cues and reward
from these motor actions is crucial for interpreting acetylcholine’s role in task learning
and performance.

44
Here we sought to constrain how acetylcholine influences sensory cortex by measuring
spatiotemporal dynamics of acetylcholine in S1, an area that undergoes remarkable
representational plasticity during learning (Kim et al., 2020). We directly image sub-
second changes in acetylcholine concentration using a GPCR Activation Based sensor
(GRAB-ACh3.0 (Jing et al., 2020)) broadly expressed on the surface of supragranular
cortical neurons. High-speed videography during active whisker-guided object location
discrimination allows dissociation of cue and stimulus from tightly coupled motor
responses, while our response design allows dissociation of choice signaling and
reward delivery from the motor action of licking. Recording across weeks of training, we
identify behavioral correlates of acetylcholine release in S1 during task performance
and how tactile discrimination learning specifically reorganizes those dynamics in the
transition from novice to master.

Results

To measure the dynamics of acetylcholine in primary somatosensory cortex
during tactile discrimination learning, we employed a go/no-go single-whisker object
localization task (Figure 4.1A (Guo et al., 2014)). We trained water-restricted head-
fixed mice (n=8 mice) to search with a whisker for the position of a thin steel pole
presented in one of two locations along an anteroposterior axis during a one second
sampling period (Figure 4.1B), and to lick for a water reward during a subsequent one
second answer period if the pole was in the posterior location. Mice were cued to the
45
presentation and removal of the pole by the sound of a pneumatic valve. Licking during
the sampling period had no consequences, while licking in the answer period
determined trial outcome and extended the duration of the pole presentation.

Mice were trimmed to their C2 whisker at least one week prior to onset of
imaging experiments. All mice reached task mastery (>70% performance for 3
consecutive sessions) within 1-3 weeks of training (mean 10.75 ± 2.02 sessions; mean
3373.75 ± 754.09 (standard error) trials Figure 4.1C). Trimming of the C2 whisker after
task mastery dropped performance to chance (Figure 4.1D), demonstrating the
whisker-dependence of the task. Expert mice initiated whisker exploration upon the pole
presentation and withdrawal cue sounds (Figure 4.1E, F, SFigure 4.1A). Since
whisking amplitude was physically restricted when the pole was in the go-associated
proximal position, we examined the whisking patterns of go trials vs. no-go trials
separately. For both positions, trials with licks during the sampling and answer periods
(i.e. Hit and False Alarm) had more sustained whisking than trials without licks (i.e. Miss
and Correct Rejection; Figure 4.1G). This difference is consistent with licking-coupled
whisker motion (Moore et al., 2014). Lick rates on Hit and False Alarm trials were
indistinguishable during the sampling period and diverged during the answer period
once water was dispensed due to reward collection on the Hit trials (Figure 4.1H).
Licking rates on Correct Rejection and Miss trials were zero by construction during the
answer period, with occasional licks outside this period.
46

Figure 4.1 Learning of whisker-guided object location discrimination and associated motor
actions.
A) Two location discrimination task design with trial outcomes. B) Single trial structure with
example licking, whisking, GRAB-ACh signal traces. C) Learning curves of 8 mice, mean ± SEM
sessions to expert. Blue circles, unbalanced go and no-go trials. Gold dash, expert threshold.
Gray dash, chance level. D) Mean performance for three early and expert sessions, per mouse.
3 mice with 1 no-whisker session, 4 mice with 3 no-whisker sessions averaged. E) Average
whisking amplitude aligned to pole in cue, 3 expert sessions / mouse. F) Same aligned to pole
out cue. G) Grand mean trial averaged whisking amplitude by trial outcome, 3 expert sessions /
mouse. H) Same for licking rates.

Acetylcholine dynamics were measured using two-photon imaging of
supragranular layers of S1. The C2 whisker barrel was targeted for viral injection of
AAV9-GRAB-ACh3.0 via intrinsic signal imaging (SFigure 4.2A). Behavioral training
commenced three weeks post injection, when strong fluorescence signals were present
in all mice (Figure 4.2A). To minimize the impact of a rapid partial dimming of
fluorescence following illumination onset (SFigure 4.2B), two-photon imaging and
illumination (15.44 fps, 940nm, 30-60mW out of objective) was continuous from the
session start, and the first 100 seconds of the session were excluded from analysis.
Phasic increases in fluorescence across the entire field of view were observed shortly
47
after most, but not all onsets of pole presentation (Figure 4.2B). These fluorescence
dynamics diverged across trial outcomes (Figure 4.2C). Hit and False Alarm trials
showed a strong, similar increase in signal during the sampling and answer periods
(example session, SFigure 4.2C; grand mean of expert sessions Figure 4.2D). This
was followed by a secondary response that, on average, persisted into the inter-trial
period. In contrast, Miss and Correct Rejection trials on average showed two short
latency, short duration increases in fluorescence following pole presentation and
withdrawal. We did not find evidence for differential regulation of acetylcholine dynamics
at the spatial scale of cortical columns, as trial type averages within the primary whisker
column versus in the surrounding columns were identical for all trial types (Figure
4.2E).

Figure 4.2 Acetylcholine release in S1 varies with trial outcome.
48
A) GRAB-ACh3.0 AAV expression in S1 barrel cortex, 3 weeks post injection. Cartoon
generated by BioRender.com. B) Phasic increase of acetylcholine release after pole in cue in
most trials. C) Acetylcholine induced fluorescence changes sorted by trial types, one expert
example session. D) Grand average, 8 mice, 3 expert sessions each. E) Comparison of grand
average acetylcholine dynamics within the C2 column (solid) and in the surround (dashed). 8
mice, 3 expert sessions each.

To investigate potential triggers of acetylcholine release in cortex, we compared
the acetylcholine dynamics across trial types to several classes of behavioral events,
including pole presentation and withdrawal cues, whisker exploration, licking, and
reward. All trial types shared a common time for pole presentation, and all trial types
showed a small, short latency (2-3 frames, 130-195ms) increase in acetylcholine
aligned to that cue (Figure 4.3A). Lick trials showed much bigger and longer transients
immediately following this initial hump. Trials without licks in the answer period had a
common pole withdrawal time, while trials with answer licks had a variable withdrawal
time. Alignment to pole withdrawal again showed a sharp upward transient of
acetylcholine across all trial types, though this was overlaid on longer duration
acetylcholine dynamics (Figure 4.3B).

Was acetylcholine release driven by the pole presentation cue per se, or was it
driven by a motor response to the cue (Figure 4.1E,F)? To test this, we examined the
covariation of cue-evoked whisking with cue-associated acetylcholine dynamics. We
restricted our analysis to no-lick trials to avoid potential confounds of licking-driven
acetylcholine responses. We sorted acetylcholine responses in no-lick trials by the
average amplitude of the whisker motion within 500 milliseconds after the pole
presentation cue (Figure 4.3C, D). A fraction of trials (23.9% mean ± 21.8% per mouse)
did not evoke whisker motion (<2° mean amplitude) upon pole presentation, with most
49
producing a range of whisking vigor (SFigure 4.3A). In trials without cue-evoked
whisking, there was no acetylcholine release following the cue (0.01% mean ∆F/F for
whisking amplitude <2°), while trials with cue-evoked whisking showed a positive
relationship between whisking amplitude and acetylcholine response (0.58% mean ∆F/F
for whisking amplitude >5°; 0.08% mean increase in ∆F/F per degree of amplitude
(Figure 4.3E, SFigure 4.3B). These findings were recapitulated in pole withdrawal-cued
whisking and acetylcholine responses (SFigure 4.3C-E). This implies that whisking,
rather than whisker-pole contact drove the responses, since whisking after pole
withdrawal rarely causes pole touches. These data demonstrate that cue-associated
acetylcholine release in S1 is directly proportional to the magnitude of the motor
response (i.e. exploratory whisking) evoked by that cue. We conclude that the motor
response to cue, rather than the cue itself, drives cue-associated acetylcholine
transients in S1.
50

Figure 4.3 Whisking drives acetylcholine release in S1.
A) Grand mean acetylcholine fluorescence change aligned to pole in cue. 8 mice, 3 expert
sessions each. B) Same to pole out cue. C) Whisking amplitude sorted by mean of 500ms post
pole in cue. No lick trials pooled from 3 expert sessions, 1 mouse. D) Same, for acetylcholine
fluorescence change sorted by whisking amplitude. E) Grand mean of acetylcholine
fluorescence change vs. whisking amplitude of 500ms after pole in cue. No lick trials from 3
early and 3 expert sessions, 8 mice.

The striking difference in acetylcholine dynamics across trial types (Figure 4.2C-
E), in particular Hit and False Alarm versus Miss and Correct Rejection, suggested that
licking could be a powerful driver of acetylcholine release in S1. We aligned trials across
expert sessions to the time of first lick, regardless of trial type or if the lick occured in the
answer period, and sorted by the number of licks in the trial (Figure 4.4A,B). Trials with
no licks were aligned to the time of the median first lick. On trials with licks, there was a
sharp increase in acetylcholine signal beginning shortly before lickport detection of the
first lick, which lasted 1-2 seconds (Figure 4.4C). This lick-associated response
51
dwarfed the cue-associated transients on the no-lick trials (Figures 4.2E, 4.4C). Licking
preceded rhythmically at a regular modal inter-lick interval of 155ms in expert mice
(Figure 4.4D). The first lick was associated with a profound increase in mean
acetylcholine over the following second, jumping from 0.35% ± 0.26% on no-lick trials to
1.60% ± 0.42% with a single lick (Figure 4.4E), an increase of 1.25% ∆F/F over
baseline. Subsequent licks yielded smaller increases in mean acetylcholine over this
period at a rate of 0.28% ∆F/F per additional lick. Similarly, the duration of the transient
was 1.46 ± 0.65s for a single lick. Subsequent licks extended this transient 116ms per
lick (Figure 4.4F), somewhat less than the inter-lick interval (Figure 4.4D). This
demonstrates that acetylcholine release in S1 is primarily coupled to the onset of goal-
directed licking, with the vigor of licking modulating the magnitude and duration of the
release.
52

Figure 4.4 Licking strongly drives acetylcholine release in S1.
A) Acetylcholine fluorescence change across trials, sorted by number of licks in trial, aligned to
first lick. One example expert session. B) Lick pattern across trials, sorted by number of licks in
trial, aligned to first lick. Same example session as A. C) Grand mean acetylcholine
53
fluorescence change aligned to first lick. 3 expert sessions, 8 mice. D) Pooled inter-lick interval
from 3 early (gray) and 3 expert (black) sessions each, 8 mice. E) Mean acetylcholine
fluorescence change in 1 second following first lick binned by number of licks in that period.
Circles, mean 3 expert sessions / mouse. Linear fit from 1-7 licks. F) Duration, onset to trough,
of the first acetylcholine transient, binned by number of licks within 2 seconds of first lick.
Circles, mean 3 expert sessions / mouse. Linear fit from 1-10 licks.

The early lick-associated acetylcholine transients were followed by a longer late
acetylcholine rebound that began 1.5-3 seconds after the first lick (Figure 4.2C-E). We
hypothesized that the late response was driven by reward delivery, as observed in BF
cholinergic neurons (Sturgill et al., 2020, Hangya et al., 2015) and BF to BLA
projections (Crouse et al., 2020). To test this, we compared the first lick aligned
responses on Hit trials (which are rewarded) and False Alarm trials (which are
unrewarded). Against our hypothesis, there was no significant difference in the
amplitude of the late acetylcholine response between Hit and False Alarm trials (Figure
4.5A), although there was a small, but significant increase on Hit trials during a one
second period following the first lick. Could this difference be explained by reward
delivery?

Rewards are only distributed upon a correct lick in the answer period. Hit trials
had higher acetylcholine levels than False Alarms during the answer period, but also
had more licks (Figure 4.5B,C). To control for increased acetylcholine release caused
by additional licking (Figure 4.4E), we pairwise compared mean acetylcholine levels of
Hit and False Alarm trials during the answer period for matched numbers of licks in the
period. Hit and False Alarm acetylcholine levels were essentially identical after
accounting for the difference in lick count (Figure 4.5D). We conclude that reward
delivery does not induce acetylcholine release in supragranular S1, consistent with prior
54
electrochemical measurements in mPFC (Parikh et al., 2007). Together, these data
establish that the main driver of acetylcholine release in supragranular S1 during tactile-
guided choice behavior is execution of the choice-signaling action, secondary drivers
are exploratory whisker motion and additional licking, while task initiation cues and
reward delivery have no direct effect.

Figure 4.5 Reward delivery does not drive acetylcholine release in S1.
A) Top: Acetylcholine fluorescence change aligned to first lick from Hit (blue) and False Alarm
(green) trials. Mean and SEM. 3 expert sessions, 8 mice for all panels. Bottom: Significance test
between Hit and False Alarm acetylcholine over time (p-value, paired t-test, corrected for
multiple comparisons). B) Mean number of licks in the answer period on Hit and False Alarm
trials. C) Distribution of lick counts in the answer period histogram for Hit (blue) and False Alarm
(green) trials. D) Top: Mean acetylcholine fluorescence change in answer period, binned by
licks in answer period for Hit (blue) and False Alarm (green). Bottom: Significance of difference
between Hit and False Alarm trials, binned by licks (p-value, paired t-test, corrected for multiple
comparisons).

55
Acetylcholine regulates attention (Klinkenberg et al., 2011), which may change
with task performance, familiarity, or learning. It follows that training might increase
acetylcholine release concurrently with improved performance. On the other hand,
training might reduce acetylcholine release, as a familiar task may require less
attentional resources to solve or reduced neural plasticity once established. To address
these possibilities, we compared the acetylcholine dynamics for the first three sessions
of the full task training and the final three expert sessions in each mouse. We found that
the initial lick-associated acetylcholine release was nearly twice as large in expert
sessions compared to early training sessions (∆F/F 3.07 ± 1.01% std expert vs. 1.67 ±
0.65% std early; Figure 4.6A). The magnitude of the increase was directly proportional
to session performance with mean ∆F/F increasing 0.53% per 10% increase in correct
rate (Figure 4.6B). This increase occurred even though expert mice licked fewer times
(Figure 4.6C,D) and at the same pace (Figure 4.6E) as compared to early sessions.
This increase of acetylcholine signal after training was not due to an increase of GRAB-
ACh sensor expression or sensitivity seen by the following internal control. On no-lick
trials, whisking-associated acetylcholine release slightly increased following the pole-
presentation cue, and slightly decreased following the pole-withdrawal cue (Figure
56
4.6F). This closely matched the change in whisking to those cues following training
(Figure 4.6G). Thus, the relationship between acetylcholine fluorescence change and
whisker motion was stable across training. Finally, we examined whether training
induced potentiation of all licks, or only the first lick (i.e. the choice-signaling lick in
expert sessions), by repeating the analysis of Figure 4.4 on early sessions of the same
cohort of mice. In contrast to experts (Figure 4.4E), the first lick in early sessions drove
a modest increase in acetylcholine from 0.14% ± 0.3% on no-lick trials to 0.50% ±
0.52% with a single lick (Figure 4.6H), an increase of 0.36% ∆F/F over baseline.
Subsequent licks caused a 0.26% increase in ∆F/F per additional lick versus 0.28%
∆F/F per lick in expert mice (Figure 4.6H). Thus, training induced a nearly 3.5x increase
in acetylcholine response over baseline specifically to the first lick, while leaving
responses to whisking and subsequent licks unchanged (Figure 4.6I). We conclude that
training in tactile-guided choice behavior induces a dramatic and selective potentiation
of acetylcholine release in supragranular S1 to a choice-signaling action and this
potentiation is correlated to improved task performance.
57
Figure 4.6 Learning selectively potentiates acetylcholine release to choice-signalling licks in
S1.
A) Grand mean acetylcholine fluorescence change aligned to first lick in early (gray) and expert
(black) sessions, 3 sessions per condition, 8 mice for all panels except B. Bands SEM. B)
Relationship between correct answer rate and acetylcholine fluorescence change within 1
second following first lick. Shade indicates mouse identity. All 51 single-whisker imaged
sessions, Fit equation r = 0.05328*x-0.0123. C) Mean lick rates for trial types in early (top) and
expert (bottom) sessions. D) Mean lick numbers per trial. E) Peak lick rate. F) Grand mean ±
SEM acetylcholine fluorescence change in no lick trials. G) Same for mean ± SEM average
whisking amplitude H) Mean acetylcholine fluorescence change in 1 second following first lick
binned by number of licks in that period. Linear fits from 1-7 licks. I) Acetylcholine fluorescence
change from whisking, subsequent lick, and first lick for early (gray) and expert (black) sessions.
Red, grand mean ± SEM.

58
Discussion

Recording and manipulation of cholinergic neurons originating in BF nuclei (Pinto
et al., 2013, Hangya et al., 2015, Kimchi et al., 2022, McGaughy et al., 2002,
Kuchibhotla et al., 2017) has established the importance of cholinergic signaling on
multiple brain functions. However, the complexity of BF organization (Kim et al., 2016,
Zaborszky et al., 2015, Li et al., 2018) has posed a challenge in linking specific
cognitive functions to acetylcholine dynamics in specific projection targets. By recording
acetylcholine dynamics directly in S1 (Figure 4.2) during whisker-guided object
localization (Figure 4.1), we discovered surprising differences in the triggers and
dynamics between previously observed BF nuclei and this cortical target essential for
tactile discrimination (O'Connor et al., 2010). Nearly all acetylcholine release in
supragranular S1 was attributable to directed motor actions (Figures 4.3, 4.4) rather
than sensory input (Figure 4.3) or reward delivery (Figure 4.5). Moreover, as task
performance improved across training, acetylcholine release to the first lick in a trial
became dramatically and specifically potentiated (Figure 4.6), paralleling the
emergence of a choice-signaling meaning to this motor action.

Together, these data support a model that acetylcholine release in sensory
cortex is driven by directed motor actions to gather information (e.g. whisking) and act
upon it (e.g. licking). This contrasts from reports that acetylcholine release is driven by
reinforcement-predictive cues (Crouse et al., 2020, Sturgill et al., 2020, Guo et al.,
2019) and reward itself (Crouse et al., 2020, Sturgill et al., 2020). This difference may
59
be because the auditory cue in our task does not predict reward, only the temporal
availability of a stimulus. In contrast to passive cue-reward association, our task
requires active exploration to gather reward-predictive information, revealing a motor
requirement for acetylcholine release. Our results also suggest that cue-induced
changes in motor behavior (e.g. increased whisking, sniffing or anticipatory licking) and
other orofacial movements (Lohani et al., 2020) may also provide a meaningful
contribution to acetylcholine release in passive association tasks.

The lack of reward-evoked acetylcholine release in S1 may be due to several
factors. One possibility is that reward activity in BF is transmitted only to particular
targets such as BLA (Crouse et al., 2020), but not S1, due to projection specificity in
subpopulations of BF cholinergic neurons (Kim et al., 2016). Second, while reward-
delivery signals have been observed in cholinergic neurons of HDB and NB (Hangya et
al., 2015), fiber photometry in those areas show reward-associated transients are small
relative to lick-evoked transients and tightly correlated with an increase of lick rate
following water delivery (Robert et al., 2021). Perhaps reward-delivery only indirectly
drives acetylcholine release via changes in licking patterns. However, this result may
depend on task and reward structure. Tasks with variable reward probability have
shown decreased cholinergic activity to highly likely rewards (Sturgill et al., 2020,
Hangya et al., 2015). In our task, licking on go trials guaranteed reward, which may
have shifted reward delivery-associated activity earlier to the time of the choice-
signaling lick, when reward becomes expected.

60
Intriguingly, acetylcholine release associated with the first lick began ramping
several hundred milliseconds prior to lick detection (Figure 4.4C). This may be due to a
combination of following factors: whisking precedes licking and drives modest and
additive acetylcholine release (Figure 4.3), decision related activity precedes motor
action by some amount of time (Shadlen et la., 1996), and the tongue requires about
150-200ms from protrusion initiation to lickport contact (Bollu et al., 2021). These are
partially counterbalanced by the indicator rise time, which is dependent on the
underlying acetylcholine concentration profile and likely on the order of tens of
milliseconds for a few percent ∆F/F change (Jing et al., 2020). Thus, we must consider
the possibility that some of the first lick-associated acetylcholine dynamics are caused
by an internal choice deliberation, rather than the initiation of motor action. This
possibility is reinforced by the observation that onset of acetylcholine release is more
closely aligned to first lick on early than on expert sessions (Figure 4.6A), suggestive of
expert-specific choice-associated dynamics overlaid on motor-triggered dynamics
common to both session classes.

The patterns of acetylcholine release in response to whisking and licking suggest
potential functions for acetylcholine on sensory cortical circuitry (Letzkus et al., 2011,
61
Urban-Ciecko et al., 2018). VIP cells in S1 are activated by whisking (Lee et al., 2013)
via acetylcholine release (Gasselin et al., 2021) leading to disinhibition of excitatory
neuron dendrites where top-down contextual information arrives in S1 (Takahashi et al.,
2016, Takahashi et al., 2020, Lacefield et al., 2019). Thus, whisking-induced
acetylcholine release could enhance integration of contextual information with sensory
input in S1 neurons, providing, for example, a potential mechanism for combining motor
and touch signals to generate location specific codes (Cheung et al., 2020) and
percepts (Cheung et al., 2019). VIP activation also gates neural plasticity in cortex
(Williams et al., 2019). Higher task performance is associated with increased
acetylcholine in V1 (Pinto et al., 2013). Our similar results in S1 (Figure 4.6B) identified
that this increase is specific to the first lick and persists for several seconds (Figure
4.4). Thus, this increase is well-poised to provide windows of enhanced neural plasticity
via VIP cell activation while the consequences of decisions are evaluated.

This work is only a step towards understanding the function of acetylcholine
dynamics in cortex and sensory processing. The specificity of acetylcholine action on
particular cortical cell types raises the question of the extent to which this reflects
differing patterns of receptor expression 54 versus preferential targeting by cholinergic
axons. While we did not see a difference in acetylcholine dynamics between center and
surround cortical columns, there could be substantial heterogeneity in release at cellular
and subcellular scales. The GRAB sensor expression across the plasma membrane of
all neuron types coupled with the intrinsic sensor kinetics made our observations well-
suited for quantification of volume transmission, but could obscure sites of fast synaptic
62
transmission. We also did not determine the extent to which the S1 acetylcholine
dynamics reflect activity from nucleus basalis (NB) versus horizontal diagonal band
(HDB) or their subdivisions which project to S1 (Kim et al., 2016). Cholinergic terminals
in mouse neocortex show laminar preferences, entering in either layer 1 or layer 6
depending on the location of originating soma in the BF (Bloem et al., 2014). The
deeper pathway could potentially convey different classes of information and serve a
distinct function from the superficial acetylcholine dynamics observed here. An intriguing
possibility is that the selective potentiation of choice-signaling action may be dissociable
by afferent source. Finally, while we quantified the triggers and timescales of
acetylcholine dynamics on cortical targets, substantial additional work is required to
determine the functional consequences of those time limits. Manipulation of local
acetylcholine dynamics and cellular targets at specific moments during task
performance and acquisition could clarify acetylcholine’s potential roles in regulation of
sensory integration and cortical plasticity.

Materials and Methods

Key Resources Table
Reagent or Resources Sources Identifier
Virus
AAV-GRAB-ACh3.0 Vigene
63
Experimental Models: Organisms/Strains
Mouse: C57BL/6J Jackson Laboratory RRID:IMSR_JAX:00064
Software and Algorithms
Matlab Math Works 2018B, 2021a
Scanbox Dario Ringach, UCLA https://scanbox.org
BControl Carlos Brody https://brodylabwiki.princet
on.edu/bcontrol/index.php?
title=Main_Page
Other
Two-photon microscope Neurolabware
NIR laser (2 photon
microscope)
Spectra-Physics InSight DS+
StreamPix Norpix RRID:SCR_015773
Pneumatic slide Festo Cat#170496
Solenoid valve Lee Company Cat#acA800-500um
CMOS camera Basler Cat#acA800-500um
Telecentric lens Edmund optics Cat#58-259

64
Animals
We used 2.5-4 months old male (n=2) and female (n=6) C57BL/6J mice
(#000664, The Jackson Laboratory). Mice were maintained on a 12:12 reversed light-
dark cycle. After water restriction, health status was assessed everyday following a
previously reported guideline (Guo et al., 2014). All procedures were performed in
accordance with the University of Southern California Institutional Animal Care and Use
Committee Protocol 20732 and 20788.

Headbar surgeries
Before each surgery, rimadyl tablet was given 0.5 mg/tablet 24 hours before
surgery. buprenorphine-SR and marcaine were injected subcutaneously at 0.5 mg/kg
and 2% right before the surgery. A customized stainless steel headbar was attached to
the skull using layers of Krazy glue (Elmer’s Products, Inc) and dental cement (Lang
Dental Mfg. Co., Inc).

Intrinsic signal imaging
Intrinsic signal imaging (ISI) was done 3 days after headbar surgery and 7 days
after the cranial window surgery. All whiskers except the C2 whisker were trimmed
before ISI. To identify the C2 barrel column, the C2 whisker was stimulated using a
Piezo stimulator when mice were under light isoflurane anesthesia (0.8-1.0%).
Comparisons of acetylcholine dynamics inside the C2 column versus surround were
based on column-sized hand-drawn ROIs centered on the ISI hotspot.

65
Cranial window and virus injection surgeries
In all mice, AAV9-hSyn-GRAB-ACh3.0 (Addgene Plasmid #121922
https://www.addgene.org/121922/ ) was injected from 1x aliquots during the cranial
window installation. A glass capillary (Wiretrol® II, Drummond) was pulled to 10-20μm in
tip diameter using a micropipette puller (Model P-97, Sutter instrument), and tip beveled
to about 30 degrees. The glass window was 2x2 mm glass hand fused to 3x3 glass
(both 0.13-0.17 mm thickness) with ultraviolet curing glue (Norland optical adhesive 61,
Norland Inc.).

Before each surgery, a rimadyl tablet was given 0.5mg/tablet 24 hours prior and
Buprenorphine-SR was injected subcutaneously at 0.5 mg/kg immediately before. A 2x2
mm square of skull whose center was the identified C2 whisker barrel region was
removed. Virus was backfilled into a pipette of mineral oil (M5904, Sigma-Aldrich). We
injected 400nl virus into the identified C2 barrel column through a single injection site
over 4 min at 300μm below pia, withdrawing after an additional two minutes. The
exposed brain region was then covered with the glass window. Targeting of the C2
column was confirmed via ISI one week after cranial window surgery, after which water
restriction commenced.

Behavioral task and training
Mice were trained in a whisker-guided Go/No-go localization discrimination task
(O'Connor et al., 2010). During training, a smooth black pole with 0.6 mm diameter (a
plunger for glass capillary, Wiretrol® II, painted with black lubricant, industrial graphite
66
dry lubricant, the B’laster Corp.) was vertically presented at two positions using a
pneumatic slide (Festo), with the posterior position rewarded (Go trials) and anterior
position non-rewarded (No-go trials). The pole came into touch range within 100 ms of
pole motion onset. Mice used a single whisker (C2) to discriminate positions. Mice
indicated their decision through licking or withhold licking during the answer period
according to pole position. Licks in the 1s sampling period were ignored. On Hit trials,
mice received water rewards on the first lick in the 1s answer period. On False Alarm
trials, based on each mouse’s learning process, each lick during the answer period re-
triggered a timeout that lasted 0-4 seconds. Miss and Correct Rejection trials were
unpunished. The behavioral task was controlled by MATLAB-based BControl software
(C. Brody, Princeton University).

We trained mice in a stepwise manner. First, we associated the timing of cue and
pole to reward to let mice learn that water can come out of the lick port and the water is
temporally associated with a pole presentation. Mice usually learned the association in
a few minutes. Then we introduced Go trials only to let the mice learn the trial structure,
which usually took 1-2 sessions for the mice to achieve high performance. After the
mice were able to get over 10 Hit trials in a row, we introduced No-go trials. During early
training, we adjusted the No-go trials probability and time-out time on False Alarm trials
to help mice learn, settling on 50% No-go probability and 0s timeout once mice did not
get discouraged and stop licking after a series of misses. Expert threshold was set at
>70% accuracy continuously for 3 sessions. After the animals became experts, we
67
trimmed the animal's last whisker (C2 whisker) to test whether the mice learned the task
in a whisker-dependent manner.

Whisker motion acquisition and analysis
Backlit whisker motion video was acquired with a CMOS camera (Basler acA800-
500um), StreamPix Software (NorPix Inc.) at 311Hz through telecentric lens (0.09X½”
GoldTL™ #58-259, Edmund optics), a CMOS camera (Basler acA800-500um). to
record whisker motion. Camera frames were triggered and synced by BControl and
Ardunio. We tracked whisker position with the Janelia Whisker Tracker
(https://www.janelia.org/open-science/whisk-whisker-tracking (Clack et al., 2012)). The
fur was masked to improve tracking quality. The whisker’s azimuthal angle was
calculated at the intersection of the mask and whisker trace. Whisking midpoint,
amplitude and phase was decomposed from this angle using the Hilbert transform, as
described in Cheung et al., 2019.

Two-photon microscopy
The two-photon microscope (Neurolabware) used a galvanometer scanner
(6215H, Cambridge Technology), Pockels cell (350-105-02 KD∗P EO modulator,
Conoptics), a resonant scanner (CRS8, Cambridge Technology), an objective (W Plan-
Apochromat 20×/1.0, Zeiss), a GaAsP photomultiplier tube (H10770B-40, Hamamatsu),
and a 510 nm emission filter (FF01-510/84-50, Semrock). We used an 80 MHz tunable
68
laser at 940 nm (Insight DS+, Spectra-Physics) for GRAB-Ach3.0 excitation. Imaging
was continuous throughout each session. The scope was controlled by a MATLAB-
based software Scanbox with custom modifications. Imaging frequency was 15.44
frames/s on the size of the FOV (512 × 796). Spatial resolution was 1.4 μm per pixel.

Data analysis and statistics
All imaging data were processed in Matlab. We excluded the first 100s for each
session due to non-linear photodynamics which stabilized after 100 seconds of
continuous excitation scanning (SFig4.2). In Figure 2B we used the mean fluorescence
intensity of the sample session as baseline. In Figure 4.2C-E we used 16 frames after
trial start as baseline. Figure 4.3A and 4.3B we used 1 second before stimuli (pole onset
and pole withdrawal) as baseline). Figure 3D we used 16 frames after trial start as
baseline. Figure 4.4A, 4.4B, 4.5A and 4.6A we used 32 frames before first lick to 16
frames before first lick as baseline. Statistical comparisons were made using paired T
Test corrected for multiple comparisons.

69
Supplemental Figures

SFigure 4.1 Pole out cue induced whisking in all trial types.
A) Average whisking amplitude aligned to pole out cue by trial outcome, 3 expert sessions /
mouse.
70

SFigure 4.2 Technical details of targeting and imaging.
71
A) Intrinsic signal imaging result showing C2 barrel column location through the cranial window.
B) Left: Raw fluorescence trace of an example session. The colored area is excluded from
analysis due to non-stationary photodynamcis. Right: Zoom of same. C) Acetylcholine
fluorescence changes averaged by trial types, same session as Figure 4.2C.

SFigure 4.3 Whisking drives acetylcholine release in S1.
A) Mean whisking amplitude of 1s post pole in cue. B) Mean acetylcholine change vs. mean
whisking amplitude over 1s following pole in cue. C) Whisking amplitude heatmap sorted by the
mean whisking amplitude of 0-500ms post pole withdrawal cue. No lick trials pooled from 3
expert sessions, one mouse. D) Acetylcholine release signal heatmap sorted by the same
72
criteria as C. E) Mean acetylcholine release signal of 500ms post pole withdrawal cue vs. mean
whisking amplitude of 500ms post pole withdrawal cue. No lick trials pooled from 3 early and 3
expert sessions, 8 mice.

SFigure 4.4 Inter-lick interval distribution comparison between early and expert sessions.
A) Inter-lick interval from early (gray) and expert (black) sessions.

73
Chapter 5: Concluding Remarks

Acetylcholine, norepinephrine, dopamine and serotonin are four important
neuromodulators in the brain that play roles in a wide range of behaviors. Chapter 1
gives a review on these four neuromodulatory systems, from their anatomies to
functions. From all these literatures, a common conclusion is that the exact function of
these neuromodulatory systems depends on specific cell type, specific input they
receive, and specific target region they innervate. So for future study, we should record
neural activity in a cell type specific or circuit specific manner. Chapter 2 gives a review
on the methods being developed to monitor the neuromodulator release in the brain,
from traditional microdialysis to recently developed genetically encoded neuromodulator
indicators. When conducting experiments, we should always keep in mind the pros and
cons of the tools we are using. Even though GPCR based neuromodulator indicators
offer us unprecedented opportunity to monitor the neuromodulator release in genetically
defined cells with high spatial and temporal resolution longitudinally, we might still fail to
capture the neuromodulator release in tens millisecond or shorter time scale. One
concern for the GPCR based neuromodulator indicators is their potential buffer effect.
So far no reported buffer effect come from the studies that used those GPCR based
neuromodulator indicators. But when we were collecting the dataset for Chapter 3, we
found mice with dopamine indicator expressions had a slower learning rate compared to
the mice expressing other indicators, or even failed to learn the task. Although these
pilot results need to be further confirmed by more well-designed experiments, this is a
74
reminder for us to consider the caveats of using these GPCR based neuromodulator
indicators.

Chapter 3 focuses on the characterization of multiple GRAB sensor variants for
acetylcholine, norepinephrine, dopamine, and serotonin in vivo performance under two
photon microscopy. We confirmed the SNR difference seen in cultured cells among
different GRAB variants was also conserved in vivo. This validated the legitimacy of the
sensor optimization process. Chapter 3 also mentioned about the comparison among
multiple cortical neuromodulator releases during tactile based associative learning.
More detailed data analysis needs to be done to this rich dataset. And Chapter 4 used
acetylcholine dataset as an example to show what exact analysis we can do to
understand the cortical neuromodulator release dynamics. We found that acetylcholine
release in S1 is highly correlated with directed motor actions to gather information and
act upon it. Similar analysis will be applied to the other three neuromodulator datasets.
This multiple cortical neuromodulator release dynamics during associative learning
dataset can offer us complementary information to the direct recording from these
neuromodulatory nuclei. Because it has been shown that these neuromodulatory nuclei
can release more than one neuromodulator. For instance, dorsal raphe has both
serotonergic and dopaminergic neurons (Mattews et al., 2016, Cho et al., 2017). So
directly monitoring different neuromodulator releases at the target region will help us
parse out the different functions of these different neuromodulators, even some of them
are released from the same nuclei.

75
These pilot results described in Chapter 3 and Chapter 4 opened multiple
potential research paths for future studies. In Chapter 4, we concluded that the actual
reward didn’t induce acetylcholine release in S1 based on the finding that we didn’t see
acetylcholine release difference between Hit and False Alarm trials that have the same
lick numbers. But we didn’t exclude the possibility that the acetylcholine release was
triggered by reward expectation. To further test this hypothesis, we can introduce catch
trials and compare the acetylcholine release between catch trials and Hit trials. We can
also change the reward probability. The acetylcholine release signal should be
positively correlated with reward probability if the acetylcholine release is triggered by
reward expectation. Whether cortical acetylcholine release has a causal effect can also
be tested. Optogenetically inhibiting the cortical acetylcholine release during sampling
period should decrease the animal's hit rate if the cortical acetylcholine release tracks
reward expectation. In the pilot studies described in Chapter 4, we couldn’t distinguish
acetylcholine release from basal forebrain and local acetylcholine producing neurons.
Thus another interesting follow-up study can be to identify the release source of cortical
acetylcholine. We can record the basal forebrain cholinergic neuron axon terminals at
S1 when animals are performing the same tactile based associative learning task, and
compare with the acetylcholine release dynamics seen from GRAB imaging in S1. More
potential study directions can be drawn from the data analysis results of the multiple
neuromodulator release dynamics in the S1 dataset described in Chapter 3. In
summary, the studies this thesis described provided us new information on multiple
neuromodulator release dynamics in S1 during tactile based associative learning. These
studies also emphasized the importance of quantitative behavioral analysis in
76
interpreting neural data. Meanwhile, studies in this thesis opened multiple interesting
research paths to further explore the function of cortical neuromodulator release.

77
References

Aigner, Thomas G., David L. Walker, and Mortimer Mishkin. "Comparison of the effects
of scopolamine administered before and after acquisition in a test of visual recognition
memory in monkeys." Behavioral and neural biology 55.1 (1991): 61-67.
Arnsten, Amy FT, and Patricia S. Goldman-Rakic. "α2-Adrenergic mechanisms in
prefrontal cortex associated with cognitive decline in aged nonhuman primates."
Science 230.4731 (1985): 1273-1276.
Aston-Jones, G., and F. E. Bloom. "Nonrepinephrine-containing locus coeruleus
neurons in behaving rats exhibit pronounced responses to non-noxious environmental
stimuli." Journal of Neuroscience 1.8 (1981): 887-900.
Bang, Sun Jung, and Thomas H. Brown. "Muscarinic receptors in perirhinal cortex
control trace conditioning." Journal of Neuroscience 29.14 (2009): 4346-4350.
Barnea, Gilad, et al. "The genetic design of signaling cascades to record receptor
activation." Proceedings of the National Academy of Sciences 105.1 (2008): 64-69.
Barter, Joseph W., et al. "Beyond reward prediction errors: the role of dopamine in
movement kinematics." Frontiers in integrative neuroscience 9 (2015): 39.
Beier, Kevin T., et al. "Circuit architecture of VTA dopamine neurons revealed by
systematic input-output mapping." Cell 162.3 (2015): 622-634.
Berke, Joshua D. "What does dopamine mean?." Nature neuroscience 21.6 (2018):
787-793.
Berridge, Craig W., and Barry D. Waterhouse. "The locus coeruleus–noradrenergic
system: modulation of behavioral state and state-dependent cognitive processes." Brain
research reviews 42.1 (2003): 33-84.
Bindra, D., and J. F. Campbell. "Motivational effects of rewarding intracranial
stimulation." Nature 215.5099 (1967): 375-376.
Bloem, Bernard, et al. "Topographic mapping between basal forebrain cholinergic
neurons and the medial prefrontal cortex in mice." Journal of Neuroscience 34.49
(2014): 16234-16246.
Bollu, Tejapratap, et al. "Cortex-dependent corrections as the tongue reaches for and
misses targets." Nature 594.7861 (2021): 82-87.
Borden, Philip M., et al. "A fast genetically encoded fluorescent sensor for faithful in vivo
acetylcholine detection in mice, fish, worms and flies." (2020).
Broks, P., et al. "Modelling dementia: effects of scopolamine on memory and attention."
Neuropsychologia 26.5 (1988): 685-700.
78
Carter, Matthew E., et al. "Tuning arousal with optogenetic modulation of locus
coeruleus neurons." Nature neuroscience 13.12 (2010): 1526-1533.
Cazettes, Fanny, et al. "Phasic activation of dorsal raphe serotonergic neurons
increases pupil size." Current Biology 31.1 (2021): 192-197.
Chaves-Coira, Irene, et al. "Basal forebrain nuclei display distinct projecting pathways
and functional circuits to sensory primary and prefrontal cortices in the rat." Frontiers in
neuroanatomy 12 (2018): 69.
Cheung, Jonathan, et al. "The sensorimotor basis of whisker-guided anteroposterior
object localization in head-fixed mice." Current Biology 29.18 (2019): 3029-3040.
Cho, Jounhong Ryan, et al. "Dorsal raphe dopamine neurons modulate arousal and
promote wakefulness by salient stimuli." Neuron 94.6 (2017): 1205-1219.
Chubykin, Alexander A., et al. "A cholinergic mechanism for reward timing within
primary visual cortex." Neuron 77.4 (2013): 723-735.
Clack, Nathan G., et al. "Automated tracking of whiskers in videos of head fixed
rodents." PLoS computational biology 8.7 (2012): e1002591.
Clarke, Samuel, and Jay A. Trowill. "Sniffing and motivated behavior in the rat."
Physiology & behavior 6.1 (1971): 49-52.
Cohen, Jeremiah Y., Mackenzie W. Amoroso, and Naoshige Uchida. "Serotonergic
neurons signal reward and punishment on multiple timescales." Elife 4 (2015): e06346.
Correia, Patricia A., et al. "Transient inhibition and long-term facilitation of locomotion by
phasic optogenetic activation of serotonin neurons." Elife 6 (2017): e20975.
Crouse, Richard B., et al. "Acetylcholine is released in the basolateral amygdala in
response to predictors of reward and enhances the learning of cue-reward
contingency." Elife 9 (2020): e57335.
Crow, T. J., and I. G. Grove-White. "An analysis of the learning deficit following
hyoscine administration to man." British Journal of Pharmacology 49.2 (1973): 322.
Croxson, Paula L., Diana A. Kyriazis, and Mark G. Baxter. "Cholinergic modulation of a
specific memory function of prefrontal cortex." Nature neuroscience 14.12 (2011): 1510-
1512.
Da Silva, Joaquim Alves, et al. "Dopamine neuron activity before action initiation gates
and invigorates future movements." Nature 554.7691 (2018): 244-248.
Dahl, Martin J., et al. "Rostral locus coeruleus integrity is associated with better memory
performance in older adults." Nature Human Behaviour 3.11 (2019): 1203-1214.
Dauer, William, and Serge Przedborski. "Parkinson's disease: mechanisms and
models." Neuron 39.6 (2003): 889-909.
79
de Jong, Johannes W., et al. "A neural circuit mechanism for encoding aversive stimuli
in the mesolimbic dopamine system." Neuron 101.1 (2019): 133-151.
Deitcher, Yair, et al. "Nonlinear relationship between multimodal adrenergic responses
and local dendritic activity in primary sensory cortices." bioRxiv (2019): 814657.
Dong, Ao, et al. "A fluorescent sensor for spatiotemporally resolved imaging of
endocannabinoid dynamics in vivo." Nature biotechnology (2021): 1-12.
Donoghue, John P., and Kristen L. Carroll. "Cholinergic modulation of sensory
responses in rat primary somatic sensory cortex." Brain research 408.1-2 (1987): 367-
371.
Ebrahimi, Shima, et al. "Central β-adrenergic receptors play an important role in the
enhancing effect of voluntary exercise on learning and memory in rat." Behavioural
brain research 208.1 (2010): 189-193.
Eggermann, Emmanuel, et al. "Cholinergic signals in mouse barrel cortex during active
whisker sensing." Cell reports 9.5 (2014): 1654-1660.
Engelhard, Ben, et al. "Specialized coding of sensory, motor and cognitive variables in
VTA dopamine neurons." Nature 570.7762 (2019): 509-513.
Feng, Jiesi, et al. "A genetically encoded fluorescent sensor for rapid and specific in
vivo detection of norepinephrine." Neuron 102.4 (2019): 745-761.
Fonseca, Madalena S., Masayoshi Murakami, and Zachary F. Mainen. "Activation of
dorsal raphe serotonergic neurons promotes waiting but is not reinforcing." Current
Biology 25.3 (2015): 306-315.
Froemke, Robert C., et al. "Long-term modification of cortical synapses improves
sensory perception." Nature neuroscience 16.1 (2013): 79-88.
Gasselin, Célia, et al. "Cell-type-specific nicotinic input disinhibits mouse barrel cortex
during active sensing." Neuron 109.5 (2021): 778-787.
Giacobini, Ezio. "Cholinergic function and Alzheimer's disease." International journal of
geriatric psychiatry 18.S1 (2003): S1-S5.
Glennon, Erin, et al. "Locus coeruleus activation accelerates perceptual learning." Brain
research 1709 (2019): 39-49.
Goard, Michael, and Yang Dan. "Basal forebrain activation enhances cortical coding of
natural scenes." Nature neuroscience 12.11 (2009): 1444-1449.
Gras, Christelle, et al. "A third vesicular glutamate transporter expressed by cholinergic
and serotoninergic neurons." Journal of Neuroscience 22.13 (2002): 5442-5451.
Guo, Wei, Blaise Robert, and Daniel B. Polley. "The cholinergic basal forebrain links
auditory stimuli with delayed reinforcement to support learning." Neuron 103.6 (2019):
1164-1177.
80
Guo, Zengcai V., et al. "Procedures for behavioral experiments in head-fixed mice."
PloS one 9.2 (2014): e88678.
Hamid, Arif A., et al. "Mesolimbic dopamine signals the value of work." Nature
neuroscience 19.1 (2016): 117-126.
Hangya, Balázs, et al. "Central cholinergic neurons are rapidly recruited by
reinforcement feedback." Cell 162.5 (2015): 1155-1168.
Harrison, Thomas C., et al. "Calcium imaging of basal forebrain activity during innate
and learned behaviors." Frontiers in neural circuits 10 (2016): 36.
Hars, Bernard, et al. "Basal forebrain stimulation facilitates tone-evoked responses in
the auditory cortex of awake rat." Neuroscience 56.1 (1993): 61-74.
Hasselmo, Michael E., and Martin Sarter. "Modes and models of forebrain cholinergic
neuromodulation of cognition." Neuropsychopharmacology 36.1 (2011): 52-73.
Herrero, Jose L., et al. "Acetylcholine contributes through muscarinic receptors to
attentional modulation in V1." Nature 454.7208 (2008): 1110-1114.
Hickey, Louise, et al. "Optoactivation of locus ceruleus neurons evokes bidirectional
changes in thermal nociception in rats." Journal of Neuroscience 34.12 (2014): 4148-
4160.
Hires, Samuel Andrew, Yongling Zhu, and Roger Y. Tsien. "Optical measurement of
synaptic glutamate spillover and reuptake by linker optimized glutamate-sensitive
fluorescent reporters." Proceedings of the National Academy of Sciences 105.11
(2008): 4411-4416.
Hobson, J. Allan, Robert W. McCarley, and Peter W. Wyzinski. "Sleep cycle oscillation:
reciprocal discharge by two brainstem neuronal groups." Science 189.4196 (1975): 55-
58.
Holets, V. R., et al. "Locus coeruleus neurons in the rat containing neuropeptide Y,
tyrosine hydroxylase or galanin and their efferent projections to the spinal cord, cerebral
cortex and hypothalamus." Neuroscience 24.3 (1988): 893-906.
Howe, Mark W., et al. "Prolonged dopamine signalling in striatum signals proximity and
value of distant rewards." nature 500.7464 (2013): 575-579.
Howe, Mark W., and Daniel A. Dombeck. "Rapid signalling in distinct dopaminergic
axons during locomotion and reward." Nature 535.7613 (2016): 505-510.
Inglis, F. M., and H. C. Fibiger. "Increases in hippocampal and frontal cortical
acetylcholine release associated with presentation of sensory stimuli." Neuroscience
66.1 (1995): 81-86.
81
Ishimatsu, Masaru, and John T. Williams. "Synchronous activity in locus coeruleus
results from dendritic interactions in pericoerulear regions." Journal of Neuroscience
16.16 (1996): 5196-5204.
Ishimura, Kazuhisa, et al. "Quantitative analysis of the distribution of serotonin-
immunoreactive cell bodies in the mouse brain." Neuroscience letters 91.3 (1988): 265-
270.
Jiang, Li, et al. "Cholinergic signaling controls conditioned fear behaviors and enhances
plasticity of cortical-amygdala circuits." Neuron 90.5 (2016): 1057-1070.
Jing, Miao, et al. "A genetically encoded fluorescent acetylcholine indicator for in vitro
and in vivo studies." Nature biotechnology 36.8 (2018): 726-737.
Jing, Miao, et al. "An optimized acetylcholine sensor for monitoring in vivo cholinergic
activity." Nature methods 17.11 (2020): 1139-1146.
Kilgard, Michael P., and Michael M. Merzenich. "Cortical map reorganization enabled by
nucleus basalis activity." Science 279.5357 (1998): 1714-1718.
Kim, Jae-Hyun, et al. "Selectivity of neuromodulatory projections from the basal
forebrain and locus ceruleus to primary sensory cortices." Journal of Neuroscience
36.19 (2016): 5314-5327.
Kim, Jinho, et al. "Behavioral and Neural Bases of Tactile Shape Discrimination
Learning in Head-Fixed Mice." Neuron 108.5 (2020): 953-967.
Kimchi, Eyal Y., et al. "Reward contingency gates selective cholinergic suppression of
amygdala neurons." bioRxiv (2022).
Klinkenberg, Inge, Anke Sambeth, and Arjan Blokland. "Acetylcholine and attention."
Behavioural brain research 221.2 (2011): 430-442.
Kruse, Andrew C., et al. "Activation and allosteric modulation of a muscarinic
acetylcholine receptor." Nature 504.7478 (2013): 101-106.
Kuchibhotla, Kishore V., et al. "Parallel processing by cortical inhibition enables context-
dependent behavior." Nature neuroscience 20.1 (2017): 62-71.
Lacefield, Clay O., et al. "Reinforcement learning recruits somata and apical dendrites
across layers of primary sensory cortex." Cell reports 26.8 (2019): 2000-2008.
Lange, Klaus W., et al. "Altered muscarinic and nicotinic receptor densities in cortical
and subcortical brain regions in Parkinson's disease." Journal of neurochemistry 60.1
(1993): 197-203.
Lee, Seung-Hee, and Yang Dan. "Neuromodulation of brain states." Neuron 76.1
(2012): 209-222.
Lee, Soohyun, et al. "A disinhibitory circuit mediates motor integration in the
somatosensory cortex." Nature neuroscience 16.11 (2013): 1662-1670.
82
Lerner, Talia N., et al. "Intact-brain analyses reveal distinct information carried by SNc
dopamine subcircuits." Cell 162.3 (2015): 635-647.
Leslie, Frances M., et al. "Noradrenergic changes and memory loss in aged mice."
Brain research 359.1-2 (1985): 292-299.
Letzkus, Johannes J., et al. "A disinhibitory microcircuit for associative fear learning in
the auditory cortex." Nature 480.7377 (2011): 331-335.
Li, Xiangning, et al. "Generation of a whole-brain atlas for the cholinergic system and
mesoscopic projectome analysis of basal forebrain cholinergic neurons." Proceedings of
the National Academy of Sciences 115.2 (2018): 415-420.
Li, Yi, et al. "Serotonin neurons in the dorsal raphe nucleus encode reward signals."
Nature communications 7.1 (2016): 1-15.
Liu, Cheng-Hang, et al. "Selective activation of a putative reinforcement signal
conditions cued interval timing in primary visual cortex." Current Biology 25.12 (2015):
1551-1561.
Liu, Zhixiang, et al. "Dorsal raphe neurons signal reward through 5-HT and glutamate."
Neuron 81.6 (2014): 1360-1374.
Lohani, Sweyta, et al. "Dual color mesoscopic imaging reveals spatiotemporally
heterogeneous coordination of cholinergic and neocortical activity." bioRxiv (2020).
Lohr, James B., and Dilip V. Jeste. "Locus ceruleus morphometry in aging and
schizophrenia." Acta Psychiatrica Scandinavica 77.6 (1988): 689-697.
Maier-Peuschel, Monika, et al. "A fluorescence resonance energy transfer-based M2
muscarinic receptor sensor reveals rapid kinetics of allosteric modulation." Journal of
Biological Chemistry 285.12 (2010): 8793-8800.
Malamud, Nathan. "Psychiatric disorder with intracranial tumors of limbic system."
Archives of Neurology 17.2 (1967): 113-123.
Manunta, Yves, and Jean-Marc Edeline. "Noradrenergic induction of selective plasticity
in the frequency tuning of auditory cortex neurons." Journal of neurophysiology 92.3
(2004): 1445-1463.
Marcinkiewcz, Catherine A., et al. "Serotonin engages an anxiety and fear-promoting
circuit in the extended amygdala." Nature 537.7618 (2016): 97-101.
Marcyniuk, B., D. M. A. Mann, and P. O. Yates. "The topography of nerve cell loss from
the locus caeruleus in elderly persons." Neurobiology of aging 10.1 (1989): 5-9.
Margules, D. L., et al. "Hypothalamic norepinephrine: circadian rhythms and the control
of feeding behavior." Science 178.4061 (1972): 640-643.
83
Marrosu, Francesco, et al. "Microdialysis measurement of cortical and hippocampal
acetylcholine release during sleep-wake cycle in freely moving cats." Brain research
671.2 (1995): 329-332.
Martins, Ana Raquel O., and Robert C. Froemke. "Coordinated forms of noradrenergic
plasticity in the locus coeruleus and primary auditory cortex." Nature neuroscience
18.10 (2015): 1483-1492.
Marvin, Jonathan S., et al. "A genetically encoded fluorescent sensor for in vivo imaging
of GABA." Nature methods 16.8 (2019): 763-770.
Matthews, Kim L., et al. "Noradrenergic changes, aggressive behavior, and cognition in
patients with dementia." Biological psychiatry 51.5 (2002): 407-416.
Matthews, Gillian A., et al. "Dorsal raphe dopamine neurons represent the experience of
social isolation." Cell 164.4 (2016): 617-631.
McCall, Jordan G., et al. "CRH engagement of the locus coeruleus noradrenergic
system mediates stress-induced anxiety." Neuron 87.3 (2015): 605-620.
McCutcheon, Robert, et al. "Defining the locus of dopaminergic dysfunction in
schizophrenia: a meta-analysis and test of the mesolimbic hypothesis." Schizophrenia
bulletin 44.6 (2018): 1301-1311.
McGaughy, Jill, et al. "Selective behavioral and neurochemical effects of cholinergic
lesions produced by intrabasalis infusions of 192 IgG-saporin on attentional
performance in a five-choice serial reaction time task." Journal of Neuroscience 22.5
(2002): 1905-1913.
Mesulam, M. M., et al. "Central cholinergic pathways in the rat: an overview based on
an alternative nomenclature (Ch1–Ch6)." Neuroscience 10.4 (1983): 1185-1201.
Mineur, Yann S., et al. "Nicotine decreases food intake through activation of POMC
neurons." Science 332.6035 (2011): 1330-1332.
Miyazaki, Katsuhiko, Kayoko W. Miyazaki, and Kenji Doya. "Activation of dorsal raphe
serotonin neurons underlies waiting for delayed rewards." Journal of Neuroscience 31.2
(2011): 469-479.
Miyazaki, Kayoko W., et al. "Optogenetic activation of dorsal raphe serotonin neurons
enhances patience for future rewards." Current Biology 24.17 (2014): 2033-2040.
Mohebi, Ali, et al. "Dissociable dopamine dynamics for learning and motivation." Nature
570.7759 (2019): 65-70.
Moore, Jeffrey D., David Kleinfeld, and Fan Wang. "How the brainstem controls
orofacial behaviors comprised of rhythmic actions." Trends in neurosciences 37.7
(2014): 370-380.
84
Muller, Arnaud, et al. "Cell-based reporters reveal in vivo dynamics of dopamine and
norepinephrine release in murine cortex." Nature methods 11.12 (2014): 1245-1252.
Nakanishi, Jun, et al. "FRET-based monitoring of conformational change of the β2
adrenergic receptor in living cells." Biochemical and biophysical research
communications 343.4 (2006): 1191-1196.
Newman, Ehren L., et al. "Cholinergic modulation of cognitive processing: insights
drawn from computational models." Frontiers in behavioral neuroscience 6 (2012): 24.
Nguyen, Quoc-Thang, et al. "An in vivo biosensor for neurotransmitter release and in
situ receptor activity." Nature neuroscience 13.1 (2010): 127-132.
O'Connor, Daniel H., et al. "Neural activity in barrel cortex underlying vibrissa-based
object localization in mice." Neuron 67.6 (2010): 1048-1061.
O'Connor, Daniel H., et al. "Vibrissa-based object localization in head-fixed mice."
Journal of Neuroscience 30.5 (2010): 1947-1967.
Okaty, Benjamin W., et al. "Multi-scale molecular deconstruction of the serotonin neuron
system." Neuron 88.4 (2015): 774-791.
Palmiter, Richard D. "Dopamine signaling in the dorsal striatum is essential for
motivated behaviors: lessons from dopamine-deficient mice." Annals of the New York
Academy of Sciences 1129.1 (2008): 35-46.
Parikh, Vinay, et al. "Prefrontal acetylcholine release controls cue detection on multiple
timescales." Neuron 56.1 (2007): 141-154.
Patriarchi, Tommaso, et al. "Ultrafast neuronal imaging of dopamine dynamics with
designed genetically encoded sensors." Science 360.6396 (2018): eaat4422.
Peñalva, Rosana G., et al. "Effect of sleep and sleep deprivation on serotonergic
neurotransmission in the hippocampus: a combined in vivo microdialysis/EEG study in
rats." European Journal of Neuroscience 17.9 (2003): 1896-1906.
Phillips, Paul EM, et al. "Subsecond dopamine release promotes cocaine seeking."
Nature 422.6932 (2003): 614-618.
Pinto, Lucas, et al. "Fast modulation of visual perception by basal forebrain cholinergic
neurons." Nature neuroscience 16.12 (2013): 1857-1863.
Quik, M., et al. "Loss of α-conotoxinMII-and A85380-sensitive nicotinic receptors in
Parkinson's disease striatum." Journal of neurochemistry 88.3 (2004): 668-679.
Reimer, Jacob, et al. "Pupil fluctuations track rapid changes in adrenergic and
cholinergic activity in cortex." Nature communications 7.1 (2016): 1-7.
Ren, Jing, et al. "Habenula “cholinergic” neurons corelease glutamate and acetylcholine
and activate postsynaptic neurons via distinct transmission modes." Neuron 69.3
(2011): 445-452.
85
Ren, Jing, et al. "Anatomically defined and functionally distinct dorsal raphe serotonin
sub-systems." Cell 175.2 (2018): 472-487.
Richardson, Russell T., and M. R. DeLong. "Context-dependent responses of primate
nucleus basalis neurons in a go/no-go task." Journal of Neuroscience 10.8 (1990):
2528-2540.
Robert, Blaise, et al. "A functional topography within the cholinergic basal forebrain for
processing sensory cues associated with reward and punishment." bioRxiv (2021).
Robinson, Donita L., et al. "Detecting subsecond dopamine release with fast-scan cyclic
voltammetry in vivo." Clinical chemistry 49.10 (2003): 1763-1773.
Rasmussen, Søren GF, et al. "Crystal structure of the human β2 adrenergic G-protein-
coupled receptor." Nature 450.7168 (2007): 383-387.
Sara, Susan J. "The locus coeruleus and noradrenergic modulation of cognition."
Nature reviews neuroscience 10.3 (2009): 211-223.
Scarr, Elizabeth, et al. "Cholinergic connectivity: it's implications for psychiatric
disorders." Frontiers in cellular neuroscience 7 (2013): 55.
Schultz, Kristin N., and Robert T. Kennedy. "Time-resolved microdialysis for in vivo
neurochemical measurements and other applications." Annu. Rev. Anal. Chem. 1
(2008): 627-661.
Schultz, Wolfram, Peter Dayan, and P. Read Montague. "A neural substrate of
prediction and reward." Science 275.5306 (1997): 1593-1599.
Schwarz, Lindsay A., and Liqun Luo. "Organization of the locus coeruleus-
norepinephrine system." Current Biology 25.21 (2015): R1051-R1056.
Sciolino, Natale R., et al. "A role for the locus coeruleus in the modulation of feeding."
BioRxiv (2019).
Shadlen, Michael N., and William T. Newsome. "Motion perception: seeing and
deciding." Proceedings of the national academy of sciences 93.2 (1996): 628-633.
Steinberg, Elizabeth E., et al. "A causal link between prediction errors, dopamine
neurons and learning." Nature neuroscience 16.7 (2013): 966-973.
Sun, Fangmiao, et al. "A genetically encoded fluorescent sensor enables rapid and
specific detection of dopamine in flies, fish, and mice." Cell 174.2 (2018): 481-496.
Sun, Fangmiao, et al. "Next-generation GRAB sensors for monitoring dopaminergic
activity in vivo." Nature methods 17.11 (2020): 1156-1166.
Swanson, L. W. "The locus coeruleus: a cytoarchitectonic, Golgi and
immunohistochemical study in the albino rat." Brain research 110.1 (1976): 39-56.
86
Takahashi, Naoya, et al. "Active cortical dendrites modulate perception." Science
354.6319 (2016): 1587-1590.
Takahashi, Naoya, et al. "Active dendritic currents gate descending cortical outputs in
perception." Nature Neuroscience 23.10 (2020): 1277-1285.
Teissier, Anne, et al. "Activity of raphé serotonergic neurons controls emotional
behaviors." Cell reports 13.9 (2015): 1965-1976.
Tremblay, Nicole, Richard A. Warren, and Robert W. Dykes. "Electrophysiological
studies of acetylcholine and the role of the basal forebrain in the somatosensory cortex
of the cat. II. Cortical neurons excited by somatic stimuli." Journal of neurophysiology
64.4 (1990): 1212-1222.
Unger, Elizabeth K., et al. "Directed evolution of a selective and sensitive serotonin
sensor via machine learning." Cell 183.7 (2020): 1986-2002.
Urban, Daniel J., et al. "Elucidation of the behavioral program and neuronal network
encoded by dorsal raphe serotonergic neurons." Neuropsychopharmacology 41.5
(2016): 1404-1415.
Urban-Ciecko, Joanna, et al. "Precisely timed nicotinic activation drives SST inhibition in
neocortical circuits." Neuron 97.3 (2018): 611-625.
Venton, B. Jill, and Qun Cao. "Fundamentals of fast-scan cyclic voltammetry for
dopamine detection." Analyst 145.4 (2020): 1158-1168.
Vilardaga, Jean-Pierre, et al. "Measurement of the millisecond activation switch of G
protein–coupled receptors in living cells." Nature biotechnology 21.7 (2003): 807-812.
Wan, Jinxia, et al. "A genetically encoded sensor for measuring serotonin dynamics."
Nature neuroscience 24.5 (2021): 746-752.
Watabe-Uchida, Mitsuko, et al. "Whole-brain mapping of direct inputs to midbrain
dopamine neurons." Neuron 74.5 (2012): 858-873.
Williams, Leena E., and Anthony Holtmaat. "Higher-order thalamocortical inputs gate
synaptic long-term potentiation via disinhibition." Neuron 101.1 (2019): 91-102.
Wu, Jiahui, et al. "Genetically encoded glutamate indicators with altered color and
topology." ACS chemical biology 13.7 (2018): 1832-1837.
Wu, Zhaofa, et al. "A sensitive GRAB sensor for detecting extracellular ATP in vitro and
in vivo." Neuron 110.5 (2022): 770-782.
Xu, Min, et al. "Basal forebrain circuit for sleep-wake control." Nature neuroscience
18.11 (2015): 1641-1647.
Yang, Ben, et al. "Locus coeruleus anchors a trisynaptic circuit controlling fear-induced
suppression of feeding." Neuron 109.5 (2021): 823-838.
87
Yao, Zizhen, et al. "A taxonomy of transcriptomic cell types across the isocortex and
hippocampal formation." Cell 184.12 (2021): 3222-3241.
Zaborszky, Laszlo, et al. "Stereotaxic probabilistic maps of the magnocellular cell
groups in human basal forebrain." Neuroimage 42.3 (2008): 1127-1141.
Zaborszky, Laszlo, et al. "Organization of the basal forebrain cholinergic projection
system: specific or diffuse?." The rat nervous system. Academic Press, 2015a. 491-507.
Zaborszky, Laszlo, et al. "Neurons in the basal forebrain project to the cortex in a
complex topographic organization that reflects corticocortical connectivity patterns: an
experimental study based on retrograde tracing and 3D reconstruction." Cerebral cortex
25.1 (2015b): 118-137.
Zhang, Hao, Shih-Chieh Lin, and Miguel AL Nicolelis. "A distinctive subpopulation of
medial septal slow-firing neurons promote hippocampal activation and theta
oscillations." Journal of neurophysiology 106.5 (2011): 2749-2763.

Abstract (if available)

Abstract Neuromodulators (acetylcholine, norepinephrine, dopamine and serotonin) play important roles in a wide range of behaviors. And the malfunction of these four neuromodulatory systems are found related to multiple brain disorders. The importance of these neuromodulators are largely appreciated, but the temporal and spatial release dynamics of these neuromodulators in the primary somatosensory cortex (S1) are still left largely unrevealed. With the recently developed GPCR based genetically-encoded neuromodulator indicators, we are able to monitor multiple neuromodulators release dynamics in S1 in awake behaving animals. This thesis reviewed latest studies on cholinergic, noradrenergic, dopaminergic and serotonergic systems, as well as the methods being developed to monitor the neuromodulator release in the brain. Besides, two projects are included in this thesis. One focused on the characterization of multiple neuromodulator indicators in vivo performance under two photon microscopy. The other one characterized the acetylcholine release dynamics during tactile-based associative learning in detail. These studies provide insight into the different roles that these different cortical neuromodulator releases play during associative learning.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Cell-type specialization of layer 5 excitatory neuron functions in tactile behavior underlying object localization

PDF

Imaging cortical noradrenergic dynamics with genetically-engineered fluorescent indicators

PDF

Transient receptor potential melastatin 8 is required for migraine-like pain behaviors in mice

PDF

Computational investigation of cholinergic modulation of the hippocampus and directional encoding of touch in somatosensory cortex

PDF

Sensory learning and anatomical plasticity in barrel cortex

PDF

Elucidating the elusive role of TRPA1 in neurogenic inflammatory cold pain

PDF

Quantitative analysis of mouse medial prefrontal cortex whole brain structural connectivity

PDF

The behavioral and neural bases of tactile object localization

PDF

Dynamical representation learning for multiscale brain activity

PDF

Motor cortical representations of sensorimotor information during skill learning

PDF

The expected and unexpected roles of TRPM8: cold pain and metabolism

PDF

Dynamic processes underlying cerebral cortical development with lifespan impact

PDF

Emergent visualization technology: evolution of PSD95.FingR

PDF

The role of the cofilin/Limk1 signaling pathway in axon growth during development and regeneration

PDF

The role of nucleotide excision repair in cisplatin chemotherapy-induced deafness

PDF

Characterizing response and plasticity in sensory cortices of the Fmr1⁻⁄⁻ mouse

PDF

Spatial and temporal precision of inhibitory and excitatory neurons in the murine dorsal lateral geniculate nucleus

PDF

Mapping multi-scale connectivity of the mouse posterior parietal cortex

PDF

Regulatory mechanisms of gamma-aminobutyric acid transporter 1 in the rat brain

PDF

The roles of TRPM8 in cold sensation: the six sides of TRPM8

Asset Metadata

Creator Zou, Jing (author)

Core Title Imaging neuromodulator dynamics in somatosensory cortex

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Degree Conferral Date 2022-08

Publication Date 08/07/2022

Defense Date 05/26/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Acetylcholine,Dopamine,neuromodulators,Norepinephrine,OAI-PMH Harvest,serotonin,somatosensory cortex

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor McKemy, David (committee chair), Arnold, Donald (committee member), Hires, Samuel Andrew (committee member)

Creator Email jingzou@usc.edu,lily.jzou@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111376229

Unique identifier UC111376229

Legacy Identifier etd-ZouJing-11077

Document Type Dissertation

Format application/pdf (imt)

Rights Zou, Jing

Type texts

Source 20220808-usctheses-batch-972 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu