NEURAL PROCESSING OF DECISION COSTS
AND AVERSIVE EVENTS
A Dissertation
Presented to the Faculty of the Graduate School
of Cornell University
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
by
Vaida Rimeikyte
August 2019
©c 2019 Vaida Rimeikyte
ALL RIGHTS RESERVED
NEURAL PROCESSING OF DECISION COSTS AND AVERSIVE EVENTS
Vaida Rimeikyte, Ph.D.
Cornell University 2019
The ability to engage in avoidance behaviors is vital to adaptively navigating
different environments, especially if these environments are unpredictable or
hostile. Despite considerable advances in understanding the neural mecha-
nisms that give rise to avoidance behaviors, some component processes remain
controversial. This dissertation focuses on understanding the neural mecha-
nisms in determining two elements of avoidance behaviors: value (whether a
particular reward is worth whatever it takes to get it) and the valence (whether
a stimulus or event is good or bad).
The first part of this dissertation focuses on understanding the neural mech-
anisms of value reduction in context of different decision costs. Specifically,
Chapter 2 examines the neural representations of delay, effort, and probability
costs, which have all been shown to devalue rewards, thereby making the re-
ward pursuit less likely. We used a computational meta-analysis to analyze the
brain activations associated with the three types of costs reported neuroimaging
literature over the past 15 years. We found that all 3 costs consistently engage
dorsal striatum and anterior insula. We also found that delay and probability,
but not effort costs, consistently engage prefrontal regions (BA46) associated
with top-down control.
The latter part of the dissertation focuses on understanding negative valence
representations in the human brain. In chapter 3 we examined the neural corre-
lates of negative valence across different type of aversive stimuli. In this study
participants were subjected to painful pressure as well as emotionally and phys-
ically aversive sounds. We found that negative valence modulates activity in
areas primarily associated with sensory processing, supporting the theory of
modality-specific affect. In chapter 4 we investigated the role of emotion reg-
ulation in the processing of aversive stimuli. We found that trait emotion reg-
ulation ability (as measured by DERS questionnaire) did not modify subjective
displeasure ratings of any aversive stimuli. In addition, DERS scores modu-
lated brain activity in pressure pain trials during pressure pain anticipation and
pain delivery, but had no effect on neural processing of emotionally aversive
or physically aversive sounds. Specifically, we found that people with lower
trait emotional regulatory capacity show less preparatory activity during pain
anticipation and instead rely on dorsolateral prefrontal cortex to regulate pain.
BIOGRAPHICAL SKETCH
Vaida was born 30 years ago in Kaunas, Lithuania. She grew up in an at the
time undeveloped suburb surrounded by fields and cows . Yet despite having
nothing to complain about, she felt the urge for going. Following that urge,
Vaida ended up leaving her home town after her Junior year in high school to
study at Red Cross Nordic United World College. This experience has proven
to be formative in many ways, not the least of which was introducing Vaida to
philosophy of mind. After spending many fruitless hours contemplating what
makes a Self, Vaida graduated high school to study Psychology at Harvard Uni-
versity. Her youthful cynicism towards ideas of free will, landed her in the lab
of Daniel Wegner, where she continued to contemplate the nature of mind, now
with some experimental constrains. While working in the lab, Vaida got inter-
ested in the nature of emotion and emotion regulation. Her curiosity on the
neural mechanisms that give rise to affective processes led her to continue her
studies at Cornell University in Affect and Cognition lab.
iii
To my parents,
who anticipated all the possible aversive events so I would not have to.
iv
ACKNOWLEDGEMENTS
Thank you to everyone who has taught about being a better scientist and being
a better human along the way. I consider myself very lucky to have met so many
of you.
Thank you to my mentors. You have taught me a lot through incisive advice
and even more through your example and presence.
Thank you to Adam Anderson for giving me the freedom and encourage-
ment to pursue my true interests in science and for pushing me not to settle for
easy answers. I hope your influence continues to shape me for years to come.
Thank you to Geoffrey Fisher for being beyond supportive and helpful. Your
insightful, straight forward advice has helped me more than I can explain.
Thank you to Melissa Warden. To me, you are a role model for what a sci-
entist should be - unerringly methodical, astonishingly insightful and innately
curious. You continue to inspire me to be a better scientist and for that, I am
beyond grateful
Thank you to my roommates - Kasi, Roli, Stephanie, Joshua, and Sara - who
have fed me, comforted me, entertained me, and supported me throughout this
process. I wouldn’t have made it without you.
To my lab mates. Your support, constructive advice, encouragement and
pictures of cute animals made all the difference.
To my family. Even from far away you make me feel connected and taken
care of.
v
TABLE OF CONTENTS
Biographical Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
1 Introduction 1
2 Weighing the costs: Neural Representation of Delay, Effort, and Prob-
ability Costs 4
2.1 ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.1 Study Selection . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3.2 ALE analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.1 Effort meta-analysis . . . . . . . . . . . . . . . . . . . . . . 13
2.4.2 Delay meta-analysis . . . . . . . . . . . . . . . . . . . . . . 13
2.4.3 Risk meta-analysis . . . . . . . . . . . . . . . . . . . . . . . 13
2.4.4 Common Cost Representations . . . . . . . . . . . . . . . . 14
2.4.5 Distinct Cost Representations . . . . . . . . . . . . . . . . . 14
2.4.6 Dissociation between Effort and Delay Costs . . . . . . . . 16
2.4.7 Overlap between Effort and Delay Costs . . . . . . . . . . 18
2.4.8 Dissociation between Effort and Risk Cost . . . . . . . . . 18
2.4.9 Overlap between Effort and Risk Costs . . . . . . . . . . . 19
2.4.10 Dissociation between Delay and Risk Costs . . . . . . . . . 19
2.4.11 Overlap between Delay and Risk Cost . . . . . . . . . . . . 19
2.5 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.1 Common Activations . . . . . . . . . . . . . . . . . . . . . . 21
2.5.2 Cost Specific Activations . . . . . . . . . . . . . . . . . . . . 22
2.5.3 Limitations and future directions . . . . . . . . . . . . . . . 25
2.6 CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3 AFFECTIVE AND SENSORY REPRESENTATIONS OF NEGATIVE
VALENCE ACROSS DIFFERENT AVERSIVE STIMULI 27
3.1 ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.1 Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.2 Stimuli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3.3 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
vi
3.3.4 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4.1 Behavioral Results . . . . . . . . . . . . . . . . . . . . . . . 35
3.4.2 FMRI Results . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.5 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.5.1 Affective representations of negative valence . . . . . . . . 40
3.5.2 Sensory representations of negative valence . . . . . . . . 41
3.5.3 Limitations and future directions . . . . . . . . . . . . . . . 41
4 Trait Differences in Emotion Regulation Modulate Neural Processing
of Painful Pressure but not of Aversive Sound Stimuli 42
4.1 ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.2 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.3 METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3.1 Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3.2 Stimuli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3.3 Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.3.4 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4 RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.4.1 Behavioral Results . . . . . . . . . . . . . . . . . . . . . . . 49
4.4.2 FMRI Results . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.5 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A Supplementary Material to Chapter 2 56
A.1 SUPPLEMENTARY TABLES . . . . . . . . . . . . . . . . . . . . . . 56
A.2 SUPPLEMENTARY FIGURES . . . . . . . . . . . . . . . . . . . . . 61
Bibliography 63
vii
LIST OF TABLES
3.1 Areas where activity scaled with subjective displeasure ratings.
Negative IADS height threshold: t(24) = 3.745, p <.001 (uncor-
rected); extent threshold, k=40. Pressure pain height threshold:
t(24) = 3.086, p <.005 (uncorrected); extent threshold, k = 40. L =
Left, R = Right. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.1 Areas modulated by DERS scores during pain anticipation.
Height threshold: t(23) = 3.819, p <.001 (uncorrected); extent
threshold, k = 50. L = Left, R = Right . . . . . . . . . . . . . . . . . 51
4.2 Areas where brain activity was modulated by DERS scores dur-
ing pain processing. Height threshold: t(23) = 3.819, p <.001
(uncorrected); extent threshold, k = 30. L = Left, R = Right. . . . . 52
A.1 Papers included in the corpus of studies . . . . . . . . . . . . . . 59
A.2 Areas reliably activated by effort costs. . . . . . . . . . . . . . . . 60
A.3 Areas reliably activated by delay costs . . . . . . . . . . . . . . . . 60
A.4 Areas reliably activated by probability costs . . . . . . . . . . . . 61
viii
LIST OF FIGURES
2.1 Effort costs reliably engaged bilateral SMA, bilateral dACC,
Midbrain, bilateral dorsal striatum, bilateral IFG, right GP, left
M1, left posterior and bilateral anterior Insular cortex. . . . . . . 14
2.2 Delay costs reliably engaged bilateral SMA, mPFC, Midbrain,
PCC, bilateral dlPFC, left anterior Insula, left Caudate, left V1,
bilateral Precuneus and left IPL. . . . . . . . . . . . . . . . . . . . 15
2.3 Probability costs reliably engaged bilateral SMA and dACC,
right OFC, right dlPFC, midbrain, bilateral Insula, bilateral Cau-
date, thalamus, right IPL. . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Left anterior insula and left Caudate were reliably engaged by
probability, effort, and delay costs. . . . . . . . . . . . . . . . . . . 16
2.5 Effort costs, compared to both delay and probablity, were more
likely to engage SMA. . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.6 Probablity costs, compared to effort and delay, were more likely
to engage ventral portions of anterior insula and right putamen. 17
2.7 Left anterior insula and left Caudate were reliably engaged by
probability, effort, and delay costs. . . . . . . . . . . . . . . . . . . 18
3.1 Experimental Procedure: Participants saw a cue and after brief,
jittered delay received pressure pain or heard either emotionally
or physically aversive sound. After another delay, participants
ranted the stimulus on pleasantness and unpleasantness. . . . . . 32
3.2 Behavioral Results: Bar graphs illustrating how pleasure and
displeasure ratings varied across different trial types . . . . . . . 36
3.3 Correlation between pleasure and displeasure ratings: Plea-
sure and displeasure ratings were significantly correlated in
physically aversive, but not in emotionally aversive or pressure
pain trials. Shading represents 95% confidence interval . . . . . . 36
3.4 Relationship between stimulus intensity and displeasure: In-
tensity significantly affected displeasure ratings in pressure pain,
but not in physically aversive sound trials . . . . . . . . . . . . . 37
3.5 Neural Correlates of Negative valence. Top panel: areas where
activity scaled positively with subjective displeasure ratings in
negative IADS A. Height threshold: t(24) = 3.745, p <.001 (un-
corrected); extent threshold, k = 40. Bottom panel: areas where
activity scaled negatively with increasing ratings of displeasure
during physically aversive ( B) and pressure pain ( C) trials.
Height threshold: t(24) = 3.086, p <.005 (uncorrected); extent
threshold ( B) k = 40 ( C) k= 20. . . . . . . . . . . . . . . . . . . . . 38
4.1 Behavioral Results: Scatter plots with regression lines showing
the lack of relationship between DERS scores and displeasure
ratings across all stimulus types . . . . . . . . . . . . . . . . . . . 50
ix
4.2 Areas where activity was negatively modulated by DERS scores
during pain anticipation. Height threshold: t(23) = 3.819, extent
threshold, k = 50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.3 Areas where brain activity was modulated by DERS scores dur-
ing pain processing. Height threshold: t(23) = 3.819, extent
threshold, k = 30 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
A.1 Compared to effort costs, delay costs were more likely to engage
bilateral dlPFC and right IFG (A). In contrast, effort costs were
more likely to engage SMA and left M1 (B). Both effort and delay
costs reliably engaged bilateral IFG, left insula and left Caudate
(C). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
A.2 Compared to probability costs. Effort costs were more likely to
engage posterior Insula, left M1 and SMA (A). In contrasts, prob-
ability costs were more likely to engage right dlPFC and ventral
portions of Insula (B). Both effort and probability costs reliably
engaged bilateral dorsal anterior Insula, bilateral Caudate and
right IPL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
A.3 Compared to delay, probability costs were more likely to engage
more anterior bilateral Insular cortex and left Caudate (A). Con-
versely, delay costs were more likely to engage bilateral dlPFC
(B). Both delay and probability costs reliably engaged dACC,
SMA, midbrain, left Insula and Caudate and right Intraparietal
Sulcus (C). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
x
CHAPTER 1
INTRODUCTION
Avoidance can often be seen as the dark side of motivation, especially when
it manifests in maladaptive ways such as anhedonia in depression [160] or ex-
treme uncertainty aversion in anxiety [21]. However avoidance is essential
for humans and non-human animals to adaptively navigate hostile or unpre-
dictable environments. Given its central role in both psychiatric illness and
healthy adaptive behaviors, avoidance is a major area of interest within the
fields of psychology and neuroscience. Yet despite a large and growing body
of research, certain questions remain. This dissertation focuses on understand-
ing the neural representations of value and valence, which both contribute to
avoidant behaviors.
A central issue in the understanding of value is how people and animals are
able to make choices between distinct rewards. The relative ease with which
humans and animals are able to compare values of distinct stimuli suggests the
must be a common mental representation of value. Over the last two decades
numerous animal and human studies have demonstrated a common neural ba-
sis for reward and value with the Orbitofrontal Cortex. However, it remains
unknown whether a similar common cost signal exists for decisions costs. Here,
we use Activation Likelihood Estimation (ALE) meta-analysis to uncover com-
mon and distinct neural representations of effort, delay, and probability costs.
Our results show that all three costs engaged portions of the dorsal striatum
and anterior insula. In addition, delay and probability costs both engaged por-
tions lateral prefrontal cortex, effort costs consistently activated on Supplemen-
tary Motor Area and Primary Motor Cortex, suggesting that overcoming effort
1
costs may require distinct mechanisms of self regulation. Together, our find-
ings suggest that despite different task requirements and divergent self-control
mechanisms, cost representations converge in insula and dorsal striatum to be
integrated in the assessment of current body state shape action selection.
In the second part of the dissertation we consider the question of neural
representations of affect. Affect colors people’s subjective experiences and per-
ceptions of the world, yet it has been thought to be processed separately from
sensory information, to be integrated at the later stages of processing. How-
ever, a growing body of research has discovered modality-specific valence rep-
resentations across perceptual and somatosensory cortices. The presence of va-
lence representations within sensory systems challenges the notion that sensory
and affective information is processed in a modular way. A similar modular
view prevails in nociception, proposing two distinct pathways for processing
affective and sensory properties of pain. To address the question of whether
modality-specific valence information is present during nociception and aver-
sive stimulus processing, we examined the brain areas in which BOLD activ-
ity parametrically varied with subjective negative valence ratings of pressure
pain stimuli as well as emotionally and physically aversive sounds. We found
that negative valence ratings were tracked by both auditory and somatosen-
sory cortices during sound and pain stimuli. Our results suggests that in addi-
tion to modality-general value and cost representations, human brain maintains
modality-specific affect information.
In chapter 4 we explore how emotion regulation shapes the affective expe-
rience or pain and other aversive stimuli. Emotion regulation techniques have
been shown to be effective in helping individuals with chronic pain manage
2
pain over time. However, it remains unclear whether trait differences in emo-
tion regulation in healthy young adults has any effect on the way pain is pro-
cessed. To address this question we examined how trait differences in emotion
regulation affected the neural processing of pressure pain as well as aversive
sound stimuli in a healthy young adult population. We found that self reported
emotion regulation ability did not affect subjective displeasure ratings, suggest-
ing limited effect of emotion regulation on subjective experience. However,
trait emotion regulation differences did modify mean pain processing activity in
pressure pain trials. Specifically, we found that participants who reported hav-
ing the most difficulties in regulating emotion, showed reduced striatal, thala-
mic and somatosensory cortex activity during pain anticipation, and increased
activity in dorsal striatum, somatosensory cortex, and cerabellum during pain
processing. These findings suggest that emotion regulation strategies may work
through modifying both affective and sensory components of pain in anticipa-
tion to the actual pain stimulus. The absence of such affective-sensory fore-
casting in individual with emotion regulation increased their reliance on dlPFC,
which may be a less adaptive strategy under more challenging circumstances.
Together these findings show that brain maintains both modality-specific
and modality-idependent representations of negative valence and cost to sup-
port different adaptive functions. Modality-general representations enable
value comparisons between distinct stimuli, while modality-specific valence in-
formation offer more granular representations of a specific object.
3
CHAPTER 2
WEIGHING THE COSTS: NEURAL REPRESENTATION OF DELAY,
EFFORT, AND PROBABILITY COSTS
2.1 ABSTRACT
Integrating costs and rewards to choose the optimal course of action is an es-
sential component of adaptive human behavior. Numerous animal and human
studies have demonstrated a common neural basis and thus a mental currency
for reward. However, it remains unknown whether this commonality applies
to costs incurred to acquire these rewards. Temporal delays, effort require-
ments and risks associated with reward are three such costs that serve to dis-
count the subjective value and bias the selection of choice options. Here, we
use Activation Likelihood Estimation (ALE) meta-analysis to uncover common
and distinct neural representations of different costs. We identified 70 studies
of different costs (22 studies of effort-based decision-making, 21 studies exam-
ining inter-temporal choice, and 36 studies on probabilistic decision making).
Our analyses revealed that all 3 costs engaged portions of the dorsal striatum
and anterior insula. In addition, while delay and, to a lesser extent, probability
costs reliably engaged prefrontal regions (BA 46) associated with top-down self-
control, effort costs distinctly relied on Supplementary Motor Area and Primary
Motor Cortex (BA 4, BA 6). Together, these findings suggest that cost represen-
tations converge in insula and dorsal striatum to be integrated in the assessment
of current body state, modulate the subjective value of choice options and bias
action selection but may be segregated according to specific cost type, support-
ing divergent self-control mechanisms.
4
2.2 INTRODUCTION
The ability to engage in motivated behavior to pursue rewards and escape pun-
ishments is essential for survival of any moving animal. By the same token, re-
duced ability to engage in motivated behavior has far-reaching consequences in
human lives. Altered ability to engage in motivated behavior is characteristic of
a wide variety of psychological disorders including depression, schizophrenia,
addiction, and Parkinsons disease [29, 149, 153, 160]. Reduced ability to engage
in motivated behavior can have severe consequences for peoples quality of life.
It also creates a large economic burden through both increased health care costs
and reduced economic opportunity. Understanding how people weigh different
costs and benefits to choose the optimal course of action is key to understanding
motivated behavior.
Findings form the fields of economics, neuroscience and psychology offer
convergent evidence that people act in accordance of maximal subjective value
of the available options [17, 71, 170]. It is rare for the available options to be op-
timal in every respect. Thus, to act in an adaptive way, people and animals must
integrate the various costs and benefits associated with divergent courses of ac-
tion. Over the last decade, a number of monkey electrophysiology and human
fMRI studies have identified several regions whose activity is consistent with
the function of representing subjective value. The BOLD activity patters Or-
bitofrontal Cortex (OFC), have been found to represent valence independent of
visual features of different stimuli, as well as independently of stimulus modal-
ities when people were passively perceiving given stimuli [28]. Medial OFC has
also been found to track stimulus value in simple choice tasks in both humans
and other primates [86, 117, 131]. A recent meta-analysis of human neuroimag-
5
ing studies has found that Orbitofrontal Cortex (OFC) ventromedial Prefrontal
Cortex (vmPFC) and Nucleus Accumbens (NAcc) were reliably engaged by sub-
jective value of both primary and monetary rewards [10].
While a large body of evidence points to the existence of a common value
signal, it remains unclear whether common cost representations exist in the
brain. Much like the necessity to be able compare different possible rewards
suggest there must be a neural signal of common currency, the necessity to
choose between different costs (e.g., waiting or working for a reward) suggests
there may be a common cost signal as well. In this way, costs and benefits could
be weighed on a similar scale to regulate action. However, vastly diverging task
demands in risk-, delay- and effort-based decision making may result in neural
representations that are in large-part segregated. Evidence from nonhuman ani-
mals demonstrate that lesions in the Anterior Cingulate Cortex (ACC) were less
likely to choose the high reward-high effort arm in a T-maze, but would do so
if they had to endure a delay. Meanwhile, rats with lesions in the OFC showed
the reverse pattern of behavior [138]. In contrast, inactivations either ACC or
OFC did not change rodents preferences for risky rewards [152]. These findings
indicate that costs might in part be processed in a cost-specific way.
In the nonhuman animal literature several notable attempts have been made
to qualitatively synthesize the commonalities and differences in neural mech-
anisms of different types of cost-benefit decision-making [5, 139, 174]. These
qualitative reviews of animal research offer converging evidence that different
decision costs may be processed by both overlapping and distinct neural mecha-
nisms. While animal lesion studies can elucidate the causal roles different brain
areas play in various types of cost-benefit decision making, brain structures in
6
humans and rodents are only in part homologous [6]. Even when direct paral-
lels can be drawn, lesions target functionality of specific structures rather than
reveal larger brain-wide circuitry afforded with human functional neuroimag-
ing.
Despite these differences, human neuroimaging studies corroborate the gen-
eral idea of separate cost-specific processing streams that converge to modulate
subjective value. Dorsal ACC seems to be specific to processing effort require-
ments and is less instrumental in processing either delay or probability costs
[22, 99]. Inter-temporal decision making engaged lateral inferior frontal gyrus,
inferior parietal lobule, and posterior cingulate cortex, while probabilistic deci-
sion making, in contrasts to the other cost benefit decision making, engaged pos-
terior parietal cortex (PPC) and portions of anterior insula (Ins) [99, 172, 121].
In addition, several qualitative reviews have examined the neural mechanisms
of inter-temporal and risk-based decision making [50, 122, 125]. These reviews
identify largely similar brain areas as instrumental for both inter-temporal and
risk-based decision making, including Insula, medial and lateral PFC, posterior
parietal cortex (PCC), OFC, as well as ventral and dorsal striatum.
The wide variety of regions implicated and discrepancies across studies
highlights the difficulty identifying distinct neural mechanisms of different
types of decision making though qualitative review. Different types of cost-
benefit decision making may engage adjacent but dissociable portions of cortex.
Thus, relying on qualitative reviews may lead to both overestimating the over-
lap and missing reliable distinctions. More than any one particular study exam-
ining different cost types, a quantitative meta-analysis can leverage the power
of convergent findings, without overlooking or overestimating reliable overlaps
7
and distinctions of different types of cost-benefit decision making. Further, met-
analyses might reveal patterns of consistency that may have not been hypothe-
sized or emphasized in the original investigations, providing a more unbiased
review of the supporting data.
Previous meta analyses have examined the component processes within a
single type of cost-benefit decision, isolating prospection, and response inhi-
bition in inter-temporal decision making [25, 173], or ambiguity and risk re-
lated processes in probabilistic decision making [77, 110]. In the present meta-
analysis, mirroring that of value representations [10], we asked whether there
is evidence for a common neural cost. Taking the lead from mesocorticolim-
bic representations of value, we examined whether an analogous system exists
for the representation of cost. We first identify areas reliably engaged in effort-
, delay- and probability- based decision making that track subjective cost and
then perform conjunction and contrast analyses to isolate common and distinct
neural correlates of each type of decision cost.
2.3 METHODS
In order to understand the neural representations of delay, effort and probability
costs, we first conducted three separate coordinate-based ALE meta-analyses,
which revealed the areas reliably engaged by each type of cost across studies.
Next, we examined the extent to which these representations engage overlap-
ping or distinct areas by conducting contrast and conjunction analyses for each
pair of costs.
8
2.3.1 Study Selection
We identified potential studies by querying PubMed database (http://pubmed.com)
with combinations of key words. To identify neuroimaging studies on cost-
benefit decision making involving effort, delay or probability cost, we used key
words reflecting the cost type and fMRI as well as related MESH terms. To re-
flect the slight differences in terms used to describe delay-based decision mak-
ing we also included terms inter-temporal and decision or choice. By the same
token, we included terms risk, uncertainty and decision or choice to identify
studies on probability-based decision making. Our initial search returned 3687
papers for effort-, 2805 papers for delay- and 3940 papers for probability-based
decision making. Additional papers were identified using in house reference
library and examining the reference sections of prior review articles and single
cost meta analyses [18, 25, 36, 173]
We selected our final corpus of studies using the following criteria. All stud-
ies used fMRI to measure BOLD signal in healthy young adults. Following this
criterion, we only included separate contrasts for healthy controls in studies ex-
amining the differences in cost-benefit decision making between a heathy and
clinical populations. We further excluded any studies that only reported Re-
gion of Interest (ROI) or Small Volume Correction analyses, because the meta-
analysis algorithm assumes that under the null hypothesis foci are randomly
distributed across the brain and ROI and SVC analyses both restrict the areas
in which foci can be reported. However, we included whole-brain exploratory
analyses from these studies if reported.
We have encountered some heterogeneity in the types of tasks based on the
queries. Because mental effort is a highly heterogeneous category, we have re-
9
stricted our sample to tasks that rely on physical effort. Studies that involve
inhibition of predominant response can sometimes be described as effort-based
(e.g., [?, ?], but are not the same as evaluating physical effort as a decision
cost, and thus were excluded from our sample. Similarly, anticipatory activity
for rewards was sometimes described as delay-related (e.g., [61, 95] , but were
excluded from the sample because these tasks involve fundamentally different
processes. All included studies had a correlational or a parametric analysis of
specific cost or a binary contrast between costlier and less costly conditions, or
a binary contrast between cost-benefit decision making and the baseline, which
was our central measure of interest. The selection criteria for cost-related cotrast
were similar to that followed by the meta-analysis of subjective value [10]. The
final corpus of studies included 21 studies (228 total foci) of delay-based deci-
sion making, 22 studies (163 total foci) of effort-based decision making, and 36
studies (396 total foci) of probabilistic choice (Supplementary table A.1).
2.3.2 ALE analysis
Single cost meta-analyses
We conducted meta-analyses of effort, delay and probability costs using the Ac-
tivtion Likelihood Estimation (ALE) algorithm as implemented in GingerALE
2.3 [44, 45, 163] in MNI space. For the studies that reported the coordinates in
Talairach, we performed a linear transformation to MNI space [83]. The ALE
method takes peak activations from each study and places a Gaussian sphere
around each peak. The width of each sphere represents the spatial uncertainty
associated with the exact location of the activation peak. The ALE algorithm
10
scales the width of each sphere by the sample number of the study, with smaller
sample studies receiving wider width spheres to reflect a greater uncertainty
in the study. This process results in a Modeled Activation (MA) map for each
study, which each voxel assigned some probability that a given task evoked
activity in that voxel. These MA maps were then added together to produce
an ALE map for each cost type. To calculate a threshold, we first randomly
sampled voxel by voxel from each MA map and adding them to create an ALE
score. Next, this procedure was iterated 104 times to create a distribution of ALE
scores to create a null distribution, which was used to threshold the ALE maps.
We compared the original ALE map values to the null distribution values for
each voxel to calculate a p-value. The voxel-wise p value was calculated by tak-
ing the number of ALE values in the null distribution that are greater or equal to
the ALE value at a given voxel and dividing it by the total number of values in
the null distribution. That is, to receive a p value of 0.05 an ALE value of an ALE
map for a given cost should be in top 5% of ALE values in the null distribution.
We used a cluster-level threshold of p <0.05, and a cluster-forming threshold of
p <0.001. The resulting maps represent the areas which were reliably engaged
across studies by different decision costs.
Contrast analyses
To identify the brain areas that were more likely to be engaged by a specific cost
we conducted a Ginger ALE contrast analysis for each pair of decision costs
[44] [], resulting in three contrast maps. We used the thresholded single-cost
ALE maps that we produced using the procedure outlined above. In addition,
a pooled image was calculated for each pair of costs using the same procedure.
11
The resulting pooled studies were randomly sub-divided into groups that were
equivalent in size to the original groups. For example, when comparing ef-
fort and probability costs, random groupings of 22 and 21 experiments were
formed. Then a difference score for each voxel was calculated by subtracting
the ALE value of the randomly generated maps from the ALE value of the orig-
inal group maps at each given voxel. This procedure was iterated 104 times
to create a null distribution of difference scores. The difference scores derived
from subtracting the thresholded ALE maps of the two costs where compared to
the null-distribution of difference scores. To adjust for multiple comparisons we
used a p value of .001 as a threshold. The difference maps were also inclusively
masked by the group main effects.
Conjunction analyses
We were interested in which areas were equally likely to be engaged by each
pair of costs. To address this question we carried out a conjunction analysis
[27] using the thesholded ALE maps produced by single-cost analyses. The
conjunction images were created by taking the minimal ALE value at any given
voxel.
12
2.4 RESULTS
2.4.1 Effort meta-analysis
Our meta-analysis of effort cost revealed a number of areas that were consis-
tently engaged across 22 studies. These areas include bilateral Supplemental
Motor Area (SMA), Left Motor Cortex (M1), Globus Pallidus, Caudate, Bilat-
eral Insular Cortex, Right Fusiform Gyrus, Cuneus, Bilateral dorsal Anterior
Cingulate Gyrus (dACC), Midbrain, Inferior Parietal Lobule (IPL), and bilateral
Inferior Frontal Gyrus (IFG) (Figure 2.1, Supplementary Table A.2).
2.4.2 Delay meta-analysis
Our meta-analysis of 22 studies examining inter-temporal decision making
showed consistent activation foci in bilateral IFG, medial Prefrontal Cortex
(mPFC), bilateral dorsolateral Prefrontal Cortex (dlPFC), left SMA, Posterior
Cingulate Cortex (PCC), bilateral Precuneus, Caudate, Insula, Midbrain and
other areas. (Figure 2.2, Supplementary Table A.3).
2.4.3 Risk meta-analysis
Our meta-analysis has shown that across 36 studies risk has reliably engaged
bilateral anterior Insula, Caudate, Putamen, Precuneus, portions of mPFC,
Fusiform Gyrus, SMA, Cerebellum, perigenual and subgenual ACC, right Su-
perior Temporal Gyrus, right dlPFC, right IPL, Thalamus, Midbrain and other
13
Figure 2.1: Effort costs reliably engaged bilateral SMA, bilateral dACC,
Midbrain, bilateral dorsal striatum, bilateral IFG, right GP, left
M1, left posterior and bilateral anterior Insular cortex.
areas (Figure 2.3, Supplementary Table A.4).
2.4.4 Common Cost Representations
A logical conjunction of the effort, delay, and probability 2-way conjunction
analyses revealed that all decision costs reliably engaged left dorsal striatum
/ head of caudate (-10, 12, 0) and dorsal portions short anterior gyrus of insular
cortex (-33, 19, 5) (Figure 2.4).
2.4.5 Distinct Cost Representations
A logical conjunction of effort ¿ probability and effort ¿ delay contrast maps
has revealed that Primary Motor Cortex ( -36, -23, 59) and Supplementary Mo-
14
Figure 2.2: Delay costs reliably engaged bilateral SMA, mPFC, Midbrain,
PCC, bilateral dlPFC, left anterior Insula, left Caudate, left V1,
bilateral Precuneus and left IPL.
Figure 2.3: Probability costs reliably engaged bilateral SMA and dACC,
right OFC, right dlPFC, midbrain, bilateral Insula, bilateral
Caudate, thalamus, right IPL.
15
Figure 2.4: Left anterior insula and left Caudate were reliably engaged by
probability, effort, and delay costs.
tor Area (0, -13, 57) were more likely to be engaged by effort costs, compared
to delay and probability costs (Figure 2.5). Similarly, the logical conjunction
of probability ¿ effort and probability ¿ delay contrast maps has shown that
probability costs, compared to other decision costs, were more likely to engage
ventral portions of anterior insula bilaterally (36, 16, -8; -39, 18, -8), as well as
right putamen (36, 16, -8) (Figure 2.6). The logical conjunction of the contrast
maps of delay ¿ effort and delay ¿ probability has shown that delay costs were
more likely to engage bilateral dorsolateral prefrontal cortex (-39, 38, 13; 50, 36,
4; Figure 2.7).
2.4.6 Dissociation between Effort and Delay Costs
Compared to delay costs, effort costs were more likely to engage left M1 and
SMA. In contrast, delay costs were more likely to engage bilateral dlPFC and
16
Figure 2.5: Effort costs, compared to both delay and probablity, were more
likely to engage SMA.
Figure 2.6: Probablity costs, compared to effort and delay, were more
likely to engage ventral portions of anterior insula and right
putamen.
17
Figure 2.7: Left anterior insula and left Caudate were reliably engaged by
probability, effort, and delay costs.
portions of right IFG (Supplementary Figure A.1).
2.4.7 Overlap between Effort and Delay Costs
A conjunction analysis revealed that both effort and delay costs reliably engaged
bilateral IFG, left Caudate, and left Insula (Supplementary Figure A.1).
2.4.8 Dissociation between Effort and Risk Cost
Our ALE contrast analysis has shown that, compared to risks, effort costs were
more likely to engage left M1, SMA and posterior insula. In contrast, risk was
more likely to activate bilateral ventral insula and right DLPFC (Supplementary
Figure A.2)
18
2.4.9 Overlap between Effort and Risk Costs
A conjunction analysis revealed that both effort and risk costs reliably engaged
bilateral Caudate, bilateral dorsal Insula, and right IPL (Supplementary Figure
A.2)
2.4.10 Dissociation between Delay and Risk Costs
Compared to risks, delays were more likely to engage left dlPFC. In contrast
risks were more likely to activate bilateral Insula and left Caudate (Supplemen-
tary Figure A.3).
2.4.11 Overlap between Delay and Risk Cost
Our conjunction analysis has revealed that both inter-temporal delays and risks
associated with rewards activated, dACC, SMA, left Insula, Caudate, right IFG,
Midbrain, and Intraparietal Sulcus (Supplementary Figure A.3)
2.5 DISCUSSION
Making tradeoffs between different costs and benefits forms the basis for envi-
ronmentally adaptive behavior. The question of how humans and nonhuman
animals make these decisions has received widespread attention in the fields of
economics, psychology, and neuroscience. Notably, attempts to understand the
neural computations that underlie the choice between different rewards lead
19
to discovery of the system of subjective valuation in humans and nonhuman
primates [57, 87, 115, 117, ?]]. However, a parallel question of how humans
and animals decide between different costs remains open. The existence of a
common valuation system for distinct rewards suggests that a similar system
may exist for comparing distinct cost, yet the notion of common cost signal has
received considerably less attention.
The research on different types of cost-benefit decision making so far has
been predominantly focused on the types of self-control required to overcome
a particular type of cost. Initial work in psychology has suggested that there
may be a common mechanism of self-control that gets exhausted no matter
what type of cost it is used to overcome [11, 69, 114]. This idea was seemingly
supported by the findings showing that childhood ability to delay gratification
was predictive of a variety of outcomes across the lifespan including academic
achievement, drug use, social skills and both psychological and physical health
[26, 106, 105, 134]. However, the notion on unitary self-control has since been
called into question after multiple researchers were not able to replicate the orig-
inal findings [24, 60].
A complex behavior such as overcoming a cost is likely to involve a plethora
of neurocognitive component processes including but not limited to working
memory, response inhibition, sustained attention, and conflict detection [9, 63].
Some of these processes may be shared across different costs in cost-benefit deci-
sion making, while others may be unique to a specific cost. In the present study
we looked in both unique and common neural activations consistently elicited
by 3 most common decision costs of effort, probability and reward.
20
2.5.1 Common Activations
We found that all three decision costs elicited activation in the dorsal striatum,
specifically in the head of caudate. Dorsomedial striatum has long been impli-
cated in goal-directed action control across species, especially when it requires
flexible adaptation to changing environment [7, 162]. In non-human primates,
neuronal spike activity in the caudate has been shown to represent both cost sig-
nals and trial outcome (i.e., whether the trial was rewarded) [41]. Furthermore,
in rodents, inhibiting the projection from the prefrontal prelimbic cortex to the
striosomes in caudate nucleus has led to reduced sensitivity to cost [49]. To-
gether these findings support the notion that caudate nucleus integrates across
inputs from discrete regions of associative cortex to represent both cost and ac-
tion outcome, in order to guide goal directed behavior.
In line with its role as a hub for integrating across disparate inputs to guide
action, caudate has also been found to track subjective value (e.g., [23, 139, 140]).
Thus, it is possible that the overlap in the caudate we find for effort, delay, and
probability costs reflects a common representation of subjective value rather
than that of cost. While some studies make rewards and costs orthogonal (e.g.,
We also found that all three types of decision costs activated dorsal portion
of the short anterior gyrus in the dorsal anterior insular cortex. Insular cortex
receives inputs from a broad range of cortical and subcortical areas has been
shown to play a key role in interoception and awareness [33, 34]. Tracking
the inherent costs of current or desired goal state is an important part of this
broader function. Dorsal anterior insula in particular has been shown to be part
of a stable functional control network along with dorsal ACC and anterior PFC,
maintaining activity across the entire task epoch [42, 40]. The connections be-
21
tween dorsal anterior insula and dorsal ACC and PFC are further corroborated
by tract tracing studies in humans and non-human primates [54, 102, 112]. Cost
representations in insula may be relayed and integrated across the control net-
work to a stable representation of a goal state for a particular task. The dorsal
portion of short angular gyrus in particular is also connected to both caudate
and NAcc, which ideally positions it for integrating cost and benefit informa-
tion [53].
2.5.2 Cost Specific Activations
As for unique activations, we found that effort costs, compared to both delay
and probability, were more likely to engage SMA and M1. Probability costs,
compared to effort and delay, were more likely to engage ventral portions of an-
terior insula and right putamen. Delay costs, compared to effort and probability,
were more likely to engage bilateral dlPFC.
While dorsal anterior insula was reliably engaged by all costs, ventral an-
terior insula was more likely to be recruited by probability, compared to delay
and effort costs. In contrast to dorsal anterior insula, which has been broadly
identified with more cognitive domain, a meta-analysis of 1,768 human neu-
roimaging studies has shown ventral anterior insula to be more likely to be
involved in emotion related processing [82]. This distinction may be a result
of some emotional aspects of risky decision making that may not be shared by
other types of cost-benefit considerations. Notably, functional resting state con-
nectivity and tract-tracing studies have shown ventral anterior insula is con-
nected to amygdala as perigenual anterior cingulate cortex, which both have
22
been implicated in processing uncertainty, especially when it involves ambigu-
ity [36, 66, 67, 88, 128]. These inputs may convey information about ambiguity
surrounding risk cost and converge on the ventral insula to modify the risk rep-
resentation.
Similarly, while all costs engaged dorsomedial striatum, risk costs were more
likely to activated dorsolateral striatum/ putamen. Consistently with our find-
ings, in a probabilistic learning task caudate has been shown to be involved in
performance monitoring and tracking cognitive control demands, while puta-
men activity was associated with monitoring outcome probabilities [20] . Con-
vergent evidence from rodents, non-human primates and humans shows that
while dorsomedial striatum contributes to early associative learning and goal
directed behavior, putamen is implicated in later/ habitual stage of learning
[8, 7]. This pattern of findings may suggest that reflect risk aversion is more
reflecsive than cost-sensitivity to either delay or effort costs.
While both delay and risk costs recruited left dlPFC, portions of bilateral
dlPFC were preferentially recruited by delay costs. Our findings are consistent
with a large body of research demonstrating that dlPFC is involved in various
aspects of delayed discounting including tracking the value of delayed option,
presence of a difficult choices, in addition to delay amount itself [50, 72, 73, 111].
Furthermore, temporary disruption of left dLPFC activity by low-frequency
repetitive transcranial magnetic stimulation resulted in more impulsive choices,
suggesting that dLPFC plays a causal role in delay discounting. However, con-
trary to our results, dlPFC has been shown to be more robustly activated by
risk than by delay discounting [172]. There may be several explanations for
this inconsistency. First, ALE analysis does not take into account the magni-
23
tude of the activation, but rather the consistency of activation across a corpus of
studies. In addition, in the study finding more robust activation for probability
discounting, participants showed a differential pattern of delay and risk pref-
erences, with risk preferences approximately normally distributed and delay
preferences skewed to the left with the mode on delay neutrality [172]. Thus,
the increase of signal in dlPFC may reflect the need to overcome more aversive
costs in that particular sample. In contrast to delay and risk costs, effort did
not consistently evoke activity in dlPFC or any other prefrontal areas associated
with top-down control. These results are in contrast to recent work implicat-
ing dlPFC in both cognitive effort anticipation and reward devaluation by both
cognitive and physical effort [167, 29]. It is possible that dlPFC is involved in
effort-based decision making only during key periods of decision making (e.g.,
anticipation), which were not captured in all the studies in our corpus. Alterna-
tively, dlPFC may be involved in value computation for effort-based decisions,
which was not captured in selecting the contrasts that primarily focused on the
costs. Our findings are somewhat in line with rodent literature, which shows
effort-based decisions relying on dACC, in contrast to delay costs, which rely
on OFC [138, 171]. However, with no rodent homologue of dlPFC, it is diffi-
cult to draw direct parallels between the mechanisms involved in overcoming
different kinds of decision costs in humans and rodents.
In line with prior research in rodents, in the single cost analysis we found
that effort costs engaged dACC. However, we did not find that dACC activ-
ity reliably distinguished between effort and either probability or delay costs.
Our findings are consistent with the account that dACC is necessary to over-
come effort costs and make the more effortful choice, but is not necessarily in-
volved in tracking increasing effort costs [53]. ACC lesions in rodents did not
24
affect animal performance on either progressive ratio task or their performance
on pressing a weighted lever [64, 146]. Thus, ACC may serve to bias choice
towards more costly option but may not track continuously increasing effort
requirements.
2.5.3 Limitations and future directions
It should be acknowledged that our estimates of the overlap both between dif-
ferent costs and within single costs may be somewhat conservative because of
the diversity of tasks and analyses. For example, with some notable exceptions
[74, 147], it is much more common to model subjective value as well as cost
continuously for delay and risk, but not for effort. More continuous parameter-
based analyses would enable to make stronger claims about areas that track
subjective costs for both individual costs as well a global cost signal. In addition
to diversity of analytic methods, our current corpus had a somewhat diverse set
of tasks. In particular, effort studies in our corpus included tasks such as chal-
lenging fine motor movements, working memory load, task switching, as well
as physical effort measured with a force grip (e.g., [35, 109, 116, 129, 150, 166].
On the one hand, the diversity of tasks gives credence to the idea that some
brain areas track specific costs, regardless of the task specifics. Despite substan-
tial overlap, mental and physical costs rely on partially distinct mechanisms
[29, 65]. Thus, it is possible that some areas that are necessary for a specific
type of effortful task may have been lost in the analysis. However, in the fu-
ture, a larger body of studies may allow for separate analyses to address these
questions.
25
In this analysis, we also could not distinguish between various processes
that are involved in the cost-benefit decision making, (e.g., the neural processes
that serve to overcome a larger costs and bias action and areas that track the
information about the levels of cost). However, as evidenced by the discrepan-
cies in effort related decision making these processes may rely on district neural
mechanisms. It may be beyond the scope of a meta-analysis to make the finer
grained distinctions. However, future work on the computational processes in-
volved in cost benefit decision making may shed light on the specific role these
structures play in tracking and overcoming various decision costs.
2.6 CONCLUSION
In the last two decades the notion of a common signal of subjective value has re-
ceived considerable attention and, subsequently, overwhelming empirical sup-
port. In the current meta analysis we asked if a similar common cost signal
could be found for commonly studied decision costs of effort, delay, and risk.
We found, despite substantial differences in the tasks, that caudate and dorsal
anterior insula were reliably engaged by all 3 decision costs. Our findings indi-
cate it a possibility for a common cost signal as well as possibilities for distinct
self-control mechanisms reflecting differential task demands.
26
CHAPTER 3
AFFECTIVE AND SENSORY REPRESENTATIONS OF NEGATIVE
VALENCE ACROSS DIFFERENT AVERSIVE STIMULI
3.1 ABSTRACT
Affect can powerfully shape people’s subjective experiences and perceptions of
the world. Traditionally, affect has been thought to be processed by a distinct
neural systems including portions of the prefrontal and orbitofrontal cortex and
the limbic system, and later integrated with sensory information from percep-
tual and somatosensory cortices. However, recent evidence from neuroimaging
studies showing valence representations in perceptual and somatosensory cor-
tices has begun to challenge this long-standing assumption. In the present study
we examined the brain areas in which BOLD activity parametrically varied with
subjective negative valence ratings of pressure pain stimuli as well as emotion-
ally and physically aversive sounds. We found that negative valence ratings
were tracked by both auditory and somatosensory cortices during sound and
pain stimuli. These findings offer additional support for modality-specific af-
fect representations.
3.2 INTRODUCTION
Emotion colors human perception, shaping the way people perceive and inter-
act with their environment. While the role of emotion in perception is com-
monly accepted, the mechanisms through which emotion shapes perception re-
27
main controversial. In his 1874 Outlines of Physiological Psychology Wundt
argued that affect, much like physical object properties such as vividness, is fun-
damental dimension of perception [179]. However, currently prevailing models
propose a more modular system where emotion is processed by and large sep-
arately, primarily in limbic system and ventromedial prefrontal/orbitofrontal
cortices (vmPFC/OFC) and modifies perception at later stages of integration
[100, 98, 126].
The idea of parallel convergent systems for emotional and sensory process-
ing is also prevalent in nociception. The dual cognitive model of pain processing
posits that sensory information about pain is be processed through the spinotha-
lamic tract which carries information through the lateral nuclei of thalamus and
to somatosensory cortices [161]. Meanwhile, the affective components of pain
are processed through the the midline and intralaminar thalamic nuclei, which
project to the limbic system, the periaqueductal grey [130, 169]. Both sensory
and affective inputs go on to converge in anterior cingulate cortex (ACC) to
shape subjective experience and guide behavior.
However, recent findings in neuroimaging have begun to challenge these
modular processing models. Neural activity patterns encoding modality-
specific valence have been found throughout the sensory cortices, including
visual, gustatory, and auditory cortices [28, 72]. The valence of the specific
stimuli were able to be decoded from the neural activity patterns confined to
sensory cortices. These findings suggest that emotion does not merely enhance
the perception of feature level properties of the perceptual stimulus, but rather
is processed in the sensory cortices alongside more basic feature-level informa-
tion.
28
In the present study, we looked and how valence information is tracked in
aversive auditory and pressure pain stimuli. We hypothesized that in addition
to corticolimbic structures commonly implicated in emotion processing, sub-
jective displeasure will parametrically modify activity in somatosensory and
auditory cortices.
3.3 METHODS
3.3.1 Participants
Twenty file participants (17 female, 8 male) were recruited form Cornell Univer-
sity and surrounding community in Ithaca, NY area. Participants were recruited
through Psychology Department Psychology Experiment Sign-Up (SONA) Sys-
tem and though posters around the Cornell University and the surrounding
area. All participants were right-handed young adults ranging in age from 18 to
29 years (M=22.16, SD=3) with normal or corrected-to-normal vision. Addition-
ally, participants were screened for current psychotropic medication use, mood
and psychiatric disorders and provided written informed consent approved by
Cornell Institutional Review Board of. Upon completion of the study, each par-
ticipant received $50.
29
3.3.2 Stimuli
Pressure pain
Participants received mechanical pressure stimulation to the left thumbnail via
a custom designed hydraulic device made to transmit controlled pressure to a 1
cm2. The pressure pain device consisted of a hard plastic cylinder that got dis-
placed by the release pressurized nitrogen gas. Before starting the scan session,
each subject was instructed on how to place and adjust the hand in the device
and received pressure stimulation at the levels of 3 kg/cm2 (non painful/mild
pressure), 5 kg/cm2 (moderately painful pressure), and 7 kg/cm2 (very painful
pressure) for 11 seconds. In each scan run participants received 7 pressure pain
stimuli: 2 stimuli of each intensity, for 11 and 8 seconds, and an additional high-
intensity painful stimulus as the first trial of each run.
Emotionally Aversive Sounds
In each run participants heard 3 emotionally aversive sounds selected from In-
ternational Affective Digital Sounds (IADS) database [19]. Participants were
informed before the study that they will hear upsetting sounds that include at-
tacks, violence, and screaming, to ensure that no participants got unduly upset.
The selected stimuli had mean valence ratings of M = 2.50, SD = .46, and mean
arousal rating of M = 6.72, SD = 0.55.
30
Physically Aversive Sounds
In addition to emotionally aversive sounds, in each run participants also heard
3 physically aversive sounds which consisted of unpleasant high pitch sounds
of 3 different intensities [52] reminiscent of nails screeching over a board for 8
seconds.
3.3.3 Procedure
The aversive stimulus task was created using PsychoPy [119]. Participants com-
pleted 5 runs of the aversive stimulus task, each lasting 11 min. 16 s. Every run
started with a trial during which participants received a 7kg/cm2 pressure pain
stimulus lasting 11 s, which was the most intense pressure pain stimulus in the
study. This was done to ensure that if participants wanted to stop the scan be-
cause of pressure pain stimulation, they could do so in the beginning of the run.
Each trial began with a cue which indicated what kind of aversive stimulus par-
ticipants would be receiving in that trial, followed by a jittered pre-stimulus de-
lay lasting 2-5.5s. The stimulus start was indicated by a change in fixation cross
500ms before the start of the stimulus. The emotionally aversive sounds lasted
5-6s, physically aversive sounds lasted 8s, and pressure pain stimuli lasted 8-
11s. The stimulus was followed by a post-stimulus delay, the duration of which
depended on the durations of previous trial components. After the post stim-
ulus delay participants rated how much pleasure and displeasure they experi-
enced on a given trial. The order of the ratings was randomized across trials,
with a variable short delay between. Participants had 5s to rate the stimulus on
each scale. The length of each trial was not dependent on participants response
31
time. The variability in response times was absorbed by inter-trial interval (ITI),
with the total length of the trial and the ITI adding up to 44s (Figure 3.1)
Figure 3.1: Experimental Procedure: Participants saw a cue and after
brief, jittered delay received pressure pain or heard either emo-
tionally or physically aversive sound. After another delay, par-
ticipants ranted the stimulus on pleasantness and unpleasant-
ness.
3.3.4 Analysis
Behavioral Data Analysis
Behavioral data analysis and visualization was performed using R Studio [137].
A repeated measure Analysis of Variance (ANOVA) was conducted to deter-
mine if there were differences in pleasure and displeasure ratings between aver-
sive stimuli types as well as between different stimuli intensities. To further
explore the pair-wise differences in pleasure and displeasure ratings between
trial types, we conducted 3 pairwise comparisons between each trial type. We
adjusted p values for multiple comparisons using Bonferroni correction. To de-
32
termine whether pleasure and displeasure ratings were correlated for each trial
type, we also ran a Kendall correlation between mean ratings. Kendall correla-
tion was chosen to account for non-normal distributions in ratings.
FMRI data acquisition and preprocessing
The images were acquired using a GE Discovery MR750 3T scanner (Gen-
eral Electric, Milwaukee, United States) with a 32-channel head coil at the
Cornell Magnetic Resonance Imaging Facility. Anatomical scans were ac-
quired with T1-weighted volumetric MRI magnetization prepared rapid gradi-
ent echo and sensitivity encoding (TR=7ms; TE=3.42ms; TI=1100ms; Flip An-
gle (FA)=7; FOV=256256mm; sampling bandwidth=25kHz; voxel size=1mm
isotropic; 176 slices; acceleration factor=2; scan time=5:25). The functional scans
were acquired using a multi-echo echo planar imaging sequence (TR=2600ms;
TEs=13.7, 30, 47ms; FA=81; matrix size=7272mm; FOV=216 x 216mm; voxel
size=3mm isotropic; 120; acceleration factor=2.5; scan time=11:16). Participants
made responses during task runs with five-button response box held in the right
hand.
The EPI data form task runs were preprocessed with Multi-Echo Indepen-
dent Components Analysis (ME-ICA) version 3 [80, 79]. In brief, ME-ICA pre-
processing pipeline identifies blood-oxygen-level-dependent (BOLD) signals as
independent components by leveraging the linear dependence of BOLD signal
on TE. Notably this linear dependence is not shared by non-BOLD sources of
artefactual signal such as motion and pulsatility. ME-ICA uses this discrep-
ancy to identify and remove artefactual signal. The de-noised time series from
all three echoes are then combined to maximize the signal-to-noise ratio. The
33
anatomical images were first skull stripped using FSL BET with default param-
eters. The data was then submitted ME-ICA processing pipeline, removing the
first four volumes of data for signal stabilization and warping the data to MNI
space using a high-resolution template (MNI caez N27). The de-noised time
series were then smoothed using AFNI 3dmerge with a 6mm FWHM kernel.
FMRI data analysis
The smoothed data from the 5 functional runs was analyzed using Ordinary
Least Squares regression as implemented in AFNI 3dDeconvove [32, 31, 56].
Cue, stimulus, and each rating period were entered into the model as sepa-
rate events and convolved with the canonical Hemodynamic Response Function
(HRF). Cue was modeled with a tent function SPMG1, while other trial events
were modelled as a duration modulated block. Displeasure ratings were en-
tered as parametric modulators for all aversive stimuli events. Pleasure ratings
were only entered as parametric modulators for painful pressure and and emo-
tionally aversive stimuli, as the pleasure and displeasure ratings for physically
aversive sound were strongly correlated and could not be modelled together
using linear regression (Figure 3.3). Stimulus intensity was also included in a
model as a modulator for physically aversive sounds and painful pressure. The
pairwise contrast between different stimuli were conducted in 3dDeconvove
using GLT SYM.
Second level analyses were conducted using one sample AFNI 3dttest++.
The second level maps were then submitted to AFNI 3dcalc for conjunction
analysis.
34
3.4 Results
3.4.1 Behavioral Results
To examine how pleasure and displeasure ratings differed across conditions,
we conducted a 3-way ANOVA analyses, which revealed a significant effect of
trial type for both pleasure (F(2,48) = 5.3, p <.005) and displeasure F(2,48) =
19.3, p <.001 (Figure 3.2). Specifically, physically aversive sounds were rated
as more unpleasant than both emotionally aversive sounds and pressure pain
(both t(24) = 4.7, adjusted p <.001). While emotionally aversive sounds were
rated as slightly more unpleasant than pressure pain, this difference was no
longer significant after correcting for multiple comparisons (t(24)=2.5, adjusted
p >.05). Physically aversive sounds were also rated as significantly less pleas-
ant compared to both emotionally aversive sounds (t(24) = 3.1, adjusted p <,05)
and pressure pain (t(24)=2.8, adjusted p <.05). There was no significant dif-
ference in pleasure ratings between emotionally aversive sounds and pressure
pain (t(24)=1.76, n.s.).
In order to determine whether pleasure and displeasure ratings were corre-
lated in each trial type, we ran a Kendall correlation analysis (Figure 3.3. We
found that pleasure and displeasure ratings were significantly correlated only
in physically aversive sound, but not in emotionally aversive or pressure pain
trials (Physically aversive r= -.4, p <.001; Emotionally aversive r = -.097, n.s.;
Pressure pain r = -.26, n.s.)
We also examined the association between objective stimulus intensity and
subjective displeasure ratings. There was a significant relationship between
35
Figure 3.2: Behavioral Results: Bar graphs illustrating how pleasure and
displeasure ratings varied across different trial types
Figure 3.3: Correlation between pleasure and displeasure ratings: Plea-
sure and displeasure ratings were significantly correlated in
physically aversive, but not in emotionally aversive or pres-
sure pain trials. Shading represents 95% confidence interval
pressure pain intensity and subjective displeasure (F(2, 28) = 27.2, p <.001), but
not between sound pain intensity and displeasure (Figure
36
Figure 3.4: Relationship between stimulus intensity and displeasure: In-
tensity significantly affected displeasure ratings in pressure
pain, but not in physically aversive sound trials
3.4.2 FMRI Results
To understand aversive stimulus-specific areas that tracked subjective displea-
sure we examined which areas were parametrically modulated by displeasure
ratings after accounting for mean stimulus related signal. During emotionally
aversive stimuli, subjective displeasure positively modulated activity in bilat-
eral auditory cortex, portions of cingulate cortex, bilateral amygdala and right
Inferior Parietal Lobule (Table 3.1, Figure 3.5 A). Meanwhile, displeasure rat-
ings during physically aversive sounds negatively modulated only Angular
Gyrus (-36,-81,25; Figure 3.5 B). Lastly, neural activity in medial Orbitofrontal
Cortex, Somatosensory Cortext and left Thalamus were negatively modulated
by subjective displeasure ratings (Table 3.1, Figure 3.5 C)
No areas were found to track negative valence in all three aversive stimulus
37
types, even when component map threshold was lowered to .05.
Figure 3.5: Neural Correlates of Negative valence. Top panel: areas
where activity scaled positively with subjective displeasure rat-
ings in negative IADS A. Height threshold: t(24) = 3.745, p
<.001 (uncorrected); extent threshold, k = 40. Bottom panel: ar-
eas where activity scaled negatively with increasing ratings of
displeasure during physically aversive ( B) and pressure pain
( C) trials. Height threshold: t(24) = 3.086, p <.005 (uncor-
rected); extent threshold ( B) k = 40 ( C) k= 20.
3.5 DISCUSSION
Over a century ago, Wundt had introduced the idea of affect as a core com-
ponent of sensory experience [179]. However, more recent theories of emo-
tion processing have adopted a more modular view, arguing that affect is pro-
38
Label Laterality X Y Z Cluster size t
Negative IADS
Superior Temporal Sulcus L -56 -52 14 419 6.10
Superior Temporal Sulcus R 58 -41 2 393 5.68
Anterior Cingulate Cortex R 2 38 7 294 4.79
Precuneus L -3 -52 57 226 4.83
Inferior Parietal Lobule L -57 -38 29 199 5.79
Amygdala R 31 -7 -14 189 5.11
Amygdala L -27 -5 -11 153 4.90
Mid Cingulate Cortex L -6 -16 39 130 5.30
Inferior Parietal Lobule R 56 -36 27 114 4.97
Posterior Cingulate cortex L -5 -48 35 76 4.47
Middle Temporal Gyrus L -52 -19 -9 72 6.16
Pressure Pain
Orbitofrontal Cortex R 1 65 -19 102 -4.55
Somatosensory Cortex R 1 -32 50 87 -3.91
Thalamus L -9 -29 7 21 -3.48
Table 3.1: Areas where activity scaled with subjective displeasure ratings.
Negative IADS height threshold: t(24) = 3.745, p <.001 (uncor-
rected); extent threshold, k=40. Pressure pain height threshold:
t(24) = 3.086, p <.005 (uncorrected); extent threshold, k = 40. L =
Left, R = Right.
cessed separately and only modifies perceptions at the later stages of integra-
tion [100, 126]. However, recent developments in analytic techniques in neu-
roimaging have revived the idea that affect may be a fundamental feature of sen-
sory processing alongside sensory object properties such as luminance or pitch
[107, 108]. In line with Wund’s theory, representations of affect were found in
visual, auditory, olfactory, gustatory, and somatosensory cortices [141, 72, 28].
Pain perception has also been argued to be comprised of a dual system, sep-
arately processing affective and sensory components of pain [169, 130]. In light
of the recent findings showing representations of affect in sensory cortices, we
examined the brain signals that tracked subjective negative valence while con-
trolling for objective stimulus intensity and mean stimulus-related activity. We
39
found that in addition to areas that have traditionally been associated with emo-
tion processing, negative valence was tracked by somatosensory and auditory
cortices in during pressure pain and aversive sounds.
3.5.1 Affective representations of negative valence
In line with with a large body of research on emotional processing, found that
negative valence during emotionally aversive sounds parametrically modified
subgenual ACC, bilateral amygdala, and MCC. These areas are broadly consis-
tent with areas which have been shown to represent core affect comprised of
valence and arousal [91, 75]. Subgenual ACC has been linked to sadness in
healthy participants, and has been shown to to be related to depressive symp-
tom severity in depression. A recent meta-analysis has shown that amygdala
is more often activated by negative emotion [90], however this may be due
to negative stimuli commonly having higher arousal. Previous research with
experimentally matched levels of has demonstrated amygdala to be sensitive
to arousal information irrespective of stimulus valence [1, 2]. Since in our
emotionally aversive stimuli we did not explicitly control for arousal, the bilat-
eral amygdala activation is likely associated with incidental the correlation be-
tween valence and arousal. We also found that OFC tracked subjective displea-
sure in pressure pain stimuli, consistent with OFC’s role in tracking stimulus-
independent valence [28, 10, 117]. However, we did not find value tracking in
the OFC during either aversive sound stimuli. This may be partially because
of the lack of explicit valence manipulation in sound stimuli. While we used
previously validated sound stimuli for physically aversive sound trials [52], the
intensity manipulation did not affect displeasure ratings in our sample. This
40
may be in part due to imperfect sound isolation in a noisy fMRI environment.
3.5.2 Sensory representations of negative valence
In addition to valence representations in across affective areas we also found
that negative valence was represented in primary and secondary auditory cor-
tices during emotionally aversive sounds as well as in somatosesory cortex in
during pressure pain stimulation. Our findings are consistent with a growing
body of research demonstrating that affect is represented in somatosensory cor-
tices, lending support to the idea that the brain does not carry only abstract
representations of affect but also represents affect in a modality-specific way,
binding it to perceptual properties of sensory objects.
3.5.3 Limitations and future directions
While our findings are broadly consistent with Wund’s account of affect as a
perceptual dimension, univariate analyses may be insufficient to establish the
presence of affective representations in somatosensory cortices. Although we
have included objective stimulus intensity and mean stimulus related activity
in our model to statistically isolate the effects of subjective displeasure, it is pos-
sible that these effects could be caused by the modulation of the sensory infor-
mation by negative valence rather than representations of valence itself. Future
work using pattern based analysis methods could help disentangle the valence
representations.
41
CHAPTER 4
TRAIT DIFFERENCES IN EMOTION REGULATION MODULATE
NEURAL PROCESSING OF PAINFUL PRESSURE BUT NOT OF
AVERSIVE SOUND STIMULI
4.1 ABSTRACT
Chronic physical pain affects a large portion of the population and has severe
impacts to individual’s quality of life as well as enormous economic and so-
cial costs. In light of the growing concern around problematic aspects of phar-
macological pain management, more and more attention is being devoted to
explore non-pharmacological ways of pain management. Emotion regulation
techniques have been shown to help individuals manage pain over time in clin-
ical samples. However, it remains unclear whether trait differences in emotion
regulation changes the way pain is processed in non-clinical samples. To ad-
dress this question we examined how trait differences in emotion regulation af-
fected the neural processing of pressure pain as well as aversive sound stimuli
in a healthy young adult population. We found that trait emotion regulation did
not affect negative valence ratings nor the neural activity tracking negative va-
lence across all 3 stimulus types. However, trait emotion regulation differences
did modify mean pain processing activity in pressure pain trials. Specifically,
we found that participants who reported having the most difficulties in regulat-
ing emotion, showed reduced striatal, thalamic and somatosensory cortex activ-
ity during pain anticipation, increased activity in dorsal striatum, somatosen-
sory cortex, and cerabellum. These findings suggest that emotion regulation
strategies may work through modifying both affective and sensory components
42
of pain. We show that subjects with lower emotion regulation capacity show
less preparatory activity during pain anticipation and instead rely on dorsolat-
eral prefrontal cortex to regulate pain.
4.2 INTRODUCTION
According to recent estimates from Center for disease control, chronic pain af-
fects around 20.4% of US adult population [38]. For an individual, chronic pain
can lead to severely reduced quality of life [59]. On a societal level, chronic
pain imposes large economic and social costs, through decreased productivity,
increased health care costs, and increased burden on caregivers. Chronic pain
can also lead to development of major depressive disorder, thereby compound-
ing the adverse consequences [15]. With high prevalence and major individual
and societal costs, pain management remains a major public health issue.
Growing concerns around addiction related to pharmacological pain man-
agement have prompted an exploration of alternative pain management strate-
gies [76]. In that vein, some techniques commonly used to regulate negative
affect have been successfully co-opted for pain management. Specifically, cog-
nitive behavioral therapy and mindfulness-based therapy has been shown to
help reduce pain across multiple studies [168, 135, 148]. While emotion regula-
tion strategies present a promising alternative for pain management, the precise
mechanisms through which pain reduction is accomplished remain elusive.
One potential explanation is that emotion regulation strategies modulate the
affective but not sensory components of pain. This model of pain regulation
has been supported by research in clinical samples showing that the presence
43
and severity of comorbid depression in patience with fibromyalgia modulated
affective components of pain but did not affect the sensory-discriminative as-
pects [55]. These findings suggest that treating comorbid depression may in
turn, modulate, the affective components of pain, thereby decreasing subjec-
tive pain experience. Additional support for this model comes from research on
using instructed suppression to modulate painful stimuli. In this study, partici-
pants were instructed to either suppress or enhance their emotional reactions to
visually aversive stimuli. Their success was measured by monitoring their fa-
cial expressions through measuring the activity of the facial corrugator muscle,
which is responsible for frowning. The researchers found that emotion regula-
tory success predicted successful pain regulation and that regulatory success in
both conditions scaled with the activity in amygdala [84].
While these findings provide valuable insights to putative mechanisms of
the relationship between emotion regulation and pain management, the role
of individual differences in capacity for emotion regulation on pain processing
in non-clinical samples is still not clear. It is possible that amygdala activity
was primarily driven by instructions to enhance the aversive stimulus. Even
healthy individuals vary in their ability to regulate emotion. However, it is
unclear if this trait difference affects the way people naturally process aversive
and painful stimuli, without external instructions.
In the present study we examined how trait variability in emotion regula-
tion capacity affected subjective experience of pain and aversive sound stimuli,
as well its relationship to the neural mechanisms involved in anticipation and
processing of different aversive stimuli.
44
4.3 METHODS
4.3.1 Participants
Twenty four participants (16 female, 8 male) were recruited form Cornell Uni-
versity and surrounding community in Ithaca, NY area. Participants were
recruited through Psychology Department Psychology Experiment Sign-Up
(SONA) System and though posters around the Cornell University and the sur-
rounding area. One additional participant was scanned but did not complete
the personality measures survey and, thus, was excluded from analysis. All
participants were right-handed young adults ranging in age from 18 to 29 years
(M=22.16, SD=3) with normal or corrected-to-normal vision. Additionally, par-
ticipants were screened for current psychotropic medication use, mood and psy-
chiatric disorders and provided written informed consent approved by Cornell
Institutional Review Board. Upon completion of the study, each participant re-
ceived $50.
4.3.2 Stimuli
Pressure pain
Participants received mechanical pressure stimulation to the left thumbnail via
a custom designed hydraulic device described in chapter 3. Before starting the
scan session, each subject received instructions on how to place and adjust the
hand in the device. Participants then received pressure stimuli at all the levels
intensity present in the task (3 kg/cm2 - non painful/mild pressure, 5 kg/cm2
45
- moderately painful pressure, and 7 kg/cm2 - very painful pressure) for 11
seconds to make sure each participant was comfortable. In each scan run par-
ticipants received 7 pressure pain stimuli: 2 stimuli of each intensity, for 11 and
8 seconds, and an additional high-intensity painful stimulus as the first trial of
each run.
Emotionally Aversive Sounds
In each run participants heard 3 emotionally aversive sounds selected from In-
ternational Affective Digital Sounds (IADS) database [19]. Participants were
informed before the study that they will hear upsetting sounds that include
attacks, violence, and screaming, to ensure that no participants got extremely
upset. The selected stimuli had mean valence ratings of M = 2.50, SD = .46, and
mean arousal rating of M = 6.72, SD = 0.55.
Physically Aversive sounds
In addition to emotionally aversive sounds, in each run participants also heard
3 physically aversive sounds which consisted of unpleasant high pitch sounds
of 3 different intensities [52] reminiscent of nails screeching over a board for 8
seconds.
4.3.3 Procedure
Participants completed 5 runs of the aversive stimulus task, each lasting 11
min. 16 s. Every run started with a trial during which participants received a
46
7kg/cm2 pressure pain stimulus lasting 11 s, which was the most intense pres-
sure pain stimulus in the study. This was done to ensure that if participants
wanted to stop the scan because of pressure pain stimulation, they could do
so in the beginning of the run. Each trial began with a cue which indicated
what kind of aversive stimulus participants would be receiving in that trial, fol-
lowed by a jittered pre-stimulus delay lasting 2-5.5s. The stimulus start was
indicated by a change in fixation cross 500ms before the start of the stimulus.
The emotionally aversive sounds lasted 5-6s, physically aversive sounds lasted
8s, and pressure pain stimuli lasted 8-11s. The stimulus was followed by a post-
stimulus delay, the duration of which depended on the durations of previous
trial components. After the post stimulus delay participants rated how much
pleasure and displeasure they experienced on a given trial. The order of the
ratings was randomized across trials, with a variable short delay between. Par-
ticipants had 5s to rate the stimulus on each scale. The length of each trial was
not dependent on participants response time. The variability in response times
was absorbed by inter-trial interval (ITI), with the total length of the trial and
the ITI adding up to 44s (Figure 3.1)
After completing the scan session participants also completed digital ver-
sions of Difficulties in Emotion Regulation Scale (DERS) on a computer outside
of the scanner [58]. The surveys were coded using Qualtrics online survey web
platform.
47
4.3.4 Analysis
Behavioral Data Analysis
Behavioral data analysis and visualization was performed using R Studio [137].
To explore the relationship between displeasure ratings of each aversive stim-
ulus type and trait emotion regulation abilities, we conducted Kenadall cor-
relation analysis. Kendall correlation was chosen to account for non-normal
distribution of variables.
FMRI data acquisition and preprocessing
The images were acquired using a GE Discovery MR750 3T scanner (General
Electric, Milwaukee, United States) with a 32-channel head coil at the Cor-
nell Magnetic Resonance Imaging Facility. Anatomical scans were acquired
with T1-weighted volumetric MRI magnetization prepared rapid gradient echo
and sensitivity encoding (TR=7ms; TE=3.42ms; TI=1100ms; Flip Angle (FA)=7;
FOV=256256mm; sampling bandwidth=25kHz; voxel size=1mm isotropic; 176
slices; acceleration factor=2; scan time=5:25). The functional scans were ac-
quired using a multi-echo echo planar imaging (EPI) sequence (TR=2600ms;
TEs=13.7, 30, 47ms; FA=81; matrix size=7272mm; FOV=216 x 216mm; voxel
size=3mm isotropic; 120; acceleration factor=2.5; scan time=11:16). Participants
made responses during task runs with five-button response box held in the
right hand. The EPI data form task runs were preprocessed with Multi-Echo
Independent Components Analysis (ME-ICA) version 3, as described in Chap-
ter 3( [80, 79]). The anatomical images were skull stripped using FSL BET. The
data was then submitted ME-ICA processing pipeline, removing the first four
48
volumes of data for signal stabilization and warping the data to MNI space us-
ing a high-resolution template (MNI caez N27). The de-noised time series were
then smoothed using AFNI 3dmerge with a 6mm FWHM kernel [32, 31, 56].
FMRI data analysis
The smoothed data from the 5 functional runs was analyzed using Ordinary
Least Squares regression as implemented in AFNI 3dDeconvove [32, 31, 56].
Cue, stimulus, and each rating period were entered into the model as sepa-
rate events and convolved with the canonical Hemodynamic Response Func-
tion (HRF). Cue was modeled with a tent function SPMG1, while other trial
events were modelled with a duration modulated block. Displeasure ratings
were entered as parametric modulators for all aversive stimuli events. Pleasure
ratings were only entered as parametric modulators for painful pressure and
and emotionally aversive stimuli.
Second level analyses were conducted using one sample two-sided AFNI
3dttest++ with DERS and scores as a centered covariate.
4.4 RESULTS
4.4.1 Behavioral Results
Upon finishing the scan session, every participant completed DERS question-
naires (DERS M=69.76, SD = 17.11)
49
To examine whether DERS scores modulated displeasure ratings in each trial
type, we ran a Kendall correlation analysis (Figure 4.1. We found that DERS
scores were not significantly correlated with subjective displeasure in any of the
trial types (all ps >.1)
Figure 4.1: Behavioral Results: Scatter plots with regression lines show-
ing the lack of relationship between DERS scores and displea-
sure ratings across all stimulus types
4.4.2 FMRI Results
Emotion regulation during aversive event anticipation
Our first aim was to examine how individual differences in emotion regula-
tion affected the brain mechanisms of aversive event anticipation. We found
that during pressure pain anticipation participants with higher DERS scores
(i.e., poorer self-reported emotion regulation skills) showed decreased activity
in broad regions of dorsal striatum (peak BOLD decrease in Putamen) and tha-
lamus, bilaterally (Table 4.1 for full list of activations). In addition, activity
in right Supramarginal Gyrus (SPMG) also scaled negatively with individual
DERS scores (Figure 4.2)
50
Label Laterality X Y Z Cluster size t
Putamen R 17 5 5 752 -5.79
Thalamus L -13 -4 3 367 -6.84
Superior Occipital Gyrus L -14 -92 3 296 -6.57
Putamen L -20 19 0 199 -5.90
Thalamus R 16 -26 6 166 -4.74
Cuneus L -6 -72 15 93 -5.20
Cerebellum L -16 -52 -52 91 -5.77
-33 -37 -37 65 -6.12
Culmen L -8 -27 -30 59 -4.47
Cerebellum L -44 -49 -52 57 -6.09
Thalamus L -19 -28 5 56 -5.46
Supramarginal Gyrus R 63 -36 31 55 -5.86
Table 4.1: Areas modulated by DERS scores during pain anticipation.
Height threshold: t(23) = 3.819, p <.001 (uncorrected); extent
threshold, k = 50. L = Left, R = Right
Surprisingly, individual differences in emotion regulation ability did not
modulate brain activity during either emotionally aversive or physically aver-
sive sound anticipation.
Figure 4.2: Areas where activity was negatively modulated by DERS
scores during pain anticipation. Height threshold: t(23) = 3.819,
extent threshold, k = 50
51
Label Laterality X Y Z Cluster size t
Insula R 30 -38 21 166 -5.54
Cerebellum L -23 -71 -50 152 6.32
Caudate R 16 12 1 141 5.74
Supramarginal Gyrus R 51 -25 39 134 6.41
Cerebellum L -1 -57 -53 86 -5.99
Insula L -18 31 4 66 -4.81
Supramarginal gyrus L -54 -25 46 60 4.57
Posterior Cingulate Cortex L -21 -39 30 56 -4.99
Dorsolateral Prefrontal Cortex L -50 14 22 51 4.66
Anterior Cingulate Cortex L -17 41 13 42 -5.72
Posterior Cingulate Cortex R 17 -43 26 35 -5.03
Medial Prefrontal Cortex R 20 40 -2 31 -5.06
Table 4.2: Areas where brain activity was modulated by DERS scores dur-
ing pain processing. Height threshold: t(23) = 3.819, p <.001
(uncorrected); extent threshold, k = 30. L = Left, R = Right.
Emotion regulation and aversive event processing
Next, we examined how DERS scores affected brain activity during aversive
stimuli processing. In contrast to the effects during pain anticipation, we found
that individuals with higher DERS scores showed increased activity in dorsal
striatum (activation peak in caudate) , bilateral SPMG. Activity in Dorsolateral
Prefrontal Cortex (DLPFC) also scaled positively with DERS scores (Figure 4.3).
DERS scores also were associated with decreased activity in bilateral posterior
insular and cingulate cortices and medial prefrontal cortex (See Table 4.2 for
full list of activations).
In line with behavioral results, DERS did not modulate brain activity that
tracked subjective displeasure during different trial types. DERS also did not
modulate brain activity during either emotionally aversive or physically aver-
sive sounds.
52
Figure 4.3: Areas where brain activity was modulated by DERS scores dur-
ing pain processing. Height threshold: t(23) = 3.819, extent
threshold, k = 30
4.5 DISCUSSION
A growing body of research suggests that emotion regulation strategies can be
adapted to be used for acute and chronic pain regulation. However, the mech-
anisms through which these techniques modulate subjective pain experience
are not well understood. In the present study we addressed this question by
53
examining how trait variability in emotion regulation in healthy young adults
modifies pain and aversive sound processing. Our results suggest that trait
emotion regulation ability modulates pain processing but not the processing of
emotionally and physically aversive sounds. Since we did not provide partici-
pants with any instructions on whether and how to regulate their reactions to
painful stimuli, this discrepancy may reflect a stronger inherent motivation to
regulate nociceptive pain.
Surprisingly, we found that in nociceptive pain trials, trait emotion regula-
tion ability did not modify subjective displeasure ratings of pain. By the same
token, it did not modify activity in brain areas that have been previously shown
to track subjective displeasure. Our findings thus were unable to demonstrate
the relationship between emotion regulation and affective dimension of pain.
Conversely, our findings seem to indicate that trait emotion regulation inter-
acts with the areas that have been traditionally thought to reflect cognitive and
sensory processing of pain.
DERS scores during anticipation of pressure pain were associated with re-
duced activity in thalamus and S1 as well as large portions of dorsal striatum
and cerabellum. While dorsal striatum does play a role in reward processing in
the context of goal directed behavior [8, 7] and therefore could be considered
to represent affect, a more likely explanation that putamen carries sensory pain
representations to guide evasive action. Previous findings show that somato-
topic pain perception information is represented in contralateral putamen [13].
However, the mere presence of sensory information does not rule out potential
representations of affect.
The the relative suppression of the activity in somatosensory cortex and tha-
54
lamus may reflect the lack of preparatory activity that may modulate the success
of regulation during pressure pain stimulus. This account is supported by our
results during the the period of stimulation. We found that during pressure
pain stimulation stimulation dorsal striatum and somatosensory cortex showed
increased activation in participants with lower self reported emotion regulation
capacity. Despite these promising results, questions remain. Specifically, fu-
ture studies should look into whether the degree of signal decrease in dorsal
striatum and somatosensory cortex is predictive of the degree of signal increase
during stimulus presentation in the same areas.
We also found that in participants with lower self reported scores of trait
emotion regulation dLPFC was more active during stimulus presentation. This
result may in part account for the lack of difference in subjective ratings of dis-
pleasure. It is possible that participants with lower trait ability for emotion reg-
ulation engage in less proative stategies, but are able to compensate though
recruitment of top-down control mechanisms, resulting in the same subjective
experience. While such neural scaffolding may result in similar subjective expe-
rience with moderate pain stimuli, it remains unclear if it would remain effec-
tive at moderating pain under more severe circumstances.
55
APPENDIX A
SUPPLEMENTARY MATERIAL TO CHAPTER 2
A.1 SUPPLEMENTARY TABLES
Paper Contrast N
Delay
Avsar et al., 2013 [3] delay>control 14
Ballard & Knutson, 2009 [6] delay duration 16
Bickel et al., 2009 [12] delay discounting>control 10
Clithero et al., 2009 [30] delay>baseline 11
Kable & Glimcher, 2007 [70] delay duration 10
Li et al., 2013 [89] delay duration 23
Liu & Feng, 2012 [93] delay discounting>baseline 18
Liu et al., 2012 [92] delay discounting>control 19
Luhmann et al., 2008 [94] delay duration 20
Luo et al., 2009 [95] longer larger>shorter smaller 37
Massar et al., 2015 [99] delay duration 23
Peters & Buchel, 2009[121] delay duration 18
Pine et al., 2009 [123] delay duration 24
Pine et al., 2010 [124] delay duration 14
Ripke et al., 2012 [133] delay>control 27
Sripada et al., 2011 [151] longer larger>shorter smaller 20
Tanaka et al., 2004 [158] longer larger>shorter smaller 20
van den Bos et al., delay discounting>control 22
2014 [164]
56
Weber & Huettel, 2008 [172] delay duration 23
Wittman et al., 2007 [175] longer larger>shorter smaller 13
Wittmann et al., 2010 [176] longer larger>shorter smaller 13
Effort
Bonelle et al., 2015 [16] effort magnitude 37
Buhler et al., 2014 [47] effort magnitude 89
Croxon et al., 2009 [35] effort magnitude 16
Dean et al., 2016 [39] hard>easy 17
Esposito et al., 2009 [46] hard>easy 8
Klein-Flugge et al, 2016 [74] effort magnitude 24
Kurniawan et al., 2013 [81] hard>easy 19
hard>easy;
Lallement et al., 2014 [62] 28
effrot discounting >control
Massar et al., 2015 [99] effort magnitude 23
McGuire & Botvinick, hard>easy 19
2010 [101]
Meyniel et al., 2013 [103] effort magnitude 19
Mulert et al., 2008 [113] hard>easy; hard 10
Otto et al., 2014 [116] effort magnitude 14
Pessiglione et al., 2007 [120] effort magnitude 18
Prevost et al., 2010 [129] effort magnitude 18
Schouppe et al., 2014 [145] hard>easy 22
Shmidt et al., 2009 [142] effort magnitude 20
Shmidt et al., 2012 [143] effort magnitude 19
Skvortsova et al., 2014 [150] effort magnitude 20
57
Vassena et al., 2014 [167] hard>easy 22
Yang et al., 2016 [181] effort discounting>control 25
Yu et al., 2014 [182] effort magnitude 21
Risk
Bach et al., 2009 [4] risky>safe 20
Bach et al., 2009 [4] risk level 20
Bjork et al., 2008 [14] risky>safe 17
d’Acremont et al., 2013 [37] risk level 23
Dreher et al., 2006 [43] risk level 31
FitzGerald et al., 2010[48] risk level 18
Galvan et al., 2013 [51] risk level 43
Hsu et al., 2005 [66] risky>safe 16
Hsu et al., 2009 [67] risk level 21
Huettel et al., 2006 [68] risky >safe 12
Kuhnen & Knutson, risk level 19
2005 [78]
Lei et al., 2017 [85] risky>safe 31
Luhmann et al., 2008 [94] risk level 20
Macoveanu et al., 2013 [96] risk level 20
Macoveanu et al., 2013 [97] risk level 22
Minati et al., 2012 [104] risk level 22
Mohr et al., 2010 [110] risk level 16
Paulus & Frank, 2006 [118] risky>safe 16
Peters & Buchel, 2009 [121] risk level 22
Preuschoff et al., 2006 [128] risk level 19
58
Preuschoff et al., 2008 [127] risk level 19
Rao et al., 2008 [132] risky>safe 14
Roy et al., 2012 [136] risky>safe 23
Schoemberg et al, 2012 [144] risk level 16
Suter et al., 2015 [154] risk level 41
Suzuki et al., 2016 [155] risk level 24
Symmonds et al., 2010 [156] risk level 16
Symmonds et al., 2011 [157] risk level 23
Tobler et al., 2007 [159] risk level 16
van Leijenhorst et al., risky>safe 14
2006 [165]
Vassena et al., 2014 [167] risky>safe 23
Vassena et al., 2014 [166] risk level 23
Weber & Huettel, 2008 [172] risk level 23
Wright et al., 2012 [177] risk level 22
Wright et al., 2013 [178] risky>safe 24
Yacubian et al., 2006 [180] risk level 42
Table A.1: Papers included in the corpus of studies
59
Label Peak Cluster size
Supplementary Motor Area -6, -8, 60 2928
L Primary Motor Cortex -34, -26, 54 2016
L Putamen -18, 12, 0 1528
R Insular Cortex 34, 20, 6 1184
L Insular Cortex -34, 22, 6 1176
R Middle Temporal Gyrus 50, -38, -6 976
L Insular Cortex -46, -26, 22 816
R Cuneus 16, -68, 38 712
R Globus Palidus 20, -2, -4 584
Dorsal Anterior Cingulate Cortex -6, 10, 34 512
L Primary Motor Cortex -52, -6, 46 424
Substantia Nigra -10, -22, -12 416
R Inferior Parietal Lobule 42, -44, 42 408
L Inferior Frontal Gyrus -50, 10, 32 400
R Inferior Frontal Gyrus 48, 10, 30 376
L Cuneus -8, -72, 38 360
Table A.2: Areas reliably activated by effort costs.
Label Peak Cluster size
R Dorsolateral Prefrontal Cortex 50, 10, 26 2584
L Middle Frontal Gyrus -40, 48, 14 1696
R Medial Frontal Gyrus 4, 8, 50 1688
R Dorsolateral Prefrontal Cortex 46, 34, 6 1472
Medial Prefrontal Cortex 2, 34, 32 1424
Posterior Cingulate Cortex 0, -30, 32 1240
L Precuneus -24, -76, 40 1232
R Precuneus 20, -76, 44 1040
L Inferior Frontal Gyrus -48, 12, 26 968
L Inferior Occipital Gyrus -12, -96, -4 904
L Angular Gyrus -44, -58, 40 856
L Insular Cortex -30, 18, 0 824
Substantia Nigra 6, -20, -12 680
L Caudate -10, 10, 4 576
L Medial Frontal Gyrus -8, 42, 22 512
L Dorsolateral Prefrontal Cortex -48, 42, 4 448
Table A.3: Areas reliably activated by delay costs
60
Label Peak Cluster size
Bilateral Caudate 10, 4, -2 13760
-10, 8, -4
L Thalamus -2, -8, 2
Dorsal Anterior Cingulate Cortex -4, 30, 34 4472
R Dorsolateral Prefrontal Cortex 44, 36, 20 3608
L Insular Cortex -32, 18, -2 3544
R Superior Parietal Lobule 30, -58, 48 3424
R Lateral Prefrontal Cortex 36, 54, 4 704
R Middle Frontal Gyrus 32 ,6, 60 680
R Precuneus 12, -74, 52 672
R Fusiform Gyrus 48, -60, -8 584
Brainstem 10, -28, -10 568
Table A.4: Areas reliably activated by probability costs
A.2 SUPPLEMENTARY FIGURES
Figure A.1: Compared to effort costs, delay costs were more likely to en-
gage bilateral dlPFC and right IFG (A). In contrast, effort costs
were more likely to engage SMA and left M1 (B). Both effort
and delay costs reliably engaged bilateral IFG, left insula and
left Caudate (C).
61
Figure A.2: Compared to probability costs. Effort costs were more likely
to engage posterior Insula, left M1 and SMA (A). In contrasts,
probability costs were more likely to engage right dlPFC and
ventral portions of Insula (B). Both effort and probability costs
reliably engaged bilateral dorsal anterior Insula, bilateral Cau-
date and right IPL.
Figure A.3: Compared to delay, probability costs were more likely to en-
gage more anterior bilateral Insular cortex and left Caudate
(A). Conversely, delay costs were more likely to engage bilat-
eral dlPFC (B). Both delay and probability costs reliably en-
gaged dACC, SMA, midbrain, left Insula and Caudate and
right Intraparietal Sulcus (C).
62
BIBLIOGRAPHY
[1] Adam K Anderson and Noam Sobel. Dissociating intensity from valence
as sensory inputs to emotion. Neuron, 39(4):581–583, August 2003.
[2] A.K. Anderson, K. Christoff, I. Stappen, D. Panitz, D. G. Ghahremani,
G. Glover, J.D.E. Gabrieli, and N. Sobel. Dissociated neural representa-
tions of intensity and valence in human olfaction. Nature Neuroscience,
6(2):196–202, January 2003.
[3] Kathy Burton Avsar, Rosalyn Eve Weller, James Edward Cox, Mered-
ith Amanda Reid, David Matthew White, and Adrienne Carol Lahti. An
fMRI investigation of delay discounting in patients with schizophrenia.
Brain and Behavior, 3(4):384–401, July 2013.
[4] D. R. Bach, B. Seymour, and R. J. Dolan. Neural Activity Associated with
the Passive Prediction of Ambiguity and Risk for Aversive Events. Journal
of Neuroscience, 29(6):1648–1656, February 2009.
[5] Matthew R. Bailey, Eleanor H. Simpson, and Peter D. Balsam. Neu-
ral substrates underlying effort, time, and risk-based decision making in
motivated behavior. Neurobiology of Learning and Memory, 133:233–256,
September 2016.
[6] Kacey Ballard and Brian Knutson. Dissociable neural representations of
future reward magnitude and delay during temporal discounting. Neu-
roImage, 45(1):143–150, March 2009.
[7] B. W. Balleine, M. R. Delgado, and O. Hikosaka. The Role of the Dor-
sal Striatum in Reward and Decision-Making. Journal of Neuroscience,
27(31):8161–8165, August 2007.
[8] Bernard W Balleine and John P O’Doherty. Human and Rodent Homolo-
gies in Action Control: Corticostriatal Determinants of Goal-Directed and
Habitual Action. Neuropsychopharmacology, 35(1):48–69, January 2010.
[9] Andrea Bari and Trevor W. Robbins. Inhibition and impulsivity: Behav-
ioral and neural basis of response control. Progress in Neurobiology, 108:44–
79, September 2013.
[10] Oscar Bartra, Joseph T. McGuire, and Joseph W. Kable. The valuation
system: A coordinate-based meta-analysis of BOLD fMRI experiments
63
examining neural correlates of subjective value. NeuroImage, 76:412–427,
August 2013.
[11] Roy F. Baumeister, Kathleen D. Vohs, and Dianne M. Tice. The strength
model of self-control. Current Directions in Psychological Science, 16(6):351–
355, December 2007.
[12] W. K. Bickel, J. A. Pitcock, R. Yi, and E. J. C. Angtuaco. Congruence of
BOLD Response across Intertemporal Choice Conditions: Fictive and Real
Money Gains and Losses. Journal of Neuroscience, 29(27):8839–8846, July
2009.
[13] U. Bingel, J. Gläscher, C. Weiller, and C. Büchel. Somatotopic represen-
tation of nociceptive information in the putamen: An event-related fMRI
study. Cerebral Cortex, 14(12):1340–1345, December 2004.
[14] James M. Bjork, Reza Momenan, Ashley R. Smith, and Daniel W. Hom-
mer. Reduced posterior mesofrontal cortex activation by risky rewards in
substance-dependent patients. Drug and Alcohol Dependence, 95(1-2):115–
128, May 2008.
[15] G. Blackburn-Munro and R. E. Blackburn-Munro. Chronic pain, chronic
stress and depression: Coincidence or consequence? Journal of Neuroen-
docrinology, 13(12):1009–1023, December 2001.
[16] Valerie Bonnelle, Sanjay Manohar, Tim Behrens, and Masud Husain. In-
dividual Differences in Premotor Brain Systems Underlie Behavioral Ap-
athy. Cerebral Cortex, page bhv247, November 2015.
[17] Peter Bossaerts and Carsten Murawski. From behavioural economics to
neuroeconomics to decision neuroscience: the ascent of biology in re-
search on human decision making. Current Opinion in Behavioral Sciences,
5:37–42, October 2015.
[18] Matthew Botvinick and Todd Braver. Motivation and Cognitive Con-
trol: From Behavior to Neural Mechanism. Annual Review of Psychology,
66(1):83–113, January 2015.
[19] M. M. Bradley and P. J. Lang. The International Affective Digitized Sounds
(2nd Edition; IADS-2): Affective ratings of sounds and instruction manual.
Technical report B-3. University of Florida, Gainesville, FL.
64
[20] Andrea Brovelli, Bruno Nazarian, Martine Meunier, and Driss Boussaoud.
Differential roles of caudate nucleus and putamen during instrumental
learning. NeuroImage, 57(4):1580–1590, August 2011.
[21] Michael Browning, Timothy E Behrens, Gerhard Jocham, Jill X OReilly,
and Sonia J Bishop. Anxious individuals have difficulty learning the
causal statistics of aversive environments. Nature Neuroscience, 18(4):590–
596, March 2015.
[22] C. J. Burke, C. Brunger, T. Kahnt, S. Q. Park, and P. N. Tobler. Neural Inte-
gration of Risk and Effort Costs by the Frontal Pole: Only upon Request.
Journal of Neuroscience, 33(4):1706–1713, January 2013.
[23] Xinying Cai, Soyoun Kim, and Daeyeol Lee. Heterogeneous Coding of
Temporally Discounted Values in the Dorsal and Ventral Striatum during
Intertemporal Choice. Neuron, 69(1):170–182, January 2011.
[24] Evan C. Carter, Lilly M. Kofler, Daniel E. Forster, and Michael E. McCul-
lough. A series of meta-analytic tests of the depletion effect: Self-control
does not seem to rely on a limited resource. Journal of Experimental Psy-
chology: General, 144(4):796–815, August 2015.
[25] R. McKell Carter, Justin R. Meyer, and Scott A. Huettel. Functional neu-
roimaging of intertemporal choice models: A review. Journal of Neuro-
science, Psychology, and Economics, 3(1):27–45, 2010.
[26] B. J. Casey, Leah H. Somerville, Ian H. Gotlib, Ozlem Ayduk, Nicholas T.
Franklin, Mary K. Askrend, John Jonides, Marc G. Berman, Nicole L. Wil-
son, Theresa Teslovich, Gary Glover, Vivian Zayas, Walter Mischel, and
Yuichi Shodae. Behavioral and neural correlates of delay of gratification
40 years later. Annals of Neurosciences, 19(1), January 2012.
[27] Svenja Caspers, Karl Zilles, Angela R. Laird, and Simon B. Eickhoff. ALE
meta-analysis of action observation and imitation in the human brain.
NeuroImage, 50(3):1148–1167, April 2010.
[28] Junichi Chikazoe, Daniel H Lee, Nikolaus Kriegeskorte, and Adam K An-
derson. Population coding of affect across stimuli, modalities and indi-
viduals. Nature Neuroscience, 17(8):1114–1122, August 2014.
[29] Trevor T.-J. Chong, Matthew Apps, Kathrin Giehl, Annie Sillence,
Laura L. Grima, and Masud Husain. Neurocomputational mecha-
65
nisms underlying subjective valuation of effort costs. PLOS Biology,
15(2):e1002598, February 2017.
[30] John A. Clithero, R. McKell Carter, and Scott A. Huettel. Local pattern
classification differentiates processes of economic valuation. NeuroImage,
45(4):1329–1338, May 2009.
[31] R. W. Cox and J. S. Hyde. Software tools for analysis and visualization of
fMRI data. NMR in biomedicine, 10(4-5):171–178, August 1997.
[32] Robert W. Cox. AFNI: Software for Analysis and Visualization of Func-
tional Magnetic Resonance Neuroimages. Computers and Biomedical Re-
search, 29(3):162–173, June 1996.
[33] A. D. Craig. How do you feel? Interoception: the sense of the physio-
logical condition of the body. Nature Reviews Neuroscience, 3(8):655–666,
August 2002.
[34] A. D. Craig. How do you feel now? The anterior insula and human
awareness. Nature Reviews Neuroscience, 10(1):59–70, January 2009.
[35] P. L. Croxson, M. E. Walton, J. X. O’Reilly, T. E. J. Behrens, and M. F. S.
Rushworth. Effort-Based Cost-Benefit Valuation and the Human Brain.
Journal of Neuroscience, 29(14):4531–4541, April 2009.
[36] M. d’Acremont and P. Bossaerts. Neurobiological studies of risk assess-
ment: A comparison of expected utility and mean-variance approaches.
Cognitive, Affective, & Behavioral Neuroscience, 8(4):363–374, December
2008.
[37] Mathieu d’Acremont, Eleonora Fornari, and Peter Bossaerts. Activity in
Inferior Parietal and Medial Prefrontal Cortex Signals the Accumulation
of Evidence in a Probability Learning Task. PLoS Computational Biology,
9(1):e1002895, January 2013.
[38] James Dahlhamer, Jacqueline Lucas, Carla Zelaya, Richard Nahin, Sean
Mackey, Lynn DeBar, Robert Kerns, Michael Von Korff, Linda Porter, and
Charles Helmick. Prevalence of chronic pain and high-impact chronic
pain among adults — united states, 2016. MMWR. Morbidity and Mortality
Weekly Report, 67(36):1001–1006, September 2018.
[39] Z. Dean, S. Horndasch, P. Giannopoulos, and C. McCabe. Enhanced
66
neural response to anticipation, effort and consummation of reward and
aversion during bupropion treatment. Psychological Medicine, 46(11):2263–
2274, August 2016.
[40] Ben Deen, Naomi B. Pitskel, and Kevin A. Pelphrey. Three Systems of
Insular Functional Connectivity Identified with Cluster Analysis. Cerebral
Cortex, 21(7):1498–1506, July 2011.
[41] TheresaM. Desrochers, Ken-ichi Amemori, and AnnM. Graybiel. Habit
Learning by Naive Macaques Is Marked by Response Sharpening of Stri-
atal Neurons Representing the Cost and Outcome of Acquired Action Se-
quences. Neuron, 87(4):853–868, August 2015.
[42] N. U. F. Dosenbach, D. A. Fair, F. M. Miezin, A. L. Cohen, K. K. Wenger,
R. A. T. Dosenbach, M. D. Fox, A. Z. Snyder, J. L. Vincent, M. E. Raichle,
B. L. Schlaggar, and S. E. Petersen. Distinct brain networks for adaptive
and stable task control in humans. Proceedings of the National Academy of
Sciences, 104(26):11073–11078, June 2007.
[43] Jean-Claude Dreher, Philip Kohn, and Karen Faith Berman. Neural Cod-
ing of Distinct Statistical Properties of Reward Information in Humans.
Cerebral Cortex, 16(4):561–573, April 2006.
[44] Simon B. Eickhoff, Danilo Bzdok, Angela R. Laird, Florian Kurth, and Pe-
ter T. Fox. Activation likelihood estimation meta-analysis revisited. Neu-
roImage, 59(3):2349–2361, February 2012.
[45] Simon B. Eickhoff, Angela R. Laird, Christian Grefkes, Ling E. Wang,
Karl Zilles, and Peter T. Fox. Coordinate-based activation likelihood esti-
mation meta-analysis of neuroimaging data: A random-effects approach
based on empirical estimates of spatial uncertainty. Human Brain Mapping,
30(9):2907–2926, September 2009.
[46] Fabrizio Esposito, Christoph Mulert, and Rainer Goebel. Combined dis-
tributed source and single-trial EEGfMRI modeling: Application to effort-
ful decision making processes. NeuroImage, 47(1):112–121, August 2009.
[47] Mira Fauth-Bhler, Evangelos Zois, Sabine Vollstdt-Klein, Tagrid Leme-
nager, Martin Beutel, and Karl Mann. Insula and striatum activity in
effort-related monetary reward processing in gambling disorder: The role
of depressive symptomatology. NeuroImage: Clinical, 6:243–251, 2014.
67
[48] Thomas H.B. FitzGerald, Ben Seymour, Dominik R. Bach, and Raymond J.
Dolan. Differentiable Neural Substrates for Learned and Described Value
and Risk. Current Biology, 20(20):1823–1829, October 2010.
[49] Alexander Friedman, Daigo Homma, LeifG. Gibb, Ken-ichi Amemori,
SamuelJ. Rubin, AdamS. Hood, MichaelH. Riad, and AnnM. Graybiel. A
Corticostriatal Path Targeting Striosomes Controls Decision-Making un-
der Conflict. Cell, 161(6):1320–1333, June 2015.
[50] Ralph Frost and Neil McNaughton. The neural basis of delay discounting:
A review and preliminary model. Neuroscience & Biobehavioral Reviews,
79:48–65, August 2017.
[51] Adriana Galvn, Tom Schonberg, Jeanette Mumford, Milky Kohno, Rus-
sell A. Poldrack, and Edythe D. London. Greater risk sensitivity of dor-
solateral prefrontal cortex in young smokers than in nonsmokers. Psy-
chopharmacology, 229(2):345–355, September 2013.
[52] Stephan Geuter, Elizabeth A. Reynolds Losin, Mathieu Roy, Lauren Y. At-
las, Liane Schmidt, Anjali Krishnan, Leonie Koban, Tor D. Wager, and
Martin A. Lindquist. Multiple brain networks mediating stimulus-pain
relationships in humans. April 2018.
[53] Jimmy Ghaziri, Alan Tucholka, Gabriel Girard, Olivier Boucher, Jean-
Christophe Houde, Maxime Descoteaux, Sami Obaid, Guillaume Gilbert,
Isabelle Rouleau, and Dang Khoa Nguyen. Subcortical structural connec-
tivity of insular subregions. Scientific Reports, 8(1):8596, December 2018.
[54] Jimmy Ghaziri, Alan Tucholka, Gabriel Girard, Jean-Christophe Houde,
Olivier Boucher, Guillaume Gilbert, Maxime Descoteaux, Sarah Lipp,
Pierre Rainville, and Dang Khoa Nguyen. The Corticocortical Struc-
tural Connectivity of the Human Insula. Cerebral Cortex, 27(2):1216–1228,
February 2017.
[55] Thorsten Giesecke, Richard H. Gracely, David A. Williams, Michael E.
Geisser, Frank W. Petzke, and Daniel J. Clauw. The relationship between
depression, clinical pain, and experimental pain in a chronic pain cohort.
Arthritis & Rheumatism, 52(5):1577–1584, May 2005.
[56] S. Gold, B. Christian, S. Arndt, G. Zeien, T. Cizadlo, D. L. Johnson,
M. Flaum, and N. C. Andreasen. Functional MRI statistical software pack-
ages: a comparative analysis. Human Brain Mapping, 6(2):73–84, 1998.
68
[57] J. A. Gottfried. Encoding Predictive Reward Value in Human Amygdala
and Orbitofrontal Cortex. Science, 301(5636):1104–1107, August 2003.
[58] Kim L. Gratz and Lizabeth Roemer. Multidimensional Assessment of
Emotion Regulation and Dysregulation: Development, Factor Structure,
and Initial Validation of the Difficulties in Emotion Regulation Scale. Jour-
nal of Psychopathology and Behavioral Assessment, 26(1):41–54, March 2004.
[59] Oye Gureje, Michael Von Korff, Gregory E. Simon, and Richard Gater.
Persistent pain and well-being. JAMA, 280(2):147, July 1998.
[60] M. S. Hagger, N. L. D. Chatzisarantis, H. Alberts, C. O. Anggono,
C. Batailler, A. R. Birt, R. Brand, M. J. Brandt, G. Brewer, S. Bruyneel,
D. P. Calvillo, W. K. Campbell, P. R. Cannon, M. Carlucci, N. P. Carruth,
T. Cheung, A. Crowell, D. T. D. De Ridder, S. Dewitte, M. Elson, J. R.
Evans, B. A. Fay, B. M. Fennis, A. Finley, Z. Francis, E. Heise, H. Hoe-
mann, M. Inzlicht, S. L. Koole, L. Koppel, F. Kroese, F. Lange, K. Lau, B. P.
Lynch, C. Martijn, H. Merckelbach, N. V. Mills, A. Michirev, A. Miyake,
A. E. Mosser, M. Muise, D. Muller, M. Muzi, D. Nalis, R. Nurwanti, H. Ot-
gaar, M. C. Philipp, P. Primoceri, K. Rentzsch, L. Ringos, C. Schlinkert,
B. J. Schmeichel, S. F. Schoch, M. Schrama, A. Schtz, A. Stamos, G. Tinghg,
J. Ullrich, M. vanDellen, S. Wimbarti, W. Wolff, C. Yusainy, O. Zerhouni,
and M. Zwienenberg. A Multilab Preregistered Replication of the Ego-
Depletion Effect. Perspectives on Psychological Science, 11(4):546–573, July
2016.
[61] A. R. Hariri, S. M. Brown, D. E. Williamson, J. D. Flory, H. de Wit, and
S. B. Manuck. Preference for Immediate over Delayed Rewards Is Asso-
ciated with Magnitude of Ventral Striatal Activity. Journal of Neuroscience,
26(51):13213–13217, December 2006.
[62] Julen Hernandez Lallement, Katarina Kuss, Peter Trautner, Bernd Weber,
Armin Falk, and Klaus Fliessbach. Effort increases sensitivity to reward
and loss magnitude in the human brain. Social Cognitive and Affective Neu-
roscience, 9(3):342–349, March 2014.
[63] Wilhelm Hofmann, Brandon J. Schmeichel, and Alan D. Baddeley. Exec-
utive functions and self-regulation. Trends in Cognitive Sciences, 16(3):174–
180, March 2012.
[64] Victoria Holec, Heather L. Pirot, and David R. Euston. Not all effort is
equal: the role of the anterior cingulate cortex in different forms of effort-
reward decisions. Frontiers in Behavioral Neuroscience, 8, 2014.
69
[65] Jay G Hosking, Stan B Floresco, and Catharine A Winstanley. Dopamine
Antagonism Decreases Willingness to Expend Physical, But Not Cogni-
tive, Effort: A Comparison of Two Rodent Cost/Benefit Decision-Making
Tasks. Neuropsychopharmacology, 40(4):1005–1015, March 2015.
[66] M. Hsu. Neural Systems Responding to Degrees of Uncertainty in Human
Decision-Making. Science, 310(5754):1680–1683, December 2005.
[67] M. Hsu, I. Krajbich, C. Zhao, and C. F. Camerer. Neural Response to Re-
ward Anticipation under Risk Is Nonlinear in Probabilities. Journal of Neu-
roscience, 29(7):2231–2237, February 2009.
[68] S. A. Huettel. Decisions under Uncertainty: Probabilistic Context Influ-
ences Activation of Prefrontal and Parietal Cortices. Journal of Neuro-
science, 25(13):3304–3311, March 2005.
[69] Michael Inzlicht and Brandon J. Schmeichel. What Is Ego Depletion? To-
ward a Mechanistic Revision of the Resource Model of Self-Control. Per-
spectives on Psychological Science, 7(5):450–463, September 2012.
[70] Joseph W Kable and Paul W Glimcher. The neural correlates of subjective
value during intertemporal choice. Nature Neuroscience, 10(12):1625–1633,
December 2007.
[71] Daniel Kahneman and Amos Tversky. Prospect Theory: An Analysis of
Decision under Risk. Econometrica, 47(2):263, March 1979.
[72] J. Kim, S. V. Shinkareva, and D. H. Wedell. Representations of modality-
general valence for videos and music derived from fMRI data. Neuroim-
age, 148:42–54, 03 2017.
[73] Soyoun Kim, Jaewon Hwang, and Daeyeol Lee. Prefrontal Coding of
Temporally Discounted Values during Intertemporal Choice. Neuron,
59(1):161–172, July 2008.
[74] M. C. Klein-Flugge, S. W. Kennerley, K. Friston, and S. Bestmann. Neu-
ral Signatures of Value Comparison in Human Cingulate Cortex during
Decisions Requiring an Effort-Reward Trade-off. Journal of Neuroscience,
36(39):10002–10015, September 2016.
[75] H. Kober, L. F. Barrett, J. Joseph, E. Bliss-Moreau, K. Lindquist, and T. D.
Wager. Functional grouping and cortical-subcortical interactions in emo-
70
tion: a meta-analysis of neuroimaging studies. Neuroimage, 42(2):998–
1031, Aug 2008.
[76] Andrew Kolodny, David T. Courtwright, Catherine S. Hwang, Peter
Kreiner, John L. Eadie, Thomas W. Clark, and G. Caleb Alexander. The
prescription opioid and heroin crisis: A public health approach to an epi-
demic of addiction. Annual Review of Public Health, 36(1):559–574, March
2015.
[77] Amy L. Krain, Amanda M. Wilson, Robert Arbuckle, F. Xavier Castel-
lanos, and Michael P. Milham. Distinct neural mechanisms of risk and
ambiguity: A meta-analysis of decision-making. NeuroImage, 32(1):477–
484, August 2006.
[78] Camelia M. Kuhnen and Brian Knutson. The Neural Basis of Financial
Risk Taking. Neuron, 47(5):763–770, September 2005.
[79] P. Kundu, N. D. Brenowitz, V. Voon, Y. Worbe, P. E. Vertes, S. J. Inati, Z. S.
Saad, P. A. Bandettini, and E. T. Bullmore. Integrated strategy for improv-
ing functional connectivity mapping using multiecho fMRI. Proceedings of
the National Academy of Sciences, 110(40):16187–16192, October 2013.
[80] Prantik Kundu, Souheil J. Inati, Jennifer W. Evans, Wen-Ming Luh, and
Peter A. Bandettini. Differentiating BOLD and non-BOLD signals in fMRI
time series using multi-echo EPI. NeuroImage, 60(3):1759–1770, April 2012.
[81] I. T. Kurniawan, M. Guitart-Masip, P. Dayan, and R. J. Dolan. Effort and
Valuation in the Brain: The Effects of Anticipation and Execution. Journal
of Neuroscience, 33(14):6160–6169, April 2013.
[82] Florian Kurth, Karl Zilles, Peter T. Fox, Angela R. Laird, and Simon B.
Eickhoff. A link between the systems: functional differentiation and inte-
gration within the human insula revealed by meta-analysis. Brain Struc-
ture and Function, 214(5-6):519–534, June 2010.
[83] Jack L. Lancaster, Diana Tordesillas-Gutirrez, Michael Martinez, Felipe
Salinas, Alan Evans, Karl Zilles, John C. Mazziotta, and Peter T. Fox.
Bias between MNI and Talairach coordinates analyzed using the ICBM-
152 brain template. Human Brain Mapping, 28(11):1194–1205, November
2007.
[84] Regina C. Lapate, Hyejeen Lee, Tim V. Salomons, Carien M. van Reekum,
Lawrence L. Greischar, and Richard J. Davidson. Amygdalar function
71
reflects common individual differences in emotion and pain regulation
success. Journal of Cognitive Neuroscience, 24(1):148–158, January 2012.
[85] Yu Lei, Lubin Wang, Pinhong Chen, Yanyan Li, Wei Han, Mingmei Ge, Liu
Yang, Shanshan Chen, Wendong Hu, Xinhuai Wu, and Zheng Yang. Neu-
ral correlates of increased risk-taking propensity in sleep-deprived people
along with a changing risk level. Brain Imaging and Behavior, 11(6):1910–
1921, December 2017.
[86] D. J. Levy and P. W. Glimcher. Comparing Apples and Oranges: Using
Reward-Specific and Reward-General Subjective Value Representation in
the Brain. Journal of Neuroscience, 31(41):14693–14707, October 2011.
[87] Dino J Levy and Paul W Glimcher. The root of all value: a neural com-
mon currency for choice. Current Opinion in Neurobiology, 22(6):1027–1038,
December 2012.
[88] Ifat Levy, Jason Snell, Amy J. Nelson, Aldo Rustichini, and Paul W. Glim-
cher. Neural Representation of Subjective Value Under Risk and Ambigu-
ity. Journal of Neurophysiology, 103(2):1036–1047, February 2010.
[89] N. Li, N. Ma, Y. Liu, X.-S. He, D.-L. Sun, X.-M. Fu, X. Zhang, S. Han, and
D.-R. Zhang. Resting-State Functional Connectivity Predicts Impulsivity
in Economic Decision-Making. Journal of Neuroscience, 33(11):4886–4895,
March 2013.
[90] K. A. Lindquist, A. B. Satpute, T. D. Wager, J. Weber, and L. F. Bar-
rett. The Brain Basis of Positive and Negative Affect: Evidence from
a Meta-Analysis of the Human Neuroimaging Literature. Cereb. Cortex,
26(5):1910–1922, May 2016.
[91] Kristen A. Lindquist, Tor D. Wager, Hedy Kober, Eliza Bliss-Moreau, and
Lisa Feldman Barrett. The brain basis of emotion: A meta-analytic review.
Behavioral and Brain Sciences, 35(3):121–143, May 2012.
[92] Lei Liu and Tingyong Feng. The neural predictors of choice preference in
intertemporal choice. Brain Research, 1436:92–100, February 2012.
[93] Lei Liu, Tingyong Feng, Jianfeng Wang, and Hong Li. The neural dis-
sociation of subjective valuation from choice processes in intertemporal
choice. Behavioural Brain Research, 231(1):40–47, May 2012.
72
[94] C. C. Luhmann, M. M. Chun, D.-J. Yi, D. Lee, and X.-J. Wang. Neural
Dissociation of Delay and Uncertainty in Intertemporal Choice. Journal of
Neuroscience, 28(53):14459–14466, December 2008.
[95] S. Luo, G. Ainslie, L. Giragosian, and J. R. Monterosso. Behavioral
and Neural Evidence of Incentive Bias for Immediate Rewards Rela-
tive to Preference-Matched Delayed Rewards. Journal of Neuroscience,
29(47):14820–14827, November 2009.
[96] Julian Macoveanu, James B. Rowe, Bettina Hornboll, Rebecca Elliott,
Olaf B. Paulson, Gitte M. Knudsen, and Hartwig R. Siebner. Playing it safe
but losing anywaySerotonergic signaling of negative outcomes in dorso-
medial prefrontal cortex in the context of risk-aversion. European Neu-
ropsychopharmacology, 23(8):919–930, August 2013.
[97] Julian Macoveanu, James B Rowe, Bettina Hornboll, Rebecca Elliott,
Olaf B Paulson, Gitte M Knudsen, and Hartwig R Siebner. Serotonin 2a re-
ceptors contribute to the regulation of risk-averse decisions. NeuroImage,
83:35–44, December 2013.
[98] Jelena Markovic, Adam K. Anderson, and Rebecca M. Todd. Tuning to
the significant: Neural and genetic processes underlying affective en-
hancement of visual perception and memory. Behavioural Brain Research,
259:229–241, February 2014.
[99] Stijn A.A. Massar, Camilo Libedinsky, Chee Weiyan, Scott A. Huettel, and
Michael W.L. Chee. Separate and overlapping brain areas encode subjec-
tive value during delay and effort discounting. NeuroImage, 120:104–113,
October 2015.
[100] Mara Mather, David Clewett, Michiko Sakaki, and Carolyn W. Harley.
Norepinephrine ignites local hotspots of neuronal excitation: How
arousal amplifies selectivity in perception and memory. Behavioral and
Brain Sciences, 39, July 2015.
[101] Joseph T. McGuire and Matthew M. Botvinick. Prefrontal cortex, cognitive
control, and the registration of decision costs. Proceedings of the National
Academy of Sciences, 107(17):7922–7926, April 2010.
[102] M.-Marsel Mesulam and Elliott J. Mufson. Insula of the old world mon-
key. III: Efferent cortical output and comments on function. The Journal of
Comparative Neurology, 212(1):38–52, November 1982.
73
[103] F. Meyniel, C. Sergent, L. Rigoux, J. Daunizeau, and M. Pessiglione. Neu-
rocomputational account of how the human brain decides when to have
a break. Proceedings of the National Academy of Sciences, 110(7):2641–2646,
February 2013.
[104] Ludovico Minati, Marina Grisoli, Silvana Franceschetti, Francesca Epi-
fani, Alice Granvillano, Nick Medford, Neil A. Harrison, Sylvie Pia-
centini, and Hugo D. Critchley. Neural Signatures of Economic Pa-
rameters During Decision-Making: A Functional MRI (fMRI), Electroen-
cephalography (EEG) and Autonomic Monitoring Study. Brain Topogra-
phy, 25(1):73–96, January 2012.
[105] Walter Mischel, Ozlem Ayduk, Marc G. Berman, B. J. Casey, Ian H. Gotlib,
John Jonides, Ethan Kross, Theresa Teslovich, Nicole L. Wilson, Vivian
Zayas, and Yuichi Shoda. Willpower over the life span: decomposing self-
regulation. Social Cognitive and Affective Neuroscience, 6(2):252–256, April
2011.
[106] Walter Mischel, Yuichi Shoda, and Philip K. Peake. The nature of adoles-
cent competencies predicted by preschool delay of gratification. Journal of
Personality and Social Psychology, 54(4):687–696, 1988.
[107] V Miskovic and AK Anderson. Modality general and modality specific
coding of hedonic valence. Current Opinion in Behavioral Sciences, 19:91–
97, February 2018.
[108] V. Miskovic, K. Kuntzelman, J. Chikazoe, and A. K. Anderson. Represen-
tation of affect in sensory cortex. Behav Brain Sci, 39:e252, Jan 2016.
[109] G. Mochizuki, T. Hoque, R. Mraz, B.J. MacIntosh, S.J. Graham, S.E. Black,
W.R. Staines, and W.E. McIlroy. Challenging the brain: Exploring the link
between effort and cortical activation. Brain Research, 1301:9–19, Novem-
ber 2009.
[110] Peter N.C. Mohr, Guido Biele, Lea K. Krugel, Shu-Chen Li, and Hauke R.
Heekeren. Neural foundations of riskreturn trade-off in investment deci-
sions. NeuroImage, 49(3):2556–2563, February 2010.
[111] John R. Monterosso, George Ainslie, Jiansong Xu, Xochitl Cordova,
Catherine P. Domier, and Edythe D. London. Frontoparietal cortical activ-
ity of methamphetamine-dependent and comparison subjects performing
a delay discounting task. Human Brain Mapping, 28(5):383–393, May 2007.
74
[112] Elliott J. Mufson and M.-Marsel Mesulam. Insula of the old world mon-
key. II: Afferent cortical input and comments on the claustrum. The Journal
of Comparative Neurology, 212(1):23–37, November 1982.
[113] Christoph Mulert, Christian Seifert, Gregor Leicht, Valerie Kirsch,
Matthias Ertl, Susanne Karch, Matthias Moosmann, Jrgen Lutz, Hans-
Jrgen Mller, Ulrich Hegerl, Oliver Pogarell, and Lorenz Jger. Single-
trial coupling of EEG and fMRI reveals the involvement of early ante-
rior cingulate cortex activation in effortful decision making. NeuroImage,
42(1):158–168, August 2008.
[114] Mark Muraven and Roy F. Baumeister. Self-regulation and depletion of
limited resources: Does self-control resemble a muscle? Psychological Bul-
letin, 126(2):247–259, 2000.
[115] M. P. Noonan, M. E. Walton, T. E. J. Behrens, J. Sallet, M. J. Buckley, and
M. F. S. Rushworth. Separate value comparison and learning mechanisms
in macaque medial and lateral orbitofrontal cortex. Proceedings of the Na-
tional Academy of Sciences, 107(47):20547–20552, November 2010.
[116] Tobias Otto, Fred R. H. Zijlstra, and Rainer Goebel. Neural correlates
of mental effort evaluation involvement of structures related to self-
awareness. Social Cognitive and Affective Neuroscience, 9(3):307–315, March
2014.
[117] Camillo Padoa-Schioppa and John A. Assad. Neurons in the orbitofrontal
cortex encode economic value. Nature, 441(7090):223–226, May 2006.
[118] Martin P. Paulus and Lawrence R. Frank. Anterior cingulate activity mod-
ulates nonlinear decision weight function of uncertain prospects. Neu-
roImage, 30(2):668–677, April 2006.
[119] Jonathan Peirce, Jeremy R. Gray, Sol Simpson, Michael MacAskill, Richard
Hchenberger, Hiroyuki Sogo, Erik Kastman, and Jonas Kristoffer Lindelv.
PsychoPy2: Experiments in behavior made easy. Behavior Research Meth-
ods, 51(1):195–203, February 2019.
[120] M. Pessiglione, L. Schmidt, B. Draganski, R. Kalisch, H. Lau, R. J. Dolan,
and C. D. Frith. How the Brain Translates Money into Force: A Neu-
roimaging Study of Subliminal Motivation. Science, 316(5826):904–906,
May 2007.
75
[121] J. Peters and C. Buchel. Overlapping and Distinct Neural Systems Code
for Subjective Value during Intertemporal and Risky Decision Making.
Journal of Neuroscience, 29(50):15727–15734, December 2009.
[122] Jan Peters and Christian Bchel. The neural mechanisms of inter-temporal
decision-making: understanding variability. Trends in Cognitive Sciences,
15(5):227–239, May 2011.
[123] A. Pine, B. Seymour, J. P. Roiser, P. Bossaerts, K. J. Friston, H. V. Curran,
and R. J. Dolan. Encoding of Marginal Utility across Time in the Human
Brain. Journal of Neuroscience, 29(30):9575–9581, July 2009.
[124] A. Pine, T. Shiner, B. Seymour, and R. J. Dolan. Dopamine, Time, and
Impulsivity in Humans. Journal of Neuroscience, 30(26):8888–8896, June
2010.
[125] Michael L Platt and Scott A Huettel. Risky business: the neuroeconomics
of decision making under uncertainty. Nature Neuroscience, 11(4):398–403,
April 2008.
[126] Gilles Pourtois, Antonio Schettino, and Patrik Vuilleumier. Brain mech-
anisms for emotional influences on perception and attention: What is
magic and what is not. Biological Psychology, 92(3):492–512, March 2013.
[127] K. Preuschoff, S. R. Quartz, and P. Bossaerts. Human Insula Activation
Reflects Risk Prediction Errors As Well As Risk. Journal of Neuroscience,
28(11):2745–2752, March 2008.
[128] Kerstin Preuschoff, Peter Bossaerts, and Steven R. Quartz. Neural Differ-
entiation of Expected Reward and Risk in Human Subcortical Structures.
Neuron, 51(3):381–390, August 2006.
[129] C. Prevost, M. Pessiglione, E. Metereau, M.-L. Clery-Melin, and J.-C.
Dreher. Separate Valuation Subsystems for Delay and Effort Decision
Costs. Journal of Neuroscience, 30(42):14080–14090, October 2010.
[130] D. D. Price. Psychological and neural mechanisms of the affective dimen-
sion of pain. Science, 288(5472):1769–1772, June 2000.
[131] Antonio Rangel and Todd Hare. Neural computations associated with
goal-directed choice. Current Opinion in Neurobiology, 20(2):262–270, April
2010.
76
[132] Hengyi Rao, Marc Korczykowski, John Pluta, Angela Hoang, and John A.
Detre. Neural correlates of voluntary and involuntary risk taking in the
human brain: An fMRI study of the balloon analog risk task (BART). Neu-
roImage, 42(2):902–910, August 2008.
[133] Stephan Ripke, Thomas Hbner, Eva Mennigen, Kathrin U. Mller, Sarah
Rodehacke, Dirk Schmidt, Mark J. Jacob, and Michael N. Smolka. Reward
processing and intertemporal decision making in adults and adolescents:
The role of impulsivity and decision consistency. Brain Research, 1478:36–
47, October 2012.
[134] Monica Larrea Rodriguez, Walter Mischel, and Yuichi Shoda. Cognitive
person variables in the delay of gratification of older children at risk. Jour-
nal of Personality and Social Psychology, 57(2):358–367, 1989.
[135] Steven Rosenzweig, Jeffrey M. Greeson, Diane K. Reibel, Joshua S. Green,
Samar A. Jasser, and Denise Beasley. Mindfulness-based stress reduction
for chronic pain conditions: Variation in treatment outcomes and role of
home meditation practice. Journal of Psychosomatic Research, 68(1):29–36,
January 2010.
[136] Amy Krain Roy, Kristin Gotimer, A.M. Clare Kelly, F. Xavier Castellanos,
Michael P. Milham, and Monique Ernst. Uncovering putative neural
markers of risk avoidance. Neuropsychologia, 49(5):937–944, April 2011.
[137] RStudio Team. RStudio: Integrated Development Environment for R. RStudio,
Inc., Boston, MA, 2015.
[138] Peter H Rudebeck, Mark E Walton, Angharad N Smyth, David M Ban-
nerman, and Matthew F S Rushworth. Separate neural pathways process
different decision costs. Nature Neuroscience, 9(9):1161–1168, September
2006.
[139] K. Samejima. Representation of Action-Specific Reward Values in the
Striatum. Science, 310(5752):1337–1340, November 2005.
[140] Samantha R. Santacruz, Erin L. Rich, Joni D. Wallis, and Jose M. Carmena.
Caudate Microstimulation Increases Value of Specific Choices. Current
Biology, 27(21):3375–3383.e3, November 2017.
[141] A. B. Satpute, J. Kang, K. C. Bickart, H. Yardley, T. D. Wager, and L. F.
Barrett. Involvement of Sensory Regions in Affective Experience: A Meta-
Analysis. Front Psychol, 6:1860, 2015.
77
[142] L. Schmidt, M.-L. Clery-Melin, G. Lafargue, R. Valabregue, P. Fossati,
B. Dubois, and M. Pessiglione. Get Aroused and Be Stronger: Emotional
Facilitation of Physical Effort in the Human Brain. Journal of Neuroscience,
29(30):9450–9457, July 2009.
[143] Liane Schmidt, Mal Lebreton, Marie-Laure Clry-Melin, Jean Daunizeau,
and Mathias Pessiglione. Neural Mechanisms Underlying Motivation
of Mental Versus Physical Effort. PLoS Biology, 10(2):e1001266, February
2012.
[144] Tom Schonberg, Craig R. Fox, Jeanette A. Mumford, Eliza Congdon,
Christopher Trepel, and Russell A. Poldrack. Decreasing Ventromedial
Prefrontal Cortex Activity During Sequential Risk-Taking: An fMRI In-
vestigation of the Balloon Analog Risk Task. Frontiers in Neuroscience, 6,
2012.
[145] N. Schouppe, J. Demanet, C. N. Boehler, K. R. Ridderinkhof, and W. Note-
baert. The Role of the Striatum in Effort-Based Decision-Making in the
Absence of Reward. Journal of Neuroscience, 34(6):2148–2154, February
2014.
[146] J. Schweimer. Involvement of the rat anterior cingulate cortex in control of
instrumental responses guided by reward expectancy. Learning & Memory,
12(3):334–342, May 2005.
[147] Kendra L Seaman, Nickolas Brooks, Teresa M Karrer, Jaime J Castrellon,
Scott F Perkins, Linh C Dang, Ming Hsu, David H Zald, and Gregory R
Samanez-Larkin. Subjective value representations during effort, probabil-
ity and time discounting across adulthood. Social Cognitive and Affective
Neuroscience, 13(5):449–459, May 2018.
[148] Marina Shpaner, Clare Kelly, Greg Lieberman, Hayley Perelman, Marcia
Davis, Francis J. Keefe, and Magdalena R. Naylor. Unlearning chronic
pain: A randomized controlled trial to investigate changes in intrinsic
brain connectivity following cognitive behavioral therapy. NeuroImage:
Clinical, 5:365–376, 2014.
[149] Bryan F. Singer, Monica Fadanelli, Alex B. Kawa, and Terry E. Robinson.
Are cocaine-seeking habits necessary for the development of addiction-
like behavior in rats? The Journal of Neuroscience, pages 2458–17, Novem-
ber 2017.
[150] V. Skvortsova, S. Palminteri, and M. Pessiglione. Learning To Mini-
78
mize Efforts versus Maximizing Rewards: Computational Principles and
Neural Correlates. Journal of Neuroscience, 34(47):15621–15630, November
2014.
[151] Chandra Sekhar Sripada, Richard Gonzalez, K. Luan Phan, and Israel
Liberzon. The neural correlates of intertemporal decision-making: Con-
tributions of subjective value, stimulus type, and trait impulsivity: Neu-
ral Correlates of Intertemporal Decision-Making. Human Brain Mapping,
32(10):1637–1648, October 2011.
[152] J. R. St. Onge and S. B. Floresco. Prefrontal Cortical Contribution to Risk-
Based Decision Making. Cerebral Cortex, 20(8):1816–1828, August 2010.
[153] G. P. Strauss, J. A. Waltz, and J. M. Gold. A Review of Reward Process-
ing and Motivational Impairment in Schizophrenia. Schizophrenia Bulletin,
40(Suppl 2):S107–S116, March 2014.
[154] Renata S. Suter, Thorsten Pachur, Ralph Hertwig, Tor Endestad, and
Guido Biele. Data from: The neural basis of risky choice with affective
outcomes, 2015. type: dataset.
[155] Shinsuke Suzuki, Emily L. S. Jensen, Peter Bossaerts, and John P.
ODoherty. Behavioral contagion during learning about another agents
risk-preferences acts on the neural representation of decision-risk. Pro-
ceedings of the National Academy of Sciences, 113(14):3755–3760, April 2016.
[156] M. Symmonds, P. Bossaerts, and R. J. Dolan. A Behavioral and Neural
Evaluation of Prospective Decision-Making under Risk. Journal of Neuro-
science, 30(43):14380–14389, October 2010.
[157] Mkael Symmonds, Nicholas D. Wright, Dominik R. Bach, and Raymond J.
Dolan. Deconstructing risk: Separable encoding of variance and skewness
in the brain. NeuroImage, 58(4):1139–1149, October 2011.
[158] Saori C Tanaka, Kenji Doya, Go Okada, Kazutaka Ueda, Yasumasa
Okamoto, and Shigeto Yamawaki. Prediction of immediate and future
rewards differentially recruits cortico-basal ganglia loops. Nature Neuro-
science, 7(8):887–893, July 2004.
[159] Philippe N. Tobler, John P. O’Doherty, Raymond J. Dolan, and Wolfram
Schultz. Reward Value Coding Distinct From Risk Attitude-Related Un-
certainty Coding in Human Reward Systems. Journal of Neurophysiology,
97(2):1621–1632, February 2007.
79
[160] Michael T. Treadway, Nicholas A. Bossaller, Richard C. Shelton, and
David H. Zald. Effort-based decision-making in major depressive disor-
der: A translational model of motivational anhedonia. Journal of Abnormal
Psychology, 121(3):553–558, 2012.
[161] R. D. Treede, A. V. Apkarian, B. Bromm, J. D. Greenspan, and F. A. Lenz.
Cortical representation of pain: functional characterization of nociceptive
areas near the lateral sulcus. Pain, 87(2):113–119, Aug 2000.
[162] Elizabeth Tricomi and Karolina M. Lempert. Value and probability cod-
ing in a feedback-based learning task utilizing food rewards. Journal of
Neurophysiology, 113(1):4–13, January 2015.
[163] Peter E. Turkeltaub, Simon B. Eickhoff, Angela R. Laird, Mick Fox, Martin
Wiener, and Peter Fox. Minimizing within-experiment and within-group
effects in activation likelihood estimation meta-analyses. Human Brain
Mapping, 33(1):1–13, January 2012.
[164] W. van den Bos, C. A. Rodriguez, J. B. Schweitzer, and S. M. McClure.
Connectivity Strength of Dissociable Striatal Tracts Predict Individual Dif-
ferences in Temporal Discounting. Journal of Neuroscience, 34(31):10298–
10310, July 2014.
[165] Linda Van Leijenhorst, Bregtje Gunther Moor, Zdea A. Op de Macks,
Serge A.R.B. Rombouts, P. Michiel Westenberg, and Eveline A. Crone.
Adolescent risky decision-making: Neurocognitive development of re-
ward and control regions. NeuroImage, 51(1):345–355, May 2010.
[166] Eliana Vassena, Ruth M. Krebs, Massimo Silvetti, Wim Fias, and Tom
Verguts. Dissociating contributions of ACC and vmPFC in reward pre-
diction, outcome, and choice. Neuropsychologia, 59:112–123, July 2014.
[167] Eliana Vassena, Massimo Silvetti, Carsten N. Boehler, Eric Achten, Wim
Fias, and Tom Verguts. Overlapping Neural Systems Represent Cognitive
Effort and Reward Anticipation. PLoS ONE, 9(3):e91008, March 2014.
[168] M. M. Veehof, H. R. Trompetter, E. T. Bohlmeijer, and K. M. G. Schreurs.
Acceptance- and mindfulness-based interventions for the treatment of
chronic pain: a meta-analytic review. Cognitive Behaviour Therapy, 45(1):5–
31, January 2016.
[169] Brent A. Vogt. Pain and emotion interactions in subregions of the cingu-
late gyrus. Nature Reviews Neuroscience, 6(7):533–544, July 2005.
80
[170] John Von Neumann and Oskar Morgenstern. Theory of games and economic
behavior. Princeton classic editions. Princeton University Press, Princeton,
N.J. ; Woodstock, 60th anniversary ed edition, 2007. OCLC: ocm78989126.
[171] Mark E. Walton, James Groves, Katie A. Jennings, Paula L. Croxson,
Trevor Sharp, Matthew F. S. Rushworth, and David M. Bannerman. Com-
paring the role of the anterior cingulate cortex and 6-hydroxydopamine
nucleus accumbens lesions on operant effort-based decision making. Eu-
ropean Journal of Neuroscience, 29(8):1678–1691, April 2009.
[172] Bethany J. Weber and Scott A. Huettel. The neural substrates of proba-
bilistic and intertemporal decision making. Brain Research, 1234:104–115,
October 2008.
[173] Michael J. Wesley and Warren K. Bickel. Remember the Future II: Meta-
analyses and Functional Overlap of Working Memory and Delay Dis-
counting. Biological Psychiatry, 75(6):435–448, March 2014.
[174] C. A. Winstanley and S. B. Floresco. Deciphering Decision Making: Vari-
ation in Animal Models of Effort- and Uncertainty-Based Choice Reveals
Distinct Neural Circuitries Underlying Core Cognitive Processes. Journal
of Neuroscience, 36(48):12069–12079, November 2016.
[175] Marc Wittmann, David S. Leland, and Martin P. Paulus. Time and deci-
sion making: differential contribution of the posterior insular cortex and
the striatum during a delay discounting task. Experimental Brain Research,
179(4):643–653, June 2007.
[176] Marc Wittmann, Kathryn L. Lovero, Scott D. Lane, and Martin P. Paulus.
Now or later? Striatum and insula activation to immediate versus de-
layed rewards. Journal of Neuroscience, Psychology, and Economics, 3(1):15–
26, 2010.
[177] N. D. Wright, M. Symmonds, K. Hodgson, T. H. B. Fitzgerald, B. Craw-
ford, and R. J. Dolan. Approach-Avoidance Processes Contribute to Dis-
sociable Impacts of Risk and Loss on Choice. Journal of Neuroscience,
32(20):7009–7020, May 2012.
[178] Nicholas D. Wright, Mkael Symmonds, and Raymond J. Dolan. Distinct
encoding of risk and value in economic choice between multiple risky
options. NeuroImage, 81:431–440, November 2013.
81
[179] W.M. Wundt and E.B. Titchener. Principles of Physiological Psychology, Vol-
ume 1 - Scholar’s Choice Edition. Scholar’s Choice, 2015.
[180] J. Yacubian. Dissociable Systems for Gain- and Loss-Related Value Predic-
tions and Errors of Prediction in the Human Brain. Journal of Neuroscience,
26(37):9530–9537, September 2006.
[181] Xin-hua Yang, Jia Huang, Yong Lan, Cui-ying Zhu, Xiao-qun Liu, Ye-fei
Wang, Eric F.C. Cheung, Guang-rong Xie, and Raymond C.K. Chan. Di-
minished caudate and superior temporal gyrus responses to effort-based
decision making in patients with first-episode major depressive disorder.
Progress in Neuro-Psychopharmacology and Biological Psychiatry, 64:52–59,
January 2016.
[182] Rongjun Yu, Dean Mobbs, Ben Seymour, James B. Rowe, and Andrew J.
Calder. The neural signature of escalating frustration in humans. Cortex,
54:165–178, May 2014.
82