University of Southern California Dissertations and Theses

The neuroscience of ambivalent and ambiguous feelings

(USC Thesis Other)

The neuroscience of ambivalent and ambiguous feelings

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2023 Anthony Gianni Vaccaro

THE NEUROSCIENCE OF AMBIVALENT AND AMBIGUOUS FEELINGS

by

Anthony Gianni Vaccaro

A Dissertation Presented to the
FACULTY OF THE DORNSIFE COLLEGE OF ARTS AND SCIENCES
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)

AUGUST 2023

ii

Acknowledgements

My first thanks is deservingly to my mentor Jonas Kaplan. I could not have asked for a better
advisor to learn from, to help improve my skills, and to cultivate am environment where I could
develop as a researcher and truly explore my interests down any tangential rabbit hole they fell.
Thank you for supporting me through my PhD, being there to bother and have a friendly chat
with, and providing the periodic existential reassurance I needed.

To Antonio: thank you for encouraging me to develop my wildest and farthest reaching ideas.
You’ve inspired me to think about all these topics across the expanse of biology, and taught me
how to craft a strong scientific idea.

Thank you to my committee members: Mary Helen Immordino-Yang, Mary Sweeney, and
Jonathan Stange for inspiring me to see my research and ideas through new lenses. You
perspectives changed the way I view my work, and expanded the areas I want to take my work to
in the future.

Thank you to John Monterosso for being such a great boss to TA for, and for fostering in me a
joy from teaching and mentoring.

Thanks to my USC friends and labmates for (almost) always being down to run off to get some
coffee and talk about the week’s current existential dread: shout out to Nina, Andrew, Sarah,
Leo, Brock, Ellen, Colin, Roshni, Chelsey, and Xiao. Another shout out to my three mentees
Shruti, Helen, and Rishab who have not only had to chat with me, but also work on my projects,
making massive contributions.

Thanks to my friends in Los Angeles for giving me a home in this city, listening to my
unrelatable anxieties, and filling my weekly schedule for the past 5 years to the brim with trivia,
games, restaurants, friendly gambling, premieres, AMC A-list, and viewing parties: Michael,
Wallis, Wenwei, Hannah, Brian, Oscar, Asia, Jacob, and Eric.

iii

Thank you to the Rochester group chat(s): Paul, Sarah, Wes, Joe, Aaron, Katrina, Nate, Dan, and
Maria; you all truly got me through pandemic lockdowns, and our weekly Zoom hang out where
we chat under the guise of pretending to play trivia is an anchor of my sanity. Additional thanks
to my good friends Zach and Chris, both completing PhD’s over the same years in different
cities, for being supportive both on matters academic and matters personal.

To my parents and brother Frankie: thank you for always believing I was the smartest person in
any room (regardless of the truth of it). I'm happy to have had your encouragement in paving my
own path, and I appreciate the valuable lesson you taught me about valuing my own opinions
and sense of what is right, and not just acquiesce to the views of others. I have to mention here
the memories of sitting in the backseat of the car at night, hearing the song Same Old Lang Syne,
and being absolutely fascinated by the ‘happy yet sad’ feelings being portrayed. Multiple times
as a kid I asked how our brains feel things and you said “I don’t know. Ask your teacher. I’m not
sure if anybody knows”. Well, here’s a small part of the answer.

My final thanks is to my original science enablers: my late grandparents Betty and Anthony
Vaccaro who, with less than high school educations, appreciated science more than anyone I
know and in a manner reserved by most people to the appreciation of art. I know both would
laugh or give a “woah” at seeing the name “Dr. Anthony Vaccaro” in writing
iv

TABLE OF CONTENTS

Acknowledgements………………………………………………………………………………..ii

List of Tables……………………………………………………………………………………...v

List of Figures……………………………………………………………………………………vii

Abstract…………………………………………………………………………………………...ix

Introduction………………………………………………………………………………………..1

Chapter 1. A theoretical roadmap for the neurobiology of mixed feelings
1a. The biology of valence at different scales……………………………………..4
1b. Bittersweet: The neuroscience of ambivalent affect…………………………21

Chapter 2: Perspective-taking is associated with increased discriminability of affective states
in ventromedial prefrontal cortex ………………………………………………………………..38

Chapter 3: Individual differences in feelings of certainty surrounding mixed emotions………..69

Chapter 4: Mixed feelings comprise a unique neural state in higher cortical regions …………111

Chapter 5: Conclusions…………………………………………………………………………138

References………………………………………………………………………………………150

v

LIST OF TABLES

Chapter 2

Table S1: Examples of Stimuli Sentence-Picture Pairs………………………………………….61
Table S2: Mean (Standard Deviation) of Rated Valence and Arousal of Stimuli (1-7 scale)…...61
Table S3 T-values From Two Sample T-Tests Comparing Valence and Arousal of Stimuli…...62
Table S4: Average Number of Participants Choosing Stimuli Category by Category of
Emotion…………………………………………………………………………………………..62

Table S5: Mean and Standard Deviations of Empathy Sub-scales………………………………62
Chapter 3
Table 1: Certainty of affect predicted by positive and negative valence, and their interaction….78

Table 2: Certainty of affect predicted by positive, negative, and mixed feelings……………….79
Table 3: Certainty of affect predicted by valence and emotional intelligence…………………..81
Table 4: Certainty of affect predicted by positive, negative, and mixed feelings (Study 2)…….87
Table 5: Certainty of affect predicted by valence and emotional intelligence (SSEIT)………....89
Table 6: Certainty of affect predicted by valence and trait clarity ……….……………………..90
Table S1: Certainty of affect predicted by valence and alexithymia (TAS)…………………….96
Table S2: Certainty of affect predicted by valence and interoceptive awareness (MAIA)……...97
Table S3: Certainty of affect predicted by mixed feelings and emotional intelligence………....98
Table S4: Certainty of affect predicted by valence and emotional intelligence subscales………99

Table S5: Certainty of affect predicted by valence and managing one’s own emotions….........101

Table S6: Certainty of affect predicted by valence, meta-mood attention, and repair…………102

Table S7: Certainty of affect predicted by valence and openness……………………………...104

Table S8: Certainty of affect predicted by valence and extraversion…………………………..105

Table S9: Certainty of affect predicted by valence and emotional stability……………………106
vi

Table S10: Certainty of affect predicted by valence and agreeableness………………………..107

Table S11: Certainty of affect predicted by valence and conscientiousness…………………...108

Ratings for Video Stimuli Used in Both Study 1 and Study 2…………………………………109

Percentage of Subjects Reporting Different Feeling Categories Across Studies 1 and 2………110

Chapter 4

Table 1: Probability of feeling state 1 second later based on current reported feeling state…...118

Table 2: Optimal Number of States in Each Region Based on Within vs. Across Correlations
in the Mean Data………………………………………………………………………………..124

vii

LIST OF FIGURES

Chapter 1

Figure 1: Conceptual model of ambivalent affect on cortical and subcortical levels……………31

Chapter 2

Figure 1: Accuracy of affect classification searchlight………………………………………….52

Figure 2: Regions where classification accuracy was significantly predicted by perspective-
taking…………………………………………………………………………………………….53

Figure 3: Classification accuracy vs. perspective-taking in peak voxels of the vmPFC and
insula……………………………………………………………………………………………..54
Figure S1: Proportion of participants that rated emotion stimuli in category……………………64

Figure S2: Distribution of null classification accuracy………………………………………….64

Figure S3: Accuracy of emotion classification searchlight using the minimum threshold
derived from permutation testing………………………………………………………………...65

Figure S4: Searchlight of Happy vs. Fear………………………………………………………..66

Figure S5: Searchlight of Happy vs. Disgust…………………………………………………….66

Figure S6: Searchlight of Sad vs. Fear…………………………………………………………...67

Figure S7: Searchlight of Sad vs. Disgust……………………………………………………….67

Figure S8: Searchlight of Fear vs. Disgust………………………………………………………68

Figure S9: Distributions of Interpersonal Reactivity Index Sub-scales…………………………68
Chapter 3
Figure 1: Certainty of affect predicted by positive and negative valence……………………….79
Figure 2: Certainty of affect predicted by valence and emotional intelligence………………….82

Figure 3: Certainty of affect predicted by positive and negative valence (Study 2)……………..88
Figure 4: Certainty of affect predicted by valence and trait emotional clarity…………………..91

viii

Chapter 4
Figure 1: Percentage of subject’s feeling positive, negative, and mixed at each timepoint of
the film…………………………………………………………………………………………118

Figure 2: Optimal Number of States in Each Region Based on Within vs. Across
Correlations in the Mean Data…………………………………………………………………125

Figure 3: Significance of Matching Rates in the Anterior Insula Using Self-Reported
Boundaries vs. Consensus Boundaries…………………………………………………………127

Figure 4: Significance of Matching Rates in the Amygdala Using Self-Reported Boundaries
vs. Consensus Boundaries……………………………………………………………………...127

Figure 5: Significance of Matching Rate in Anterior and Posterior Cingulate Using
Self-Reported Boundaries vs. Consensus Boundaries………………………………………….128

Figure 6: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Insular Sub-Regions………………………………………...129

Figure 7: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Ventromedial Prefrontal Cortex and Amygdala…………….130

Figure 8: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Cingulate Regions…………………………………………...131

Chapter 5

Figure 1: The generation of an ambiguous or ambivalent feeling……………………………...145
Figure 2: The neurobiology of ambivalence in regards to affective experience in addiction…..146

Figure 3: The resolution of ambivalence across timescales in the development of empathetic
artificial intelligence……………………………………………………………………………148

ix

Abstract

Mixed feelings or ambivalence, the simultaneous experience of positivity and negativity,
are common and universal experiences. Despite this, there is no research in affective
neuroscience on the topic. This is in large part due to many hurdles that exist in studying them.
Mixed feelings do not easily fit into most prominent theories on the nature of emotion; most
theories conceptualize positive and negative valence as a being on a bipolar spectrum. Some
researchers contend that mixed feelings do not truly exist, and are instead merely an illusion of
consciousness. Furthermore, it is difficult to define what makes a feeling truly “mixed”, when
there is ongoing debate as to what would even make feelings distinct from each other in various
emotion theories. Moving past this issue, mixed feelings are difficult to induce in experimental
settings, and even when induced, can be associated with subjects feeling they “do not know”
how they feel. This dissertation consists of novel theories, and 3 experimental studies, which aim
to tackle the prominent issues involved in researching these complex emotional experiences. The
studies shed light on how cognitive traits relate to the distinctiveness of emotions in the cortex,
the relationship between uncertainty and ambivalence, and how various brain regions differ in
their representation of valence. This work shows not only that ambivalent and ambiguous
feelings are researchable in neuroscience, but that expanding the field of affective neuroscience
into this area is necessary and fruitful.

Keywords: Mixed feelings; ambivalence; mixed emotions; affective neuroscience; uncertainty;
bittersweetness
1

INTRODUCTION
“Ambivalence is my favorite thing to write about, because it's the way I feel,
and I think the way most people feel.”
-Stephen Sondheim
How do you feel right now, on scale from 1-9, with 1 being most negative and 9 being
most positive? This simple judgement is perhaps the most common proxy for feeling used in the
study of affect. It is simple, comprehendible to a reader, and a well-controlled measure for
associating with other variables. The importance of valence is largely unquestioned in the study
of affect, even between competing theories of the nature of emotion. However, the seeming
simplicity of this measure, and its use across the spectrum of emotion theorists, does not mean it
carries no theoretical assumptions of its own. Whenever we ask a participant to rate how they
feel on this bipolar scale, we are making a huge theoretical assumption: that the degree of
positivity we feel is the exact inverse of how negative we feel. Simply, we never feel both
positively and negatively at the same time. At a minimum, we are assuming positivity and
negativity inherently cancel each other out in our consciousness.
Feeling is interesting not because of its reducible simplicity, but because of its chaotic
and messy nature. If our feelings never conflicted, never confused us, and we always felt
confident about exactly what we are feeling, we would not feel such a strong need to discuss
them, and explore them. The utility of these internal signals being conscious at all would be
lessened. By excluding the possibility of ambivalence and uncertainty in our study of emotion,
we veer away from studying the reality of what people experience much of the time.
2

While the literature often use terminology that sidesteps directly calling out mixed
feelings, the idea of conflict in the psyche can be traced back to the roots of the field. Breuer and
Freud were focused on largely on ideas of inner conflict, and that psychological disruption
mostly arose from conflicting desires (Eagle, 2017). Their explanations and solutions to this
issue were irreparably incorrect; but the observation loosely touched on the phenomena of mixed
feelings as a prevalent source of confusion and distress. Historically, the development of
different schools of clinical psychology has largely been driven by various ways in which
conflict and ambivalence are conceptualized, and how they can be resolved (Jonas et al., 2000;
Sincoff, 1990; Thagard et al., 2023).
The field of psychology has primarily approached the question of mixed feelings by
asking questions about the dimensionality of valence, the quality of affective states that makes
them positive/desirable or negative/undesirable. With valence, the main question is whether
positive and negative valence exist on a single bipolar spectrum, or instead are two independent
dimensions (Barrett & Russell, 1998; Briesemeister et al., 2012; Larsen, 2017; Larsen &
McGraw, 2014). If valence is a unidimensional bipolar spectrum, this means that mixed feelings
cannot exist, in the sense that any affective state would have to be either positive or negative; it
could never simultaneously be both. In a unidimensional model, mixed feelings could only be
implemented as a rapid vacillation between positivity and negativity, too quick to subjectively
perceive (Feldman Barrett & Russell, 1998; Larsen, 2017; Russell & Carroll, 1999; Young,
1918). If, alternatively, valence is comprised of two independent dimensions, then truly
concurrent positivity and negativity would be possible (Cacioppo & Berntson, 1994;
Schimmack, 2001).
3

The overarching goal of this dissertation is to pave the path for research on mixed
feelings from a neurobiological perspective. The first 3 chapters are aimed at addressing issues in
studying mixed feelings. In Chapter 1, I present an in depth literature review on the
neurobiology of valence, and then present my own theories of what functional neuroanatomy
may underly mixed feelings- this addresses the lack of a neuroscientific theory of emotion that
can accommodate mixed feelings. The second half of this chapter is already published in
Perspectives on Psychological Science (Vaccaro et al., 2020). In Chapter 2, I demonstrate that
we can use fMRI, machine learning, and cognitive measures to explore what processes may lead
to affective states being distinct from each other. This chapter is already published in Social
Cognitive Affective Neuroscience (Vaccaro et al., 2022). In Chapter 3, I present a study
investigating the relationship between being uncertain of how you feel, and experiencing mixed
feelings- two concepts that are often mistakenly blended into one, or used as an alternative
explanation for the existence of mixed feelings. Finally in Chapter 4, I present an fMRI study
investigating whether various cortical regions can predict individualized feeling dynamics even
when experiencing mixed feelings, and attempting to find supporting evidence for my theory that
some regions may represent mixed feelings as a consistent and unique state, whereas others do
not. Chapter 5 serves as a conclusion for the dissertation as a whole, where I highlight the
importance of the findings, and discuss how essential the understanding of ambivalent and
ambiguous feelings is to some of the most important issues in our modern and future society.
4

CHAPTER 1: A THEORETICAL ROADMAP FOR THE NEUROBIOLOGY OF MIXED
FEELINGS

1a. The biology of valence at different scales

The fact is that the field of biopsychology uses the term valence for any behavior, feeling,
computation, or belief that is in any manner considered “positive” or “negative”; valence is used
hedonically, motivationally, and behaviorally (Berridge, 2019). Valence may potentially be
operationalized in different ways at these different levels; this gives us room to question how
properties of valence may differ depending on the scale and substrate. As is done in the
psychology of self-report affect ratings, we can ask whether valence is bipolar or represented as
two independent structures at these different scales. Additionally important is the question of
whether there is just one representation of valence, if multiple valences of different domains or
of different appraisal types co-exist, or if multiple valences do exist but converge into one
“macro-valence”. Finally, are certain perceptual or neural signals inherently valenced, or do
other factors shape the affective stance that is taken on whether a signal is positive or negative?
These questions may be too large to answer definitively, but I believe that some evidence in the
literature, particularly when affect is treated as a homeostatic process, shed light on some of
these questions.
Emotion theories

Before addressing any issue in affective neuroscience, we must first deal with the
problem which leads to essentially all problems in affective neuroscience: the definition of
relevant terms. For instance, what is an emotion, and what is a feeling? This seemingly simple
question has many different elements to it and there is unfortunately no widespread agreement on
the answers (Adolphs et al., 2019 2019). The chief elements debated in affective neuroscience
5

are whether emotions are innate or learned, categorical or dimensional, and conscious or
unconscious. Generally, available answers fall along a spectrum of positions, but two major
theories anchor each end: constructionist theories and basic emotion theories.
For constructionist theorists, emotions are learned, dimensional instead of categorical,
and conscious (Barrett, 2006a). On the other side of the spectrum is what is historically referred
to as the ‘basic’ emotion camp, which posits that emotions are innate, categorical, and have
strong unconscious components. The ‘basic’ emotion camp has not coalesced as strongly in
terms of having shared positions on the extent of innateness, specific categories, and the extent
of unconscious processing in emotions as the constructionist camp (A. Celeghin et al., 2017
Viola, & Tamietto, 2017). However, this has led to some interesting developments where basic
emotion theorists have begun to carve out different aspects of emotional experience as operating
differently- though their definitions are not yet consistent across researchers (Adolphs, 2016).
Largely, modern iterations of basic emotion theories are centered more around the proposal of
innate functional circuitries, which correspond to distinct types of affect which can be viewed as
universal, than identifying specific “natural kind” categories (Adolphs & Andler, 2018;
Nummenmaa & Saarimäki, 2019). The core criticism from this theoretical arm towards
constructionism is that constructionist conclusions are an act of throwing out the innate
emotional baby with the bathwater of variability. The suggestion that any innate universal
categories are ruled out by variability is a strong claim, and possibly a fully unfalsifiable one
(Majeed, 2022; Scarantino, 2012; Scarantino & Griffiths, 2011) .
Despite the vast differences between these theories, both use the concept of valence: the
sense of positivity or negativity inherent in affect. Both hold that valence is an irreducible part of
our conscious experience of affect (Barrett, 2006b; Carruthers & Veillet, 2017; Panksepp, 1992;
6

Panksepp, 2003). This increases the importance of understanding the biological mechanisms that
underlie this concept.

Homeostasis as valence: lessons from nerve-less organisms
As will be seen in this review, there are an abundance of ways valence can be considered,
even when constrained by requiring a link to biological processes. The closest attempt to a
general grouping of what valence is may be in the link to homeostasis. Simply put, positive
valence is anything that relates to promoting an organism’s well-being, while negative valence is
anything that relates to harming it (Carvalho & Damasio, 2021; Damasio, 2018; Lyon &
Kuchling, 2021). This is a definition of valence that would successfully group an array of
different behaviors and sensations across levels of biology and psychology(Berridge, 2019). Not
only can it be applied to pure approach and avoidance behavior, it can be applied to non-neural
organisms (Ginsburg & Jablonka, 2021; Lyon, 2015; Lyon & Kuchling, 2021).
Bacteria and other single-celled organisms monitor their environments for life promoting,
and harmful, stimuli. Much like how affect would work in a neural organism, these molecular
pathways allow the organisms to behave optimally in response to the positive (Larsen & Green)
or negative (Samson et al., 2016) conditions of the environment. An example used by Lyon &
Kuchling in their recent article is the pH sensing in Bacillus subtillis (Lyon & Kuchling, 2021).
These bacteria must navigate themselves into environments that are neither too acidic nor too
alkaline. To do this, they actually have separate molecular pathways for sensing acidity and
alkaline environments (Tohidifar et al., 2020 & Rao, 2020). To balance between them and
successfully navigate to the most neutral environment possible, the bacteria becomes more
sensitive to alkalinity when in an acidic environment, and vice versa. Instead of a unidimensional
mechanisms for pH sensing, the bacteria uses two separate ones which flexibly compete with
7

each other. Bacillus subtillis is not the only organism that uses this method, which has been
referred to by other scientists as a “push-pull” mechanism (Bi & Sourjik, 2018; Yang & Sourjik,
2012). Remarkably, these systems can also have push-pull effects with systems in other domains.
In Escherichia coli, chemical sensing pathways can affect the sensitivity of temperature sensing
pathways, and vice versa, allowing navigation when two environment factors are in conflict with
each other in regards to the organism’s needs (Hu & Tu, 2014; Paulick et al., 2017).
Of course, none of this behavior is conscious; there is no true feeling of positivity or
negativity in bacteria. Yet the organizational trends of these processes at lower levels appear to
also exist at higher levels. And why wouldn’t they? These systems of organization offered
evolutionary success for these organisms. Nervous systems simply allow organisms to better
capitalize on the responsiveness, integration, and flexibility that the chemical systems which
preceded them were using (Damasio, 2018). For example, the independent sensing, but
competing mechanisms, of chemotaxis in many of these organisms is remarkably similar to the
organization of competing emotional action programs in the brainstem: separate pathways for
carrying out the behavior, but which reciprocally inhibit each other (Berridge & Grill, 1983;
Vaccaro et al., 2020). Secondly, even without a nervous system, some level of integration of
different information appears to be present. Finally, what external stimuli are relatively
“positive” and “negative” to the organism is ever-changing. Valence is not some intrinsic
chemical property or specific receptor that makes the sensing of these stimuli behaviorally
positive or negative, but instead the dynamic state of the organisms—in this case its integrated
homeostastic state—determine positivity and negativity. This flexibility is necessary, because
agents rarely if ever encounter just a singular pair of positive and negative stimuli in isolation:
there will always inherently be conflict and adjusting. This interestingly shows that even at this
8

so called “primitive” level of biology, valenced behavior is arising due to context and need
dependent framings of stimuli, rather than a permanently defined quality of a stimulus or
receptor. This system also appears to prioritize summating the affective information of all
potential homeostatic needs and harms to “decide” on one behavior, efficiently resolving
scenarios with mixed motivations. This seems to suggest a preference in living systems for an
ultimate bipolar coding of homeostatic information, as the cell must either approach or avoid.

Coding of valence by neuron populations
Within the nervous systems, researchers have investigated how populations of neurons
encode valence. The majority of these studies take place in rodent models, and focus on affective
regions of interest, such as the amygdala and nucleus accumbens (NAcc) (Berridge, 2019;
Pignatelli & Beyeler, 2019; Reynolds & Berridge, 2002; Tye, 2018). Much research has focused
on identifying the spatial organization of valences which encode positive and negative valence.
There has been some progress in finding pathways and populations of these nuclei which
correspond to positive versus negative behavior. In certain regions (such as the amygdala, NAcc,
and ACC) small clusters of neurons, and sometimes individual neurons, appear to specifically
fire either in response to positive or negative stimuli (Monosov, 2017; Paton, Belova, Morrison,
& Salzman, 2006; Ray, Moaddab, & McDannald, 2021; Xiu et al., 2014). However, finding
larger clusters of neurons with valence-specific responses, such as sub-nuclei or consistently
identifiable regions of neurons, has been more elusive. Much like how cognitive neuroimaging
has progressed from searching for thinking of functions as constrained to singular regions, it may
be that valence should not be thought to exist in specific spatial populations. Evidence suggests
that neurons which respond selectively to positive or negative valence are often interspersed
9

together, along with neurons which appear more flexible in how they responds to valence
(Caracheo, Grewal, & Seamans, 2018; Namburi, Al-Hasani, Calhoon, Bruchas, & Tye, 2016).
With neurons interspersed in this manner, the question becomes how do these neurons
work together to coordinate affective valenced behavior? Tye identifies four major patterns for
how individual neurons come together to implement valence: labeled lines, divergent paths,
opposing components, and neuromodulatory gain (Tye, 2018). Labeled lines is the simplest, and
invokes the thinking of modules permanently dedicated to one cognitive function (K. C.
Berridge, 2019). Simply put, in cases of labelled lines there are separate pathways of neurons for
positive valence and for negative valence. This separation means that minimal computational
effort is used in determining valence, and in how the sensory experience of valence influences
subsequent behavior. Labelled line pathways are frequently found within sensory systems such
as the olfactory and gustatory systems system (Root, Mejias-Aponte, Qi, & Morales, 2014; Tye,
2018; Wang et al., 2018). Many viewpoints of the amygdala’s layout are suggestive of a
labelled-line system, especially when these connections are related to the quick processing of
potentially valenced sensory information or to conditional learning (Namburi et al., 2015; Pessoa
& Adolphs, 2010; Wang et al., 2018). A labelled-line system does not even require initial spatial
separation. A very recent study found distinct neurons in the basolateral amygdala that innately
responded to positive and negative valenced stimuli, and would continue to demonstrate this
activity to newly conditioned appetitive and aversive stimuli (Xian Zhang et al., 2021). These
neurons could be told apart by different genetic markers, and projected to distinct areas of the
striatum demonstrating separate positive and negative valenced pathways to facilitate behaviors.
It is pointed out that labelled line pathways when operating alone do not provide much
flexibility (Tye, 2018). Interestingly, Tye brings up the homeostatic example of salt potentially
10

being positively or negatively valenced depending on the state of the organism. Similarly to the
single cell organisms, we are seeing a need for these responses to be shaped by flexible
“attitudes” towards external stimuli, rather than being dictated by a stimulus that always leads to
the same phenomenal quality when perceived. It is also possible that even the previously
reported pathways have more heterogeneous in their responses than thought, as the study did find
some evidence that certain genetic factors may explain the degree of specialization for positive
vs. negative valence (Ju & Beyeler, 2021), and even the predominantly positively valenced
pathway could contain subpopulations, that if activated, respond to negative valence and lead to
the reduction of positive behavior (Shen et al., 2019). Other types of pathways are needed in
conjunction in order to afford an organism this more diverse repertoire of responses.
Divergent paths begin with a computation of whether a sensory stimulus is positive or
negative, then the result of this computation sends information to an either positively or
negatively valenced pathway. The importance of this neural setup is primarily to allow initially
neutral stimuli to be flexibly learned as either positive or negative (Tye, 2018). For pathways to
be considered divergent paths, neurons must then project to separate pathways depending on
whether the stimulus is learned as positive or negative. While in labelled-line approaches all
neurons have distinct valences they respond to, in divergent paths valence is more related to the
projections than to the sensing neurons. Studies on associative learning in the basolateral
amygdala show evidence of this motif, with projections to the nucleus accumbens supporting
learned positive valence, and projections to the central amygdala supporting learned negative
valence (Beyeler et al., 2018; Namburi et al., 2015; Tye et al., 2011), The prefrontal cortex may
also exhibit pathways like these, with projections to the nucleus accummbens being for reward
while pathways to the periaqueductal grey facilitate avoidance (Britt et al., 2012; Otis et al.,
11

2017; Rozeske et al., 2018; Vander Weele et al., 2018). The mechanisms by which affective
information is diverted to either positive or negative conditioning is unclear, but single neurons
in the orbitofrontal cortex have been seen to develop conditioned responses to both rewarding
and aversive stimuli, suggesting some of these computations could potentially happen at even the
single cell level (Morrison & Salzman, 2009). The opposing components motif is, as Tye says,
highly complimentary to the divergent pathways (Tye, 2018). In this motif, both positive and
negative inputs project to the same target where the different signals can be integrated and
weighted. Essentially, it is the reverse order of the divergent pathways. The main example given
are projections from the hypothalamus. Many pathways originating from the hypothalamus are
seen to have neurons primarily responsible for positively valenced behavior, and negatively
valenced behavior (Kempadoo et al., 2013; Nieh et al., 2016; Olds & Milner, 2020). When these
reach their target, they provide information about the weighting of positive and negative
homeostatic factors simultaneously: this allows these downstream targets to implement both
valences when processing what behaviors to enact.
The neuromodulatory motif is the most flexible, allowing entire populations of neurons to
shift their valenced functions depending on transmitter and receptor related changes. This motif
may be compatible with, and best explained by, recent thinking by Berridge on how neuronal
activity is related to valence. Traditionally, we have thought of the relationship between valence
and neurons in a static spatial manner- certain neurons or pathways are devoted to either positive
or negative valence in a permanent manner. An alternative to this approach would be what
Berridge calls an “affective modes” hypothesis (K. C. Berridge, 2019). In this approach, different
populations of neurons may respond to positive, negative, or some combination of both valences
depending on different conditions and contexts. The changes in function induced by different
12

neuromodulatory reactions. These neuromodulatory shifts also appear to be affected by the
external environment. In one study showing this effect, they looked at how different
environments affected the valenced responses of injecting glutamate receptor antagonists at
different sites of the medial shell of the nucleus accumbens (Reynolds & Berridge, 2008). When
the rats were in a quiet dark environment, the majority of injections sites lead to an increase in
appetitive behavior. When in a standard lab environment, some of these injections sites started to
instead lead to an increase in defensive behaviors, and some lead to rats displaying both
appetitive eating behaviors, and behaviors which indicate fear. In a purposefully bright and noisy
environment, the proportion of injection sites leading to defensive or ambivalent behaviors again
increased. The flexibility of affective modules to respond in context dependent manners may be
a particularly efficient way to ready appropriate behaviors. It is interesting to note in this
example that sites increasing both appetitive eating and defensive fearful behaviors could also be
interpreted as valence not being the defining factor of how neurons facilitate behavior- eating
and defensive behaviors are not united on one spectrum just because of their opposite valences,
showing that multiple behavior relevant valences are being encoded. Berridge points out that
how environments change the affective modes of these sites is unclear, but likely has to do with
larger systems level processing (Richard & Berridge, 2011). Larger scale brain activity may have
downstream effects on the smaller scale, encoding the affective context and shifting how
subpopulations behavior in response to stimuli.

Neuroimaging and systems level valence processing
A few meta-analyses have attempted to describe valence at the systems level using fMRI.
The main questions of these analyses have been 1) how is valence represented? And 2) Is
valence at the systems level bipolar, independent, or something else altogether?
13

The importance of this question is highly relevant for neuroimaging methodology, as the
choice of how to model valence effects results. This point was made in a recent study by Mattek
and colleagues (2020). In the study, they used valenced stimuli from different domains to test
two possible models of valence on the data: one model where positive and negative valence are
one linear dimensions –with arousal as a second dimension consistent with constructionist views-
and another where positivity and negativity are separate. The study found that the areas which
were significant for positive and negative valence when valence was modelled as two
independent scales (accomplished by contrasting positivity and ambiguous valence with negative
and neutral, and negative and ambiguous valence with positive and neutral) were not the same as
when valence was modelled as one bipolar scale (Mattek, Burr, Shin, Whicker, & Kim, 2020). In
fact, only one area was found to be significant for bipolar valence: superior parietal lobule. When
valence was modelled as bipolar, multiple regions were found significant for arousal. However,
the vast majority of these regions were found to be significant for either positivity or negativity
in the independent dimension analysis. This study demonstrates a dilemma in affective
neuroscience: our analyses may not even show the same regions being related to valence if we
use different models. Neuroimaging in this way may be particularly susceptible to the effects of
our a priori models of emotion.
The study by Mattek and colleagues is one experiment; but what happens if we look at a
large-scale meta-analysis of valence studies in an attempt to pick a best model? A 2015 meta-
analysis by Lindquist and colleagues aimed to gather as many fMRI studies on valence as
possible in order to test different models of valence. The authors gathered 397 fMRI and positron
emission topography studies (containing 6827 participants) (Lindquist, Satpute, Wager, Weber,
& Barrett, 2015). Three models of how valence may be represented in the cortex were assessed:
14

the first was bipolar valence. For the purpose of testing, bipolar valence was to be defined as
regions which responded in a consistent fashion of positivity>neutral>negativity or the reverse,
while excluding regions which responded equally strong to positivity and negativity. The authors
then, due to not being able to find contrasts in the literature of neutral>positivity or
neutral>negativity, and not wanting to use deactivation results, redefined bipolarity in their
model as any voxel that showed more activity during positivity>negativity than
positivity>neutral, or the reverse for negativity. This analysis found clusters in the ACC and
mPFC which responded more frequently to positivity>negativity than positivity>neutral, but no
clusters for more frequent negativity>positivity than negativity>neutral. The authors took this as
limited evidence of bipolarity (more on that later). The second model tested was independent
regions which respond to positivity or negativity exclusively: the bivalent hypothesis. This was
defined a voxels which respond exclusively to either positivity or negativity, so once again
valence general voxels were excluded. The authors state that in order to be more conservative
with this analysis, they added the additional criteria that for a region to support independent
positivity, for example, it must also be more likely to be significant during positivity>negativity
than positivity>neutral. No such region was found. Finally, they tested if there were valence
general regions which appeared more often during either positive>neutral contrasts,
negative>neutral contrast, or positive>negative contrasts: they defined this as a valence general
affective workspace model. The amygdala and insula were more likely to respond to negativity
than positivity, and no valence general regions were found to respond more often to positivity.
The authors of the study overall state they found a little evidence for bipolarity of valence
in ACC and mPFC, no evidence for independent positivity and negativity, and plenty of evidence
for a valence general workspace. However, there are methodological issues with this study. At
15

the surface level, it is questionable whether it is appropriate to apply a more conservative
approach to modelling bivalence after having used a more lenient model of bipolarity than was
originally intended. Secondly, merging this many studies without a more in-depth analysis of the
specific modalities of valence may not be appropriate. Current research suggests that valence is
represented both in domain-specific and domain-general ways- as previously mentioned valence
in the cortex appears to be heterogeneous (Gao & Shinkareva, 2021; Miskovic & Anderson,
2018; Shinkareva et al., 2014; Vaessen, van de Heijden, & de Gelder, 2019). Without knowing
the proportions of different modalities in this meta-analysis, we cannot interpret the consistencies
or lack of consistencies across studies. Most importantly, we must address the logical framework
of how evidence of valence representation was treated in this study. The methods used rely on
spatial separation to understand valence. However, this is a methodologically dated view of the
representation of affect in the brain, in which non-constructionism views are erroneously implied
to be solely locationist. (Loaiza, 2021; Scherer, 2012). Neither the bipolar nor the bivalent
hypotheses in this study were afforded what the affective workspace was allowed implicitly: the
ability to consider relating valence to functional patterns of the brain rather than permanent
functions of neuronal tissue. This once again brings us back to Berridge’s point at the neuronal
level of valence being flexibly implemented rather than tied to specific permanently devoted
cells. The authors actually did do one post-hoc test for a multivariate functional separation of
valence: a whole-brain support vector model to try and distinguish positive and negative valence;
the results were not above chance (Lindquist et al., 2015), although this analysis probably could
have been improved with feature-selection (only using the most informative voxels rather than
using every voxel as a variable), and was limited by the nature of only using voxels that had been
significant in univariate analyses. Valence is likely represented on a full-scale cortical level
16

rather than in a specific region if it is a singular represented construct (L. F. Barrett & Bliss-
Moreau, 2009; Carruthers, 2018; Wager et al., 2008). The only possible counter-argument to this
would be some findings that the orbitofrontal cortex appears to integrate valenced information
from multiple sources to represent “overall” valence (FitzGerald, Seymour, & Dolan, 2009; Levy
& Glimcher, 2012). However, I think this may be better interpreted as the orbitofrontal cortex’s
important role in utilizing affect to aid decision-making- it is recruiting all the goal-relevant
information it needs to enact the optimal affective process (Bechara, Damasio, & Damasio, 2000;
A. R. Damasio, 1996).

Multivariate pattern analysis shows domain specific valence
The advent of multivariate pattern analysis (MVPA) has allowed us to tie mental states to
patterns of activity rather than static consistent regions that must be tied to a function (Haxby,
Connolly, & Guntupalli, 2014). This has been an especially big development in the study of
affect, allowing us to distinguish between affective states in ways univariate methods had failed
to (J. Kim et al., 2015; Kragel & LaBar, 2015; Saarimaki et al., 2018). What happens if we take
these new methodologies and apply them to representations of valence? More recent MVPA
studies have been able to successfully classify positive and negative valence. The regions
implicated in making these classifications of positive and negative valence possible appear to be
the same ones traditionally implicated in univariate studies of affect, such as the amygdala,
orbitofrontal cortex, medial prefrontal cortex, and precuneus (Bush, Privratsky, Gardner,
Zielinski, & Kilts, 2018; Habes et al., 2013; Jongwan Kim, Shinkareva, & Wedell, 2017;
Jongwan Kim et al., 2020). In most of these studies, the best classification accuracy results from
the patterning of activity across multiple regions, rather than just in one. Whether these exact
patterns represent something universal about valence is unclear: research has suggested that the
17

specific patterns of activity yielded by MVPA are specific to the type of stimulus used to induce
them (Bush, Gardner, et al., 2018; Shinkareva et al., 2014). In an example of this problem,
researchers took a model derived from classifying how negative people felt pictures were and
applied it to a dataset of individuals experiencing thermal pain (Chang, Gianaros, Manuck,
Krishnan, & Wager, 2015). While in the original study the model was successful in predicting
how highly negative individuals rated the picture, it failed to classify how intensely negative
individuals rated their pain: it could only classify whether a stimulus was painful or a negative
picture. Furthermore, a tested model that was successful at predicting the how negative pain was
rated was not successful at classifying negative picture ratings. These results demonstrate that
these patterns are not evidence of a shared system for valence across stimuli and domains – they
may instead be capturing something specific to the distinctiveness of these different negative
experiences. Cross-modality work may help bridge this gap to understand if at least some
representations of valence are general – and if positive and negative feelings share features
across modalities. Work looking to classify valence across audio and visual domains has shown
success (Gao & Shinkareva, 2021; Jongwan Kim et al., 2017; Shinkareva, Gao, & Wedell,
2020).

Valence may not be the principle organizer of affect in the cortex
It is also possible that valence is not the defining dimension of how affective states are
organized in the cortex (Kragel & LaBar, 2016). Kragel and LeBar point out in their review
paper that while the previously discussed meta-analysis of valence (Lindquist et al., 2015) failed
to distinguish positive and negative valence using MVPA on their dataset, a large meta-analysis
classifying emotion categories (fear, anger, disgust, sadness, and happiness) which included 148
studies and 2159 participants had been successful with an average accuracy of 66% (Wager et
18

al., 2015). The combination of these meta-analyses, according to Kragel and LeBar, may suggest
valence is not a defining dimension of how affect is organized in the cortex. The authors of the
meta-analysis would rebut this, as they interpreted their results as supportive of a constructionist
view of emotion, because emotion categories were predicted by patterns across regions instead of
specific regions. This, however, is again conflating a locationist view as a requirement for ‘basic’
emotion categories.
These meta-analyses do not mean that valence means nothing to the organization of
affective states in the cortex: even if categories are more consistent, valence may be an
applicable dimension to those categories. In a study by Saaramaki and colleagues, 16 different
affective states were classified using fMRI data (2018). When a classifier was trained to
distinguish all 16 apart, 14 were successfully classifiable above chance (longing and shame were
not) (Saarimaki et al., 2018). The authors also decided to test whether the neural patterns
underlying these states were more similar if they were subjectively perceived to be more similar.
Participants were asked to rate how similar each possible pair of the 16 emotions were, and the
extent of similarity was compared to neural pattern similarity. Neurally, classification found 4
clusters of similar emotions as follows:
1. Love, Gratitude, Happiness, Pride, Longing; 2) Surprise, Neutral; 3) Sadness, Disgust,
Fear, Shame; 4) Despair, Guilt, Contempt, Anger.
Experientially, the 4 clusters were:
1. Love, Gratitude, Happiness, Pride; 2) Surprise, Neutral; 3) Sadness, Longing; 4) Fear,
Despair, Shame, Guilt, Anger, Contempt, Disgust
It is clear that both the neural clusters, and experiential clusters, appear to be organized
somewhat by valence. Interestingly, longing, which some consider a mixed feeling (Farrell,
19

2006; Lomas, 2017; Vaccaro et al., 2020), was classified with positive feelings neurally but
negative feelings experientially. This analysis overall shows that valence likely still plays a role
in the cortical differences between emotions, but it is not the only relevant dimensions.

Considering mixed feelings in light of the biology of valence
Surprisingly, the construct that has the most agreement as at least being important to
affect (Dukes et al., 2021) – valence – may be the source of the biggest issues in resolving mixed
feelings (J. T. Larsen, 2017). Overall evidence demonstrates that valence may be implemented
differently on different scales of biology. On the single-cellular organism level, we see that
valence can be defined in terms of promoting life or potentially harming it, and the need to
translate potentially conflicting sensory stimuli into one action choice. While these “push-pull”
mechanisms are computing potentially positive and negative stimuli concurrently, they defer to
an ultimate bipolar decision to approach or avoid. As we move up the biological hierarchy, the
complexity of these representations for valenced stimuli increases. The degree of independence
between positivity and negativity also increases with more biological complexity at the neuronal
level – and with this, it becomes less reasonable to think of valence as entirely bipolar. At the
level of neuroimaging, this is even more so the case and the degree of heterogeneity in valence
systems appears to relegate valence to becoming only part of affect, rather than being its core
differentiating feature.
What appears to be shared for valence across all these scales is the flexibility of
representation- the lack of a one-to-one relationship between a stimulus, encoding cell, or
pathway for positive or negative valence. Biological features that correspond to valence appear
to be strongly affected by context rather than having permanently devoted relationships to
valence; this is true at all levels. Sensory signals in single-cell organisms are not inherently
20

coded as relating to positive or negative valence, and instead translate into valenced action
depending on the homeostatic state of the organism. On the neuronal level, many pathways
flexibly encode either positive or negative valence depending on neuromodulatory changes –
with these modulations depending on the organism’s environmental context. On the
neuroimaging level, we see valence flexibly encoded in distributed brain activity, and often times
in domain-specific manners.
Perhaps the most puzzling consideration here, and why valence has not received much
scrutiny in the literature as a concept, is the reality that the most accessible measure of feeling is
not as clear of a concept as we may have hoped. It would be methodologically convenient if
valence was tied to a specific stimulus, neurotransmitter, cell group, or neural pathway. A
domain general representation of valence would be the most intimate measurement of affect we
could get. We do not have this, and this raises the question as to what relating valence to biology
is actually telling us. It may be time for us to displace valence as a universal driver of affect, and
move on to considering it as a much looser and variable concept (Charland, 2005; Harmon-
Jones, 2019). The solution may be to think of valence specifically in the context it is being
measured in – its domain specificity, whether its towards a behavior, decision, perceptual
judgment, or subjective experience, and the timescale it is measured across. Valence can still be,
and is still, important, but it may not be the infallible consistent center-piece of measuring
emotion. Thus, an understanding of mixed feelings can benefit from dividing levels of the
biological hierarchy and considering the different processes occurring at each separately.

21

1b. Bittersweet: The Neuroscience of Ambivalent Affect

Anthony G. Vaccaro, Jonas T. Kaplan, and Antonio Damasio

Published in Perspectives on Psychological Science

Expanding the study of affect to ambivalent states

Think back to when you graduated from school. There was probably a sense of
excitement for finally reaching that point, perhaps an enthusiasm for wherever you would be
heading next. But at this same moment and in moments leading up to it, there may well have
been a reflective feeling, lining the edges of this enthusiasm with an uncertainty. Things would
never be the same – the ordinary moments of your current life were about to change. Although
you wanted to graduate and seemed to be moving on to something you wished for, there was also
a shadow around the places, things, and people you were leaving, and were fond of. There was
joy and laughter but also enough sadness to bring some tears. You were feeling bittersweet.
Most research on affective experience addresses states of positive or negative valence,
such as fear, joy, sadness, disgust, or anger (A. Celeghin et al., 2017; Critchley & Garfinkel,
2017; Croy et al., 2011; Ekman, 1992; Russell, 2003). Yet human affect also includes less
frequent states which appear to have both positive and negative valences. These ambivalent
states remain understudied and the mechanisms behind their appearance have not been
elucidated.
Examples of ambivalent states include sentiments such as bitter-sweetness and longing
(also referred to by the German term Sehnsucht) and nostalgia. The feeling that occurs during
cue-induced craving in addiction has also been viewed by some as a mixed positive and negative
state (Cartwright & Stritzke, 2008; Veilleux et al., 2013). Such feelings may be described and
22

experienced somewhat differently across individuals, cultures and contexts, but the fundamental
condition of feeling simultaneously positive and negative appears to be universal (Lomas, 2017;
Miyamoto et al., 2010; Rafaeli et al., 2007; Scheibe et al., 2011). Interestingly, ambivalent
affective states appear to have less clear neural correlates than their univalent counterparts, and
experiments in psychology have yielded varied and conflicting interpretations as to how they
may be constructed. This should not be surprising given that the most influential theoretical
frameworks in affective neuroscience have focused largely on univalenced states. As a
consequence, we do not know whether ambivalent states result from a rapid sequence of two
different emotions or from the actual mixture of the two (Kreibig & Gross, 2017; Larsen, 2017;
Norris et al., 2010; Young, 1918). Here we rely on current work in psychology and neuroscience,
to discuss the potential psychological and neural foundations of ambivalent affective states and
consider how they may be investigated.

Emotion, feeling, and valence in the homeostatic view of affect
Before we proceed with our presentation, it is important to establish the meaning of three
terms within our framework: emotion, valence, and feeling. Definitions of emotion differ
according to whether they are viewed, for example, as innate vs. acquired, as states described as
categorical or in terms of dimension, and according to whether they necessarily include the
conscious experience associated with them (Adolphs, 2016; Barrett, 2017; LeDoux & Brown,
2017; Vandekerckhove & Panksepp, 2011). In our view, an emotion is a set of physiological
changes produced by a largely innate “action program” (Damasio, 2018; Damasio & Carvalho,
2013). The changes are triggered by stimuli, simple as well as complex, that directly or indirectly
threaten or strengthen homeostasis, and thus threaten or favor the integrity of the organism. The
23

physiological changes are largely carried out by subcortical structures, which act to alter the
chemical and visceral parameters of the interior. In our perspective emotions do not include the
conscious experience which co-occurs and follows them. For conscious experience
accompanying emotion we prefer the specific term emotional feelings, which are supported by
processes of interoception, and also require cortical processes.
All affective states can be classified in terms of valence, which corresponds to positive or
negative affective valuation of any experience. Emotions can be said to have valence in the sense
that they relate to events that take the organism toward or away from homeostasis, but valence
truly blossoms in the realm of feeling. Conscious experiences of emotive states tend to be clearly
evaluated as positive or negative. Feelings elicited by a specific emotion tend to have a specific
valence.

Ambivalent affect may result from both vacillation and mixing but at different
physiological levels
The main question concerning ambivalent affect concerns whether it is best viewed as a
genuinely mixed, simultaneously positive and negative experience, or rather as due to a rapid
vacillation between two distinctively valenced processes that occur so rapidly that the distinct
components cannot be experiences separately (Larsen, 2017; Young, 1918). One way to address
this question is to consider the construct of valence, and specifically how positive and negative
valence relate to each other. If valence is a unidimensional construct, where an affect’s valence
lies somewhere on a single positive-negative spectrum, ambivalent affect must be a rapid
vacillation as the two ends cannot co-exist (Barrett & Russell, 1998; Larsen, 2017; Russell &
Carroll, 1999). If positive and negative valence instead exist independently as separate
24

constructs, ambivalent feelings can involve feeling positivity and negativity concurrently
(Cacioppo & Berntson, 1994; Larsen & McGraw, 2014; Schimmack, 2001). Psychological
research on ambivalent affect has used the valence issue as a way of investigating simultaneity
versus vacillation. Studies using post-experience rating scales as well as in-the-moment button
presses to signify the presence of positive and negative feelings, support the position that
subjects do experience these sentiments simultaneously in certain kinds of situations (Larsen &
Green, 2013; Larsen & McGraw, 2011; Larsen et al., 2001; Larsen et al., 2004; Moeller et al.,
2018a; Schimmack, 2007; Schneider & Schwarz, 2017). Several psychometric studies have
suggested that separate positivity and negativity dimensions have stronger explanatory power
than a unidimensional model of valence, and also that positivity and negativity are not inversely
correlated with each other as they would have to be as part of a unidimensional construct (An et
al., 2017; Briesemeister et al., 2012; Cacioppo & Berntson, 1994; Colombetti, 2005; Moeller et
al., 2018a). Nonetheless, some have contradicted this position by pointing out that the rapid
alternation between opposite valences may be too fast to be subjectively detected. If that is the
case, self-report would reflect both emotional feelings which occurred during the narrow
timeframe, and accounting for this rapid vacillation can lead to similar predicted data for a
unidimensional model of valence (Barrett & Bliss-Moreau, 2009; Russell, 2017; Russell &
Carroll, 1999). The idea that simultaneous contradicting feelings is an illusion of consciousness
has been around since the early 1900’s, when early researchers concluded that while people
certainly believed they were experiencing mixed feelings, the nature of affect meant that the
feelings would not be simultaneous (Johnston, 1908; Young, 1918).
Research on the neural substrates of valence provides evidence that positive and negative
valence are physiologically separate states. Several studies using neuroimaging data suggest that
25

rather than a single locus of changing activity, both independent regions and overlapping regions
of the brain are responsible for the representation of positivity and negativity (Berridge, 2019;
Lane et al., 1997; Lindquist et al., 2015). While unidimensional models can account for much
variance in the data, they are not as satisfactory when compared to models which separate
positive and negative constructs (Colibazzi et al., 2010; Lewis et al., 2006; Viinikainen et al.,
2010; Viinikainen et al., 2012).
An important possibility this debate has failed to consider is that both concurrence and
vacillation may be involved albeit at different physiological levels. A complete affective
experience comprises multiple levels of processing related to different systems of the brain
hierarchy (Berridge, 2019; Immordino-Yang, 2010; Man et al., 2017; Norris et al., 2010; Tye,
2018). In our perspective, it is critical to distinguish between the largely subcortical level where
emotional action programs are triggered, and the cortical level, on which the experience of
emotional feelings largely depend. These distinct yet integrated components of the affective
process may contribute different experiential contents concerning the same situation. While the
neural substrates for feelings would accommodate bivalent representations, the neural substrates
for emotions would not. Ambivalent affect may not be reducible to the simple alternatives of
vacillation or simultaneity. Instead, vacillation of opposite action programs and simultaneity of
experience might co-exist albeit at different hierarchical levels of neural structure, and affective
process, respectively the levels of emotion and feeling.

26

Drawing on child development data to describe the cognitive processes behind ambivalent
affect
Stages of child development provide a window into the discrete functions necessary for
ambivalent affect. At 4-5 years of age children generally report experiencing only one feeling at
a time relative to a situation (Larsen et al., 2007; Smith et al., 2015). Four to five year olds also
show limited ability to perceive ambivalent affect in situational stimuli, and cannot be trained to
do so (Holm et al., 2002; Kestenbaum & Gelman, 1995; Peng et al., 1992; Wintre & Vallance,
1994). By ~7-8 years of age children report feeling opposite valences simultaneously; this is
shown through self-report and through tasks that require the prediction of feelings which may
occur in different hypothetical situations (Harter & Buddin, 1987; Wintre & Vallance, 1994).
Children do not begin to report a subjective sense of conflict concerning these opposing feelings
until ~10-11 years old (Choe et al., 2005; Donaldson & Westerman, 1986; Whitesell & Harter,
1989; Zajdel et al., 2013). This trajectory aligns with the development of affective processes,
during which subcortical-prefrontal and prefrontal-associative cortical pathways develop to
support more complex utilization, representation, and control of affect and emotion (Casey et al.,
2019; Goldsmith & Davidson, 2004; Tottenham, 2014). The frontal lobe transitions from
operating in a localized manner to operating within distributed functional networks such as the
broader central executive and salience networks (Fair et al., 2009). Moreover, there is evidence
that the insula’s effective connectivity to these networks increases with maturation (Uddin et al.,
2011).
These developmental facts support our hypothesis that ambivalent affect is a complex
phenomenon which requires multiple levels of processing, where different types of information
of potentially different valence are produced and eventually integrated (Filippetti et al., 2019;
27

Fischer & Bidell, 2007; Immordino-Yang, 2010; Kuhlmann et al., 2016). Moreover, the fact that
ambivalent and univalent feelings emerge at different stages of development is consistent with
the idea that simultaneity and vacillation of valences would depend on different neural
substrates.

Reciprocal inhibition in brainstem nuclei prevents simultaneously opposite emotions while
also allowing rapid switches between them
In our perspective emotions arise in part from distinct action programs largely
coordinated by brainstem nuclei, in what has been characterized as the ‘lateral emotional motor
system’ (Bandler & Shipley, 1994; Denton et al., 2009; Venkatraman et al., 2017). Through
descending pathways of the brainstem, distinct emotive action programs are executed in response
to homeostatic needs (Ekman et al., 1983; Panksepp, 1998; Rainville et al., 2006; Venkatraman
et al., 2017). Action programs can be adjusted when the influence of the cerebral cortex as a
consequence, for example, of interoceptive feedback, cognitive regulation, or endocrine changes
from the medial emotional motor system. These action programs are not identical with their
corresponding conscious feelings, and in fact they may even be misinterpreted in introspection
(Brewer et al., 2016; H. D. Critchley, 2009; Critchley et al., 2013; Davidson, 1993; Jung et al.,
2017; Venkatraman et al., 2017). Emotive action programs are responses to something specific,
calling for a primarily consistent and coordinated response across various bodily systems. While
the exact patterns of autonomic response may be hard to categorize universally,
psychophysiological research has revealed supportive evidence that the autonomic nervous
system produces distinguishable patterns for distinct emotional states across individuals, genders,
28

and cultures, though more work is needed on this topic (Levenson et al., 1990; Levenson et al.,
1992; Scherer & Wallbott, 1994; Tsai et al., 2002).
A fact of major relevance for a discussion of ambivalent affect concerns the neural
substrates for conflicting emotive action programs. These programs are innervated through a
system of reciprocal inhibition. Examples of this are seen for maternal vs. predatory behavior
and fight/flight vs. freeze responses in functional and lesion studies focused on the
periaqueductal gray (Bandler & Shipley, 1994; De Oca et al., 1998; Fanselow, 1994; Sukikara et
al., 2006). Reciprocal inhibition allows for rapid behavioral switches as circumstances change,
and this flexibility is impaired by lesions to the relevant brainstem nuclei (Bandler & Shipley,
1994; Strigo & Craig, 2016; Sukikara et al., 2006). This suggests that opposite emotions at the
level of the brainstem rapidly vacillate rather than mix, just as the fixed-action patterns of rats do
in response to simultaneous palatable and aversive tastes (Berridge, 2019; Berridge & Grill,
1983, 1984).
We acknowledge that there is some uncertainty regarding whether the above findings
truly rule out all possible overlap between basic positive and negative reactions at the level of the
viscera. Changes to physiological states do not instantaneously resolve once the emotional action
program is discontinued, and so it is possible that some of these features may overlap temporally
even if they are not being enacted at the same time (for example increased heart rate resulting
from one emotion could remain, while breathing has already slowed as a result of another
emotion). This could create an emergent distinct physiological profile for a mixed emotion in the
moment (Kreibig et al., 2013). However, we doubt that an emergent unique pattern is a
satisfactory explanation for mixed feelings. Mixes of truly opposite affective states (such as
happiness and sadness) would likely be physiologically neutral in this circumstance, as the
29

switch between opposite action programs would lead to largely cancelled out autonomic
changes.

Interoceptive signals resulting from rapid action program switches produce a unified
feeling moment
The processes discussed so far have concerned ambivalent affect as a sequence of
vacillating events across time. But the stark contrast between ambivalent affect and a typical
instance of one affective state following another comes from the sense of simultaneity. The
subjective experience of co-occurring emotions is a defining element of mixed feelings,
regardless of the extent to which they are truly simultaneous. For ambivalence to be experienced,
there must be representations of both emotions at some level of the nervous system, even though
the respective emotive action-programs may not be executed at the same time.
A region well situated to play a role in this process is the anterior insula. The insula
receives information about the homeostatic state of the body from the brainstem nuclei including
the nucleus of the solitary track, and is known to be involved in conscious interoception
(Critchley et al., 2004). According to Craig (2009), the anterior insula assembles experiences of
discrete moments in time to form a series of “global emotional experiences”, each of which
integrates information from somatic states across a period of time (Whittman, 2013). While on
average these moments represent ~125ms of somatic states (Picard & Craig, 2009) the true
length of each individual moment is variable in temporal length, but is limited by the amount of
salient information it gathers. In other words, each discrete “global emotional moment” collects
information from the body and “fills up” when its capacity has been reached. This can make time
appear to go by faster or slower depending on how quickly these moments “fill-up”. This
30

phenomenon is observable in instances of emotional time dilation, where affective experiences
alter the subjective perception of time (Droit-Volet & Gil, 2009; Stetson et al., 2007).
There is additional and more impressive evidence that the insula is important for the
sense of time from the study of seizures with ecstatic auras, which in lesions of the anterior
insula induce euphoria, a feeling of timelessness, and a feeling of heightened or ‘overloaded’
consciousness (Gschwind & Picard, 2016; Picard & Craig, 2009; Picard & Kurth, 2014). Thus
the anterior insula’s buffered representation of states allows for the integration and comparison
of affective information across time. Indeed, insular activity connected with the integration of
rapidly changing emotions appears to be associated with processing ambivalence (Critchley et
al., 2000; Szabó et al., 2017; Wilson-Mendenhall et al., 2014). The emotional reappraisals
involved in ambivalent affect may be in rapid enough succession to be integrated within the
same emotional moment, and thus we experience both emotions as a unique and integrated
emotional feeling.
Further cognitive elaboration of this moment may occur outside the insula. Just as the
posterior insula projects current information about bodily states to the anterior insula for
processing into a summated moment, the anterior insula projects to other brain regions where the
contextual relevance and the valences of these moments can be considered. With ambivalent
affect, afferent pathways to the orbitofrontal cortex (OFC) and anterior cingulate cortex (ACC)
possibly contribute to this process, and the summated moment may be evaluated as a source of
conflicting affective information (Medford & Critchley, 2010; Roy et al., 2012; Rushworth &
Behrens, 2008; Seth et al., 2012; Sutherland et al., 2013; White et al., 2010).

31

Figure 1: Conceptual model of ambivalent affect on cortical and subcortical levels

Figure 1: An initial stimulus triggers not only the first emotion action program, but also triggers
the recall of facts and ideas associated with the stimulus. Recalled material, in turn can trigger a
second emotion of a different valence. As a result of the reciprocal inhibitory innervation in
brainstem nuclei, this results in the inhibition of the initial action program, after which the two
emotions relevant to the situation rapidly alternate. While this plays out as a sequence of events
on the level of emotions/action-programs, at the subjective level the sequence is experienced as a
single integrated feeling.

Ambivalent affect relies on memory and reappraisal processes
It is no coincidence that many ambivalent affective states such as nostalgia, longing, and
Sehnsucht are driven largely by memories and the feelings associated with them (Oba et al.,
2016; Scheibe & Freund, 2008; van Tilburg et al., 2018). Many ambivalent affects such as
nostalgia and longing involve recall of specific memories of positive or negative valence
followed by counterfactual thinking leading to reappraisal of the recalled contents (Epstude &
Roese, 2008; van Tilburg et al., 2018). This is especially likely with the recall of events or
individuals that currently are perceived as psychologically or temporally distant and, in many
instances permanently so (Farrell, 2006; Scheibe et al., 2007). Evidence for the importance of
memory’s active role in the affective experience of nostalgia has been demonstrated in research
32

using fMRI, in which nostalgic images as compared to visually and conceptually similar images
evoked increased and sustained co-activation of the hippocampus and ventral striatum (Oba et
al., 2016). In this perspective, the memory first confers a primarily positive emotion, but a rapid
reappraisal reveals that the topic should also trigger a negative emotion. Once this happens, the
emotions begin to vacillate rapidly due to the presence of both the image of the initial stimuli and
its reappraisal both being cortically present; one image does not ‘delete’ the other, at this level, it
just becomes more prominent in an ebb and flow. Possible evidence of this is seen in a study
aiming to cluster various affective states by their subjective and neural similarities. Using
hierarchical clustering analysis of whole-brain neural patterns and of experiential traits for 15
different categories of affect, it was found that longing was the only category of affect to be
clustered with positive affect neurally, but with negative affect in terms of experience (Saarimaki
et al., 2018). Of note, a multivariate-pattern analysis of this same neural data was unable to
distinguish longing from other affective states. Along with shame, longing was one of only two
states – among 15 – that could not be successfully distinguished from the others at an above
chance level. It is possible that the neural representations of this state may have an uncommonly
high amount of temporal fluctuations due to the ebb and flow of continuous reappraisals.
Accordingly, we could hypothesize that a construct such as regional homogeneity, which
measures the similarity between the time-series of voxels and their nearest neighbors, would be
decreased during ambivalent states. Such evidence is in fact available for bittersweet feelings in
a study of lovelorn individuals. The data reveal decreased regional homogeneity of the anterior
cingulate cortex associated with the length of time spent in love (Song et al., 2015).

33

Conflict is critical to the function of ambivalent affect
The perception of conflict may be a crucial element in the unique experience of
ambivalent affective states, distinguishing them from a simple sum of two simultaneous
experiences (Berrios et al., 2015b; Mejia & Hooker, 2017; Schulte et al., 2012). Insula and OFC
projections to lateral prefrontal and anterior cingulate regions are likely to contribute to the
experience of conflict. A unique sub-region of the OFC, which is known for responding to
affective conflict as compared to either positivity or negativity (Becker et al., 2014; Schulte et
al., 2012; Simmons et al., 2006), has been found to be similarly responsive to bittersweetness
(Schulte et al., 2012) and longing (Saarimaki et al., 2018) (t-map available on NeuroVault). This
suggests that conflict turns these states into more than the positive and negative sums of their
parts. While the OFC is suited for representing multiple features of an affective state, the regions
to which it projects are instead optimized for the more conclusive evaluations involved in
carrying out decisions (Rolls & Grabenhorst, 2008). It is in these regions, the ACC, the
dorsolateral PFC, and the anterior PFC, that upcoming decisions are considered and the degree of
uncertainty surrounding the action is weighed (Hsu et al., 2005; Luttrell et al., 2016; Nohlen et
al., 2014; Stolyarova, 2018; Vaccaro & Fleming, 2018). Activity in the ACC has been shown to
relate to feelings of ambivalence and conflict (Cunningham et al., 2004; Nohlen et al., 2014;
Preckel et al., 2014; Saunders et al., 2017).
The conflict within ambivalent affect allows us to be motivated not only by what is
present but also by representations of what could be (Fong, 2006; Rothman, 2011; Seth et al.,
2012). Due to counterfactual thinking or even to mere recall, we often think of non-current
scenarios about which we may feel differently when they are compared to the current scenario;
this happens often with nostalgia, longing and Sehnsucht. As this occurs, nostalgia, longing, and
34

Sehnsucht lead us to adapt our decision-making in ways that we probably would not if we had
solely experienced negative affect. For example, ambivalent affect has been found to be
beneficial in working towards social, economic, and health-related goals, in which the actions
required for progress may cause negative feelings, but in the long-term will lead to positive
feelings (Fredrickson & Branigan, 2005; Garland et al., 2010; Hershfield et al., 2013; Scheibe &
Freund, 2008). Being able to represent such future positive feelings motivates the individual to
continue with the actions necessary to obtain the goal in spite of their negative feeling. Along
these lines, ambivalent affect can also provide a sense of comfort, promoting both psychological
and physiological resilience in the face of negative mood or the traumatic elements of memories
(Braniecka et al., 2014; Larsen et al., 2003). This effect that has been well-documented for
nostalgia (Matsunaga et al., 2013; Newman et al., 2019; Routledge et al., 2013; Zhou et al.,
2008; Zhou et al., 2012). Interestingly and consistently, nostalgia and Sehnsucht/longing are
elicited primarily during negative moods and as a response to negative emotional states (Kotter-
Grühn et al., 2009; Sedikides et al., 2008; Wildschut et al., 2006). While only limited work has
been done on these topics, the current research on nostalgia supports these interpretations as it
consistently shows links between trait-level susceptibility to negative affect, susceptibility to
nostalgia, and the neurobiological correlates of affect representation and memory (Barrett &
Janata, 2016; Luo et al., 2017; Oba et al., 2016; Trost et al., 2012).
The behavior-expanding nature of ambivalent affect is not always beneficial. In instances
of longing and Sehnsucht in which the desire is definitively unobtainable, the feeling is more
pervasive and associated with lowered perceived well-being (Kotter-Gruhn et al., 2009; Scheibe
et al., 2011). This same effect can also occur in nostalgia when it represents a strong, undesired
change in self-identity or the current state of being (Iyer & Jetten, 2011). Nostalgia itself was
35

considered a serious psychiatric condition in centuries past – a malady bearing someone’s over
attachment to what was once loved, and an inability to adapt to a new environment, eventually
leading to “catatonia” and sometimes death (Rosen, 2009; Roth, 1991). While nostalgia is no
longer classified as a psychiatric disorder, elements of prolonged ambivalent affect may also
relate to varied psychopathologies such as post-traumatic stress disorder, obsessive-compulsion
disorder, depression, and addiction (Bhar & Kyrios, 2007; Jerg-Bretzke et al., 2013; Losada et
al., 2018; Veilleux et al., 2013; Verplanken, 2012). Still, ambivalence is not purely negative in
the clinical setting. There is some evidence that deficits related to affective intelligence, such as
alexithymia and reduced perspective taking, are associated with less awareness of ambivalent
feelings in the self and others for various populations (Albano et al., 2013; Beck et al., 2012;
Bird & Cook, 2013; O'Kearney et al., 2017; Schultz et al., 2005). Furthermore, while non-
dialectical thinkers tend to exhibit ambivalent affect more so during primarily positive life events
than during primarily negative ones, dialectical thinkers exhibit ambivalent affect equally during
both positive and negative life events (Hui et al., 2009; Miyamoto et al., 2010). These findings
show that ambivalent affect represents a marker of complexity in affect representation and
interpretation, especially when considered developmentally (Albanese et al., 2010; Schultz et al.,
2005; Steele et al., 1999; Zajdel et al., 2013). Finally, increased experiences of ambivalent
affective states are also associated with psychotherapy-induced improvements, demonstrating
their complicated relationship with temporal change (Adler & Hershfield, 2012).

Future directions
Ambivalent affective experiences may well remind us of why we have conscious affect in
the first place. Standard affective experiences guide our decision-making in an advantageous
36

manner, one likely reason why affect was selected in evolution (Damasio, 2018; Denton et al.,
2009). But ambivalent affective experiences appear to have their own advantages. The
bittersweet affects discussed earlier are valuable because they counter the disadvantages that
standard affective experiences can produce in a complex social world. On occasion, it is
advantageous to decide against or ignore our immediate feelings. Given how important affect is
in motivating our decisions, the especially nuanced representations that ambivalent affect allow
would be an asset.
While ambivalent affective experiences appear to be universal, the culturally-specific
variants are evidence for conceptual fine-tuning leading to divergent and culturally novel
feelings (Lomas, 2017; Perlovsky, 2009). For example, there is evidence to show that semantic
word categories influence the neural representation and experience of affect; the presence or
existence of a word to label an affective experience shapes that experience as a more consistent
entity within its category (Brooks et al., 2017). But there is no reason to assume that language is
the only type of conceptual information that can contribute to the shaping of novel affective
experiences. Sociocultural and technological shifts in how we interact with the world lead to new
feelings which, once labeled converge conceptually. Furthermore, the process by which we
cognitively embody and integrate homeostatic signals into feelings is strongly influenced by
psychosocial development during childhood and even infancy (Fotopoulou & Tsakiris, 2017;
Seth & Friston, 2016). With how dependent ambivalent affect is on the affective-interoceptive
processes we develop during childhood, the experience of these states is likely to be highly
influenced by our upbringing (Lane & Schwartz, 1987).
Affective neuroscience has devoted well deserved attention to how core affect manifests
and influences our behavior. Nonetheless, in an ambiguous world we also need to understand the
37

more complex and nuanced feelings which are often at odds in situations that society paints as
binary: situations such as the nostalgia for a relationship that has since finished or for certain
political and moral issues. The cultural ideals of decisiveness and definitiveness should not force
us to simplify feelings. When does having something to miss or to long for help us persevere or
else compromise our functional capacity? How does a feeling of longing form when what we
long for is unclear or abstract? When does the triggering of nostalgia help us cope with a
changing unfamiliar world, and when does it bias the reconsolidation of our memories and future
decision-making in maladaptive ways? Affective neuroscience can help us explore these clinical
and fundamental questions.
The study of ambivalent feelings can also answer important questions on the general
science of affect e.g. whether positive and negative valence are represented neurally as part of a
unidimensional construct or as two separate dimensions, how affective states should be
contrasted or classified, whether we must have a distinct word for feeling distinct experiences,
and whether some affects are more basic or complex than others. Ambivalent feelings are more
difficult to characterize on an objective basis than the plain variety. Still, the fact that they
constitute a significant aspect of our affective life warrants their investigation as we move
affective neuroscience closer to being a science of the human experience.
38

Chapter 2: Perspective-taking is associated with increased discriminability of
affective states in ventromedial prefrontal cortex
Published: Vaccaro, A. G., Heydari, P., Christov-Moore, L., Damasio, A., & Kaplan, J. T.
(2022). Social Cognitive and Affective Neuroscience, 17(12), 1082-1090.
Abstract
Recent work using multivariate-pattern analysis (MVPA) on functional magnetic
resonance imaging (fMRI) data has found that distinct affective states produce correspondingly
distinct patterns of neural activity in the cerebral cortex. However, it is unclear whether
individual differences in the distinctiveness of neural patterns evoked by affective stimuli
underlie empathic abilities such as perspective-taking (PT). Accordingly, we examined whether
we could predict PT tendency from the classification of blood-oxygen-level-dependent (BOLD)
fMRI activation patterns while participants (n = 57) imagined themselves in affectively charged
scenarios. We used an MVPA searchlight analysis to map where in the brain activity patterns
permitted the classification of four affective states: happiness, sadness, fear and disgust.
Classification accuracy was significantly above chance levels in most of the prefrontal cortex
and in the posterior medial cortices. Furthermore, participants’ self-reported PT was positively
associated with classification accuracy in the ventromedial prefrontal cortex and insula. This
finding has implications for understanding affective processing in the prefrontal cortex and for
interpreting the cognitive significance of classifiable affective brain states. Our multivariate
approach suggests that PT ability may rely on the grain of internally simulated affective
representations rather than simply the global strength.

39

Introduction
Contemporary neuroscience research has highlighted the complex relationship between
neural activity and affective states. We define affective states to include both emotions, which in
our view comprise physiological and motoric responses to stimuli in the environment that are
relevant to the homeostatic welfare of the organism, and feelings, the conscious perceptions of
emotion-related changes in the body. Affective states appear to involve interactions between
cortical and subcortical regions, as well with the viscera (Hugo D Critchley, 2009; Damasio &
Carvalho, 2013; Kober et al., 2008; Smith & Lane, 2015; Tettamanti et al., 2012; Vaccaro et al.,
2020).
Recently, our understanding of emotion has progressed from considering solely the
experience and mechanisms of individual experience, to understanding that the consideration of
other’s minds plays a large role in one’s own affective experience (Christov-Moore & Iacoboni,
2016; Decety & Meltzoff, 2011; C Lamm et al., 2016). While the core mechanisms of affect
were once viewed and researched as a primarily private, intraindividual processes, there is a
growing consensus that the development of affect is inescapably linked to sociality: this places
empathetic processes as more core to individual affect than they were originally considered
(Dukes et al., 2021; Fotopoulou & Tsakiris, 2017; Parkinson & Manstead, 2015). Empathy is
multifaceted construct combining cognitive processes that allow us to understand the internal
states of others, and affective processes that allow us to share in the internal states of others.
These include aversive reactions to others’ distress (personal distress), concern for others’
welfare (empathic concern), feeling and understanding the experiences of hypothetical or absent
others (fantasizing) and taking others’ perspectives (perspective-taking) (Davis, 1983). Research
has found that systems involved in understanding one’s own emotions are also involved in
40

understanding the affective states of others (Ochsner et al., 2004). Empathizing with another’s
feelings recruits affective brain regions involved in representing one’s own affective state (C.
Lamm et al., 2016; Singer et al., 2004) and there is evidence that impaired affective experience
(as in psychopathy) may limit empathic abilities (Blair et al., 2002). For example, participants
administered an analgesic were impaired in their ability to recognize and respond to others’ pain
(Mischkowski et al., 2019).
One feature of affective experience that is relevant to both empathizing with others and to
representing one’s own state is the extent of differentiation among affective states. Previous
studies have shown that individuals who are more successful at judging the affective states of
others experience more differentiated categories of affect (Erbas et al., 2016; Israelashvili et al.,
2019). Having, and being able to simulate, affective states which are categorically discernable
may facilitate skills such as mentalizing, empathy, and perspective-taking due to the perceived
increased in clarity of what one is feeling, and therefore what that feeling means functionally
(Eckland et al., 2018; Hill & Updegraff, 2012; Thompson & Boden, 2019). Therefore, we
hypothesize that increased neural differentiation of affective states may be associated with
greater empathy.
Recent studies using multivariate pattern analysis (MVPA) find that distinct affective
states may be associated with specific patterns of neural activity within a network of brain
regions (Alessia Celeghin et al., 2017; Nummenmaa & Saarimäki, 2019; Scarantino, 2012).
MVPA studies have demonstrated that discrete, induced affective states can be accurately
distinguished from each other (i.e., classified) using patterns of BOLD activation in fMRI data
(Kassam et al., 2013; Kragel & LaBar, 2015a; Saarimaki et al., 2016). The most commonly
studied of these states in MVPA studies are sadness, disgust, fear, happiness, and, to a lesser
41

extent, anger, all classically considered “basic emotions” (A. Celeghin et al., 2017; Saarimaki et
al., 2016). However, other more subtle affective states have also been studied, such as shame,
envy, contempt, pride, guilt, and longing (Kassam et al., 2013; Kragel & LaBar, 2015a;
Saarimaki et al., 2018), though these may be less easily classified than their more “basic” cousins
(Saarimaki et al., 2018).
Cortical regions found to contribute most to classification accuracy in MVPA studies
tend to be consistent with those found in univariate analyses of affective processing. These
regions include the medial prefrontal cortex (mPFC), inferior frontal gyrus (IFG), posterior
medial cortex (PMC), insula, and amygdalae (Kim et al., 2015; Peelen et al., 2010; Saarimaki et
al., 2018; Saarimaki et al., 2016; Sachs et al., 2018). A key region involved in both judging
another’s affective state and representing one’s own affective state is the mPFC (Seitz et al.,
2006). Specifically, the ventral areas of mPFC have been shown to play a selective role in
affective perspective-taking as compared to cognitive perspective-taking (or general theory of
mind) (Corradi-Dell’Acqua et al., 2014; Meghan L Healey & Murray Grossman, 2018; Hynes et
al., 2006).
Empathy is often an implicit part of paradigms used to study emotion differentiation. In
order to experimentally invoke affective states inside the fMRI scanner, it is common to present
subjects with affect-provoking stimuli and also to engage subjects in voluntary mental
simulation. For instance, Saarimaki et al. (2018) used narratives that describe the lead-up to an
emotional event along with a guided imagery technique to evoke 14 different affective states. It
can be difficult or impractical to design stimuli that effectively induce genuine affect. Tasks
often involve explicitly asking participants to imagine themselves in emotional scenarios based
on visual or audio imagery. Paradigms such as this require participants to access their concepts
42

of emotion. Interestingly, this naturally creates individual variability where some individuals can
easily generate strong feelings from retrieving emotion concepts while others cannot. This
difference in the ability or motivation to deliberately simulate affective states from concepts is
similar to what has been proposed in the somatic marker hypothesis for the vmPFC, generating
feelings “as-if” one is in a scenario (Damasio, 1996). It is possible that this individual variability
presents itself in the distinctiveness of the neural states evoked in response to different cues. In
line with previous work on the overlap between empathy and the representation of one’s own
affective states, a trait level measure of empathy may be associated with these differences.
In our study, participants underwent an affect induction paradigm in which they viewed
pictures of situations invoking fear, happiness, sadness, and disgust, alongside captions
describing the scenario from a first-person perspective. It is likely that this type of paradigm
evokes both emotions and feelings, thus it allows us to investigate the correlates of affective
states as a whole, but not to differentiate between neural patterns for emotion and for feeling. We
ran two sets of MVPA analyses on the evoked neural data to investigate the classification
accuracy of emotions from patterns of fMRI activity. In the first, we examined which regions’
activation was most informative for classifying the four evoked affective states. We
hypothesized that the mPFC would have the highest classification accuracy. In the second
analysis, we attempted to predict individual differences in empathic ability from the
classification accuracy of participants’ patterns of neural activation during emotion induction.
We hypothesized that individual differences in empathic ability would be reflected in the
distinctiveness of neural patterns of activation evoked by different emotions. Given the mPFC’s
prominence in MVPA studies of emotions, as well as its role in affective perspective-taking, we
43

hypothesized that classification accuracy for emotions in this region would show the highest
correspondence with empathic abilities.

Methods
Healthy adult participants (n=57) were recruited as part of two different studies, all
recruited through flyers from the University of Southern California and surrounding Los Angeles
area. Thirty-six participants’ data (18 female, age = 24.21±8.68, range = 18-52) were collected in
the first study. In a second study, 21 more participants’ data (11 female, age = 22.67±6.45, range
= 18-42) were collected. Since the two studies used the same experimental paradigm and stimuli,
with slight differences detailed below, the data were combined to increase statistical power. All
participants were right-handed, had normal or corrected-to-normal vision, no history of
neurological or psychiatric conditions, All participants gave informed consent in accordance with
the Institutional Review Board approval guidelines approved by the University of Southern
California. Because the second study involved a more comprehensive battery of additional
behavioral measures not used in the analyses of this paper, behavioral data for those participants
was collected on a different day. For this reason, 2 of the 21 participants did not provide
behavioral data collected before the university shutdown due to COVID-19, leaving us with 19
participants from this second study for analyses relating to the empathy measures (9 female, age
= 22.32±6.14, range = 18-42): a total of 55 participants between the two studies for this second
analysis.

44

Interpersonal Reactivity Index
Behavioral measures of empathy were acquired through the Interpersonal Reactivity
Index (Davis, 1983). This self-report measure consists of four seven-item subscales: (1)
perspective-taking (PT): the ability of the participant to take on the point of view of another
individual, (example: “Before criticizing somebody, I try to imagine how I would feel if I were
in their place)”, (2) fantasy (FS): the tendency of the participant to identify with fictitious
characters (example: “I really get involved with the feelings of characters in a novel”), (3)
empathic concern (EC): the presence of the participant’s feeling of compassion or concern for
others (example: “I am often quite touched by things I see happen”, and (4) personal distress
(PD): the presence of the participant’s feeling of discomfort or anxiety for others (example:
“When I see someone who badly needs help in an emergency, I go to pieces”) (Davis, 1983). For
each participant, each subscale score was assessed separately, resulting in four distinct scores per
participant.

Stimuli
Stimuli were presented as one photo in the center of the screen with anecdotal descriptive
text underneath each photo. Photos were first gathered from a subset of images in the
International Affective Pictures Set (IAPS; Bradley & Land, 2007) covering the affective
categories of happiness, fear, sadness, and disgust. Text sentences from the Affective Norms for
English Text (ANET; Bradley & Lang, 2007) were also chosen in these four categories. Stimuli
captions are written in the 2
nd
person, telling the subject what they were experiencing (example:
“As you leave the concert, a drunk vomits all over your jacket, soaking it.”). Pictures from the
IAPS were then matched with a corresponding piece of text from the ANET that described a
45

situation associated with the picture. For example, a picture of a snarling dog was combined with
the caption “The dog strains forward, snarling, and suddenly leaps out at you.” (See Table S1 for
examples).
For pictures that did not have appropriately matching text from ANET or text that did
not have appropriately matching images from the IAPS, text/images were written to fit or
acquired from the web. These new images were rated for valence and arousal by 51 participants
in an earlier study (Table S2 and S3). Subjects were also asked to indicate what category of
affective state each photo/text combination corresponded to. For every stimuli selected for the
study, the expected category (among happy, sad, fear, disgust and neutral) was the most
commonly picked category by the subjects (see Figure S1 and Table S4). In addition to these
emotional stimuli, non-emotional / neutral stimuli were used as a control, and a fixation cross
was used as a rest. Since our goal was to predict emotional states with MVPA, analysis of the
neutral images is not presented here.

Functional Neuroimaging: fMRI
fMRI Design
Stimuli were presented in an event-related design using MATLAB’s Psychtoolbox. In
Study 1, 60 stimuli (photo + text) were randomly presented during 4 functional runs (15 stimuli
per run, 6 mins per run). Each stimulus was presented for 12 seconds followed by a 12-second
fixation cross in between each trial as a “rest” period. In study 2, 45 stimuli were randomly
presented during 3 functional runs. In both studies, participants were instructed to lay still,
observe the displayed photograph, read the text, and attempt to embody the described emotional
situation as strongly as possible for each stimulus.
46

fMRI Data Acquisition
All scanning was completed on a 3-T Siemens Prisma System Scanner at the USC
Dornsife Cognitive Neuroimaging Center using a 32-channel head coil. Anatomical images were
acquired with a T1-weighted magnetization-prepared rapid gradient-echo (MPRAGE) sequence
(repetition time [TR]/echo time [TE]=2300/2.26, voxel size 1 mm isotropic voxels, flip angle
9°). Functional images were acquired with a T2*-weighted gradient echo sequence (repetition
time [TR]/echo time [TE]= 2000/25 ms, 41 transverse 3-mm slices, flip angle 90°). A T2-
weighted volume was acquired for blind review by an independent neuroradiologist, in
compliance with the scanning center’s policy and local IRB guidelines. T2-weighted scans were
not analyzed by the researchers for any purpose in this study.

fMRI Analysis
Preprocessing and GLM
Data were first processed using FEAT, FSL’s implementation of the General Linear
Model (GLM) (FMRIB Software Library, Smith et al., 2004) to generate voxelwise z-statistic
maps showing voxels that responded significantly to each emotion type for each participant.
Those z-statistic maps were then used for classification analysis. Data preprocessing was
conducted in FSL (FMRIB Software Library, Smith et al., 2004) using brain extraction, slice-
time and motion correction using MCFLIRT, spatial smoothing (5mm), and high-pass temporal
filtering (sigma=50s). The functional data were registered to each participant’s own anatomical
image and the anatomical data were registered to the standard MNI Brain (Montreal
Neurological Institute) using FSL’s FNIRT tool (Jenkinson & Smith, 2001). The data were
47

modeled with a separate regressor for each of the 4 emotions (happy, sad, fear, disgust), one for
the neutral condition, the temporal derivatives of all task regressors, and 6 motion parameters to
account for residual motion effects. These same smoothed standard space z-maps were used for
both the emotion discrimination analysis and for the individual subject searchlights.

MVPA analysis
Emotion discrimination analysis
All MPVA analyses were conducted using PyMVPA (Hanke et al., 2009). A whole brain
searchlight analysis was conducted to identify regions whose activation patterns allowed us to
classify between the 4 emotions across all subjects’ data. The input data to the classifier was a
single 4D image file combining z-stat maps (normalized to MNI standard space using FNIRT)
for each affective state, functional run and participant (resulting in a total of 828 concatenated
images – 36 participants x 4 runs x 4 emotions for study 1 combined with 21 participants x 3
runs x 4 emotions for study 2). For every voxel in the brain a sphere centered on that voxel
(radius 5 voxels) was used to train and test a linear support vector machine (SVM) using a leave-
one-out cross-validation. In other words, in each iteration the classifier was trained on all the
participants’ data except one, and then tested on the remaining participant’s data leading to 55
cross-validation folds. The resulting average accuracy over all iterations, after leaving each
participant out once, was mapped to the center voxel of the sphere, ultimately resulting in a
cross-participant map of classification accuracies for every voxel. For the SVM regularization
parameter C, we used the default in PyMVPA, which chooses this parameter automatically
according to the norm of the data.

48

Empathy correlation analysis
To correlate the scales of the IRI with individual classification accuracy, we first
computed whole-brain searchlights within every individual participant’s data. We ran
searchlights on each individual subject’s data in standard space. For each participant, a sphere
(radius = 3 voxels) centered on every voxel in the brain was used to iteratively assess
classification accuracy throughout the brain. For each sphere, the SVM classifier was trained on
all but one of the emotions’ functional scanning runs and tested on the left out run. The resulting
accuracy values were mapped to the center voxel of the sphere and resulting whole-brain voxel-
wise accuracy maps were warped into the standard space. This created an accuracy map for each
participant where a voxel’s value represented the classification accuracy of the 3 voxel sphere
surrounding that voxel. We chose a smaller sphere size than the between-subjects analysis to
increase our spatial resolution. Furthermore, these individual subject searchlights required
significantly less computing power than the between subjects analysis, so we were less restricted
by computing limitations. To identify relationships between these individual participant
searchlight maps and individual differences in self-reported empathy, we used FSL’s Randomise
tool (with FWE correction). We created a series of regressors using participants’ demeaned
scores on each of the IRI’s subscales (perspective-taking, empathic concern, personal distress,
and fantasy). These regressors were then related to the searchlight accuracy values using
Randomise’s nonparametric permutation testing approach (Winkler et al., 2014). At each voxel,
the accuracies across subjects are randomly associated with the regressor-values (the empathy
sub-scales) and a test statistic is computed. This permutation procedure is repeated 5000 times to
generate a null distribution. One of the resulting maps is then a map of 1 minus the p-value for
49

the association between the regressor and the classification accuracy, determined by comparison
to the null distribution. These maps were corrected for familywise error using FSL’s threshold-
free cluster enhancement (TFCE) algorithm. We restricted our interpretation of the resulting
maps to regions that significantly predicted emotions in the group searchlight analysis by
masking these maps with the regions found to significantly distinguish affective states at the
group level in the initial affect discrimination analysis. Therefore, the final resulting map shows
voxels where both a) classification was above chance across all subjects and b) classification
varied significantly with the empathy scores.

Statistical thresholding
To determine an appropriate threshold for significance, we employed two parallel
methods to correct for multiple comparisons in each of our analyses. For the affect
discrimination analysis, we first used permutation testing to create a null distribution of
classification accuracies within a simulated searchlight sphere by shuffling affect labels. We ran
151,801 permutations of the classification within a single sample searchlight sphere. This
allowed us to create a distribution of classification values that might occur in a given searchlight
sphere assuming the null hypothesis that the four affect conditions cannot be distinguished from
patterns of activation. To account for multiple comparisons in the correlation analysis, we first
performed a resel-wise Bonferroni correction: we determined the total number of 5 voxel radius
independent spheres which could fit in the standard brain (~725) and divided 0.05 by this
number to determine our alpha (6.89 *10
-5
) (Kaplan & Meyer, 2012). Therefore, a classification
accuracy for which greater than or equal values appeared in our distribution less than 10 times
would be considered significantly above chance. In our simulated null distribution, where there
50

were four emotions and an expected chance accuracy of 0.25, a significant above chance
classification accuracy was determined to be > 0.30592. The maximum value found in our null
distribution was 0.311 (Figure S2). For the second correction method in both analyses, we used
a voxel-wise Bonferroni correction (.05 / n tests) and determined the corresponding accuracy
value using the binomial distribution. With this method, we established that a Bonferroni
corrected significance of .05 would correspond to an accuracy threshold of 0.3815. To enhance
the replicability of our findings and isolate the most informative regions, we opted in both
analyses to use the more conservative threshold (permutation-corrected results can be found in
the appendix: Figure S3). This threshold is extremely conservative given that it is based on a full
voxel-wise Bonferroni correction, and is also greater than any of the classification results in our
over 150,000 permutations.
In post-hoc analyses, we wanted to determine if our results in the regions that
significantly distinguished between the 4 emotions were driven by one emotion being uniquely
distinguishable compared to the others. To do this, we ran pair-wise searchlight analyses
classifying between each of 6 possible pairs of emotions. If one specific emotion was driving our
findings, only pairs with that one emotion would show a similar spatial pattern of significant
classification to our initial 4-way searchlight emotion discrimination analysis.

Results
Maps of the statistical results for this study can be found on the Open Science Foundation
website for this study, https://osf.io/8wcnz/.

51

Affect discrimination searchlight
Multiple regions significantly predicted affect classification, even using the more
conservative (voxel-wise Bonferroni-corrected) threshold (compared with the 4-way chance
level of 0.25). These included vmPFC (x = -9, y = 55, z = -16; accuracy = .444), anterior
prefrontal cortex (x= -9, y = 55, z = 12; accuracy = .436), dorsomedial prefrontal cortex (x = 14,
y = 36, z = 46; accuracy = .419), bilateral insula (left x = -40, y = -11, z = 9; accuracy = .441,
right x = 44, y = -6, z = 8; accuracy = .420), bilateral amygdala (left x = -28, y = -10, z= -12;
accuracy = .412; right x = 26, y = -10, z= -12; accuracy = .417) posterior cingulate cortex (x = -
2, y = -52, z = 14; accuracy = .425), bilateral temporal gyrus (x = -38, y = -54, z = -8; accuracy
= .446) (x = 33, y = -43, z = -10; accuracy = .448), and bilateral superior parietal lobule (left x =
-32, y = -69, z = 37; accuracy = .428; right x = 33, y = -63, z = 30; accuracy = .426).
These spatial patterns, especially the high classification values in medial frontal and
parietal areas, were mirrored in all the post-hoc pair-wise classification searchlights except
happy vs. sad, which did not reach a significant threshold of .68 accuracy (see Supplementary
Figures S4-S8). This suggests that our 4-way classification results did not result from one
emotion being distinguishable compared to the rest, or one pair of emotions being especially
distinguishable compared to other combinations.

52

Figure 1: Accuracy of affect classification searchlight

Classification accuracy and IRI measures
(See Table S5 and Figure S9 for descriptive statistics and distributions of subscales).
Fantasy, personal distress, and empathic concern were not significantly related to classification
accuracy. Perspective-taking was significantly related to classification accuracy (p<0.05 with
FWE correction) in a large area of vmPFC (x = -15, y = 48, z = -7), as well as in bilateral, though
predominantly left, insula (x = -39, y = -7, z = 7) (x = 43, y = -5, z = -6) (See Figure 2 and 3).

53

Figure 2: Regions where classification accuracy was significantly predicted by perspective-
taking

54

Figure 3: Classification accuracy vs. perspective-taking in peak voxels of the vmPFC and
insula

Discussion
MVPA has become a popular method in affective neuroscience over the past decade
(Coutanche, 2013; Morawetz et al., 2016; Oosterwijk et al., 2017). While it has been
demonstrated repeatedly that neural patterns associated with affective states can be distinguished
from each other, the interpretation of what these patterns tell us about categories, and how they
come to be, is contentious (Clark-Polner et al., 2017; Coutanche, 2013; Gessell et al., 2021b;
Kragel & LaBar, 2016). In our study we used a novel approach to increase our power of
inference in these often “black-box” like scenarios. We demonstrate that an individual trait may
be related to differences in the distinguishability of neural patterns corresponding to affective
states. This method provides a new way of exploring and theorizing why neural patterns can be
distinguished in this way, and demonstrates a new avenue of inference for testing how personal
55

traits and cognitive functions might be associated with how different neural states are discernible
from each other. In this case, we found that self-reported perspective-taking is related to the
discernibility of the affective states our participants were asked to imagine. Our study has
implications for both the use of MVPA in general, and our understanding of how empathetic
traits relate to affect.

Relating MVPA performance to individual differences
It has long been recognized that connecting an MVPA result directly to some measured
aspect of the underlying behavior or psychology being studied is important (Naselaris et al.,
2011; Raizada & Kriegeskorte, 2010). Otherwise, successfully decoded neural signals may
reflect information that is either not used at all by the brain, or is not relevant to aspects of the
psychology in question (Kriegeskorte & Douglas, 2019; Ritchie et al., 2019). Nevertheless, it is
uncommon for MVPA studies to relate classifier performance to individual differences. Using
procedures similar to ours, Coutanche et al. (2011) found that MVPA classification of visual
stimuli correlated with Autism Spectrum Disorder symptom severity across subjects. Studies in
the auditory domain have found that MVPA classification accuracy in a musical instrument
identification task (Ogg et al., 2019) and a speaker identification task (Aglieri et al., 2021; Bonte
et al., 2014) correlate with between-subject differences in accuracy on those task. Raizada et al.
(2010) found that the performance of a classifier at distinguishing neural patterns related to the
perception of different phonemes was related to subjects’ performance in discriminating the
same phonemes, and that the separability of neural patterns during a numerical discrimination
task was related to arithmetic performance. Similarly, Meyer et al. (2010) showed that classifier
56

discrimination of imagined auditory stimuli was correlated with the reported vividness of
imagination.
Yet while linking MVPA performance to individual differences can make headway in
showing the psychological relevance of classification performance, many of the limitations that
exist in other studies attempting to relate individual differences to neuroimaging analysis still
remain. For example, to optimize studying individual differences, the states induced should
produce enough variability across individuals while still maintaining high identifiability of
stimulus representations (Finn et al., 2017). In terms of identifiability of neural patterns, MVPA
tends to be more sensitive than univariate analysis. However it also tends to be less sensitive to
between-subject variability in mean activation levels (Davis et al., 2014). It remains to be
determined whether the variability across people in classifier performance is optimal for the
study of individual differences. Furthermore, while the trait-relevance of classifier performance
is informative, interpretation must always be careful, in that such relationships can always be
mediated by additional, unmeasured variables.
Decoding of affective states correlates with perspective taking
We found multiple brain regions whose activity permitted us to robustly classify evoked
affective states. Many of these brain regions have allowed successful classification of affect in
prior studies: the posterior cingulate, insula, temporal gyrus, medial prefrontal, and ventral
medial prefrontal regions have all been implicated in previous MVPA studies of emotions and
affective context (Bush et al., 2018; Oosterwijk et al., 2017; Paquette et al., 2018; Saarimaki et
al., 2016; Sachs et al., 2018; Skerry & Saxe, 2014).
The vmPFC has long been implicated in implementing emotion and feeling (Bechara et
al., 1994; Bechara et al., 1999; Roy et al., 2012; Winecoff et al., 2013). A theoretical link
57

between affect and the vmPFC is provided by the somatic marker hypothesis (Damasio, 1996).
This view posits that the vmPFC is a key region for incorporating emotion-related signals from
the body, consciously experienced as feelings, into our cognitive decision-making process.
According to the SMH, the vmPFC also permits us to bypass pure bodily input in order to
simulate these feelings even without the direct presence of a trigger (Poppa & Bechara, 2018).
By this account, the vmPFC could serve to simulate the affective experience being evoked
(Keysers & Gazzola, 2007; Schacter et al., 2017). Indeed, studies on affective perspective-taking
vs. cognitive perspective-taking have implicated the vmPFC (Meghan L Healey & Murray
Grossman, 2018; Sebastian et al., 2012). The vmPFC has also been implicated in affective
simulation in studies in which participants imagined affective contexts that could happen to them
in the future (D’Argembeau et al., 2008) or that have happened in the past (Benoit et al., 2016;
Benoit et al., 2019). Correspondingly, lesions to vmPFC appear to hamper both the ability to
simulate future affective scenarios and imaginary ones (Bertossi et al., 2016; Bertossi et al.,
2017).
Importantly, we show here that the vmPFC’s role in affective perspective-taking is related
not just to the level of neural activity invoked, but that people who report being perspective-
takers show more distinct affect-related neural patterns. This could suggest that affective
perspective-taking may relate to simulation of the affective experience that is more specific to
the target emotion, reflected functionally in more easily classifiable representations of affective
context. If simulated affective states are more generalized and overlapping, as opposed to
detailed and nuanced, their associated neural patterns can be expected to be less differentiable
Perceived emotional actions, interoception, and affective contexts of other individuals can all be
successfully classified using vmPFC activity (Oosterwijk et al., 2017). Notably, areas where
58

accuracy was significantly associated with perspective-taking showed little overlap with regions
typically implicated in univariate studies of perspective-taking: regions such as visual cortex,
temporal-parietal junction, and dorsolateral prefrontal cortex (Bukowski, 2018; Decety &
Grèzes, 2006; Meghan L. Healey & Murray Grossman, 2018). Accuracy in these areas did not
correlate with our perspective-taking measure. Additionally, even though it is conceivable that
our visual and descriptive prompts would support a largely visual simulation of the scene, or that
improved affective representation might result from increased attention, neither primary visual
regions nor parietal and temporal visuospatial regions provided significant classification
accuracy. Visual, temporal-parietal, and dorsal prefrontal regions may be important for the
general process of perspective-taking, but more related to the general effort involved in
performing the task rather than its success. In a study where individuals were instructed to take
the perspective that an image of a painful stimuli was occurring to either themselves or another
person (van der Heiden et al., 2013), the main effects of the different perspective-taking
conditions were evident in regions such as supramarginal gyrus, temporal gyrus, frontal gyrus,
and ventrolateral PFC. However, when the effects of good vs. bad perspective takers were
analyzed, all these regions except ventrolateral PFC were not significant predictors, and instead
the left insula, postcentral gyrus, and vmPFC differed between the groups. Perhaps the insula and
vmPFC are specifically important for generating a fine-grained affective simulation, and this
granularity is reflected in more distinct neural patterns.
We note also that while the IRI includes items that ask participants to judge their ease or
difficulty at empathizing (“I sometimes find it difficult to see things from the "other guy's" point
of view.”) many of the questions are instead about the tendency to take someone else’s
perspective (“I sometimes try to understand my friends better by imagining how things look from
59

their perspective.”) or the motivation to do so (“I try to look at everybody's side of a
disagreement before I make a decision.”). As such, we cannot distinguish which aspects of
perspective taking are directly related to increased classifier accuracy. It is possible that
participants who score higher on the IRI-PT are motivated to engage in the task with more effort
and that the increased pattern separation is related to increased effort in the task. Indeed, the IRI-
PT is correlated with the Mind Reading Motivation scale, a measure specifically aimed at the
motivation to think about other people’s minds (Carpenter et al., 2016).
Future directions
Our searchlight analysis found large regions of the brain, many previously implicated in
affect-related processing, that significantly classified affective states. Despite this, only the
accuracy in the vmPFC and insula were related to perspective-taking. This leaves the question of
what traits, cognitive processes, or noise factors could potentially explain individual differences
in classification accuracy in other regions. It is possible that other cognitive traits, such the
ability to interpret other’s affective intent or bodily perception, may explain how accurate other
regions such as temporal gyrus (Wicker et al., 2003) and the parietal lobule (Engelen et al., 2015)
are at distinguishing affective states in MVPA studies. Furthermore, the affective states we asked
participants to imagine are amongst the most common, “basic emotions”. These affective
experiences are most commonly described as categorical in daily life and therefore may be more
easily discriminated. It remains unclear whether the same mechanisms, perspective-taking and
affective simulation in the vmPFC, would be as accurate discriminating mixed feelings, or if
other neural regions and cognitive processes may be differentially important for non-typical
affective states.
60

Our study highlights the importance of investigating the neural substrate of individual
differences in perspective-taking overall, especially across domains. While some brain regions
may be involved broadly in perspective-taking across participants, univariate approaches may
miss more specific regions which truly distinguish successful vs unsuccessful perspective-taking.
In our affective simulation task, these were the vmPFC and insula, though these likely may be
different for perspective-taking tasks involving other domains.

Conclusion
In our study, we found a relationship between trait perspective-taking ability and the
classification accuracy of an individual’s vmPFC and insular activity for distinguishing task-
evoked affective states. The value of these findings is important because it shows that the
discriminability of signal in these regions, which exhibit high classification accuracy among
affective states overall, is associated with a task-relevant personal trait, namely perspective
taking. More work is needed, however, to explore what underlying functional properties underlie
successful classification (Anderson et al., 2012; Carlson & Wardle, 2015).
Methodologically, we show that searchlight MVPA can be used to uncover the
mediating traits and processes which explain why a particular region of the brain contributes to
classification accuracy. Connecting MVPA results directly to individual traits or behaviors
greatly enhances their interpretability. These findings reflect the strength of both multivariate
analysis and the study of individual differences: both seek to gain information from what differs
between participants, rather than averages and commonalities across participants.

61

Supplementary Material
Table S1: Examples of Stimuli Sentence-Picture Pairs
Stimuli Category Image Description IAPS or ANET ID

1 Happy man jumping for joy in store ANET_8500

2 Happy child laughing at table with family ANET_2560

3 Fear dog with teeth bared IAPS 1301

4 Fear robber in a mask ANET_6370

5 Disgust bug on slice of pizza IAPS 7380

6 Disgust man throwing up on other man IAPS 9321

7 Sad old man with dying woman IAPS 2205

8 Sad woman and man crying at factory IAPS 2456

9 Neutral grocery store aisle ANET_2520

10 Neutral flashlight shining on map ANET_7040

Stimuli Sentence
1 You've just won 10 million dollars; you jump up and down, screaming.
2 You lounge around the crowded table, laughing with your family.
3 The dog strains forward, snarling, and suddenly leaps out at you.
4 You freeze as a masked figure looms over you and ties you to the bed.
5 You gag, seeing a roach moving slowly over the surface of the pizza.
6 As you leave the concert, a drunk vomits all over your jacket, soaking it.
7 After fifty years of marriage, your wife has died, leaving you alone.
8
Today you found out you lost your job, and you won't be able to buy food for your
children.
9 You walk through the grocery aisles, adding necessary items to your cart.
10 You hold the flashlight steady in order to get a better look at the map.

Table S2: Mean (Standard Deviation) of Rated Valence and Arousal of Stimuli (1-7 scale)
Emotion Valence Arousal
Happy 6.94 (.633) 5.09 (.844)
Fear 2.96 (1.10) 6.81 (.541)
Disgust 2.72 (1.01) 5.70 (.994)
Sad 2.67 (.665) 4.72 (.557)
Neutral 5.49 (.892) 4.49 (.590)

62

Table S3 T-values From Two Sample T-Tests Comparing Valence and Arousal of Stimuli
Valence Happy Fear Disgust Sad Neutral
Happy x 10.852** 12.231** 16.094** 4.593**
Fear x x 0.551 0.794 -6.174**
Disgust x x x 0.164 -7.093**
Sad x x x x -8.768**
Neutral x x x x x
Arousal Happy Fear Disgust Sad Neutral
Happy x -5.937* -1.602 1.27 2.037
Fear x x 3.413* 9.232** 10.059**
Disgust x x x 2.961* 3.623*
Sad x x x x 1.002
Neutral x x x x x
*p<.01
**p<.001

Table S4: Average Number of Participants Choosing Stimuli Category by Category of Emotion

Chosen Category

Intended Happy Fear Disgust Sadness Neutral
Happy 30.33 0.58 0.42 0.75 11
Fear 2.75 27.67 1.5 6.08 5.17
Disgust 1.83 4 31.67 8.5 5.67
Sadness 0 6 6 32 6
Neutral 14.67 1 1.5 1.42 26.42
Intended Anger Anxiety Shame Love Surprise
Happy 0.33 1.75 0.42 15.83 5.66
Fear 3.08 24.58 0.5 0.25 6.67
Disgust 2.17 8 3.67 1.33 2.17
Sadness 12 19 4 1 0
Neutral 0.67 5.58 1.75 2.83 1
Table S5: Mean and Standard Deviations of Empathy Sub-scales
IRI Sub-Scale Mean Standard Deviation
Empathic Concern 20.31 4.118
Perspective Taking 20.42 4.18
Personal Distress 10.8 4.266
Fantasy 18.24 4.823

63

Stimuli Presentation
Subjects were given the following instructions: “When the experiment begins, you will
see a series of pictures and words. Each picture will appear on the screen in front of you with a
short sentence below it. Please read the statement, look at the picture, and try to feel the emotions
and feelings described by the scenario as strongly as possible. The picture will stay on the screen
for several seconds. When the picture disappears, do your best to put it out of your mind. After a
short period, the next picture will appear.”

Comparison of individual differences in subjects from study 1 and study 2
Since the 36 subjects from the study 1 data had 4 functional runs, while the 19 subjects
from study 2 had 3 functional runs, we ran an analysis to compare the classification accuracy
between the two groups. This way, we can determine if our findings are due to different numbers
of trials across our dataset. To compare individual differences in classification accuracy between
3-run and 4-run subjects, we used FSL’s Randomize tool to compare the searchlight maps of
both cohorts of subjects. We then tested the overlap between these voxels that differed and the
voxels in our main result: the voxels significant for perspective-taking correlating with emotion
discrimination accuracy in regions where emotion discrimination accuracy was significant
overall.
81% of voxels that show correlation between perspective-taking and accuracy did not
significantly differ between subjects with three trials and subjects with four trials. Furthermore,
we have tested for differences in perspective taking scores between the two groups, and found no
difference (M1=20.89 M2=20.17, p=.54). These show that our findings are not due to the
64

varying number of trials or differences in perspective-taking between the two groups of
participants.

Figure S1: Proportion of participants that rated emotion stimuli in category

Figure S2: Distribution of null classification accuracy

65

Figure S3: Accuracy of emotion classification searchlight using the minimum threshold
derived from permutation testing

66

Figure S4: Searchlight of Happy vs. Fear

Figure S5: Searchlight of Happy vs. Disgust

67

Figure S6: Searchlight of Sad vs. Fear

Figure S7: Searchlight of Sad vs. Disgust

68

Figure S8: Searchlight of Fear vs. Disgust

Figure S9: Distributions of Interpersonal Reactivity Index Sub-scales

69

CHAPTER 3: INDIVIDUAL DIFFERENCES IN FEELINGS OF CERTAINITY
SURROUNDING MIXED EMOTIONS

Individual differences in feelings of certainty surrounding mixed emotions
Anthony G Vaccaro, Shruti Shakthivel, Helen Wu, Rishab Iyer, Jonas Kaplan

Abstract
Ambivalence is not the same as uncertainty. However, there is variability in how uncertain
individuals feel when experiencing mixed emotions. One way to explore these differences is by
directly investigating the relationship between self-reported affective certainty and mixed
emotions trial-to-trial. In two samples, we tested the relationship between the intensity of co-
occurring mixed feelings and self-reported certainty of those feelings, as well as whether this
relationship was moderated by emotional intelligence and interoceptive awareness. In the first
sample of 140 participants, we found a significant negative relationship between the intensity of
mixed feelings and affective certainty, and that this relationship was lessened in those with
higher emotional intelligence. We next conducted a pre-registered online study with 310
participants in a sample more demographically representative of the United States, and tested
Trait Meta-Mood as a potential moderator for the relationship between mixed feelings and
uncertainty. We replicated our finding that uncertainty was predicted by higher intensity of co-
occurring positive and negative affect, but the relationship was not moderated by emotional
intelligence. Trait meta-mood, however, did moderate this relationship. Individual differences in
subjective certainty of mixed feelings may be relevant to explaining why mixed feelings are
found to be associated with both positive and negative well-being in different individuals.

70

Introduction
Most research in affective science measures valence on a bipolar scale, from negative to
positive (Barrett & Russell, 1998). This approach conflicts with observations that people often
report positive and negative feelings concurrently (Larsen & McGraw, 2014) and often prevents
the detection of these states even when they occur (Schneider & Schwarz, 2017). This
methodology reflects a common oversight in the literature on affect. Without independent
measures of positivity and negativity, ambivalence - the simultaneous experience of positive and
negative - can be mistaken as neutrality in most experimental designs. More concerning, certain
methods and viewpoints conflate ambivalence with uncertainty over how one feels. But
ambivalence is not the same as uncertainty, and it has been shown that one can be quite certain
about ambivalent attitudes (Hong & Lee, 2010; Schneider & Schwarz, 2017). To distinguish
these two phenomena, we refer to the simultaneous experience of positive and negative feelings
as ambivalence, and the uncertainty about one’s feelings as affective uncertainty.
It is well-established that there are substantial individual differences in who experiences
mixed feelings and how they are experienced (Barford & Smillie, 2016; Hui et al., 2009 2009;
Rafaeli et al., 2007 2007). As mixed feelings do not occur very early in the lifespan, they are
likely linked to more developed emotion categorization and emotional intelligence (Vaccaro et
al., 2020 2020). It is possible that some individuals may not properly identify mixed feelings in
themselves, or are less confident in identifying what they are feeling when conflicting feelings
occur. Research suggests that there may be vast differences from person to person in how mixed
feelings relate to positive or negative effects on well-being. In some individuals, mixed feelings
are related to coping, future improvements in mental health, or progression towards goals; in
71

others, they lead to distress and can even be a sign of worse mental health (Berrios et al., 2018b
2017, 2018; Braniecka et al., 2014 & Wytykowska, 2014; Jerg-Bretzke et al., 2013 & Traue,
2013; Losada et al., 2018; Oh, 2022; Verplanken, 2012).
The stark differences found across these studies could be attributed to the difficulty in
discerning ambivalence from affective uncertainty. Part of the reason for the misconception that
ambivalence is affective uncertainty comes from their co-occurrence, but ambivalence which is
associated with uncertainty has different consequences than ambivalence with certainty. It has
been found that ambivalence with and without certainty of those positions leads to different long-
term stability of those feelings (Luttrell et al., 2020 2020; Van Harreveld et al., 2009 Nordgren,
& Van Der Pligt, 2009). Behavioral studies on attitudes have also shown that uncertainty can be
manipulated while ambivalence stays stable, supporting the separability of these constructs
(Petrocelli et al., 2007 2007; Rucker & Petty, 2004). Neuroimaging studies have further shown
that ratings of affective certainty and ambivalence track onto different cortical regions (Luttrell
et al., 2016 & Cunningham, 2016).
How affective certainty relates to cognitive abilities and personality traits is unclear.
Though limited, there is some preliminary work on how individuals assign confidence to the
identification of their own affective states (Vaccaro & Fleming, 2018). This area of work is
largely limited by methodological issues. In most metacognitive paradigms, the goal is to
understand how measures of confidence relate to objective task performance. Without
“objective” markers of affect, it is unclear how to quantify someone’s accuracy in identifying
their own feelings (Fleming & Lau, 2014; Katyal & Fleming, 2023). In the extreme version of
this view, there is even the possibility that one can “misidentify” their own feelings. Even if it is
impossible to judge a true accuracy in feeling identification, the variability between subjects in
72

when and how they report feeling certain of their feelings may shed light on cognitive abilities
and traits.
A related area of work is the study of “emotion granularity”, where paradigms help
identify how discrete subjects’ categorization of their own emotions are. Studies of emotion
granularity have found relationships between low granularity and traits such as alexithymia,
internalizing symptomatology, and anxiety (Aaron et al., 2018 & Park, 2018; Nook, 2021).
However, these paradigms rely on the assumption that less consistency in choosing a distinct
emotion category is always related to worse affective labeling and conceptualization - implying
that it is not possible for an individual to report co-occurring emotion categories and be accurate
when doing so. To study complex feelings, an agnostic first-step is to assume positive and
negative valences may co-occur, and to first analyze subjects’ own ratings of their certainty,
before the granularity of those feelings can be understood.
The traits that explain differences in how certain people are of how they feel may fall
under the umbrella of emotional intelligence. The term emotional intelligence has been used to
explain the socio-affective skills, not explained by traditional cognitive intelligence, that help
individuals thrive and navigate the problems we face in our daily lives. Research on emotional
intelligence has traditionally described the construct as a set of inter-related skills such as
perspective-taking, empathy, emotion perception, emotion regulation, affect labelling, and
mental clarity of one’s own feelings (Beck et al., 2012; Goldsmith & Davidson, 2004; Hall et al.,
1998; Salovey et al., 1995; Schultz et al., 2005). Some, or multiple, of these abilities and traits
may lead to some individuals feeling more certain of their current feelings, even when
experiencing situations which cause complex series of emotions. However, there has not been
consensus in the literature on which skills and traits may overlap with each other, and whether
73

the trait and skill based measures in the lab truly translate to the phenomena outside the lab we
are attempting to explain.
Differences in affective certainty for mixed feelings could explain the diversity in how
and when people experience these states. In previous work, we suggested that individuals with
less interoceptive accuracy and emotional intelligence may be less likely to experience mixed
feelings, and less able to identify how they feel when they do experience them (Vaccaro et al.,
2020). However, no studies have directly tested the relationship between affective certainty and
emotional intelligence. Furthermore, studies on traits such as alexithymia and meta-mood, which
are important for the awareness of feelings, have called upon mixed feelings to be an important
future direction of study but these studies have not yet been conducted (Aaron et al., 2018 &
Park, 2018; Bailen et al., 2019 2019).
Part of the reason mixed feelings have traditionally been understudied is the difficulty in
inducing them in lab settings (Berrios et al., 2015a). A promising type of stimuli is video clips,
which are more robustly able to convey the context required for mixed feelings (Berrios et al.,
2018b; Samson et al., 2016; Smith et al., 2015). In particular, video clips from film and
television may serve to be more useful than the emotional stimuli that have been developed over
the years for laboratory purposes. Conflicting feelings make for engaging plots, and the audience
desire to be ‘moved’ has led to a plethora of film content made specifically with the goal of
inducing mixed feelings (Eder, 2016; Hanich et al., 2014; Oliver, 2008)

STUDY 1: Piloting of Certainty Measure and Mixed Feeling Induction/Measurement
We set out to investigate both the relationship between mixed feelings and certainty, and
how personality traits may moderate this relationship. We first conducted this study with 140
74

undergraduate student volunteers at University of Southern California, in part to pilot whether
subjects would report meaningful variation in affective certainty on a trial by trial basis.
Furthermore, we used this initial study to test various possible methods of calculating the
“mixed-ness” of feelings that have been used in previous literature (Berrios et al., 2015b 2015).

Hypotheses
H1: We expected a general trend that most people would express less affective certainty when
reporting greater mixed feelings, and that this effect would interact with the following individual
factors:
H2: Higher self-reported emotional intelligence would lead to a weakening in the negative
relationship between ambivalence and certainty
H3: Higher alexithymia scores would lead to a strengthening of the negative relationship
between ambivalence and certainty
H4: Higher interoceptive awareness would lead to a weakening of the negative relationship
between ambivalence and certainty

Methods
Participants
Data was collected from 140 undergraduates (92 female, 48 male; mean age= 20.21) at the
University of Southern California who received course credit for their participation.

Stimuli
75

Each subject watched 4-5 videos meant to induce mixed feelings out of 45 potential videos.
These video stimuli ranged in length from approximately 2-10 minutes long, and consisted of
excerpts from movies, television, and animated short films. Four authors on the study chose
these videos as most likely to induce mixed feelings out of a potential videos that were being
considered for a stimuli set.

Procedure
Subjects first completed demographic measures of age, gender identity, race, and ethnicity. Next,
each subject was shown 4-5 different videos. After watching each video, subjects were asked to
complete the following ratings:
1) How familiar were you with this clip before watching? (“Never seen it before”, “Seen it
once before”, “Seen it a few times”, “Seen it many times before”) “How certain are you
that you know how this video made you feel?” (0%, 20%, 40%, 60%, 80%, 100%).
2) How positive did the clip make you feel (1-5 Likert scale)
3) How negative did the clip make you feel (1-5 Likert scale)
4) Check off every emotion you felt from the clip? [Happy, Excited, Love, Amusement,
Fearful, Sad, Angry, Disgusted, Nostalgic, Bittersweet, Conflicted, Awe]

After watching all videos, subjects completed the following measures of individual differences:
1)The Schutte Self-Report Emotional Intelligence Test (Schutte et al., 1998)- a self-report
measure of 4 aspects of emotional intelligence: emotion perception, utilizing emotion, managing
self-relevant emotions, and managing other’s emotions
76

2)The Toronto Alexithymia Scale (Bagby et al., 1994 1994)- a measure of difficulty naming and
describing emotions
3) The Multidimensional Assessment of Interoceptive Awareness (MAIA) (Mehling et al.,
2018)- a trait measure of introspective ability of one’s own bodily sensations

Analysis
Measures of mixed feelings
Part of the difficulty in studying mixed feelings is operationalizing their measurement. In this
first study, we used four different measurements of mixed feelings in order to determine their
comparability and suitability:
1) An interaction effect between positive and negative valence (Berrios et al., 2018a)
2) The minimum value between positive and negative valence in each trial (Kreibig &
Gross, 2017)
3) The Griffin formula for ambivalence: (Positive+Negative)/2 - |Positive-Negative|
(Thompson et al., 1995)
4) The Gradual Threshold (GTM) model of ambivalence: Min(Positive, Negative)
0.5
–
Max(Positive, Negative
)^(1/Min(Positive, Negative))
(Priester & Petty, 1996)

Constructed models
To test each hypothesis, we constructed 4 mixed-effects linear models: one using each of the 4
methods of calculating mixed feelings. The outcome variable of every model was the affective
certainty during the trial, with the 6 options between 0 and 100% being coded as 1-6. Every
model included a random intercept to account for the effect of individual subjects. Models
77

always included positive and negative valence as additional variables to the measure of mixed
feelings.

The models for testing the first hypothesis are as follows:
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Positiveᵢ×Negativeᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Min_Mixᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Griffin_Mixᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃GTM_Mixᵢ + uᵢ
Where uᵢ represents the random effect of each subject.

The models for testing hypotheses 2-4 about the different trait measures, as follows:
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Traitᵢ + β₄Positiveᵢ×Negativeᵢ +
β₅Positiveᵢ×Traitᵢ + β₆Negativeᵢ×Traitᵢ + β₇Positiveᵢ×Negativeᵢ×Traitᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Traitᵢ + β₃Positiveᵢ×Traitᵢ + β₄Negativeᵢ +
β₅Negativeᵢ×Traitᵢ + β₆Min_Mixᵢ + β₇Min_Mixᵢ×Traitᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Traitᵢ + β₃Positiveᵢ×Traitᵢ + β₄Negativeᵢ +
β₅Negativeᵢ×Traitᵢ + β₆Griffin_Mixᵢ + β₇Griffin_Mixᵢ×Traitᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Traitᵢ + β₃Positiveᵢ×Traitᵢ + β₄Negativeᵢ +
β₅Negativeᵢ×Traitᵢ + β₆GTM_Mixᵢ + β₇GTM_Mixᵢ×Traitᵢ + uᵢ
Where uᵢ represents the random effect of each subject.

78

Results
All of the models had a total of 674 observations at the first level, and 140 groups in the second
level of the model: one for each individual subject.

Uncertainty and mixed feelings
In all models testing the relationship between mixed feelings and certainty, greater mixed
feelings predicted less certainty, whereas increases in both positive and negative feelings
independently predicted increased certainty. (Tables 1 and 2; Figure 1)

Table 1: Certainty of affect predicted by positive and negative valence, and their
interaction
Model 1 (Pos-Neg Interaction)
Predictors Estimates CI p
(Intercept) 1.59 1.04 – 2.15 <0.001
Positive 0.77 0.62 – 0.93 <0.001
Negative 0.70 0.51 – 0.88 <0.001
Positive * Negative -0.18 -0.24 – -0.12 <0.001
Random Effects
σ
2
0.96
τ
00

Subject
0.56
ICC 0.37
N
Subject
140
Observations 674
Marginal R
2
/ Conditional R
2
0.143 / 0.458

Results of mixed effects-model (674 observations, 140 groups) predicting certainty of knowing
how you feel. Significance of predictors determined with t-tests using Satterhwaitte’s method
CI= 95% confidence interval.
79

Figure 1: Certainty of affect predicted by positive and negative valence

Predicted certainty of knowing how you feel in a trial, as predicted by positive and negative
valence ratings. When positive ratings are high and negative are low, and vice versa, certainty is
generally higher. When positive and negative ratings are both high concurrently, certainty
begins to decrease.

Table 2: Certainty of affect predicted by positive, negative, and mixed feelings
Model 2 (Min) Model 3 (Griffin_Mix) Model 4 (GTM)
Predictors Estimates CI Statistic p Estimates CI Statistic p Estimates CI Statistic p
(Intercept) 2.62 2.23 – 3.01 13.19 <0.001 2.33 1.92 – 2.73 11.20 <0.001 2.62 2.23 – 3.01 13.19 <0.001
Positive 0.53 0.44 – 0.62 11.58 <0.001 0.50 0.41 – 0.58 11.43 <0.001 0.41 0.33 – 0.48 10.02 <0.001
Negative 0.47 0.36 – 0.59 8.12 <0.001 0.45 0.34 – 0.56 8.11 <0.001 0.35 0.26 – 0.44 7.42 <0.001
Mix -0.49 -0.64 – -
0.35
-6.72 <0.001

Mixed GTM

-0.14 -0.18 – -
0.10
-6.82 <0.001

Mixed Griffin

-0.25 -0.32 – -
0.18
-6.72 <0.001
Random Effects
80

σ
2
0.96 0.96 0.96
τ00 0.53 Subject 0.51 Subject 0.53 Subject
ICC 0.35 0.35 0.35
N 140 Subject 140 Subject 140 Subject
Observations 674 674 674
Marginal R
2
/
Conditional
R
2

0.151 / 0.452 0.151 / 0.448 0.151 / 0.452

Results of mixed-effects-model (674 observations, 140 subjects) predicting certainty of knowing
of how you feel. Significance of predictors determined with t-tests using Satterhwaitte’s method
CI= 95% confidence interval.

Individual differences and the relationship between mixed feelings and uncertainty
While interoception and alexithymia were predictive of certainty in general, they did not have a
significant interaction with mixed feelings in any model (Tables S1 and S2). Emotional
intelligence did not significantly interact with the relationship between mixed feelings and
certainty in the models using the minimum, gradual threshold, or Griffin methods of calculating
mixed feelings (Table S3). Yet, the three-way interaction between positive feelings, negative
feelings, and emotional intelligence significantly predicted increased certainty (Table 3). A
closer examination of this result showed that while increasingly intense co-occurring positive
and negative feelings were associated with decreases in certainty, higher emotional intelligence
led to a weakening of this effect (Figure 1).

81

Table 3: Certainty of affect predicted by valence and emotional intelligence

Predictors Estimates CI Statistic p
(Intercept) 1.69 1.14 – 2.24 6.03 <0.001
Positive 0.75 0.60 – 0.91 9.71 <0.001
Negative 0.68 0.49 – 0.87 7.19 <0.001
SSEIT 0.96 0.41 – 1.51 3.45 0.001
Positive * Negative -0.18 -0.24 – -0.12 -5.82 <0.001
Positive * SSEIT -0.20 -0.35 – -0.05 -2.64 0.008
Negative * SSEIT -0.23 -0.41 – -0.04 -2.38 0.017
(Positive * Negative) *
SSEIT
0.07 0.01 – 0.13 2.23 0.026
Random Effects
σ
2
0.96
τ
00

Subject
0.49
ICC 0.34
N
Subject
140
Observations 674
Marginal R
2
/ Conditional R
2
0.201 / 0.472

Results of mixed-effects-model (674 observations, 140 groups) predicting certainty of knowing of
how you feel. Significance of predictors determined with t-tests using Satterhwaitte’s method
CI= 95% confidence interval. SSEIT= Shutte Self-Report Emotional Intelligence Test

82

Figure 2: Certainty of affect predicted by valence and emotional intelligence
Predicted certainty of knowing how you feel in a trial, as predicted by the subject’s standardized
score on the Schutte Self-Report Emotional Intelligence Test (SSEIT) and their positive and
negative valence ratings. When positive ratings are high and negative are low, and vice versa,
certainty is generally higher. When positive and negative ratings are both high concurrently,
certainty begins to decrease, but this effect is lessened in individuals with relatively higher
emotional intelligence scores.

Exploring the link between emotional intelligence and certainty of mixed feelings
We next explored the 4 sub-scales of the SSEIT to determine if any specific aspect of emotional
intelligence was driving this effect. We found that the Managing Own Emotions subscale and the
Utilizing Emotions subscale had the same three-way interaction effect with positive and negative
feelings that the total SSEIT score had, whereas the Perceiving Emotions and Managing Other’s
Emotions subscales did not (Table S4).

83

STUDY 2: Pre-registered Follow-up Study
We decided to investigate whether our findings would replicate in a larger, more
representative sample of the United States population. A new sample was recruited using the
Prolific online subject pool service. In this study, we added a measure of trait meta-mood, to
investigate whether trait-levels of emotional awareness were related to the trial-by-trial measure
of uncertainty we measured. We also reduced the measures of mixed feelings we used, and
reduced the sample of videos to those most effective at inducing feelings in general, whether
those be positive, negative, or mixed feelings. We pre-registered this study at OSF Registries
(https://osf.io/anbvj)

Pre-registered Hypotheses
Hypothesis 1: Participants will be more uncertain of how they feel when reporting a co-
occurrence of positive and negative valence (mixed feelings).
Hypothesis 2a: Participants with higher emotional intelligence will exhibit a weaker relationship
between mixed feelings and uncertainty.
Hypothesis 2b: Higher scores on the "managing self-relevant emotions" sub-scale of the
emotional intelligence scale specifically will lead to a weaker relationship between mixed
feelings and uncertainty.
Hypothesis 3: The three measures of the meta-mood scale (Clarity, Repair, and Attention) will
all lead to a weaker relationship between mixed feelings and uncertainty.

84

Methods
Participants
Data was collected on Prolific from 311 subjects (149 female, 154 male, 5 non-binary, 2 who
reported ‘other’, and 1 who preferred not to say) in the United States with ages ranging from 18-
65 (mean= 35.08, standard deviation= 11.82). Subjects were paid $7.50 for the 50 minute study.

Procedure
Each subject watched 4 random videos meant to induce mixed feelings out of 28 potential
videos, with each video being evenly distributed across the total sample. After each video,
subjects answered the following questions (see Pages 110 and 111 for normative data about
this stimulus set):
1) How familiar were you with this clip before watching? (“Never seen it before”, “Seen it
once before”, “Seen it a few times”, “Seen it many times before”) This will allow us to
control for film familiarity if this has a strong effect on how mixed people’s feelings are.
2) “How certain are you that you know how this video made you feel?” (0%, 20%, 40%,
60%, 80%, 100%). Including this question during stimuli validation will allow us to pilot
whether this question generates variability in responses.
3) How positive did the clip make you feel (1-5 Likert scale)
4) How negative did the clip make you feel (1-5 Likert scale)
5) In addition, subjects were be able to check off categorical labels for emotions they felt
during the clip. These categorical labels will allow analyses of how emotion labelling
(such as labelling a feeling as bittersweet) relates to individual differences and certainty.

85

After watching all videos, subjects completed the following measures of individual differences,
the first two for our pre-registered hypotheses, and the following 2 for exploratory purposes:
1) The Schutte Self-Report Emotional Intelligence Scale
2) Trait Meta-Mood Scale (Salovey et al., 1995)- A measure of various traits related to
identifying, understanding, and reflecting on one’s own mood
3) The Multidimensional Assessment of Interoceptive Awareness
4) The Ten-Item Personality Index (Gosling et al., 2003)

Analysis
Data exclusion
Following pre-registered procedures, subjects' data was excluded entirely if their total time
taking the survey is less than the total length of all the videos they watched. Observations were
excluded from the Hypothesis 1 analyses if any of certainty, positive valence, or negative
valence is missing. Additionally, in all other analyses, subjects' data were excluded from the
analysis if they did not complete the entire relevant scale. 43 subjects clicked to start the study
but did not complete it, and 7 subjects timed out and did not complete the study. Data was
automatically collected till 311 participants completed the survey.

Measures of mixed feelings
Due to the high levels of similarity between the various measures of mixed feelings in Study 1,
we proceeded with only two: the interaction between positive and negative valence, and the
Griffin formula.

86

Models
The models for testing the first hypothesis are as follows:
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Positiveᵢ×Negativeᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Griffin_Mixᵢ + uᵢ
Where uᵢ represents the random effect of each subject.

The models for testing the rest of the hypotheses about the different trait measures, are as
follows:
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Negativeᵢ + β₃Traitᵢ + β₄Positiveᵢ×Negativeᵢ +
β₅Positiveᵢ×Traitᵢ + β₆Negativeᵢ×Traitᵢ + β₇Positiveᵢ×Negativeᵢ×Traitᵢ + uᵢ
• Certaintyᵢ = β₀ + β₁Positiveᵢ + β₂Traitᵢ + β₃Positiveᵢ×Traitᵢ + β₄Negativeᵢ +
β₅Negativeᵢ×Traitᵢ + β₆Griffin_Mixᵢ + β₇Griffin_Mixᵢ×Traitᵢ + uᵢ
Where uᵢ represents the random effect of each subject.

Model fit was determined using restricted maximum likelihood (REML), and t-tests to evaluate
the effect of variables in the models used Sattherwaite’s method.

Results
Uncertainty and mixed feelings
We again found that mixed feelings had a significant negative relationship with certainty in both
models, whereas positive and negative feelings individually were positively associated with
certainty.

87

Table 4: Certainty of affect predicted by positive, negative, and mixed feelings (Study 2)
Model 1 (Pos-Neg Interaction) Model 2 (Griffin_Mix)
Predictors Estimates CI Statistic p Estimates CI Statistic p
(Intercept) 2.88 2.48 – 3.29 14.02 <0.001 3.59 3.30 – 3.88 24.08 <0.001
Positive 0.60 0.50 – 0.70 11.88 <0.001 0.35 0.29 – 0.40 12.56 <0.001
Negative 0.43 0.31 – 0.56 6.86 <0.001 0.20 0.14 – 0.27 6.04 <0.001
Positive * Negative -0.12 -0.16 – -0.09 -6.59 <0.001

Mix

-0.19 -0.24 – -0.14 -8.07 <0.001
Random Effects
σ
2
0.73 0.72
τ00 0.40 Subject 0.39 Subject
ICC 0.36 0.35
N 311 Subject 311 Subject
Observations 1237 1237
Marginal R
2
/
Conditional R
2

0.127 / 0.438 0.140 / 0.443

Results of mixed-effects-model (1237 observations, 311 subjects) predicting certainty of knowing
of how you feel. CI= 95% confidence intervals. Significance of variables determined with t-tests
using Satterhwaitte’s method

88

Figure 3: Certainty of affect predicted by positive and negative valence (Study 2)

Predicted certainty of knowing how you feel in a trial, as predicted by positive and negative
valence ratings. When positive ratings are high and negative are low, and vice versa, certainty is
generally higher. When positive and negative ratings are both high concurrently, certainty
begins to decrease.

Individual differences and the relationship between mixed feelings and uncertainty
While emotional intelligence itself was predictive of certainty, there was no significant
interaction with mixed feelings in this sample (Table 5). The managing one’s own emotions
subscale was also not a significant predictor, nor was its interaction with mixed feelings (Table
S5).

89

Table 5: Certainty of affect predicted by valence and emotional intelligence (SSEIT)
Model 1 (Pos-Neg Interaction) Model 2 (Griffin_Mix)
Predictors Estimates CI Statistic p Estimates CI Statistic p
(Intercept) 2.99 2.58 – 3.40 14.28 <0.001 3.66 3.37 – 3.95 24.46 <0.001
Positive 0.58 0.48 – 0.68 11.13 <0.001 0.33 0.28 – 0.39 11.90 <0.001
Negative 0.42 0.29 – 0.54 6.44 <0.001 0.20 0.13 – 0.27 5.79 <0.001
SSEIT 0.47 0.09 – 0.86 2.43 0.015 0.36 0.08 – 0.65 2.48 0.013
Positive * Negative -0.12 -0.16 – -
0.08
-6.28 <0.001

Positive * SSEIT -0.08 -0.18 – 0.02 -1.62 0.105

Negative * SSEIT -0.08 -0.20 – 0.03 -1.42 0.157

(Positive * Negative) *
SSEIT
0.02 -0.01 – 0.06 1.29 0.197

Mix

-0.19 -0.23 – -0.14 -7.76 <0.001
Mix * SSEIT

0.01 -0.04 – 0.05 0.39 0.695
SSEIT * Positive

-0.04 -0.10 – 0.01 -1.49 0.138
SSEIT* Negative

-0.03 -0.09 – 0.04 -0.85 0.395
Random Effects
σ
2
0.73 0.72
τ00 0.37 Subject 0.37 Subject
ICC 0.34 0.34
N 311 Subject 311 Subject
Observations 1237 1237
Marginal R
2
/
Conditional R
2

0.166 / 0.450 0.177 / 0.455
Results of mixed-effects-model (1237 observations, 311 subjects) predicting certainty of knowing
of how you feel. Significance of predictors determined with t-tests using Satterhwaitte’s method
CI= 95% confidence interval. SSEIT= Shutte Self-Report Emotional Intelligence Test

Of the meta-mood subscales, the clarity subscale had a significant interaction with mixed
feelings, in which high scores were associated with a decrease in the negative relationship
90

between mixed feelings and certainty (Table 6). Attention and mood repair did not have a
significant interaction effect (Tables S6).

Table 6: Certainty of affect predicted by valence and trait clarity
Model 1 (Pos-Neg Interaction) Model 2 (Griffin_Mix)
Predictors Estimates CI Statistic p Estimates CI Statistic p
(Intercept) 2.87 2.47 – 3.27 14.02 <0.001 3.57 3.28 – 3.86 24.21 <0.001
Positive 0.60 0.50 – 0.70 11.88 <0.001 0.35 0.30 – 0.40 12.72 <0.001
Negative 0.44 0.31 – 0.56 6.86 <0.001 0.20 0.14 – 0.27 6.04 <0.001
Clarity 1.06 0.66 – 1.46 5.22 <0.001 0.71 0.42 – 1.01 4.70 <0.001
Positive * Negative -0.12 -0.16 – -0.08 -6.48 <0.001

Positive * Clarity -0.19 -0.29 – -0.10 -3.85 <0.001

Negative * Clarity -0.23 -0.35 – -0.11 -3.77 <0.001

(Positive * Negative) *
Clarity
0.05 0.02 – 0.09 2.85 0.004

Mix

-0.18 -0.23 – -0.13 -7.62 <0.001
Mix * Clarity

0.05 0.01 – 0.10 2.26 0.024
Clarity* Positive

-0.08 -0.14 – -0.03 -2.85 0.004
Clarity * Negative

-0.11 -0.18 – -0.05 -3.33 0.001
Random Effects
σ
2
0.73 0.72
τ00 0.33 Subject 0.33 Subject
ICC 0.31 0.31
N 308 Subject 308 Subject
Observations 1225 1225
Marginal R
2
/ Conditional
R
2

0.188 / 0.442 0.197 / 0.449

Results of mixed-effects-model (1225 observations, 308 subjects) predicting certainty of knowing
of how you feel. Significance of variables determined with t-tests using Satterhwaitte’s method
CI= 95% confidence interval.

91

Figure 4: Certainty of affect predicted by valence and trait emotional clarity

Predicted certainty of knowing how you feel in a trial, as predicted by the subject’s standardized
Clarity score on the Trait Meta-Mood Scale (at -1.5 standard deviations from the mean, the
average, and 1.5 standard deviations above the mean) and their positive and negative valence
ratings. When positive ratings are high and negative are low, and vice versa, certainty is
generally higher. When positive and negative ratings are both high concurrently, certainty
begins to decrease, but this effect is lessened in individuals with relatively higher clarity scores.

Exploratory analyses of the Big 5:
Extraversion, Openness, and Emotional Stability did not significantly predict certainty in
any model (Tables S7-S9). Agreeableness significantly predicted greater certainty, but did not
significantly interact with positive, negative, or mixed valence (Table S10). Greater
Conscientiousness significantly predicted greater certainty. Additionally, there was a significant
negative interaction between negativity and conscientiousness in predicting certainty (Table
S11).

92

Discussion
Mixed feelings and certainty
Consistently in both samples, we find that mixed feelings are associated with affective
certainty. Specifically, as concurrent positive and negative feelings increase in intensity,
certainty decreases. Research on the relationship between attitudes and behaviors has shown that
individuals who feel ambivalent are less likely to engage in behaviors that are consistent with
their stronger positive or negative attitude (Armitage & Conner, 2000; van Harreveld et al.,
2015; Van Harreveld et al., 2009). However, our data shows that there is variability in the
strength of the relationship between these ambivalent judgements and affective uncertainty.
Previous work has shown that even when subjects report that they hold both positive and
negative attitudes (referred to in the literature as “objective ambivalence”), this does not always
associate with feeling uncertain (referred to as subjective ambivalence) or an internal sense of
conflict (Ng et al., 2022) One prior study found that the slope of the relationship between
objective and subjective ambivalence varied by a magnitude of four- to five-fold across
individuals (Simons et al., 2018). Variability in ambivalence ratings is better explained by
variance in individual subject ratings than by how often specific stimuli are rated as ambivalent,
which demonstrates that this variability comes from differences in cognitive-affective processes,
not from ambiguity in stimuli.
In our study, rather than asking subjects to reflect on their sense of subjective conflict, we
asked them to reflect on how confident they felt that they had accurately identified their own
feelings. This process is by nature metacognitive. The neuroimaging literature has shown
overlaps in brain regions consistently activated for metacognitive judgements of perceptual tasks
and for mentalizing tasks (traditionally related to one’s own beliefs and emotions), suggesting
93

some shared mechanisms (Vaccaro & Fleming, 2018), though recent work has shown distinct
neural representations of uncertainty for each (Jiang et al., 2022). Our results show there is
sufficient variability across individuals in certainty ratings, and that this can partially be
explained by trait differences.
Future work may benefit from linking affect to an objective marker in order to find a
proxy for metacognitive accuracy. It may be possible to calculate a measure similar to
metacognitive accuracy if feelings are linked to either a biomarker, or to a specific behavioral
effect or bias in a task (Katyal & Fleming, 2023). Regarding mixed feelings, prior work has
shown improvement on creativity-related tasks when experiencing mixed feelings as opposed to
solely positive and negative ones, making this a potential behavioral marker that can be studied
in laboratory settings (Fong, 2006; Kung & Chao, 2019). Physiological studies have also had
success in demonstrating distinct features for mixed emotions as compared to univalent ones
(Kreibig et al., 2013, 2015).

Emotional intelligence
In Study 1, emotional intelligence as measured by the SSEIT interacted with the negative
association between certainty and mixed feelings. Specifically higher emotional intelligence was
associated with a weaker version of the effect. The effect was mainly driven by the managing
emotion subscale of the SSEIT. However, this finding wasn’t replicated in our pre-registered
analysis - which had a broader age demographic. A recent study had a seemingly contradictory
finding: higher emotional intelligence was associated with experiencing less mixed feelings in
workplace scenarios, and this effect was mostly driven by the Managing Emotion ability
(Robinson et al., 2020) The authors suggest that the relationship is driven by high emotion-
94

management down-regulating conflicting affect, leading to more univalent states. In our study,
we specifically aimed to induce mixed feelings. It is possible that individuals with high
emotional intelligence do often regulate mixed feelings away, but when scenarios truly call for
them, they are more certain that they feel them, and that they should be feeling them. Further
research should explore this hypothesis in greater detail.

Trait meta-mood and trial-by-trial certainty
In our second study we found that the clarity sub-scale of the Trait Meta-Mood scale
interacted with the negative relationship between mixed feelings and certainty, in line with prior
work (Salovey et al., 1995). Our results demonstrate a link between two similar constructs:
certainty and clarity. Our results suggest that our trial-by-trial certainty measure is probing the
same cognitive function that is traditionally assessed on a trait-level basis. The clarity sub-scale
has frequently been found to blunt the negative effect of stressors on general and life satisfaction
(Extremera et al., 2009; Extremera & Fernández-Berrocal, 2005; Ramos et al., 2007).
Interestingly, many studies on mixed feelings have focused on life satisfaction, often finding
positive effects (Berrios et al., 2018b; Sedikides et al., 2016) while others have found vast
individual differences and even negative effects (Berrios et al., 2017; Newman & Sachs, 2022;
Newman et al., 2019; Oh, 2022; Scheibe et al., 2011; Verplanken, 2012). Clarity may be an
important factor in explaining these individual differences, where reported mixed feelings that
stem from a lack of clarity about one’s affective state are related to distress, whereas clear mixed
feelings can be a sign of processing complicated affective scenarios. Previous studies have found
that mixed feelings which subjects’ feel certain about are more stable overtime than ones for
which they feel uncertain (Luttrell et al., 2020; Luttrell et al., 2016; Tormala & Rucker, 2007).
95

Future studies could test whether this holds true with a paradigm where subjects re-view the
stimuli on later dates. This could give an additional measure of attitude stability, as well as be
used to investigate whether uncertain subjects may become more certain of ambivalent feelings
overtime with additional processing of the stimulus. In this way, the clarity trait could represent a
speed of processing complex affective scenarios, leading to less occurrence of uncertain feelings
in daily life.

Conclusions
This study demonstrates variation in the experiences of mixed feelings both between
subjects, and between individual trials. The results support the anecdotal link between
experiencing ambivalence and being uncertain of how exactly you are feeling. Our analyses
show that various factors moderate this effect, and that some individuals can experience intense
mixed feelings while still feeling fully certain about their affective state. While emotional
intelligence was not a consistent moderator across both samples, certainty ratings are shown to
be a meaningful insight into self-awareness for affect. With this baseline knowledge, more
advanced techniques of measuring metacognition for emotion can be developed. Finally, mixed
feelings are a prime candidate for demonstrating the important interplay between metacognition
and emotion and this relationship should be studied further.
96

97

98

99

100

101

102

103

104

105

106

107

108

Copyright 2023 Anthony Gianni Vaccaro

110

111

CHAPTER 4: MIXED FEELINGS COMPRISE A UNIQUE NEURAL STATE IN
HIGHER CORTICAL REGIONS

Abstract

Affective neuroscience using fMRI has often shied away from focusing on complex affective
experiences such as mixed feelings, leaving a large gap in our understanding of neural
mechanisms. In this study, we had two main goals: 1) to determine if individualized feeling
transitions during a dynamic, naturalistic experience can be predicted with fMRI 2) to determine
if various brain regions and networks differ in their representation of mixed feelings as a state of
consistent neural patterns. Twenty subjects were scanned with fMRI while watching an animated
short film chosen to induce bittersweet mixed feelings. Outside the scanner, subjects then re-
watched the video and were asked to label when they felt positively, negatively, and mixed,
when they were watching the video in the scanner. Hidden-Markov Models were applied to each
subject’s data to determine how accurately transition points predicted from regional activity align
with the subjects’ labels. We find that even with considerable variation in the timing and types of
feelings across subjects, specific affective regions such as the ventral anterior insula, amygdala,
and anterior cingulate could predict the onsets of new feeling states. Furthermore, we found that
insular cortex had unique and consistent neural signatures for univalent states, but not for mixed
valence. Contrarily, ventromedial prefrontal cortex and the anterior cingulate had consistent
neural signatures for both univalent and mixed valence states. This study demonstrates the
feasibility of studying individualized feelings with fMRI, and is the first to show direct evidence
for regional differences in the representation of mixed feelings.

112

Introduction
Mixed feelings, reported as experiencing positivity, negativity, and often a sense of
conflict simultaneously, are common, and occur across cultures (Larsen & McGraw, 2014;
Lomas, 2017; Miyamoto et al., 2010; Moeller et al., 2018b). Events that trigger them are often
ascribed significant meaningfulness (Abeyta & Routledge, 2017; Oliver & Woolley, 2010; Shirai
& Kimura, 2022). Despite their ubiquity, mixed feelings are largely absent from affective
neuroscience research. One reason mixed feelings are rarely studied is their lack of fit with the
dominant ways of thinking about emotion in neuroscience. In the constructionist approach,
bipolar valence is an irreducible dimension of affect, and therefore, positivity and negativity
cannot truly co-occur (Barrett & Bliss-Moreau, 2009; Larsen, 2017; Russell, 2017). In
frameworks that view emotions as stemming from discrete functional biological states (Chang et
al.), the core affective systems, and traditionally defined emotion categories, are all solely
positive or negative (Adolphs, 2016; Kragel & LaBar, 2015b; Nummenmaa & Saarimaki, 2017).
When affective neuroscience research bases experiments in service of finding evidence for or
against these two theories, important topics such as mixed feelings tend to fall between the gaps
(Lench et al., 2013; Vaccaro et al., 2020).
An additional problem with researching mixed feelings is the challenges they pose to our
common methods of research. Traditional emotion research in fMRI relies on the ability to have
subjects feel similarly when experiencing the same stimuli. This increases our analytic power by
allowing us to control for differences between stimuli in order to isolate the quality we are
interested in, and by averaging our data across subjects. Mixed feelings are not as easy to
consistently induce across subjects, especially in the laboratory setting (Berrios et al., 2015a;
Kreibig & Gross, 2017) as they generally require contexts which allow for the contrast between
113

the present view and another point in time (Grossmann & Ellsworth, 2017). Mixed feelings
research can benefit from, and likely requires, utilizing approaches and stimuli that allow the
consideration of individualized experiences rather than consistent responses across participants
(Moore & Martin, 2022; Oh & Tong, 2022; Rafaeli et al., 2007). The recent trend of using more
naturalistic types of stimuli such as film, and finding creative ways to analyze them, is the
perfect environment for encouraging research on mixed feelings (Morgenroth et al.; Saarimäki,
2021). Film stimuli are excellent for inducing mixed feelings as they are dynamic enough to
generate the proper context for emotional conflict between the current moment and either the
past or future context (Grall & Finn, 2022). Perhaps the classic example of inducing mixed
feelings is the bittersweet ending in media (Ersner-Hershfield et al., 2008; Janicke-Bowles et al.,
2021). Research has found that viewers may even value the meaningfulness, and controlled
ability to contemplate life, that comes from bittersweet feelings in entertainment more so than
the often presumed ability to induce purely pleasurable feelings (Hall, 2015; Hall & Bracken,
2011; Oliver, 2008; Oliver & Woolley, 2010; Wulf et al., 2018). Mixed feelings induced by
media may also facilitate coping with ambivalence- research has found that engaging with
narratives that induce these feelings can reduce delay-discounting (Slater et al., 2019) and induce
greater reflection and acceptance of complex emotional issues (Greenwood & Long, 2015; Khoo,
2016).
Hidden-Markov models are particularly suited to analyzing brain activity during film
stimuli. With HMM, we assume that regional activity in the brain shifts through a number of
distinct states, each with its own consistent and unique neural signature (Baldassano et al., 2017).
With this type of analysis, we can test hypotheses about when state shifts occur in relation to the
continuous stimuli, as well as regional-based hypotheses as to the features of the stimuli relevant
114

to shifts in states (Adolphs, 2016; Saarimäki, 2021). HMM approaches have found that state
changes in regions such as posterior medial cortex and prefrontal cortex align well with
meaningful narrative events in film, whereas auditory and visual regions align instead with
lower-level sensory features (Baldassano et al., 2017; Lee et al., 2021; Yates et al., 2022). Recent
studies have begun to apply HMM analyses to fMRI studies of affective experience. In one such
study, states of the ventromedial prefrontal cortex were found to associate with affective states
while watching a TV drama, though the timing of state onsets varied across subjects (Chang et
al., 2021). The heterogeneity of these state change times highlights the importance of analyzing
the neural dynamics of affect on an individualized basis. However, this study assumed the same
number of state changes for each participant, and did not link subject’s individual affective
experience to their specific data. These two limitations raised by the authors are important, and
need to be tested for feasibility if affective neuroscience is to study experiences as individual as
mixed feelings.
In a previous theoretical paper we made various hypotheses about the neurobiological
foundations of mixed feelings (Vaccaro et al., 2020). Our model posits that even if on the
brainstem level, and in subcortical limbic structures, emotions rapidly vacillate, integrative
processes in the anterior insula, then elaborated in higher cortical regions, would result in a
unified feeling. With this model in mind, we may expect differences in the regions where state
changes can accurately predict state changes during a film that tends to induce complex and
mixed feelings. Specifically, insular cortex, ventromedial prefrontal cortex, and cingulate cortex
may be predictive due to their roles in integrating various sources of affective information
(Barbas, 2000; Craig, 2009b). Furthermore, our framework predicts regional differences in
whether mixed feelings are associated with a consistent neural state. If some functional
115

circuitries encode positivity and negativity as mutually exclusive states (Wilson et al.), the neural
patterns when subjects are experiencing mixed feelings should be in flux- this would lead to the
regional patterns at different timepoints of experiencing mixed feelings to be no more correlated
with each other than they are with the regional patterns during the positive and negative states
they are switching between. However, if regions represent the feeling of mixed valence as an
integrated and unique state (as we hypothesize with the anterior insula, ventromedial prefrontal
cortex, anterior cingulate, and posterior cingulate), the regional pattern during mixed feelings
should be correlated with other timepoints of itself than with these positive and negative states
(Vaccaro et al., 2020).
With the present study, we had two main goals: 1) to determine if individualized feeling
transitions during a dynamic, naturalistic experience can be predicted with fMRI 2) to determine
if various brain regions and networks differ in their representation of mixed feelings as a state of
consistent neural patterns. For the first, we used HMM as a data-driven approach to see if
predicted regional state shifts would occur at similar timepoints to individual’s emotion
annotations. For the second, we used subjects’ individual emotion annotations as a ground truth
to analyze whether mixed, positive, and negative states were neurally consistent and unique from
each other.

Methods
Participants
20 right-handed, English-speaking subjects (mean age= 25.7, SD= 7.2; 12 female, 8
male) participated in the study. None had any history of neurological trauma. All provided
informed consent as approved by the USC Institutional Review Board.
116

Scanning parameters
All fMRI scanning was completed on a 3-T Siemens Prisma System Scanner at the USC
Dornsife Cognitive Neuroimaging Center using a 32-channel head coil. Anatomical images were
acquired with a T1-weighted magnetization-prepared rapid gradient-echo (MPRAGE) sequence
(repetition time [TR]/echo time [TE]=2300/2.26, voxel size 2 mm isotropic voxels, flip angle
9°). Functional images were acquired with a T2*-weighted gradient echo sequence (repetition
time [TR]/echo time [TE]= 2000/25 ms, 41 transverse 3-mm slices, flip angle 90°). A T2-
weighted volume was acquired for blind review by an independent neuroradiologist, in
compliance with the scanning center’s policy and local IRB guidelines.

Procedure
While in the scanner, subjects watched the Oscar-nominated animated short One Small
Step (Chesworth & Pontillas, Taiko Studios, 2018). The film tells the story of a young girl who
dreams of being an astronaut, and her shoe-making father who supports her dreams. The father
encourages her dream as a child by making her astronaut boots, and when she gets older she
begins to study astrophysics in college. She struggles in her courses and is initially rejected from
the astronaut program. Throughout this struggle, every time she comes home her father is sitting
at the kitchen table with food ready for her. One day, she returns home and he is not there- we
then see her at his grave, crying. Later, when sorting through his belongings, she finds the old
astronaut boot he had made for her. This rekindles her motivation; she begins to excel at school,
and gets accepted to the astronaut program. We finally see her launch in a rocket ship to the
117

moon, and when she takes the first step onto the surface, the scene pans to her as a child wearing
the astronaut boots playing with her father on her bed.
After scanning, subjects re-watched the video and performed a feeling labelling task
using a custom JavaScript application (http://www.jonaskaplan.com/cinemotion/). Subjects were
instructed to reflect back on when they watched the video in the scanner, and to press buttons to
indicate how they were feeling during that initial watching. Subjects were able to turn feeling
labels on and off to indicate stretches of time where they felt “Positive”, “Negative”, and
“Mixed”, and also had an additional button to indicate any period of time they had cried.
On a second to second basis, subjects were overall most likely to remain in whatever
feeling state they were currently in. Interestingly, negative feelings were nearly never directly
followed by positive feelings the next second, and vice versa. Instead, if changing to a different
state, positive and negative feelings were likely to be followed by either a period of neutrality or
mixed feelings. Mixed feelings and periods of neutrality had relatively similar probability rates
of being followed by any of the other states.

118

Figure 1: Percentage of subject’s feeling positive, negative, and mixed at each timepoint of
the film

Table 1: Probability of feeling state 1 second later based on current reported feeling state

Positive Negative Mixed Neutral
Positive 0.63 0.03 0.23 0.25
Negative 0.01 0.47 0.15 0.16
Mixed 0.2 0.3 0.45 0.23
Neutral 0.17 0.19 0.18 0.36

fMRI preprocessing
Data was pre-processed using fMRIPrep (Esteban et al., 2019). fMRIPrep implements
brain extraction, slice-time correction, standard motion correction, spatial smoothing, and high-
pass temporal filtering. Additionally ICA-AROMA was run for the removal of motion related
components. For each preprocessed subject, we regressed out the effect of white matter, grey
matter, cerebrospinal fluid from the timecourse data. Finally, all subjects’ data were trimmed to
be 454 TRs long, removing the opening 12 seconds of scanning and final 6 seconds. This helped
119

align the neural data while accounting for both the initial 6 seconds delay before the video began,
and 6 seconds of hemodynamic response function delay.

Feeling labels
Each subject’s feeling labelling data was annotated to create individualized transition
points for them. At each second of the clip, we determined which feeling state the subject was in
based on the labels that were turned on. Positive and negative feelings were defined as moments
where either label was on exclusively, while mixed feelings were defined as any moment the
mixed feeling button was on, as subjects differed in whether they tended to use the mixed button
as mutual exclusively to the other two, or tended to turn all three on together when experiencing
mixed feelings. If no buttons were on, this was considered a neutral feeling period. The feeling
labels were smoothed with a 5 second sliding window to account for mistaken button presses,
small gaps of time between new labels, and delays between actual feeling onset and response
time. Timepoints (to the nearest second) where feeling labels changed were then used as that
subject’s specific transitions points.

Regions of interest
In each of our analyses, we extracted voxel-wise timeseries from 10 regions of interest.
We investigated the dorsal anterior, ventral anterior, and posterior subdivisions of the insula
derived from functional connectivity patterns in Deen et al, 2011. We also analyzed the insula as
a whole by merging these three subdivisions (Deen et al., 2011). The anterior cingulate and
amygdala regions were obtained from the was obtained from the Harvard Cortical and
Subcortical atlases. The posterior cingulate region was taken from Shirer et al., 2012. Finally, we
120

used the planum temporale from the Harvard-Oxford cortical atlas as auditory cortex, Brodmann
area 17 thresholded at 50 as early visual cortex.

Matching the mean data’s regional optimal boundaries to individual subject data
All HMM models in this study were fit using the BrainIak python package (Kumar et al.,
2021). The analysis fits an HMM model to the data by iteratively estimating the neural event
signatures for a specific number of events (from here on referred to as k) and the temporal event
structure, till the model converges on a high-likelihood solution (Baldassano et al., 2017). In our
first analysis, we explored the optimal number of states in various regions, agnostic to any
consideration of emotion. With the mean time-course data of all subjects, in each of our regions
of interest, we ran HMM models with different k values ranging from 2-45. For each k-value (the
number of states), we calculated the correlations of spatial signal patterns within predicted
boundaries vs. across boundaries using the method originated in Baldassano, et al., 2017.
Specifically, we calculated the correlation of every timepoints with the timepoint occurring 5
seconds later. The “within” correlation was determined as all of the timepoint correlations where
both timepoints were within the boundaries of a predicted state. The “across” correlation were
between timepoints outside a state and timepoints inside the state. If the period of time indicated
by the state boundaries is a neurally consistent state in that region, timepoints within the state
should be more spatially correlated with each other than with timepoints outside the state, despite
being separated by the same temporal distance. Whichever k-value led to the greatest positive
difference of the within correlation minus the across correlation was considered the optimal
model for that region.
121

Using this optimal k value from the mean data, we then tested whether this same value,
when applied to individual subjects’ data, would find similar boundary locations. In each region,
for each subject, we used the best k value from the mean data in that region. We then compared
the subject’s HMM determined transition points with the transition points from the mean data
using a forward moving algorithm. The first neural boundary found within 5 TRs (TR=1 second)
of the feeling boundary was considered a match. We calculated a matching accuracy rate for
each subject in each region using this method.
To test for statistical significance, we first generated null distributions for every subject in
every region by randomly shuffling the neural events, and randomly generating transition
timepoints. We fit an HMM to these randomly shuffled data, and tested whether the randomly
generated timepoints matched within 5 TRs of the HMM predicted ones, repeating this entire
process 10,000. After the 10,000 permutations, we compared the subject’s true match rate to the
mean of their null distribution to get a metric of how far their match rate was from chance. All
these differences were then averaged across subjects to get one metric of the average difference
between subjects’ match rates and their null match rates in a specific region. We repeated this
process using all 10,000 permutations of each subject's null distribution, resulting in an
aggregated null distribution for the average difference metrics in each region. By comparing the
actual average difference metrics to these aggregated null distributions, we were able to
determine whether the differences between subjects' neurally predicted boundaries were
consistent across regions at a level that was significantly greater than chance.

122

Matching individualized feeling boundaries to predicted neural state changes
For each subject, in each region of interest, we fit a separate HMM model. The K value
for each subject was picked to be equal to the number of feeling transitions plus 1, so the model
would fit the same number of state transitions as were reported subjectively while labelling the
video.
We then compared the HMM-determined transition points with that subject’s specific
feeling transition points using a forward moving algorithm. The first neural boundary found
within 5 TRs (TR=1 second) of the feeling boundary was considered a match. This resulted in an
overall fractional match rate for each region, in each subject, of what proportion of the feeling
boundaries matched neurally predicted boundaries.
To account for the variability in the number of feelings subjects reported, and subjects’
respective neural differences, we generated null distributions for every subject in every region
using the same approach as the previous analysis. Overall, this allowed us to test whether the
neurally predicted boundaries in a region of interest matched up with subjects’ specifically
reported feeling transitions at a rate that was significantly better than chance.

Matching consensus feeling boundaries to predicted neural state changes
We additionally tested the matching rate of each subject’s data with “consensus” emotion
boundaries- an approach more similar to previous HMM fMRI studies (Baldassano et al., 2017;
Chang et al., 2021; Sachs et al., 2023). We defined consensus emotion boundaries as timepoints
where 50% of subjects agreed on their current emotional state for at least 5 seconds. The HMM
analysis, and null distributions, were then fit in the same manner as the analysis of individualized
123

feeling boundaries. Using Wilcoxon signed-rank tests, we then compared the regional matching
accuracy rates for individualized feelings vs. consensus boundaries.

Neural consistency of feelings in different regions
In our final analysis, we aimed to test whether specific regions differed in the consistency
of their neural states corresponding to positive, negative, and mixed feelings. To test this, we
used each subjects’ feeling labels to determine periods of time where they reported feeling
positively, negatively, and mixed. To quantify the consistency of neural states within subject-
reported feeling state boundaries, we used a similar logic to the within vs. across correlation
analysis for each feeling type. Specifically, we compared the difference in the average
correlation of each timepoint of a feeling type with all other timepoints of that feeling type, with
the average correlation of those timepoints correlated with all other timepoints that were of
different feeling types (i.e. the correlation of all timepoints where a subject indicated mixed
feelings with other timepoints where they reported mixed feelings, vs. the correlation of all
mixed feeling timepoints with all positive and negative time points). We will refer to this metric
as neural consistency. We specifically excluded timepoints where subjects did not report any
feeling to avoid our results being driven largely by differences in how correlated affect
timepoints are with neutral timepoints- our interest is in the distinguishability of different types
of affective states. To create null distributions for each subject in each region, we shuffled the
order of every TR 1000 times and calculated this correlation difference each time with the
randomized TR’s. We then calculated a test-statistic by subtracting the subject’s true correlation
difference with the mean of the null distribution. Finally, to create overall null distributions for
each region, we used the same method as in the previous analyses: we calculated an average null
124

test-statistic across all subjects for all 1000 of the permutations, and compared the true average
test-statistic across the subjects to this distribution. Overall, this allowed us to test for an effect
size of each feeling type being neurally more consistent, or less consistent, in each region of
interest at a level that is higher than what would be expected by chance. In other words, we
asked: at a timepoint labelled with a specific feeling type, is the pattern of voxels in a region
more correlated with other timepoints that have the same feeling label than timepoints without
that label?

Results
Overall optimal boundaries in each region
The number of states that led to the highest within vs across correlation difference
differed in reach region. As expected, the optimal fit for sensory cortices involved a higher
number of states than those for other regions. The k values displayed in Table 2 and Figure 2
were used in the next analysis.
Table 2 and Figure 2: Optimal Number of States in Each Region Based on Within vs.
Across Correlations in the Mean Data
Region of Interest k
Dorsal Anterior Insula 17
Ventral Anterior Insula 15
Posterior Insula 3
Ventromedial Prefrontal Cortex 6
Amygdala 5
Anterior Cingulate 3
Posterior Cingulate 16
Visual Cortex (V1) 27
Auditory Cortex 28

125

Matching the regional optimal boundaries to individual subjects’ data
These mean-data derived boundaries significantly matched with boundaries predicted
from individual subjects’ data in the dorsal anterior insula (35% accuracy, p=0.046), ventral
anterior insula (42.1% match, p<0.0001), posterior insula (40.7% match, p<0.0001), anterior
cingulate (10% match, p=0.016), posterior cingulate (37.3% match, p<0.0001) visual cortex
(33.4% match, p=0.005), and auditory cortex (32.1% match, p=0.032). Boundaries predicted
from the mean of the group’s data did not match individually predicted boundaries in the
amygdala (10% match, p=0.19) or ventromedial prefrontal cortex (10% match, p=0.78).

Matching individual’s feeling transitions to predicted neural state changes
HMM-predicted boundaries matched subjects’ individual boundaries in the full insula
(p=0.02 33.2%) ventral anterior insula (29.3% match, p=0.019), amygdala (30% match,
p=0.009), and anterior cingulate (29% match, p=0.02) at rates significantly better than would be
expected by chance. The dorsal anterior insula (27.9% match, p=0.06), posterior insula (22.2%
match, p=0.80), ventromedial prefrontal cortex (26.9% match, p =0.13), posterior cingulate
0
5
10
15
20
25
30
Dorsal Anterior Insula
Ventral Anterior Insula
Posterior Insula
Ventromedial…
Amygdala
Anterior Cingulate
Posterior Cingulate
Visual Cortex (V1)
Auditory Cortex
Optimal number of states in each region of
interest
126

(23.2% match, p=0.35), visual cortex (21% match, p=0.65), and auditory cortex (26% match,
p=0.18) did not significantly match with individuals’ self-report.

Matching consensus feeling transitions to predicted neural state changes
HMM-predicted boundaries matched consensus feeling transitions in the full insula
(22.3%, p=0.049) and posterior cingulate (19.2% match, p=0.04) at rates significantly better than
would be expected by chance. No other regions reached significance.

Comparing individual feeling matching rates to consensus feeling matching rates
Matching rates were significantly greater for individualized feelings vs. consensus feeling
boundaries in the full insula (average match 33.2% vs 22.3%, p=0.011), of dorsal anterior insula
(27.9% vs. 14.3%, p=0.011), ventral anterior insula (29.3% vs. 19.3%, p=0.034), amygdala
(30.0% vs. 17.8%, p=0.037), anterior cingulate (29% vs. 17.1%, p=0.022), and ventromedial
prefrontal cortex (26.9% vs. 16.4%, p=0.026). There was no significant difference between
individual matching and consensus matching in the other regions.

127

Figure 3: Significance of Matching Rates in the Anterior Insula Using Self-Reported
Boundaries vs. Consensus Boundaries

Average difference of the boundary matching rate when matching data to individual feelings vs.
consensus feelings. Significance was determined by comparison with 10,000 permutations of
shuffling TRs and boundaries. Green dot indicates a matching rate significantly greater than the
null mean. Green bar indicates a significant difference between matching rates.

Figure 4: Significance of Matching Rate in the Amygdala Using Self-Reported Boundaries
vs. Consensus Boundaries

Average difference of the boundary matching rate when matching data to individual feelings vs.
consensus feelings. Significance was determined by comparison with 10,000 permutations of
128

shuffling TRs and boundaries. Green dot indicates a matching rate significantly greater than the
null mean. Green bar indicates a significant difference between matching rates.

Figure 5: Significance of Matching Rate in Anterior and Posterior Cingulate Using Self-
Reported Boundaries vs. Consensus Boundaries

Average difference of the boundary matching rate when matching data to individual feelings vs.
consensus feelings. Significance was determined by comparison with 10,000 permutations of
shuffling TRs and boundaries. Green dot indicates a matching rate significantly greater than the
null mean. Green bar indicates a significant difference between matching rates.

Neural consistency of positive, negative, and mixed feelings in different regions

In the dorsal anterior insula, timepoints reported as positive were more significantly
correlated with each other than with negative or mixed timepoints (p=0.002) This was not the
case for negative (p=0.523) or mixed (p=0.872). In both the ventral anterior insula and posterior
insula, neural consistency was significant for negative feelings (p=0.002 and p=0.032
respectively), but not for positive (p=0.419 and p=0.383) or mixed (p=0.121 and p=0.073).
Analysis of the insular cortex as a singular region found significant neural consistency for
positive (p=0.007) and negative states (p=0.002), but not mixed (p=0.691).In the amygdala, only
negative states (p<0.001) were significantly more neurally consistent than would be expected by
129

chance (p=0.326 for positive; p=0.841 for mixed). The vmPFC was significant for all three
feeling states (p<0.001 for all), as was the ACC (p<0.001 positive; p=0.024 negative; p<0.001
mixed). The PCC reached significance for negative (p<0.001) and mixed (p=0.04) but not for
positive (p=0.784). In auditory cortex positive (p<0.001) and mixed (p<0.001) reached
significance, whereas negative did not (p=0.118). In visual cortex, all three reached significance
(p<0.001).

Figure 6: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Insular Sub-Regions

130

Average difference between neural consistency of positive (red), negative (blue) and mixed
(purple) timepoints across subjects, and the mean of individual subjects’ null distributions.
Neural consistency is calculated as the average correlation between timepoints of the same
feeling type subtracted by the average correlation of timepoints of that feeling type with other
feeling types. Null distributions are made by 1000 permutations of shuffling TRs in each subject.
Significance at p<0.05 is indicated by the true mean dot being green. In the top image, the
dorsal anterior, ventral anterior, and posterior masks are show in orange, blue, and green
respectively.

Figure 7: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Ventromedial Prefrontal Cortex and Amygdala

Average difference between neural consistency of positive (red), negative (blue) and mixed
(purple) timepoints across subjects, and the mean of individual subjects’ null distributions.
Neural consistency is calculated as the average correlation between timepoints of the same
feeling type subtracted by the average correlation of timepoints of that feeling type with other
feeling types. Null distributions are made by 1000 permutations of shuffling TRs in each subject.
Significance at p<0.05 is indicated by the true mean dot being green. Amygdala mask is shown
in green and ventromedial prefrontal cortex mask is shown in orange.

131

Figure 8: Average Difference in Neural Consistency from Null Distribution for Positive,
Negative, and Mixed Feelings in Cingulate Regions

Average difference between neural consistency of positive (red), negative (blue) and mixed
(purple) timepoints across subjects, and the mean of individual subjects’ null distributions.
Neural consistency is calculated as the average correlation between timepoints of the same
feeling type subtracted by the average correlation of timepoints of that feeling type with other
feeling types. Null distributions are made by 1000 permutations of shuffling TRs in each subject.
Significance at p<0.05 is indicated by the true mean dot being green. Anterior cingulate mask is
shown in green and the posterior cingulate mask is shown in orange.

Discussion

In this study, we aimed to use individualized feelings dynamics to understand the neural
basis of mixed feelings. In the process, we show evidence that focusing on the relationship
between individual affect and individual’s data, rather than using assuming consensus emotional
features based on the stimuli, leads to unique results. Overall, we found that brain state changes
in the insular cortex (in particular anterior regions), anterior cingulate, amygdala, and
ventromedial prefrontal cortex match significantly better with transitions that are defined by
individual’s self-report than by a consensus measure of the stimuli’s affective features. All of
132

these regions have been classically associated with interoception and salience processing,
representing crucial aspects of affective experience (Critchley & Garfinkel, 2017; Damasio &
Carvalho, 2013; Saarimaki et al., 2016; Seeley, 2019). Importantly, these findings cannot be
explained by a general lack of consistency in any aspect of neurocognitive processing, or by the
self-reported emotion transitions reflecting simply low-level stimulus features: our analysis on
the consistency of optimal boundaries found significant consistency in the timing of brain state
changes across subjects in many regions, but these boundaries were not the same as the self-
reported affect-related ones, did not match at the same rates, and significantly matched in many
regions that did not end up significantly matching with self-report, such as the posterior insula,
visual cortex, and auditory cortex.
While we often use rely on assumed consistency of emotional features across subjects
due to time constraints in collecting self-report data, and to increase the power of our studies,
defining stimuli in this manner likely shifts what our measures are studying further from the
neural correlates of conscious emotional experience, and closer to other cognitive and stimuli-
based features. An important advantage of naturalistic stimuli, when we have the corresponding
self-report data, is that we are encouraged to treat the “affect” in our study not just as a feature of
the stimuli, but as a feature that is also defined by the subject themselves (Adolphs, 2016;
Saarimäki, 2021). Interestingly, the one region where the matching rate was consistently greater
using consensus boundaries than self-reported boundaries was the posterior cingulate, which had
a significant matching rate on its own, but the direct comparison with the individualized
matching rates did not reach significance. The posterior cingulate has been implicated in many
studies of narrative processing, and appears to have neural dynamics which correspond more
closely with higher meaning than with low-level sensory features (Baldassano et al., 2017;
133

Dehghani et al., 2017; Kaplan et al., 2017; Wu et al., 2023). It is possible that when defining
emotion transitions based on consensus, the boundaries we end up with more closely correspond
with scene changes and narratively important moments in the film- smoothing over emotional
idiosyncrasies between subjects. These narrative understandings may be consistent even when
affect is not.
Our analyses of the neural consistency of positive, negative, and mixed feelings had,
fittingly, mixed results in terms of our initial hypotheses. We proposed that since the insular
cortex was involved in integrating our current affective moment it would have consistent neural
states during mixed feelings as the key region involved in integrating the positive and negative
inputs. The subcortical structures, such as the amygdala, that project to it would not have
consistent neural states during mixed feelings, and the anterior cingulate and prefrontal regions,
which the insular cortex projects to, would fully flesh out the phenomenological experience of
mixed feelings and thus have a neurally consistent state for it (Vaccaro et al., 2020). The key
deviation from our hypotheses was the finding that the insular cortex had consistent neural
signatures for both positive and negative states, but not for mixed. The insular cortex has a role
in integrating sources of affective information from multiple sources, and has a strong
relationship with forming our subjective sense of the present moment (Craig, 2009a, 2009b; Kent
& Wittmann, 2021; Picard & Kurth, 2014). While we hypothesized that the dynamics on the
temporal scale of seconds reflected in BOLD activity would be consistent and unique in this
region when integrating mixed valence, with the unique pattern of activity reflecting how this
integration process differs when low-level affective processes vacillate between positive and
negative, it is possible that the neural fluctuations on the milliseconds scale are still reflected in a
more in-flux BOLD response.
134

The anterior cingulate and ventromedial prefrontal cortex, the two regions we proposed
as being crucial for elaborating on the experience of mixed feelings, had significant neural
consistency during mixed feelings, as well as during positive and negative feelings. The
juxtaposition between positive and negative feelings arising during same situation often leads to
a sense of conflict (Berrios et al., 2015b; Mejia & Hooker, 2017). Activity in the anterior
cingulate has long been linked to the experience of conflict and ambivalence, likely related to its
proposed role in error monitoring (Cunningham et al., 2004; Luttrell et al., 2016). In the context
of mixed feelings such as bittersweetness, the anterior cingulate may similarly allow the
utilization of information about conflicting goals to be used in making complex decisions and
self-regulation (Kruschwitz et al., 2018; Nohlen et al., 2014; Yang, Wildschut, et al., 2022). The
posterior cingulate was also had a consistent neural signature for mixed feelings and negative
feelings, however, it was not significantly neurally consistent for negative feelings, leaving room
for the potential argument that the stability of the mixed feeling state is mostly driven by the
component negativity. The posterior cingulate has been implicated in many studies of nostalgia
as being important for the contextual and autobiographical processes involved in setting the stage
for the bittersweetness of nostalgic remembrance (Yang, Izuma, et al., 2022; Yang, Wildschut, et
al., 2022).
The ventromedial prefrontal cortex may play an important role in the representation of
mixed feelings. Previous studies have found the orbitofrontal portion of ventromedial prefrontal
cortex to show unique activations in response to conflicting affective information (Becker et al.,
2014; Rolls & Grabenhorst, 2008; Simmons et al., 2006). More broadly, the ventromedial
prefrontal cortex has been proposed to integrate affective information from the body with other
sources of information, such as memories and conceptual knowledge (Barbas, 2000; Rolls &
135

Grabenhorst, 2008; Roy et al., 2012). The integration of these various modalities may play a
pivotal role in the generation of complex feelings. The one small-scale fMRI study on reactions
to bittersweet clips, with a sample of 10 subjects, study, found a unique sub-region of higher
activation when compared to positive and negative clips (Schulte et al., 2012). The high
distinctiveness of the three types of feelings in ventromedial prefrontal cortex also aligns with
the somatic marker hypothesis (Damasio, 1996). Even if the various physiological patterns in
insular cortex and lower are distinctively positive or negative, the vmPFC may facilitate the
learning and representation of this fluctuating somatic pattern, leading to an emotional marker
that is stable, yet representing mixed valence (Dunning et al., 2017; Panksepp, 2005)
Sensory cortices also were also strongly neurally consistent for affective states, with the
auditory cortex being significant for positive and mixed states, while the visual cortex was
significant for all three. Previous studies have found that the valence and emotion category of
visuo-auditory stimuli can be predicted remarkably well from patterns of activity in lower-level
sensory cortices (Bo et al., 2021; Ethofer et al., 2009; Kragel et al., 2019; Sachs et al., 2018).
Specific visual and auditory cues may have associations with specific types of affect. These
types of features may additionally highlight a missing element of using media as ‘naturalistic’
stimuli. Recent reviews have pointed out that labelling media stimuli as ‘naturalistic’ may lead
us to fail to consider the deliberate techniques used in media construction to induce specific
experiences (Grall & Finn, 2022; Schmaelzle & Huskey, 2023). For sensory features qualities
such as volume, brightness, etc, the inclusion of these features as regressors in neuroimaging
analyses may help future studies understand how sensory features contribute to forming affective
experience.
136

Beyond the sensory domain, these reviews also highlight that content analysis for why
certain narrative and scene elements are designed the specific ways they are in pieces of media
may help further parse out the specific cognitive processes involved in the induced affective
experience. Future studies could investigate whether specific neural processes relate to the
cognition involved in developing the narrative experience of the ‘bittersweet ending’ we used as
a tool for inducing mixed feelings. Furthermore, it has been suggested that the poignancy and
induction of mixed feelings in media is largely related to the ability to resonant with meaningful
elements of the viewer’s life (Klimmt & Rieger, 2021). The level of autobiographical relevance
may modulate various neurocognitive processes, and could be investigated in future studies
which integrate personal interviews.

Conclusions
Our findings suggest that mixed feeling states are neurally consistent in many cortical
regions, indicating that they may not simply be the result of fluctuation in feeling, and are not
just noise in self-report. Furthermore, these results underscore the importance of considering
both individualized experiences when investigating the neural correlates of affect, as it may
capture a different and crucial element of affective processing. There are difficulties and pitfalls
to studying parts of affect as individualized as mixed feelings, but these obstacles should not
stand in the way of building a complete picture of the neurobiology of affect. The fact that
mixed, positive, and negative feelings are associated patterns of activity that are unique from
each other, and consistent, in anterior cingulate and ventromedial prefrontal cortex is important
evidence for a neurobiological state related to mixed feelings that is not characterized as simply
being a switching back and forth between a positive and negative state. While hypotheses related
to whether the psychology of mixed feelings involves vacillation or a stable mixed state may
137

never be 100% falsifiable (Larsen, 2017), our data shows a neurobiological correlate to a
sufficiently stable state which was defined through self-report. On at least this level of analysis,
we can make a strong case for viewing mixed valence as a relevant concept for affective
neuroscience.
138

CHAPTER 5: CONCLUSIONS

Overview

The studies presented here demonstrate that mixed and ambiguous feelings are
researchable, and that creative analysis approaches can help mitigate the issues that are
traditionally feared with studying them. In Chapter 1, I explored the concept of valence on
various biological scales. Overall, I find that definitions of valence do not center well around a
singular concept, but that different biological levels differ in how they handle an ambivalent
world based on how detailed, varied, and independent their encodings of positivity and
negativity are. I then take the leap to propose how these lessons might inform where ambivalence
fits into our neurobiological models of affect (Vaccaro et al., 2020).
In Chapter 2, I aimed to tackle an important problem in relating affective categories to
neurobiological data. While multivariate methods have begun to show promise at identifying
biological patterns which may correspond to distinct affective concepts, these methods also leave
a problem of making the “fuzzy boundaries” between categories difficult to interpret (Gessell et
al., 2021a; Kragel & LaBar, 2015b, 2016; Nummenmaa & Saarimäki, 2019; Ritchie et al., 2019).
Without the ability to explain categorical boundaries in affect, we can become lost in the
theoretical woods before the chance to explore what it means for categories of affect to “mix”. I
conduct a study that utilizes a new approach, centered on individual differences in dispositional
experiences, to explain variance in how distinct the neurobiological patterns associated with
affective categories are. The results showed that in ventromedial prefrontal cortex and insular
cortex subject-level differences in categorical distinctiveness can be explained by their
perspective-taking tendency (Vaccaro et al., 2022). The study also hints that many of our
neuroimaging analyses that use abstract stimuli may not be measuring the neurobiology of affect
139

as much as they are measuring the ability of subject’s to generate affect willfully, by taking the
perspective we are asking them to.
In Chapter 3, I research the relationship between mixed feelings and the common refrain
of “I don’t know how I feel”. The results show that as the intensity of mixed feelings increases,
in general, subjects’ uncertainty of how they feel also increases. However, individual’s with
certain traits, such as a tendency for emotional clarity or emotional intelligence, appear to be
more certain of their mixed feelings.
Finally, in Chapter 4, the lessons of the previous 3 chapters are put together in an
experiment that aims to test the theories put forward in Vaccaro, et al., 2020. By using a poignant
animated film to assure the induction of strong feelings, and analyzing all the data on an
individual basis in regards to both the neurobiology and the psychological experience, I find that
the onsets of individual feelings can be predicted from neural activity with a margin of error of
just 5 seconds. Furthermore, the analyses show that mixed feelings are not represented as a
consistent and unique feeling in the insular cortex, contrary to our original hypothesis. Aligned
with our previous hypotheses, the ventromedial prefrontal cortex and anterior cingulate
represented mixed feelings as a consistent and unique pattern compared to positivity and
negativity. These data are the first to show a temporally consistent state in the cortex that we can
point to as relevant to the experience of mixed feelings.

Affective pathways to ambiguity and ambivalence
Throughout this dissertation, the studies and data make an argument for the fields of
neuroscience and psychology to make room for mixed and ambiguous feelings in their theories
and models. Historically, the idea of emotional ambiguity has been used to sidestep the
140

possibility of mixed feelings: we either do not have enough information to know what we are
feeling, or all the information does not belong to a discrete conceptual emotion depending on
your viewpoint of affect (Fang et al., 2018; Hoemann et al., 2017; Kafetsios & Hess, 2019;
Mattek et al., 2017; Murray et al., 2023; Tay & Kuykendall, 2017; Young, 1918). To further
complicate this matter, mixed feelings often do feel subjectively ambiguous to us (Livet, 2010;
Mejia & Hooker, 2017; Vazard, 2022). But sometimes they don’t. Put together, the data I have
collected may suggest an explanation for why even though ambivalence and emotional
ambiguity are distinct, they are so closely related.
In chapter 2 of this dissertation (Vaccaro et al., 2022), the data show that individuals who
can better take the perspective of the abstract stimuli, and willingly simulate that state, have
more distinctive neural patterns in ventromedial prefrontal cortex underlying those different
feelings. In line with the somatic marker hypothesis and simulationist views of empathy, better
perspective takers are better able to self-induce these affective states based on learned
associations of the bodily/internal that underlie them (Benoit et al., 2019; Damasio, 1996; Decety
& Grèzes, 2006; Hynes et al., 2006). Studies have shown the importance of this ability in
decision-making through its use in weighing and imagining the affective consequences of future
scenarios, and have found distinct representations for differently valenced imagined future affect
(Bechara et al., 1994; Bertossi et al., 2017; Devitt et al., 2022; Janowski et al., 2013).
The individuals who were not as able to perspective-take failed to “generate” enough
information to represent a consistent and stable representation of the targeted affective state, and
so the patterns were less distinguishable. In settings where these affective states naturally arise,
perhaps their representations would have been more distinct, triggered by multiple internal and
perceptual signals that inform them of the current state of their body, and allow for adequate
141

integration of that information into a recognizable, and familiar, somatic state (Chang et al.,
2021; Delgado et al., 2016; O'Doherty et al., 2001; Roy et al., 2012; Todd et al., 2020). Studies
have also linked activity in ventromedial prefrontal cortex with the vividness of emotional states,
further demonstrating the link between the integration of affective information and the
instantiation of distinct represented patterns (D’Argembeau et al., 2008; Lin et al., 2016; Todd et
al., 2015). However, when we are discussing affective categories such as pleasure, sadness, fear,
and disgust, even the integration of information from various cognitive and physiological
streams is relatively ‘simple’. These feelings are consistently brought about by consistent fixed-
action programs in brainstem mechanisms, and behaviorally motivate us in consistent manners
through affective modules (Bandler & Shipley, 1994; Berridge, 2019; De Oca et al., 1998;
Mobbs et al., 2020; Stephens et al., 2010; Venkatraman et al., 2017). Subcortical regions have
additionally been found to be more selectively activated for different categories of affect than
cortical regions which tend to be active during a wider range of affective states (Saarimaki et al.,
2023). We experience these bodily events from extremely early in development, and as soon as
we have the cognitive ability to do so, can develop consistent representations of these feelings
(Atzil & Gendron, 2017; Fischer et al., 1990; Fotopoulou & Tsakiris, 2017; Gergely & Watson,
1999; Ruba & Repacholi, 2020; Saxbe et al., 2013).
The physiological and cognitive signals related to mixed feelings, in contrast, are
complex (Kreibig et al., 2013; Voutilainen et al., 2014). Fixed action-programs rapidly vacillate
(Berntson & Cacioppo, 2008; Berridge & Grill, 1983; Kim et al., 2016; Miller & Kraeling, 1952;
Redgrave et al., 1999). Affective modules in subcortical nuclei are modulated in manners that are
less consistent with what we would consider a single valence, with appetitive and aversive
motivations coinciding (Faure et al., 2010; Norman et al., 2011; Reynolds & Berridge, 2002;
142

Reynolds & Berridge, 2008). Chapter 4 of my dissertation demonstrated that while the
spatiotemporal patterns of the insular cortex for solely positive or solely negative states were
consistent and unique, the mixed feeling state was not. With the insula integrating information
from all these lower level inputs, with all their inherent vacillation, and unusual combinations of
nuclei modulations, the spatial patterns are not consistent overtime: they are a mess. This messy
nature may increase the chances of experiencing this state as ambiguous, as more cortical
cognitive processing will likely be needed to formulate it as a consistent feeling. Chapter 3 of my
dissertation shows that while the stronger univalent affect leads to less uncertainty, stronger
intensities of co-occurring mixed valence has the opposite effect. Research has suggested that
insular cortex and cingulate cortex become more active when processing ‘atypical’ affective
experiences, and this may directly relate to the complexity of these incoming signals (Murray et
al., 2023; Nohlen et al., 2014; Wilson-Mendenhall et al., 2014).
The anterior and posterior cingulate are the key players in sorting out the nature of these
experiences. The anterior cingulate has been suggested to be involved in the monitoring and
detection of motivational conflict; its role in processing ambivalence has been extensively
documented (Cunningham et al., 2004; Etkin et al., 2011; Joyce & Barbas, 2018; Nohlen et al.,
2014). The region has key bidirectional connections to the insular cortex, ventromedial
prefrontal cortex, and posterior cingulate, which enable it to engage with a wide-range regions ,
in facilitating evaluation and decision-making under conflicting and ambiguous scenarios (Cauda
et al., 2011; Joyce & Barbas, 2018; Stolyarova, 2018). In my study, the anterior cingulate
represented mixed feelings, as well as positive and negative ones, as stable and unique states,
which we had originally hypothesized due to the nature of mixed feelings (Vaccaro et al., 2020).
Less discussed is the neural evidence for distinctions between ambivalence processing and
143

feelings of certainty: studies have demonstrated that while the anterior cingulate is more related
to the general processing of ambivalence, during similar tasks increased posterior cingulate
activity is associated with increased certainty, and less felt choice ambiguity (Bach et al., 2011;
Luttrell et al., 2016; Nohlen et al., 2014; Sun et al., 2017). I will take the leap to speculate that
the posterior cingulate is not actively tracking subjective certainty; it is possible that its
correlation with higher certainty is driven by the cognitive processes needed to successfully
contextualize these messy and less typical internal signals.
The ability to link these complex bodily signals to contextual information from
autobiographical memory, determines whether we experience the state as ambiguous or certain
ambivalence. If we have the conceptual and autobiographical information to contextualize this
bodily signal, we can experience a consistent ambivalent state (Bradley et al., 2022; Klasen et al.,
2011; Lin et al., 2016; Takano & Nomura, 2022; Thagard et al., 2023; Vaccaro et al., 2022;
Wilson-Mendenhall et al., 2014; Yang, Wildschut, et al., 2022). I hypothesize that this state
would be clearly distinguishable using classification methods, and neurally consistent like the
mixed feeling state found in my study. If we do not have the relevant information needed to point
to this bodily signal as a consistent somatic marker, we experience a sense of ambiguity, which I
hypothesize may involve less distinguishable and consistent spatiotemporal patterns for affective
states (Glenn et al., 2020; Grupe & Nitschke, 2013; Hennings et al., 2019; Hogeveen et al., 2021;
Lee et al., 2017; Liemburg et al., 2012; Sharpe et al., 2015; Simmons et al., 2006). Large-scale
classification studies of affect categories have found that patterns for feelings traditionally
viewed as mixed are less distinguishable in midline and prefrontal regions than more ‘basic’
states (Kassam et al., 2013; Saarimaki et al., 2018). I hypothesize that the inconsistency of these
patterns may reflect two different elements 1) the differences between subjects in the formation
144

of these types of feelings through distinct autobiographical information 2) the ambiguity of
experience in some of these subjects leading to a less spatiotemporally consistent pattern of
activity. These hypotheses could be tested by relating subjective uncertainty of affective states to
classification accuracy of individual subjects’ mixed feelings. If classification accuracy of mixed
valence states in ventromedial prefrontal cortex (especially compared to positive and negative
ones) is associated with the degree of certainty, it would show a link between subjective
ambiguity and representational ambiguity. Additionally, during these less ambiguous mixed
states, we may expect heightened connectivity of the posterior cingulate with anterior cingulate
and ventromedial prefrontal cortex, reflecting the subjects’ ability to contextualize their complex
feeling. The ability to contextualize such complex feelings may be important for well-being, as
somatic representations of these states may facilitate emotion regulation through the
ventromedial prefrontal cortex’s connections with subcortical descending pathways (Babalian et
al., 2018; Kruschwitz et al., 2018; Venkatraman et al., 2017; Weissman et al., 2018).

145

Figure 1: The generation of an ambiguous or ambivalent feeling

Visual summary of the neurocognitive processes leading to either ambiguous, or clear
ambivalent, affective states. Made with BioRender.com

Beyond improving our knowledge of emotion, these experiences are important because of
their relevance and ubiquity in daily life, and in explaining the complex issues that affect us.
Naturally, future research on these experiences can benefit some of our most pressing societal
issues.

Health and well-being

It is frequently mentioned in various chapters that the relationship between mixed
feelings and well-being is unclear. Studies in recent years have increasingly looked at the role of
mixed feelings in emotional well-being, with a possible trend emerging that mixed feelings that
146

are developed intentionally may be beneficial, whereas mixed feelings which occur frequently in
response to life stressors may cause emotional distress (Newman et al., 2019; Oh, 2022).
Furthermore, large life transitions often bring about mixed feelings, and much how well we
handle these transitions is directly related to our experience and conceptualization of this
ambivalence (Agarwal et al., 2020; Dabb et al., 2022; Kotter-Grühn et al., 2009; Scheibe &
Freund, 2008). Studying mixed feelings can bridge a divide between the most complex
experiences we face and neuroscience. In one recent example of this, a clinician and a
neuroscientist discussed how the dominant models neurobiological models of addiction do not
adequately relate to the affective experiences reported by their patients (Vandaele & Daeppen,
2022). They instead point to our theory on how ambivalence fits into affective neuroscience as
pointing out a promising model where neuroscience research would align more with clinical
relevance.

Figure 2: The neurobiology of ambivalence in regards to affective experience in addiction

Neurobiological model of ambivalence from Vaccaro, Kaplan, Damasio, 2020 applied to the
context of addiction in Vandaele & Daeppen, 2022. Figure by Vandaele & Daeppen

147

Emotional effects of media
There is massive pressure to understand media from an affective perspective (Goldenberg
& Gross, 2020; Wahl-Jorgensen, 2019). Social media has been associated with increased mental
health issues and feelings of social isolation in adolescents, while also being shown to facilitate
feelings of social connection (Beyens et al., 2020). At the same time that internet gaming
disorder is being added to the ICD-11 and viewed as a negative coping mechanism (Blasi et al.,
2019; King & Potenza, 2019; Plante et al., 2019), other research has found some games to have a
positive effect on well-being, appearing to be a positive form of escapism (Johannes et al., 2021;
Wulf et al., 2020; Wulf et al., 2022). Beyond an explanation of differential effects between
individuals, I think an important aspect of understanding the affective experiences brought upon
by media is that they are often complex and mixed. Behaviors of escapism and meeting social
needs through digital spaces inherently both increase positive affect and exacerbate pre-existing
negative affect by being a partially effective substitution for various needs of connectedness and
emotional safety (Hussain et al., 2021; Jarman et al., 2021; O’Day & Heimberg, 2021; Roberts &
David, 2020). These contrasting parts of the emotional experience can be difficult not only for us
to measure, but for the individuals involved to feel they have a full grasp over how they feel
(Halfmann & Reinecke, 2021; Lopez-Fernandez et al., 2022; Lupinacci, 2021; Yan et al., 2022).
If neuroscience is to be useful in this space, we have to use models of affective neuroscience
which allow for ambivalence.

148

Affective technology
In recent years, there has been an increase in technology that is designed to recognize our
emotional experience, and use that information for its task completion (Braun et al., 2019;
Chakriswaran et al., 2019; Marín-Morales et al., 2020). In fact, many of the technologies being
currently developed aim to use physiological measures to understand our affective experience
(Bota et al., 2019; Torres et al., 2020; Wang et al., 2022). Affective neuroscience should inform
the development of these technologies, and our understanding of their limitations (Immordino-
Yang & Singh, 2011; MacLean, 2022; Wiltshire & Fiore, 2014). However, mixed and
ambiguous feelings are a large oversight in affective neuroscience- and they will be even more
relevant in the real-world contexts where these technologies will be used. For artificial systems
to use affective information about us in a manner which benefits us, they will have to understand
the complexities of our experiences, and the systems themselves will have to handle complex
ambivalent information in a manner similar to how we do (Christov-Moore et al., 2022; Seïler &
Craig, 2016).

Figure 3: The resolution of ambivalence across timescales in the development of empathetic
artificial intelligence

Figure from Christov-Moore, et al., 2022

149

Ending thought

The biggest point I hope to get across with this dissertation is that mixed and ambiguous
feelings are not a fluke of self-report: they are a necessary area of study, and in many ways, a
model of how we should study the mind. There has long been debate about whether psychology
or neuroscience is more fitting for the study of the mind (Block, 2007; Marshall, 2009; Miller &
Keller, 2000). On one hand, the goal of neuroscience can be viewed skeptically as reductive
elimination of psychology- and conscious insights tell us nothing that neurobiology cannot. On
the other, if emergent psychological properties can never be fully reduced to neurobiology, it
calls into question what insights we gain from neuroscience at all. With mixed and ambiguous
feelings, we can actually put aside the or to embrace an and. We are presented with a concept
that is defined by our conscious experience, and yet, our conscious experience informs us that we
are missing some aspect of it- our grasp on that quality feels incomplete. In a way, our
neurobiological study of this phenomena is not a manner of disproving our experiential concept,
as can often feel the case in the struggle to define affect: it’s answering a question raised by it.

150

REFERENCES

Aaron, R. V., Snodgress, M. A., Blain, S. D., & Park, S. (2018). Affect labeling and other
aspects of emotional experiences in relation to alexithymia following standardized
emotion inductions. Psychiatry Res, 262, 115-123.
https://doi.org/10.1016/j.psychres.2018.02.014
Abeyta, A. A., & Routledge, C. (2017). Nostalgia as a psychological resource for a meaningful
life. The Happy Mind: Cognitive Contributions to Well-Being, 427-442.
Adler, J. M., & Hershfield, H. E. (2012). Mixed emotional experience is associated with and
precedes improvements in psychological well-being. PLoS One, 7(4), e35633-e35633.
https://doi.org/10.1371/journal.pone.0035633
Adolphs, R. (2016). How should neuroscience study emotions? by distinguishing emotion states,
concepts, and experiences. Social cognitive and affective neuroscience, 12(1), 24-31.
https://doi.org/10.1093/scan/nsw153
Adolphs, R., & Andler, D. (2018). Investigating emotions as functional states distinct from
feelings. Emotion Review, 10(3), 191-201.
Adolphs, R., Mlodinow, L., & Barrett, L. F. (2019). What is an emotion? Curr Biol, 29(20),
R1060-r1064. https://doi.org/10.1016/j.cub.2019.09.008
Agarwal, S., Guntuku, S. C., Robinson, O. C., Dunn, A., & Ungar, L. H. (2020). Examining the
phenomenon of quarter-life crisis through artificial intelligence and the language of
Twitter. Frontiers in Psychology, 11, 341.
Aglieri, V., Cagna, B., Velly, L., Takerkart, S., & Belin, P. (2021). FMRI-based identity
classification accuracy in left temporal and frontal regions predicts speaker recognition
performance. Sci Rep, 11(1), 489. https://doi.org/10.1038/s41598-020-79922-7
Albanese, O., De Stasio, S., Di Chiacchio, C., Fiorilli, C., & Pons, F. (2010). Emotion
comprehension: The impact of nonverbal intelligence. The Journal of genetic psychology,
171(2), 101-115.
Albano, S., Piccardi, L., Pizzamiglio, M. R., Volpe, C., & D'Amico, S. (2013). Narrative
discourse and sociocognitive abilities of a child with Cri-du-Chat syndrome. J Genet
Psychol, 174(1), 51-72. https://doi.org/10.1080/00221325.2011.639409
An, S., Ji, L. J., Marks, M., & Zhang, Z. (2017). Two Sides of Emotion: Exploring Positivity and
Negativity in Six Basic Emotions across Cultures. Front Psychol, 8, 610.
https://doi.org/10.3389/fpsyg.2017.00610
Anderson, A., Han, D., Douglas, P. K., Bramen, J., & Cohen, M. S. (2012). Real-time functional
MRI classification of brain states using markov-svm hybrid models: peering inside the rt-
151

fMRI black box. In Machine learning and interpretation in neuroimaging (pp. 242-255).
Springer.
Armitage, C. J., & Conner, M. (2000). Attitudinal ambivalence: A test of three key hypotheses.
Personality and Social Psychology Bulletin, 26(11), 1421-1432.
Atzil, S., & Gendron, M. (2017). Bio-behavioral synchrony promotes the development of
conceptualized emotions. Current Opinion in Psychology, 17, 162-169.
Babalian, A., Eichenberger, S., Bilella, A., Girard, F., Szabolcsi, V., Roccaro, D., Alvarez-
Bolado, G., Xu, C., & Celio, M. R. (2018). The orbitofrontal cortex projects to the
parvafox nucleus of the ventrolateral hypothalamus and to its targets in the ventromedial
periaqueductal grey matter. Brain Struct Funct. https://doi.org/10.1007/s00429-018-
1771-5
Bach, D. R., Hulme, O., Penny, W. D., & Dolan, R. J. (2011). The known unknowns: neural
representation of second-order uncertainty, and ambiguity. Journal of neuroscience,
31(13), 4811-4820.
Bagby, R. M., Parker, J. D., & Taylor, G. J. (1994). The twenty-item Toronto Alexithymia
Scale—I. Item selection and cross-validation of the factor structure. Journal of
psychosomatic research, 38(1), 23-32.
Bailen, N. H., Wu, H., & Thompson, R. J. (2019). Meta-emotions in daily life: Associations with
emotional awareness and depression. Emotion, 19(5), 776.
Baldassano, C., Chen, J., Zadbood, A., Pillow, J. W., Hasson, U., & Norman, K. A. (2017).
Discovering event structure in continuous narrative perception and memory. Neuron,
95(3), 709-721. e705.
Bandler, R., & Shipley, M. T. (1994). Columnar organization in the midbrain periaqueductal
gray: modules for emotional expression? Trends in Neurosciences, 17(9), 379-389.
https://doi.org/https://doi.org/10.1016/0166-2236(94)90047-7
Barbas, H. (2000). Connections underlying the synthesis of cognition, memory, and emotion in
primate prefrontal cortices. Brain Res Bull, 52(5), 319-330.
https://www.ncbi.nlm.nih.gov/pubmed/10922509
Barford, K. A., & Smillie, L. D. (2016). Openness and other Big Five traits in relation to
dispositional mixed emotions. Personality and Individual Differences, 102, 118-122.
Barrett, F. S., & Janata, P. (2016). Neural responses to nostalgia-evoking music modeled by
elements of dynamic musical structure and individual differences in affective traits.
Neuropsychologia, 91, 234-246. https://doi.org/10.1016/j.neuropsychologia.2016.08.012
Barrett, L. F. (2006a). Are emotions natural kinds? Perspectives on Psychological Science, 1(1),
28-58.
152

Barrett, L. F. (2006b). Valence is a basic building block of emotional life. Journal of Research in
Personality, 40(1), 35-55.
Barrett, L. F. (2017). The theory of constructed emotion: an active inference account of
interoception and categorization. Social cognitive and affective neuroscience, 12(1), 1-23.
https://doi.org/10.1093/scan/nsw154
Barrett, L. F., & Bliss-Moreau, E. (2009). Affect as a Psychological Primitive. Adv Exp Soc
Psychol, 41, 167-218. https://doi.org/10.1016/s0065-2601(08)00404-8
Barrett, L. F., & Russell, J. A. (1998). Independence and bipolarity in the structure of current
affect. Journal of Personality and Social Psychology, 74(4), 967-984.
https://doi.org/10.1037/0022-3514.74.4.967
Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S. W. (1994). Insensitivity to future
consequences following damage to human prefrontal cortex. Cognition, 50(1-3), 7-15.
https://doi.org/10.1016/0010-0277(94)90018-3
Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Different contributions of the
human amygdala and ventromedial prefrontal cortex to decision-making. J Neurosci,
19(13), 5473-5481. https://doi.org/10.1523/JNEUROSCI.19-13-05473.1999
Beck, L., Kumschick, I. R., Eid, M., & Klann-Delius, G. (2012). Relationship between language
competence and emotional competence in middle childhood. Emotion, 12(3), 503-514.
https://doi.org/10.1037/a0026320
Becker, M. P., Nitsch, A. M., Schlosser, R., Koch, K., Schachtzabel, C., Wagner, G., Miltner, W.
H., & Straube, T. (2014). Altered emotional and BOLD responses to negative, positive
and ambiguous performance feedback in OCD. Soc Cogn Affect Neurosci, 9(8), 1127-
1133. https://doi.org/10.1093/scan/nst095
Benoit, R. G., Davies, D. J., & Anderson, M. C. (2016). Reducing future fears by suppressing the
brain mechanisms underlying episodic simulation. Proceedings of the National Academy
of Sciences, 113(52), E8492-E8501.
https://doi.org/https://doi.org/10.1073/pnas.1606604114
Benoit, R. G., Paulus, P. C., & Schacter, D. L. (2019). Forming attitudes via neural activity
supporting affective episodic simulations. Nature communications, 10(1), 1-11.
https://doi.org/https://doi.org/10.1038/s41467-019-09961-w
Berntson, G. G., & Cacioppo, J. T. (2008). The neuroevolution of motivation. Handbook of
motivation science, 188-200.
Berridge, K. C. (2019). Affective valence in the brain: modules or modes? Nature Reviews
Neuroscience. https://doi.org/10.1038/s41583-019-0122-8
153

Berridge, K. C., & Grill, H. J. (1983). Alternating ingestive and aversive consummatory
responses suggest a two-dimensional analysis of palatability in rats. Behavioral
Neuroscience, 97(4), 563-573. https://doi.org/10.1037/0735-7044.97.4.563
Berridge, K. C., & Grill, H. J. (1984). Isohedonic tastes support a two-dimensional hypothesis of
palatability. Appetite, 5(3), 221-231.
Berrios, R., Totterdell, P., & Kellett, S. (2015a). Eliciting mixed emotions: a meta-analysis
comparing models, types, and measures. Front Psychol, 6, 428.
https://doi.org/10.3389/fpsyg.2015.00428
Berrios, R., Totterdell, P., & Kellett, S. (2015b). Investigating goal conflict as a source of mixed
emotions. Cogn Emot, 29(4), 755-763. https://doi.org/10.1080/02699931.2014.939948
Berrios, R., Totterdell, P., & Kellett, S. (2017). Individual differences in mixed emotions
moderate the negative consequences of goal conflict on life purpose. Personality and
Individual Differences, 110, 18-22.
Berrios, R., Totterdell, P., & Kellett, S. (2018a). Silver linings in the face of temptations: how
mixed emotions promote self-control efforts in response to goal conflict. Motivation and
Emotion, 42, 909-919.
Berrios, R., Totterdell, P., & Kellett, S. (2018b). When feeling mixed can be meaningful: The
relation between mixed emotions and eudaimonic well-being. Journal of Happiness
Studies, 19, 841-861.
Bertossi, E., Aleo, F., Braghittoni, D., & Ciaramelli, E. (2016). Stuck in the here and now:
Construction of fictitious and future experiences following ventromedial prefrontal
damage. Neuropsychologia, 81, 107-116.
https://doi.org/10.1016/j.neuropsychologia.2015.12.015
Bertossi, E., Candela, V., De Luca, F., & Ciaramelli, E. (2017). Episodic future thinking
following vmPFC damage: impaired event construction, maintenance, or narration?
Neuropsychology, 31(3), 337. https://doi.org/10.1037/neu0000345
Beyens, I., Pouwels, J. L., van Driel, I. I., Keijsers, L., & Valkenburg, P. M. (2020). The effect of
social media on well-being differs from adolescent to adolescent. Scientific Reports,
10(1), 10763.
Bhar, S. S., & Kyrios, M. (2007). An investigation of self-ambivalence in obsessive-compulsive
disorder. Behav Res Ther, 45(8), 1845-1857. https://doi.org/10.1016/j.brat.2007.02.005
Bi, S., & Sourjik, V. (2018). Stimulus sensing and signal processing in bacterial chemotaxis.
Current opinion in microbiology, 45, 22-29.
Bird, G., & Cook, R. (2013). Mixed emotions: the contribution of alexithymia to the emotional
symptoms of autism. Transl Psychiatry, 3, e285. https://doi.org/10.1038/tp.2013.61
154

Blair, R., Mitchell, D., Richell, R., Kelly, S., Leonard, A., Newman, C., & Scott, S. (2002).
Turning a Deaf Ear to Fear: Impaired Recognition of Vocal Affect in Psychopathic
Individuals. Journal of abnormal psychology, 111, 682-686.
https://doi.org/10.1037//0021-843X.111.4.682
Blasi, M. D., Giardina, A., Giordano, C., Coco, G. L., Tosto, C., Billieux, J., & Schimmenti, A.
(2019). Problematic video game use as an emotional coping strategy: Evidence from a
sample of MMORPG gamers. Journal of Behavioral Addictions, 8(1), 25-34.
Block, N. (2007). Consciousness, accessibility, and the mesh between psychology and
neuroscience. Behavioral and Brain Sciences, 30(5-6), 481-499.
Bo, K., Yin, S., Liu, Y., Hu, Z., Meyyappan, S., Kim, S., Keil, A., & Ding, M. (2021). Decoding
neural representations of affective scenes in retinotopic visual cortex. Cerebral Cortex,
31(6), 3047-3063.
Bonte, M., Hausfeld, L., Scharke, W., Valente, G., & Formisano, E. (2014). Task-dependent
decoding of speaker and vowel identity from auditory cortical response patterns. J
Neurosci, 34(13), 4548-4557. https://doi.org/10.1523/JNEUROSCI.4339-13.2014
Bota, P. J., Wang, C., Fred, A. L., & Da Silva, H. P. (2019). A review, current challenges, and
future possibilities on emotion recognition using machine learning and physiological
signals. IEEE Access, 7, 140990-141020.
Bradley, M. M., Sambuco, N., & Lang, P. J. (2022). Neural correlates of repeated retrieval of
emotional autobiographical events. Neuropsychologia, 169, 108203.
Braniecka, A., Trzebinska, E., Dowgiert, A., & Wytykowska, A. (2014). Mixed emotions and
coping: the benefits of secondary emotions. PLoS One, 9(8), e103940.
https://doi.org/10.1371/journal.pone.0103940
Braun, M., Schubert, J., Pfleging, B., & Alt, F. (2019). Improving driver emotions with affective
strategies. Multimodal Technologies and Interaction, 3(1), 21.
Brewer, R., Cook, R., & Bird, G. (2016). Alexithymia: a general deficit of interoception. Royal
Society open science, 3(10), 150664.
Briesemeister, B. B., Kuchinke, L., & Jacobs, A. M. (2012). Emotional valence: A bipolar
continuum or two independent dimensions? SAGE Open, 2(4), 2158244012466558.
Brooks, J. A., Shablack, H., Gendron, M., Satpute, A. B., Parrish, M. H., & Lindquist, K. A.
(2017). The role of language in the experience and perception of emotion: a
neuroimaging meta-analysis. Soc Cogn Affect Neurosci, 12(2), 169-183.
https://doi.org/10.1093/scan/nsw121
Bukowski, H. (2018). The Neural Correlates of Visual Perspective Taking: a Critical Review.
Current Behavioral Neuroscience Reports, 5(3), 189-197.
https://doi.org/10.1007/s40473-018-0157-6
155

Bush, K. A., Privratsky, A., Gardner, J., Zielinski, M. J., & Kilts, C. D. (2018). Common
functional brain states encode both perceived emotion and the psychophysiological
response to affective stimuli. Scientific Reports, 8(1), 1-10.
https://doi.org/10.1038/s41598-018-33621-6
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes and evaluative space:
A critical review, with emphasis on the separability of positive and negative substrates.
Psychological Bulletin, 115(3), 401.
Carlson, T. A., & Wardle, S. G. (2015). Sensible decoding. Neuroimage, 110, 217-218.
https://doi.org/https://doi.org/10.1016/j.neuroimage.2015.02.009
Carpenter, J. M., Green, M. C., & Vacharkulksemsuk, T. (2016). Beyond perspective-taking:
Mind-reading motivation. Motivation and Emotion, 40(3), 358-374.
https://doi.org/10.1007/s11031-016-9544-z
Carruthers, P., & Veillet, B. (2017). Consciousness operationalized, a debate realigned.
Consciousness and Cognition, 55, 79-90.
Cartwright, F., & Stritzke, W. G. (2008). A multidimensional ambivalence model of chocolate
craving: construct validity and associations with chocolate consumption and disordered
eating. Eating behaviors, 9(1), 1-12.
Carvalho, G. B., & Damasio, A. (2021). Interoception and the origin of feelings: A new
synthesis. BioEssays, 43(6), 2000261.
Casey, B. J., Heller, A. S., Gee, D. G., & Cohen, A. O. (2019). Development of the emotional
brain. Neuroscience Letters, 693, 29-34.
https://doi.org/https://doi.org/10.1016/j.neulet.2017.11.055
Cauda, F., D'Agata, F., Sacco, K., Duca, S., Geminiani, G., & Vercelli, A. (2011). Functional
connectivity of the insula in the resting brain. Neuroimage, 55(1), 8-23.
https://doi.org/10.1016/j.neuroimage.2010.11.049
Celeghin, A., Diano, M., Bagnis, A., Viola, M., & Tamietto, M. (2017). Basic Emotions in
Human Neuroscience: Neuroimaging and Beyond. Front Psychol, 8, 1432.
https://doi.org/10.3389/fpsyg.2017.01432
Celeghin, A., Diano, M., Bagnis, A., Viola, M., & Tamietto, M. (2017). Basic Emotions in
Human Neuroscience: Neuroimaging and Beyond [Review]. Frontiers in Psychology,
8(1432). https://doi.org/10.3389/fpsyg.2017.01432
Chakriswaran, P., Vincent, D. R., Srinivasan, K., Sharma, V., Chang, C.-Y., & Reina, D. G.
(2019). Emotion AI-driven sentiment analysis: A survey, future research directions, and
open issues. Applied Sciences, 9(24), 5462.
Chang, L. J., Jolly, E., Cheong, J. H., Rapuano, K. M., Greenstein, N., Chen, P.-H. A., &
Manning, J. R. (2021). Endogenous variation in ventromedial prefrontal cortex state
156

dynamics during naturalistic viewing reflects affective experience. Science Advances,
7(17), eabf7129.
Charland, L. C. (2005). Emotion experience and the indeterminacy of valence. Emotion and
consciousness, 231-254.
Choe, K. S., Keil, F. C., & Bloom, P. (2005). Children's understanding of the Ulysses conflict.
Dev Sci, 8(5), 387-392. https://doi.org/10.1111/j.1467-7687.2005.00426.x
Christov-Moore, L., & Iacoboni, M. (2016). Self-other resonance, its control and prosocial
inclinations: Brain–behavior relationships. Human brain mapping, 37(4), 1544-1558.
https://doi.org/10.1002/hbm.23119
Christov-Moore, L., Reggente, N., Vaccaro, A., Schoeller, F., Pluimer, B., Douglas, P. K.,
Iacoboni, M., & Kaplan, J. (2022). How to Prevent Robotic Sociopaths: A Neuroscience
Approach to Artificial Ethics.
Clark-Polner, E., Johnson, T. D., & Barrett, L. F. (2017). Multivoxel Pattern Analysis Does Not
Provide Evidence to Support the Existence of Basic Emotions. Cereb Cortex, 27(3),
1944-1948. https://doi.org/10.1093/cercor/bhw028
Colibazzi, T., Posner, J., Wang, Z., Gorman, D., Gerber, A., Yu, S., Zhu, H., Kangarlu, A.,
Duan, Y., Russell, J. A., & Peterson, B. S. (2010). Neural systems subserving valence
and arousal during the experience of induced emotions. Emotion, 10(3), 377-389.
https://doi.org/10.1037/a0018484
Colombetti, G. (2005). Appraising valence. Journal of consciousness studies, 12(8-9), 103-126.
Corradi-Dell’Acqua, C., Hofstetter, C., & Vuilleumier, P. (2014). Cognitive and affective theory
of mind share the same local patterns of activity in posterior temporal but not medial
prefrontal cortex. Social cognitive and affective neuroscience, 9(8), 1175-1184.
https://doi.org/10.1093/scan/nst097
Coutanche, M. N. (2013). Distinguishing multi-voxel patterns and mean activation: why, how,
and what does it tell us? Cogn Affect Behav Neurosci, 13(3), 667-673.
https://doi.org/10.3758/s13415-013-0186-2
Coutanche, M. N., Thompson-Schill, S. L., & Schultz, R. T. (2011). Multi-voxel pattern analysis
of fMRI data predicts clinical symptom severity. Neuroimage, 57(1), 113-123.
https://doi.org/10.1016/j.neuroimage.2011.04.016
Craig, A. D. (2009a). Emotional moments across time: a possible neural basis for time
perception in the anterior insula. Philosophical Transactions of the Royal Society B:
Biological Sciences, 364(1525), 1933-1942. https://doi.org/10.1098/rstb.2009.0008
Craig, A. D. (2009b). How do you feel--now? The anterior insula and human awareness. Nat Rev
Neurosci, 10(1), 59-70. https://doi.org/10.1038/nrn2555
157

Critchley, H. D. (2009). Psychophysiology of neural, cognitive and affective integration: fMRI
and autonomic indicants. Int J Psychophysiol, 73(2), 88-94.
https://doi.org/10.1016/j.ijpsycho.2009.01.012
Critchley, H. D. (2009). Psychophysiology of neural, cognitive and affective integration: fMRI
and autonomic indicants. International journal of psychophysiology, 73(2), 88-94.
https://doi.org/10.1016/j.ijpsycho.2009.01.012
Critchley, H. D., Eccles, J., & Garfinkel, S. N. (2013). Interaction between cognition, emotion,
and the autonomic nervous system. Handb Clin Neurol, 117, 59-77.
https://doi.org/10.1016/B978-0-444-53491-0.00006-7
Critchley, H. D., Elliott, R., Mathias, C. J., & Dolan, R. J. (2000). Neural activity relating to
generation and representation of galvanic skin conductance responses: a functional
magnetic resonance imaging study. J Neurosci, 20(8), 3033-3040.
https://www.ncbi.nlm.nih.gov/pubmed/10751455
Critchley, H. D., & Garfinkel, S. N. (2017). Interoception and emotion. Curr Opin Psychol, 17,
7-14. https://doi.org/10.1016/j.copsyc.2017.04.020
Critchley, H. D., Wiens, S., Rotshtein, P., Ohman, A., & Dolan, R. J. (2004). Neural systems
supporting interoceptive awareness. Nat Neurosci, 7(2), 189-195.
https://doi.org/10.1038/nn1176
Croy, I., Olgun, S., & Joraschky, P. (2011). Basic emotions elicited by odors and pictures.
Emotion, 11(6), 1331-1335. https://doi.org/10.1037/a0024437
Cunningham, W. A., Raye, C. L., & Johnson, M. K. (2004). Implicit and explicit evaluation:
FMRI correlates of valence, emotional intensity, and control in the processing of
attitudes. J Cogn Neurosci, 16(10), 1717-1729.
https://doi.org/10.1162/0898929042947919
D’Argembeau, A., Xue, G., Lu, Z.-L., Van der Linden, M., & Bechara, A. (2008). Neural
correlates of envisioning emotional events in the near and far future. Neuroimage, 40(1),
398-407. https://doi.org/10.1016/j.neuroimage.2007.11.025
Dabb, C., Dryer, R., Brunton, R. J., Yap, K., & Roach, V. J. (2022). Paternal pregnancy-related
anxiety: Systematic review of men's concerns and experiences during their partners'
pregnancies. Journal of Affective Disorders.
Damasio, A. (2018). The Strange Order of Things: Life, Feeling, and the Making of Cultures.
Knopf Doubleday Publishing Group.
https://books.google.com/books?id=VDotDQAAQBAJ
Damasio, A., & Carvalho, G. B. (2013). The nature of feelings: evolutionary and neurobiological
origins. Nat Rev Neurosci, 14(2), 143-152. https://doi.org/10.1038/nrn3403
158

Damasio, A. R. (1996). The somatic marker hypothesis and the possible functions of the
prefrontal cortex. Philos Trans R Soc Lond B Biol Sci, 351(1346), 1413-1420.
https://doi.org/10.1098/rstb.1996.0125
Davidson, T. L. (1993). The nature and function of interoceptive signals to feed: toward
integration of physiological and learning perspectives. Psychol Rev, 100(4), 640-657.
https://www.ncbi.nlm.nih.gov/pubmed/8255952
Davis, T., LaRocque, K. F., Mumford, J. A., Norman, K. A., Wagner, A. D., & Poldrack, R. A.
(2014). What do differences between multi-voxel and univariate analysis mean? How
subject-, voxel-, and trial-level variance impact fMRI analysis. Neuroimage, 97, 271-283.
https://doi.org/10.1016/j.neuroimage.2014.04.037
De Oca, B. M., DeCola, J. P., Maren, S., & Fanselow, M. S. (1998). Distinct regions of the
periaqueductal gray are involved in the acquisition and expression of defensive
responses. Journal of neuroscience, 18(9), 3426-3432.
Decety, J., & Grèzes, J. (2006). The power of simulation: Imagining one's own and other's
behavior. Brain Research, 1079(1), 4-14.
https://doi.org/https://doi.org/10.1016/j.brainres.2005.12.115
Decety, J., & Meltzoff, A. N. (2011). Empathy, imitation, and the social brain. In Empathy:
Philosophical and psychological perspectives (pp. 58-81). Oxford University Press.
Deen, B., Pitskel, N. B., & Pelphrey, K. A. (2011). Three systems of insular functional
connectivity identified with cluster analysis. Cerebral Cortex, 21(7), 1498-1506.
Dehghani, M., Boghrati, R., Man, K., Hoover, J., Gimbel, S. I., Vaswani, A., Zevin, J. D.,
Immordino-Yang, M. H., Gordon, A. S., Damasio, A., & Kaplan, J. (2017). Decoding the
neural representation of story meanings across languages. Human brain mapping, 38(12),
6096-6106. https://doi.org/10.1002/hbm.23814
Delgado, M. R., Beer, J. S., Fellows, L. K., Huettel, S. A., Platt, M. L., Quirk, G. J., & Schiller,
D. (2016). Viewpoints: dialogues on the functional role of the ventromedial prefrontal
cortex. Nature Neuroscience, 19(12), 1545-1552.
Denton, D. A., McKinley, M. J., Farrell, M., & Egan, G. F. (2009). The role of primordial
emotions in the evolutionary origin of consciousness. Conscious Cogn, 18(2), 500-514.
https://doi.org/10.1016/j.concog.2008.06.009
Devitt, A. L., Thakral, P. P., & Schacter, D. L. (2022). Decoding the emotional valence of future
thoughts. Cognitive neuroscience, 13(1), 10-14.
Donaldson, S. K., & Westerman, M. A. (1986). Development of children's understanding of
ambivalence and causal theories of emotions. Developmental Psychology, 22(5), 655.
159

Droit-Volet, S., & Gil, S. (2009). The time–emotion paradox. Philosophical Transactions of the
Royal Society B: Biological Sciences, 364(1525), 1943-1953.
https://doi.org/10.1098/rstb.2009.0013
Dukes, D., Abrams, K., Adolphs, R., Ahmed, M. E., Beatty, A., Berridge, K. C., Broomhall, S.,
Brosch, T., Campos, J. J., & Clay, Z. (2021). The rise of affectivism. Nature Human
Behaviour, 1-5.
Dunning, D., Fetchenhauer, D., & Schlösser, T. (2017). The varying roles played by emotion in
economic decision making. Current Opinion in Behavioral Sciences, 15, 33-38.
Eagle, M. N. (2017). Inner conflict in Freudian theory. In Psychoanalytic perspectives on conflict
(pp. 1-20). Routledge.
Eckland, N. S., Leyro, T. M., Mendes, W. B., & Thompson, R. J. (2018). A multi-method
investigation of the association between emotional clarity and empathy. Emotion, 18(5),
638.
Eder, J. (2016). Films and existential feelings. Projections, 10(2), 75-103.
Ekman, P. (1992). An argument for basic emotions. Cognition & emotion, 6(3-4), 169-200.
Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous system activity
distinguishes among emotions. Science, 221(4616), 1208-1210.
https://www.ncbi.nlm.nih.gov/pubmed/6612338
Engelen, T., de Graaf, T. A., Sack, A. T., & de Gelder, B. (2015). A causal role for inferior
parietal lobule in emotion body perception. Cortex, 73, 195-202.
https://doi.org/10.1016/j.cortex.2015.08.013
Epstude, K., & Roese, N. J. (2008). The functional theory of counterfactual thinking. Personality
and social psychology review : an official journal of the Society for Personality and
Social Psychology, Inc, 12(2), 168-192. https://doi.org/10.1177/1088868308316091
Erbas, Y., Sels, L., Ceulemans, E., & Kuppens, P. (2016). Feeling me, feeling you: The relation
between emotion differentiation and empathic accuracy. Social Psychological and
Personality Science, 7(3), 240-247.
Ersner-Hershfield, H., Mikels, J. A., Sullivan, S. J., & Carstensen, L. L. (2008). Poignancy:
mixed emotional experience in the face of meaningful endings. Journal of Personality
and Social Psychology, 94(1), 158.
Esteban, O., Markiewicz, C. J., Blair, R. W., Moodie, C. A., Isik, A. I., Erramuzpe, A., Kent, J.
D., Goncalves, M., DuPre, E., & Snyder, M. (2019). fMRIPrep: a robust preprocessing
pipeline for functional MRI. Nature methods, 16(1), 111-116.
Ethofer, T., Van De Ville, D., Scherer, K., & Vuilleumier, P. (2009). Decoding of emotional
information in voice-sensitive cortices. Current biology, 19(12), 1028-1033.
160

Etkin, A., Egner, T., & Kalisch, R. (2011). Emotional processing in anterior cingulate and medial
prefrontal cortex. Trends in cognitive sciences, 15(2), 85-93.
https://doi.org/10.1016/j.tics.2010.11.004
Extremera, N., Durán, A., & Rey, L. (2009). The moderating effect of trait meta-mood and
perceived stress on life satisfaction. Personality and Individual Differences, 47(2), 116-
121.
Extremera, N., & Fernández-Berrocal, P. (2005). Perceived emotional intelligence and life
satisfaction: Predictive and incremental validity using the Trait Meta-Mood Scale.
Personality and Individual Differences, 39(5), 937-948.
Fair, D. A., Cohen, A. L., Power, J. D., Dosenbach, N. U. F., Church, J. A., Miezin, F. M.,
Schlaggar, B. L., & Petersen, S. E. (2009). Functional Brain Networks Develop from a
“Local to Distributed” Organization. PLOS Computational Biology, 5(5), e1000381.
https://doi.org/10.1371/journal.pcbi.1000381
Fang, X., Sauter, D. A., & Van Kleef, G. A. (2018). Seeing Mixed Emotions: The Specificity of
Emotion Perception From Static and Dynamic Facial Expressions Across Cultures. J
Cross Cult Psychol, 49(1), 130-148. https://doi.org/10.1177/0022022117736270
Fanselow, M. S. (1994). Neural organization of the defensive behavior system responsible for
fear. Psychonomic bulletin & review, 1(4), 429-438.
Farrell, P. (2006). Portuguese SAUDADE and other emotions of absence and longing. Semantic
primes and universal grammar, 235-258.
Faure, A., Richard, J. M., & Berridge, K. C. (2010). Desire and dread from the nucleus
accumbens: cortical glutamate and subcortical GABA differentially generate motivation
and hedonic impact in the rat. PLoS One, 5(6), e11223.
https://doi.org/10.1371/journal.pone.0011223
Feldman Barrett, L., & Russell, J. A. (1998). Independence and bipolarity in the structure of
current affect. Journal of personality and social psychology, 74(4), 967.
Filippetti, M. L., Kirsch, L. P., Crucianelli, L., & Fotopoulou, A. (2019). Affective certainty and
congruency of touch modulate the experience of the rubber hand illusion. Scientific
Reports, 9(1), 2635. https://doi.org/10.1038/s41598-019-38880-5
Finn, E. S., Scheinost, D., Finn, D. M., Shen, X., Papademetris, X., & Constable, R. T. (2017).
Can brain state be manipulated to emphasize individual differences in functional
connectivity? Neuroimage, 160, 140-151.
Fischer, K. W., & Bidell, T. R. (2007). Dynamic Development of Action and Thought. In
Handbook of Child Psychology. https://doi.org/doi:10.1002/9780470147658.chpsy0107
Fischer, K. W., Shaver, P. R., & Carnochan, P. (1990). How emotions develop and how they
organise development. Cognition and Emotion, 4(2), 81-127.
161

Fleming, S. M., & Lau, H. C. (2014). How to measure metacognition. Frontiers in Human
Neuroscience, 8, 443.
Fong, C. T. (2006). The Effects of Emotional Ambivalence on Creativity. The Academy of
Management Journal, 49(5), 1016-1030. https://doi.org/10.2307/20159814
Fotopoulou, A., & Tsakiris, M. (2017). Mentalizing homeostasis: The social origins of
interoceptive inference. Neuropsychoanalysis, 19(1), 3-28.
https://doi.org/10.1080/15294145.2017.1294031
Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden the scope of attention and
thought-action repertoires. Cogn Emot, 19(3), 313-332.
https://doi.org/10.1080/02699930441000238
Garland, E. L., Fredrickson, B., Kring, A. M., Johnson, D. P., Meyer, P. S., & Penn, D. L.
(2010). Upward spirals of positive emotions counter downward spirals of negativity:
insights from the broaden-and-build theory and affective neuroscience on the treatment of
emotion dysfunctions and deficits in psychopathology. Clin Psychol Rev, 30(7), 849-864.
https://doi.org/10.1016/j.cpr.2010.03.002
Gergely, G., & Watson, J. S. (1999). Early socio-emotional development: Contingency
perception and the social-biofeedback model. Early social cognition: Understanding
others in the first months of life, 60, 101-136.
Gessell, B., Geib, B., & De Brigard, F. (2021a). Multivariate pattern analysis and the search for
neural representations. Synthese, 199(5-6), 12869-12889.
Gessell, B., Geib, B., & De Brigard, F. (2021b). Multivariate pattern analysis and the search for
neural representations. Synthese, 1-21.
Ginsburg, S., & Jablonka, E. (2021). Evolutionary transitions in learning and cognition.
Philosophical Transactions of the Royal Society B, 376(1821), 20190766.
Glenn, D. E., Fox, N. A., Pine, D. S., Peters, M. A., & Michalska, K. J. (2020). Divergence in
cortical representations of threat generalization in affective versus perceptual circuitry in
childhood: Relations with anxiety. Neuropsychologia, 142, 107416.
Goldenberg, A., & Gross, J. J. (2020). Digital emotion contagion. Trends in cognitive sciences,
24(4), 316-328.
Goldsmith, H. H., & Davidson, R. J. (2004). Disambiguating the components of emotion
regulation. Child Dev, 75(2), 361-365. https://doi.org/10.1111/j.1467-8624.2004.00678.x
Gosling, S. D., Rentfrow, P. J., & Swann Jr, W. B. (2003). A very brief measure of the Big-Five
personality domains. Journal of Research in personality, 37(6), 504-528.
162

Grall, C., & Finn, E. S. (2022). Leveraging the power of media to drive cognition: A media-
informed approach to naturalistic neuroscience. Social cognitive and affective
neuroscience, 17(6), 598-608.
Greenwood, D., & Long, C. R. (2015). When movies matter: Emerging adults recall memorable
movies. Journal of Adolescent Research, 30(5), 625-650.
Grossmann, I., & Ellsworth, P. C. (2017). What are mixed emotions and what conditions foster
them? Life-span experiences, culture and social awareness. Current Opinion in
Behavioral Sciences, 15, 1-5.
Grupe, D. W., & Nitschke, J. B. (2013). Uncertainty and anticipation in anxiety: an integrated
neurobiological and psychological perspective. Nature Reviews Neuroscience, 14(7),
488-501.
Gschwind, M., & Picard, F. (2016). Ecstatic Epileptic Seizures: A Glimpse into the Multiple
Roles of the Insula. Frontiers in Behavioral Neuroscience, 10, 21-21.
https://doi.org/10.3389/fnbeh.2016.00021
Halfmann, A., & Reinecke, L. (2021). Binge-watching as case of escapist entertainment use. The
Oxford handbook of entertainment theory, 181-203.
Hall, A. E. (2015). Entertainment-oriented gratifications of sports media: Contributors to
suspense, hedonic enjoyment, and appreciation. Journal of Broadcasting & Electronic
Media, 59(2), 259-277.
Hall, A. E., & Bracken, C. C. (2011). I really liked that movie. Journal of Media Psychology.
Hall, L., Haggerty, D., Cooper, J., Golden, C., & Dornheim, L. (1998). Development and
validation of a measure of emotional intelligence. Personality and Individual Differences,
25, 167-177.
Hanich, J., Wagner, V., Shah, M., Jacobsen, T., & Menninghaus, W. (2014). Why we like to
watch sad films. The pleasure of being moved in aesthetic experiences. Psychology of
Aesthetics, Creativity, and the Arts, 8(2), 130.
Harmon-Jones, E. (2019). On motivational influences, moving beyond valence, and integrating
dimensional and discrete views of emotion. Cognition and Emotion, 33(1), 101-108.
Harter, S., & Buddin, B. J. (1987). Children's understanding of the simultaneity of two emotions:
A five-stage developmental acquisition sequence. Developmental Psychology, 23(3),
388-399. https://doi.org/10.1037/0012-1649.23.3.388
Healey, M. L., & Grossman, M. (2018). Cognitive and Affective Perspective-Taking: Evidence
for Shared and Dissociable Anatomical Substrates. Frontiers in neurology, 9, 491-491.
https://doi.org/10.3389/fneur.2018.00491
163

Healey, M. L., & Grossman, M. (2018). Cognitive and affective perspective-taking: evidence for
shared and dissociable anatomical substrates. Frontiers in neurology, 9, 491.
https://doi.org/10.3389/fneur.2018.00491
Hennings, A. C., McClay, M., Lewis-Peacock, J. A., & Dunsmoor, J. E. (2019). Neural
reinstatement guides context-dependent emotional memory retrieval. bioRxiv, 785899.
Hershfield, H. E., Scheibe, S., Sims, T. L., & Carstensen, L. L. (2013). When Feeling Bad Can
Be Good: Mixed Emotions Benefit Physical Health Across Adulthood. Soc Psychol
Personal Sci, 4(1), 54-61. https://doi.org/10.1177/1948550612444616
Hill, C. L., & Updegraff, J. A. (2012). Mindfulness and its relationship to emotional regulation.
Emotion, 12(1), 81.
Hoemann, K., Gendron, M., & Barrett, L. F. (2017). Mixed emotions in the predictive brain.
Curr Opin Behav Sci, 15, 51-57. https://doi.org/10.1016/j.cobeha.2017.05.013
Hogeveen, J., Krueger, F., & Grafman, J. (2021). Association between alexithymia and impaired
reward valuation in patients with fronto-insular damage. Emotion, 21(1), 137.
Holm, O., Greaker, E., & Strömberg, A. (2002). Experiences of Longing in Norwegian and
Swedish 4- and 5-Year-Old Children. The Journal of Psychology, 136(6), 608-612.
https://doi.org/10.1080/00223980209604822
Hong, J., & Lee, A. Y. (2010). Feeling mixed but not torn: The moderating role of construal
level in mixed emotions appeals. Journal of Consumer research, 37(3), 456-472.
Hsu, M., Bhatt, M., Adolphs, R., Tranel, D., & Camerer, C. F. (2005). Neural systems
responding to degrees of uncertainty in human decision-making. Science, 310(5754),
1680-1683. https://doi.org/10.1126/science.1115327
Hu, B., & Tu, Y. (2014). Behaviors and strategies of bacterial navigation in chemical and
nonchemical gradients. PLOS Computational Biology, 10(6), e1003672.
Hui, C. M., Fok, H. K., & Bond, M. H. (2009). Who feels more ambivalence? Linking dialectical
thinking to mixed emotions. Personality and Individual Differences, 46(4), 493-498.
https://doi.org/https://doi.org/10.1016/j.paid.2008.11.022
Hussain, U., Jabarkhail, S., Cunningham, G. B., & Madsen, J. A. (2021). The dual nature of
escapism in video gaming: A meta-analytic approach. Computers in Human Behavior
Reports, 3, 100081.
Hynes, C. A., Baird, A. A., & Grafton, S. T. (2006). Differential role of the orbital frontal lobe in
emotional versus cognitive perspective-taking. Neuropsychologia, 44(3), 374-383.
https://doi.org/https://doi.org/10.1016/j.neuropsychologia.2005.06.011
Immordino-Yang, M. H. (2010). Toward a microdevelopmental, interdisciplinary approach to
social emotion. Emotion Review, 2(3), 217-220.
164

Immordino-Yang, M. H., & Singh, V. (2011). Perspectives from social and affective
neuroscience on the design of digital learning technologies. New perspectives on affect
and learning technologies, 233-241.
Israelashvili, J., Oosterwijk, S., Sauter, D., & Fischer, A. (2019). Knowing me, knowing you:
emotion differentiation in oneself is associated with recognition of others’ emotions.
Cognition and Emotion.
Iyer, A., & Jetten, J. (2011). What's left behind: Identity continuity moderates the effect of
nostalgia on well-being and life choices. J Pers Soc Psychol, 101(1), 94-108.
https://doi.org/10.1037/a0022496
Janicke-Bowles, S. H., Bartsch, A., Oliver, M. B., & Raney, A. A. (2021). Traditional
Conceptualizations of Eudaimonic Entertainment. The Oxford handbook of entertainment
theory, 363.
Janowski, V., Camerer, C., & Rangel, A. (2013). Empathic choice involves vmPFC value signals
that are modulated by social processing implemented in IPL. Social cognitive and
affective neuroscience, 8(2), 201-208. https://doi.org/10.1093/scan/nsr086
Jarman, H. K., Marques, M. D., McLean, S. A., Slater, A., & Paxton, S. J. (2021). Motivations
for social media use: Associations with social media engagement and body satisfaction
and well-being among adolescents. Journal of Youth and Adolescence, 50(12), 2279-
2293.
Jenkinson, M., & Smith, S. (2001). A global optimisation method for robust affine registration of
brain images. Medical image analysis, 5(2), 143-156. https://doi.org/10.1016/s1361-
8415(01)00036-6
Jerg-Bretzke, L., Walter, S., Limbrecht-Ecklundt, K., & Traue, H. C. (2013). Emotional
ambivalence and post-traumatic stress disorder (PTSD) in soldiers during military
operations. Psychosoc Med, 10, Doc03. https://doi.org/10.3205/psm000093
Jiang, S., Wang, S., & Wan, X. (2022). Metacognition and mentalizing are associated with
distinct neural representations of decision uncertainty. PLoS Biology, 20(5), e3001301.
Johannes, N., Vuorre, M., & Przybylski, A. K. (2021). Video game play is positively correlated
with well-being. Royal Society open science, 8(2), 202049.
Johnson, D. R., Jasper, D. M., Griffin, S., & Huffman, B. L. (2013). Reading narrative fiction
reduces Arab-Muslim prejudice and offers a safe haven from intergroup anxiety. Social
Cognition, 31(5), 578-598.
Johnston, C. H. (1908). The Combination of Feelings. Harvard Psychological Studies, 2, 159-
191.
Jonas, K., Broemer, P., & Diehl, M. (2000). Attitudinal ambivalence. European review of social
psychology, 11(1), 35-74.
165

Joyce, M. K. P., & Barbas, H. (2018). Cortical Connections Position Primate Area 25 as a
Keystone for Interoception, Emotion, and Memory. J Neurosci, 38(7), 1677-1698.
https://doi.org/10.1523/JNEUROSCI.2363-17.2017
Jung, W. M., Ryu, Y., Lee, Y. S., Wallraven, C., & Chae, Y. (2017). Role of interoceptive
accuracy in topographical changes in emotion-induced bodily sensations. PLoS One,
12(9), e0183211. https://doi.org/10.1371/journal.pone.0183211
Kafetsios, K., & Hess, U. (2019). Seeing mixed emotions: Alexithymia, emotion perception bias,
and quality in dyadic interactions. Personality and Individual Differences, 137, 80-85.
Kaplan, J., Gimbel, S. I., Dehghani, M., Immordino-Yang, M. H., Sagae, K., Wong, J. D.,
Tipper, C. M., Damasio, H., Gordon, A. S., & Damasio, A. (2017). Processing narratives
concerning protected values: A cross-cultural investigation of neural correlates. Cerebral
Cortex, 27(2), 1428-1438.
Kaplan, J., & Meyer, K. (2012). Multivariate pattern analysis reveals common neural patterns
across individuals during touch observation. Neuroimage, 60(1), 204-212.
https://doi.org/10.1016/j.neuroimage.2011.12.059
Kassam, K. S., Markey, A. R., Cherkassky, V. L., Loewenstein, G., & Just, M. A. (2013).
Identifying emotions on the basis of neural activation. PLoS One, 8(6), e66032.
https://doi.org/https://doi.org/10.1371/journal.pone.0066032
Katyal, S., & Fleming, S. (2023). Construct validity in metacognition research: balancing the
tightrope between rigor of measurement and breadth of construct.
Kent, L., & Wittmann, M. (2021). Time consciousness: the missing link in theories of
consciousness. Neuroscience of Consciousness, 2021(2), niab011.
Kestenbaum, R., & Gelman, S. A. (1995). Preschool children's identification and understanding
of mixed emotions. Cognitive Development, 10(3), 443-458.
https://doi.org/https://doi.org/10.1016/0885-2014(95)90006-3
Keysers, C., & Gazzola, V. (2007). Integrating simulation and theory of mind: from self to social
cognition. Trends in cognitive sciences, 11(5), 194-196.
https://doi.org/10.1016/j.tics.2007.02.002
Khoo, G. S. (2016). Contemplating tragedy raises gratifications and fosters self-acceptance.
Human Communication Research, 42(2), 269-291.
Kim, J., Pignatelli, M., Xu, S., Itohara, S., & Tonegawa, S. (2016). Antagonistic negative and
positive neurons of the basolateral amygdala. Nature Neuroscience, 19(12), 1636-1646.
Kim, J., Schultz, J., Rohe, T., Wallraven, C., Lee, S.-W., & Bülthoff, H. H. (2015). Abstract
representations of associated emotions in the human brain. Journal of neuroscience,
35(14), 5655-5663. https://doi.org/10.3389/fnhum.2018.00419
166

King, D. L., & Potenza, M. N. (2019). Not Playing Around: Gaming Disorder in the
International Classification of Diseases (ICD-11). Journal of Adolescent Health, 64(1), 5-
7. https://doi.org/10.1016/j.jadohealth.2018.10.010
Klasen, M., Kenworthy, C. A., Mathiak, K. A., Kircher, T. T., & Mathiak, K. (2011).
Supramodal representation of emotions. Journal of neuroscience, 31(38), 13635-13643.
Klimmt, C., & Rieger, D. (2021). Biographic resonance theory of eudaimonic media
entertainment. The Oxford handbook of entertainment theory, 383-402.
Kober, H., Barrett, L. F., Joseph, J., Bliss-Moreau, E., Lindquist, K., & Wager, T. D. (2008).
Functional grouping and cortical–subcortical interactions in emotion: A meta-analysis of
neuroimaging studies. Neuroimage, 42(2), 998-1031.
https://doi.org/https://doi.org/10.1016/j.neuroimage.2008.03.059
Kotter-Gruhn, D., Scheibe, S., Blanchard-Fields, F., & Baltes, P. B. (2009). Developmental
emergence and functionality of Sehnsucht (life longings): the sample case of involuntary
childlessness in middle-aged women. Psychol Aging, 24(3), 634-644.
https://doi.org/10.1037/a0016359
Kotter-Grühn, D., Wiest, M., Zurek, P. P., & Scheibe, S. (2009). What is it we are longing for?
Psychological and demographic factors influencing the contents of Sehnsucht (life
longings). Journal of Research in Personality, 43(3), 428-437.
https://doi.org/https://doi.org/10.1016/j.jrp.2009.01.012
Kragel, P. A., & LaBar, K. S. (2015a). Multivariate neural biomarkers of emotional states are
categorically distinct. Social cognitive and affective neuroscience, 10(11), 1437-1448.
https://doi.org/10.1093/scan/nsv032
Kragel, P. A., & LaBar, K. S. (2015b). Multivariate neural biomarkers of emotional states are
categorically distinct. Soc Cogn Affect Neurosci, 10(11), 1437-1448.
https://doi.org/10.1093/scan/nsv032
Kragel, P. A., & LaBar, K. S. (2016). Decoding the nature of emotion in the brain. Trends in
cognitive sciences, 20(6), 444-455.
Kragel, P. A., Reddan, M. C., LaBar, K. S., & Wager, T. D. (2019). Emotion schemas are
embedded in the human visual system. Science Advances, 5(7), eaaw4358.
Kreibig, S. D., & Gross, J. J. (2017). Understanding Mixed Emotions: Paradigms and Measures.
Curr Opin Behav Sci, 15, 62-71. https://doi.org/10.1016/j.cobeha.2017.05.016
Kreibig, S. D., Samson, A. C., & Gross, J. J. (2013). The psychophysiology of mixed emotional
states. Psychophysiology, 50(8), 799-811. https://doi.org/10.1111/psyp.12064
Kreibig, S. D., Samson, A. C., & Gross, J. J. (2015). The psychophysiology of mixed emotional
states: Internal and external replicability analysis of a direct replication study.
Psychophysiology, 52(7), 873-886. https://doi.org/10.1111/psyp.12425
167

Kriegeskorte, N., & Douglas, P. K. (2019). Interpreting encoding and decoding models. Curr
Opin Neurobiol, 55, 167-179. https://doi.org/10.1016/j.conb.2019.04.002
Kruschwitz, J. D., Waller, L., List, D., Wisniewski, D., Ludwig, V. U., Korb, F., Wolfensteller,
U., Goschke, T., & Walter, H. (2018). Anticipating the good and the bad: A study on the
neural correlates of bivalent emotion anticipation and their malleability via attentional
deployment. Neuroimage, 183, 553-564.
https://doi.org/10.1016/j.neuroimage.2018.08.048
Kuhlmann, M., Hofmann, M. J., Briesemeister, B. B., & Jacobs, A. M. (2016). Mixing positive
and negative valence: Affective-semantic integration of bivalent words [Article].
Scientific Reports, 6, 30718. https://doi.org/10.1038/srep30718
https://www.nature.com/articles/srep30718#supplementary-information
Kumar, M., Anderson, M. J., Antony, J. W., Baldassano, C., Brooks, P. P., Cai, M. B., Chen, P.-
H. C., Ellis, C. T., Henselman-Petrusek, G., & Huberdeau, D. (2021). BrainIAK: The
brain imaging analysis kit. Aperture neuro, 1(4).
Kung, F. Y., & Chao, M. M. (2019). The impact of mixed emotions on creativity in negotiation:
An interpersonal perspective. Frontiers in Psychology, 9, 2660.
Lamm, C., Bukowski, H., & Silani, G. (2016). From shared to distinct self-other representations
in empathy: evidence from neurotypical function and socio-cognitive disorders.
Philosophical transactions of the Royal Society of London. Series B, Biological sciences,
371(1686), 20150083-20150083. https://doi.org/10.1098/rstb.2015.0083
Lamm, C., Bukowski, H., & Silani, G. (2016). From shared to distinct self–other representations
in empathy: evidence from neurotypical function and socio-cognitive disorders.
Philosophical Transactions of the Royal Society B: Biological Sciences, 371(1686),
20150083. https://doi.org/10.1098/rstb.2015.0083
Lane, R. D., Reiman, E. M., Ahern, G. L., Schwartz, G. E., & Davidson, R. J. (1997).
Neuroanatomical correlates of happiness, sadness, and disgust. Am J Psychiatry, 154(7),
926-933. https://doi.org/10.1176/ajp.154.7.926
Lane, R. D., & Schwartz, G. E. (1987). Levels of emotional awareness: a cognitive-
developmental theory and its application to psychopathology. American Journal of
Psychiatry, 144(2), 133-143. https://doi.org/10.1176/ajp.144.2.133
Larsen, J. T. (2017). Holes in the Case for Mixed Emotions. Emotion Review, 9(2), 118-123.
https://doi.org/10.1177/1754073916639662
Larsen, J. T., & Green, J. D. (2013). Evidence for mixed feelings of happiness and sadness from
brief moments in time. Cognition & emotion, 27(8), 1469-1477.
Larsen, J. T., Hemenover, S. H., Norris, C. J., & Cacioppo, J. T. (2003). Turning adversity to
advantage: On the virtues of the coactivation of positive and negative emotions. In A
psychology of human strengths: Fundamental questions and future directions for a
168

positive psychology. (pp. 211-225). American Psychological Association.
https://doi.org/10.1037/10566-015
Larsen, J. T., & McGraw, A. P. (2011). Further evidence for mixed emotions. J Pers Soc
Psychol, 100(6), 1095-1110. https://doi.org/10.1037/a0021846
Larsen, J. T., & McGraw, A. P. (2014). The case for mixed emotions. Social and Personality
Psychology Compass, 8(6), 263-274.
Larsen, J. T., McGraw, A. P., & Cacioppo, J. T. (2001). Can people feel happy and sad at the
same time? Journal of Personality and Social Psychology, 81(4), 684.
Larsen, J. T., McGraw, A. P., Mellers, B. A., & Cacioppo, J. T. (2004). The agony of victory and
thrill of defeat: Mixed emotional reactions to disappointing wins and relieving losses.
Psychol Sci, 15(5), 325-330. https://doi.org/10.1111/j.0956-7976.2004.00677.x
Larsen, J. T., To, Y. M., & Fireman, G. (2007). Children's understanding and experience of
mixed emotions. Psychol Sci, 18(2), 186-191. https://doi.org/10.1111/j.1467-
9280.2007.01870.x
LeDoux, J. E., & Brown, R. (2017). A higher-order theory of emotional consciousness.
Proceedings of the National Academy of Sciences, 114(10), E2016.
https://doi.org/10.1073/pnas.1619316114
Lee, C. S., Aly, M., & Baldassano, C. (2021). Anticipation of temporally structured events in the
brain. Elife, 10, e64972.
Lee, J. Y., Lindquist, K. A., & Nam, C. S. (2017). Emotional granularity effects on event-related
brain potentials during affective picture processing. Frontiers in Human Neuroscience,
11, 133.
Lench, H. C., Bench, S. W., & Flores, S. A. (2013). Searching for evidence, not a war: Reply to
Lindquist, Siegel, Quigley, and Barrett (2013).
Levenson, R. W., Ekman, P., & Friesen, W. V. (1990). Voluntary facial action generates
emotion-specific autonomic nervous system activity. Psychophysiology, 27(4), 363-384.
https://www.ncbi.nlm.nih.gov/pubmed/2236440
Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V. (1992). Emotion and autonomic
nervous system activity in the Minangkabau of west Sumatra. J Pers Soc Psychol, 62(6),
972-988. https://www.ncbi.nlm.nih.gov/pubmed/1619551
Lewis, P. A., Critchley, H., Rotshtein, P., & Dolan, R. (2006). Neural correlates of processing
valence and arousal in affective words. Cerebral Cortex, 17(3), 742-748.
Liemburg, E. J., Swart, M., Bruggeman, R., Kortekaas, R., Knegtering, H., Ćurčić-Blake, B., &
Aleman, A. (2012). Altered resting state connectivity of the default mode network in
alexithymia. Social cognitive and affective neuroscience, 7(6), 660-666.
169

Lin, W.-J., Horner, A. J., & Burgess, N. (2016). Ventromedial prefrontal cortex, adding value to
autobiographical memories. Scientific Reports, 6(1), 1-10.
Lindquist, K. A., Satpute, A. B., Wager, T. D., Weber, J., & Barrett, L. F. (2015). The Brain
Basis of Positive and Negative Affect: Evidence from a Meta-Analysis of the Human
Neuroimaging Literature. Cerebral Cortex, 26(5), 1910-1922.
https://doi.org/10.1093/cercor/bhv001
Livet, P. (2010). Rational choice, neuroeconomy and mixed emotions. Philos Trans R Soc Lond
B Biol Sci, 365(1538), 259-269. https://doi.org/10.1098/rstb.2009.0177
Loaiza, J. R. (2021). Emotions and the problem of variability. Review of Philosophy and
Psychology, 12(2), 329-351.
Lomas, T. (2017). The value of ambivalent emotions: a cross-cultural lexical analysis.
Qualitative Research in Psychology, 1-25.
https://doi.org/10.1080/14780887.2017.1400143
Lopez-Fernandez, O., Romo, L., Kern, L., Rousseau, A., Graziani, P., Rochat, L., Achab, S.,
Zullino, D., Landrø, N. I., & Zacarés, J. J. (2022). Perceptions underlying addictive
technology use patterns: insights for cognitive-behavioural therapy. International journal
of environmental research and public health, 19(1), 544.
Losada, A., Marquez-Gonzalez, M., Vara-Garcia, C., Gallego-Alberto, L., Romero-Moreno, R.,
& Pillemer, K. (2018). Ambivalence and guilt feelings: Two relevant variables for
understanding caregivers' depressive symptomatology. Clin Psychol Psychother, 25(1),
59-64. https://doi.org/10.1002/cpp.2116
Luo, Y. L. L., Welker, K. M., Way, B., DeWall, N., Bushman, B. J., Wildschut, T., & Sedikides,
C. (2017). 5-HTTLPR polymorphism is associated with nostalgia proneness: The role of
neuroticism. Social Neuroscience, 1-8. https://doi.org/10.1080/17470919.2017.1414717
Lupinacci, L. (2021). ‘Absentmindedly scrolling through nothing’: liveness and compulsory
continuous connectedness in social media. Media, Culture & Society, 43(2), 273-290.
Luttrell, A., Petty, R. E., & Briñol, P. (2020). The interactive effects of ambivalence and
certainty on political opinion stability. Journal of Social and Political Psychology, 8(2),
525-541.
Luttrell, A., Stillman, P. E., Hasinski, A. E., & Cunningham, W. A. (2016). Neural dissociations
in attitude strength: Distinct regions of cingulate cortex track ambivalence and certainty.
J Exp Psychol Gen, 145(4), 419-433. https://doi.org/10.1037/xge0000141
Lyon, P. (2015). The cognitive cell: bacterial behavior reconsidered. Frontiers in microbiology,
6, 264.
170

Lyon, P., & Kuchling, F. (2021). Valuing what happens: a biogenic approach to valence and
(potentially) affect. Philosophical Transactions of the Royal Society B, 376(1820),
20190752.
MacLean, K. E. (2022). Designing affective haptic experience for wellness and social
communication: where designers need affective neuroscience and psychology. Current
Opinion in Behavioral Sciences, 45, 101113.
Majeed, R. (2022). Does the Problem of Variability Justify Barrett’s Emotion Revolution?
Review of Philosophy and Psychology, 1-21.
Man, V., Nohlen, H. U., Melo, H., & Cunningham, W. A. (2017). Hierarchical Brain Systems
Support Multiple Representations of Valence and Mixed Affect. Emotion Review, 9(2),
124-132. https://doi.org/10.1177/1754073916667237
Marín-Morales, J., Llinares, C., Guixeres, J., & Alcañiz, M. (2020). Emotion recognition in
immersive virtual reality: From statistics to affective computing. Sensors, 20(18), 5163.
Marshall, P. J. (2009). Relating psychology and neuroscience: Taking up the challenges.
Perspectives on Psychological Science, 4(2), 113-125.
Matsunaga, M., Bai, Y., Yamakawa, K., Toyama, A., Kashiwagi, M., Fukuda, K., Oshida, A.,
Sanada, K., Fukuyama, S., Shinoda, J., Yamada, J., Sadato, N., & Ohira, H. (2013).
Brain-immune interaction accompanying odor-evoked autobiographic memory. PLoS
One, 8(8), e72523. https://doi.org/10.1371/journal.pone.0072523
Mattek, A. M., Wolford, G. L., & Whalen, P. J. (2017). A mathematical model captures the
structure of subjective affect. Perspectives on Psychological Science, 12(3), 508-526.
Medford, N., & Critchley, H. D. (2010). Conjoint activity of anterior insular and anterior
cingulate cortex: awareness and response. Brain Structure and Function, 214(5-6), 535-
549.
Mehling, W. E., Acree, M., Stewart, A., Silas, J., & Jones, A. (2018). The multidimensional
assessment of interoceptive awareness, version 2 (MAIA-2). PLoS One, 13(12),
e0208034.
Mejia, S. T., & Hooker, K. (2017). Mixed Emotions Within the Context of Goal Pursuit. Curr
Opin Behav Sci, 15, 46-50. https://doi.org/10.1016/j.cobeha.2017.05.015
Meyer, K., Kaplan, J. T., Essex, R., Webber, C., Damasio, H., & Damasio, A. (2010). Predicting
visual stimuli on the basis of activity in auditory cortices. Nat Neurosci, 13(6), 667-668.
https://doi.org/10.1038/nn.2533
Miller, G. A., & Keller, J. (2000). Psychology and neuroscience: Making peace. Current
Directions in Psychological Science, 9(6), 212-215.
171

Miller, N. E., & Kraeling, D. (1952). Displacement: greater generalization of approach than
avoidance in a generalized approach-avoidance conflict. Journal of Experimental
Psychology, 43(3), 217.
Mischkowski, D., Crocker, J., & Way, B. M. (2019). A Social Analgesic? Acetaminophen
(Paracetamol) Reduces Positive Empathy. Frontiers in Psychology, 10, 538-538.
https://doi.org/10.3389/fpsyg.2019.00538
Miyamoto, Y., Uchida, Y., & Ellsworth, P. C. (2010). Culture and mixed emotions: co-
occurrence of positive and negative emotions in Japan and the United States. Emotion,
10(3), 404-415. https://doi.org/10.1037/a0018430
Mobbs, D., Headley, D. B., Ding, W., & Dayan, P. (2020). Space, time, and fear: survival
computations along defensive circuits. Trends in cognitive sciences, 24(3), 228-241.
Moeller, J., Ivcevic, Z., Brackett, M. A., & White, A. E. (2018a). Mixed emotions: Network
analyses of intra-individual co-occurrences within and across situations. Emotion, 18(8),
1106-1121. https://doi.org/10.1037/emo0000419
Moeller, J., Ivcevic, Z., Brackett, M. A., & White, A. E. (2018b). Mixed Emotions: Network
Analyses of Intra-Individual Co-Occurrences Within and Across Situations. Emotion.
https://doi.org/10.1037/emo0000419
Moore, M. M., & Martin, E. A. (2022). Taking stock and moving forward: A personalized
perspective on mixed emotions. Perspectives on Psychological Science, 17(5), 1258-
1275.
Morawetz, C., Bode, S., Baudewig, J., Jacobs, A. M., & Heekeren, H. R. (2016). Neural
representation of emotion regulation goals. Hum Brain Mapp, 37(2), 600-620.
https://doi.org/10.1002/hbm.23053
Morgenroth, E., Vilaclara, L., Muszynski, M., Gaviria, J., Vuilleumier, P., & Van De Ville, D.
Probing Neurodynamics of Experienced Emotions-A Hitchhiker's Guide to Film Fmri.
Murray, R. J., Kreibig, S. D., Pehrs, C., Vuilleumier, P., Gross, J. J., & Samson, A. C. (2023).
Mixed emotions to social situations: An fMRI investigation. Neuroimage, 271, 119973.
Naselaris, T., Kay, K. N., Nishimoto, S., & Gallant, J. L. (2011). Encoding and decoding in
fMRI. Neuroimage, 56(2), 400-410. https://doi.org/10.1016/j.neuroimage.2010.07.073
Newman, D. B., & Sachs, M. E. (2022). Variation in bittersweet nostalgic feelings and their
divergent effects on daily well-being. Emotion.
Newman, D. B., Sachs, M. E., Stone, A. A., & Schwarz, N. (2019). Nostalgia and well-being in
daily life: An ecological validity perspective. Journal of Personality and Social
Psychology.
172

Ng, W. J. R., See, Y. H. M., & Wallace, L. E. (2022). When Objective Ambivalence Predicts
Subjective Ambivalence: An Affect–Cognition Matching Perspective. Personality and
Social Psychology Bulletin, 01461672221102015.
Nohlen, H. U., van Harreveld, F., Rotteveel, M., Lelieveld, G. J., & Crone, E. A. (2014).
Evaluating ambivalence: social-cognitive and affective brain regions associated with
ambivalent decision-making. Soc Cogn Affect Neurosci, 9(7), 924-931.
https://doi.org/10.1093/scan/nst074
Nook, E. C. (2021). Emotion differentiation and youth mental health: Current understanding and
open questions. Frontiers in Psychology, 12, 700298.
Norman, G. J., Norris, C. J., Gollan, J., Ito, T. A., Hawkley, L. C., Larsen, J. T., Cacioppo, J. T.,
& Berntson, G. G. (2011). Current Emotion Research in Psychophysiology: The
Neurobiology of Evaluative Bivalence. Emotion Review, 3(3), 349-359.
https://doi.org/10.1177/1754073911402403
Norris, C. J., Gollan, J., Berntson, G. G., & Cacioppo, J. T. (2010). The current status of research
on the structure of evaluative space. Biological Psychology, 84(3), 422-436.
https://doi.org/10.1016/j.biopsycho.2010.03.011
Nummenmaa, L., & Saarimaki, H. (2017). Emotions as discrete patterns of systemic activity.
Neurosci Lett. https://doi.org/10.1016/j.neulet.2017.07.012
Nummenmaa, L., & Saarimäki, H. (2019). Emotions as discrete patterns of systemic activity.
Neuroscience Letters, 693, 3-8. https://doi.org/10.1016/j.neulet.2017.07.012
O'Doherty, J., Kringelbach, M. L., Rolls, E. T., Hornak, J., & Andrews, C. (2001). Abstract
reward and punishment representations in the human orbitofrontal cortex. Nature
Neuroscience, 4(1), 95.
O'Kearney, R., Salmon, K., Liwag, M., Fortune, C. A., & Dawel, A. (2017). Emotional Abilities
in Children with Oppositional Defiant Disorder (ODD): Impairments in Perspective-
Taking and Understanding Mixed Emotions are Associated with High Callous-
Unemotional Traits. Child Psychiatry Hum Dev, 48(2), 346-357.
https://doi.org/10.1007/s10578-016-0645-4
O’Day, E. B., & Heimberg, R. G. (2021). Social media use, social anxiety, and loneliness: A
systematic review. Computers in Human Behavior Reports, 3, 100070.
Oba, K., Noriuchi, M., Atomi, T., Moriguchi, Y., & Kikuchi, Y. (2016). Memory and reward
systems coproduce 'nostalgic' experiences in the brain. Soc Cogn Affect Neurosci, 11(7),
1069-1077. https://doi.org/10.1093/scan/nsv073
Ochsner, K. N., Knierim, K., Ludlow, D. H., Hanelin, J., Ramachandran, T., Glover, G., &
Mackey, S. C. (2004). Reflecting upon feelings: an fMRI study of neural systems
supporting the attribution of emotion to self and other. Journal of cognitive neuroscience,
16(10), 1746-1772.
173

Ogg, M., Moraczewski, D., Kuchinsky, S. E., & Slevc, L. R. (2019). Separable neural
representations of sound sources: Speaker identity and musical timbre. Neuroimage, 191,
116-126. https://doi.org/10.1016/j.neuroimage.2019.01.075
Oh, V. Y. (2022). Torn between valences: Mixed emotions predict poorer psychological well-
being and job burnout. Journal of Happiness Studies, 23(5), 2171-2200.
Oh, V. Y., & Tong, E. M. (2022). Specificity in the Study of Mixed Emotions: A Theoretical
Framework. Personality and Social Psychology Review, 26(4), 283-314.
Oliver, M. B. (2008). Tender affective states as predictors of entertainment preference. Journal
of Communication, 58(1), 40-61.
Oliver, M. B., & Woolley, J. K. (2010). Tragic and poignant entertainment: The gratifications of
meaningfulness as emotional response. In The Routledge handbook of emotions and mass
media (pp. 148-161). Routledge.
Oosterwijk, S., Snoek, L., Rotteveel, M., Barrett, L. F., & Scholte, H. S. (2017). Shared states:
using MVPA to test neural overlap between self-focused emotion imagery and other-
focused emotion understanding. Social cognitive and affective neuroscience, 12(7), 1025-
1035. https://doi.org/10.1093/scan/nsx037
Panksepp, J. (1992). A critical role for "affective neuroscience" in resolving what is basic about
basic emotions. Psychol Rev, 99(3), 554-560.
https://www.ncbi.nlm.nih.gov/pubmed/1502276
Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions.
Oxford University Press.
Panksepp, J. (2003). At the interface of the affective, behavioral, and cognitive neurosciences:
Decoding the emotional feelings of the brain. Brain and cognition, 52(1), 4-14.
Panksepp, J. (2005). On the embodied neural nature of core emotional affects. Journal of
consciousness studies, 12(8-9), 158-184.
Paquette, S., Takerkart, S., Saget, S., Peretz, I., & Belin, P. (2018). Cross-classification of
musical and vocal emotions in the auditory cortex. Ann. NY Acad. Sci, 1423, 329-337.
https://doi.org/10.1111/nyas.13666
Parkinson, B., & Manstead, A. S. (2015). Current emotion research in social psychology:
Thinking about emotions and other people. Emotion Review, 7(4), 371-380.
Paulick, A., Jakovljevic, V., Zhang, S., Erickstad, M., Groisman, A., Meir, Y., Ryu, W. S.,
Wingreen, N. S., & Sourjik, V. (2017). Mechanism of bidirectional thermotaxis in
Escherichia coli. Elife, 6, e26607.
174

Peelen, M. V., Atkinson, A. P., & Vuilleumier, P. (2010). Supramodal representations of
perceived emotions in the human brain. Journal of neuroscience, 30(30), 10127-10134.
https://doi.org/https://doi.org/10.1523/JNEUROSCI.2161-10.2010
Peng, M., Johnson, C., Pollock, J., Glasspool, R., & Hams, P. (1992). Training Young Children
to Acknowledge Mixed Emotions. Cogn Emot, 6(5), 387-401.
https://doi.org/10.1080/02699939208409693
Perlovsky, L. (2009). Language and emotions: Emotional Sapir–Whorf hypothesis. Neural
Networks, 22(5), 518-526. https://doi.org/https://doi.org/10.1016/j.neunet.2009.06.034
Petrocelli, J. V., Tormala, Z. L., & Rucker, D. D. (2007). Unpacking attitude certainty: attitude
clarity and attitude correctness. Journal of Personality and Social Psychology, 92(1), 30.
Picard, F., & Craig, A. D. (2009). Ecstatic epileptic seizures: A potential window on the neural
basis for human self-awareness. Epilepsy & Behavior, 16(3), 539-546.
https://doi.org/10.1016/j.yebeh.2009.09.013
Picard, F., & Kurth, F. (2014). Ictal alterations of consciousness during ecstatic seizures.
Epilepsy Behav, 30, 58-61.
Pignatelli, M., & Beyeler, A. (2019). Valence coding in amygdala circuits. Current Opinion in
Behavioral Sciences, 26, 97-106.
Plante, C. N., Gentile, D. A., Groves, C. L., Modlin, A., & Blanco-Herrera, J. (2019). Video
games as coping mechanisms in the etiology of video game addiction. Psychology of
Popular Media Culture, 8(4), 385.
Poppa, T., & Bechara, A. (2018). The somatic marker hypothesis: Revisiting the role of the
‘body-loop’in decision-making. Current Opinion in Behavioral Sciences, 19, 61-66.
https://doi.org/10.1016/j.cobeha.2017.10.007

Preckel, K., Scheele, D., Eckstein, M., Maier, W., & Hurlemann, R. (2014). The influence of
oxytocin on volitional and emotional ambivalence. Social cognitive and affective
neuroscience, 10(7), 987-993.
Priester, J. R., & Petty, R. E. (1996). The gradual threshold model of ambivalence: relating the
positive and negative bases of attitudes to subjective ambivalence. Journal of personality
and social psychology, 71(3), 431.
Rafaeli, E., Rogers, G. M., & Revelle, W. (2007). Affective synchrony: individual differences in
mixed emotions. Pers Soc Psychol Bull, 33(7), 915-932.
https://doi.org/10.1177/0146167207301009
Rainville, P., Bechara, A., Naqvi, N., & Damasio, A. R. (2006). Basic emotions are associated
with distinct patterns of cardiorespiratory activity. Int J Psychophysiol, 61(1), 5-18.
https://doi.org/10.1016/j.ijpsycho.2005.10.024
175

Raizada, R. D., & Kriegeskorte, N. (2010). Pattern-information fMRI: new questions which it
opens up, and challenges which face it. International Journal of Imaging Systems
Technology, 20, 31-41.
Raizada, R. D., Tsao, F. M., Liu, H. M., Holloway, I. D., Ansari, D., & Kuhl, P. K. (2010).
Linking brain-wide multivoxel activation patterns to behaviour: Examples from language
and math. Neuroimage, 51(1), 462-471.
https://doi.org/10.1016/j.neuroimage.2010.01.080
Ramos, N. S., Fernandez-Berrocal, P., & Extremera, N. (2007). Perceived emotional intelligence
facilitates cognitive-emotional processes of adaptation to an acute stressor. Cognition and
Emotion, 21(4), 758-772.
Redgrave, P., Prescott, T. J., & Gurney, K. (1999). The basal ganglia: a vertebrate solution to the
selection problem? Neuroscience, 89(4), 1009-1023.
Reynolds, S. M., & Berridge, K. C. (2002). Positive and Negative Motivation in Nucleus
Accumbens Shell: Bivalent Rostrocaudal Gradients for GABA-Elicited Eating, Taste
“Liking”/“Disliking” Reactions, Place Preference/Avoidance, and Fear. The Journal of
Neuroscience, 22(16), 7308-7320. https://doi.org/10.1523/jneurosci.22-16-07308.2002
Reynolds, S. M., & Berridge, K. C. (2008). Emotional environments retune the valence of
appetitive versus fearful functions in nucleus accumbens. Nature Neuroscience, 11(4),
423-425.
Ritchie, J. B., Kaplan, D. M., & Klein, C. (2019). Decoding the Brain: Neural Representation
and the Limits of Multivariate Pattern Analysis in Cognitive Neuroscience. Br J Philos
Sci, 70(2), 581-607. https://doi.org/10.1093/bjps/axx023
Roberts, J. A., & David, M. E. (2020). The social media party: Fear of missing out (FoMO),
social media intensity, connection, and well-being. International Journal of Human–
Computer Interaction, 36(4), 386-392.
Robinson, M. D., Irvin, R. L., Persich, M. R., & Krishnakumar, S. (2020). Bipolar or
independent? Relations between positive and negative affect vary by emotional
intelligence. Affective Science, 1, 225-236.
Rolls, E. T., & Grabenhorst, F. (2008). The orbitofrontal cortex and beyond: from affect to
decision-making. Prog Neurobiol, 86(3), 216-244.
https://doi.org/10.1016/j.pneurobio.2008.09.001
Rosen, G. (2009). Nostalgia: a ‘forgotten’ psychological disorder. Psychological Medicine, 5(4),
340-354. https://doi.org/10.1017/S003329170005697X
Roth, M. S. (1991). Dying of the past: Medical studies of nostalgia in nineteenth-century France.
History and Memory, 3(1), 5-29.
176

Rothman, N. B. (2011). Steering sheep: How expressed emotional ambivalence elicits
dominance in interdependent decision making contexts. Organizational Behavior and
Human Decision Processes, 116(1), 66-82.
https://doi.org/https://doi.org/10.1016/j.obhdp.2011.05.004
Routledge, C., Wildschut, T., Sedikides, C., & Juhl, J. (2013). Nostalgia as a Resource for
Psychological Health and Well-Being. Social and Personality Psychology Compass,
7(11), 808-818. https://doi.org/doi:10.1111/spc3.12070
Roy, M., Shohamy, D., & Wager, T. D. (2012). Ventromedial prefrontal-subcortical systems and
the generation of affective meaning. Trends in cognitive sciences, 16(3), 147-156.
https://doi.org/10.1016/j.tics.2012.01.005
Ruba, A. L., & Repacholi, B. M. (2020). Beyond language in infant emotion concept
development. Emotion Review, 12(4), 255-258.
Rucker, D. D., & Petty, R. E. (2004). When resistance is futile: Consequences of failed
counterarguing for attitude certainty. Journal of Personality and Social Psychology,
86(2), 219.
Rushworth, M. F. S., & Behrens, T. E. J. (2008). Choice, uncertainty and value in prefrontal and
cingulate cortex [Review Article]. Nature Neuroscience, 11, 389.
https://doi.org/10.1038/nn2066
Russell, J. A. (2003). Core affect and the psychological construction of emotion. Psychol Rev,
110(1), 145-172. https://www.ncbi.nlm.nih.gov/pubmed/12529060
Russell, J. A. (2017). Mixed emotions viewed from the psychological constructionist
perspective. Emotion Review, 9(2), 111-117.
Russell, J. A., & Carroll, J. M. (1999). On the bipolarity of positive and negative affect. Psychol
Bull, 125(1), 3-30. https://www.ncbi.nlm.nih.gov/pubmed/9990843
Saarimäki, H. (2021). Naturalistic stimuli in affective neuroimaging: A review. Frontiers in
Human Neuroscience, 15, 675068.
Saarimaki, H., Ejtehadian, L. F., Glerean, E., Jaaskelainen, I. P., Vuilleumier, P., Sams, M., &
Nummenmaa, L. (2018). Distributed affective space represents multiple emotion
categories across the human brain. Soc Cogn Affect Neurosci, 13(5), 471-482.
https://doi.org/10.1093/scan/nsy018
Saarimaki, H., Gotsopoulos, A., Jaaskelainen, I. P., Lampinen, J., Vuilleumier, P., Hari, R.,
Sams, M., & Nummenmaa, L. (2016). Discrete Neural Signatures of Basic Emotions.
Cereb Cortex, 26(6), 2563-2573. https://doi.org/10.1093/cercor/bhv086
Saarimaki, H., Nummenmaa, L., Volynets, S., Santavirta, S., Aksiuto, A., Sams, M.,
Jaaskelainen, I. P., & Lahnakoski, J. (2023). Cerebral Topographies of Perceived and Felt
Emotions. bioRxiv, 2023.2002. 2008.521183.
177

Sachs, M. E., Habibi, A., Damasio, A., & Kaplan, J. (2018). Decoding the neural signatures of
emotions expressed through sound. Neuroimage, 174, 1-10.
https://doi.org/10.1016/j.neuroimage.2018.02.058
Sachs, M. E., Ochsner, K., & Baldassano, C. (2023). Brain state dynamics reflect emotion
transitions induced by music. bioRxiv, 2023.2003. 2001.530528.
Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. P. (1995). Emotional
Attention, clarity, and repair: exploring emotional intelligence using the trait meta-mood
scale.
Samson, A. C., Kreibig, S. D., Soderstrom, B., Wade, A. A., & Gross, J. J. (2016). Eliciting
positive, negative and mixed emotional states: A film library for affective scientists.
Cognition and Emotion, 30(5), 827-856.
Saunders, B., Lin, H., Milyavskaya, M., & Inzlicht, M. (2017). The emotive nature of conflict
monitoring in the medial prefrontal cortex. Int J Psychophysiol, 119, 31-40.
https://doi.org/10.1016/j.ijpsycho.2017.01.004
Saxbe, D. E., Yang, X.-F., Borofsky, L. A., & Immordino-Yang, M. H. (2013). The embodiment
of emotion: language use during the feeling of social emotions predicts cortical
somatosensory activity. Social cognitive and affective neuroscience, 8(7), 806-812.
Scarantino, A. (2012). Functional specialization does not require a one-to-one mapping between
brain regions and emotions. Behavioral and Brain Sciences, 35(3), 161.
https://doi.org/10.1016/j.neubiorev.2015.08.003
Scarantino, A., & Griffiths, P. (2011). Don’t give up on basic emotions. Emotion Review, 3(4),
444-454.
Schacter, D. L., Benoit, R. G., & Szpunar, K. K. (2017). Episodic future thinking: Mechanisms
and functions. Current Opinion in Behavioral Sciences, 17, 41-50.
https://doi.org/10.1016/j.cobeha.2017.06.002
Scheibe, S., Blanchard-Fields, F., Wiest, M., & Freund, A. M. (2011). Is longing only for
Germans? A cross-cultural comparison of Sehnsucht in Germany and the United States.
Developmental Psychology, 47(3), 603-618. https://doi.org/10.1037/a0021807
Scheibe, S., & Freund, A. M. (2008). Approaching Sehnsucht (Life Longings) from a Life-Span
Perspective: The Role of Personal Utopias in Development. Research in Human
Development, 5(2), 121-133. https://doi.org/10.1080/15427600802034868
Scheibe, S., Freund, A. M., & Baltes, P. B. (2007). Toward a developmental psychology of
Sehnsucht (life longings): The optimal (utopian) life. Developmental Psychology, 43(3),
778.
178

Scherer, K. R. (2012). Neuroscience findings are consistent with appraisal theories of emotion;
but does the brain" respect” constructionism? Behavioral and Brain Sciences, 35(3), 163-
164.
Scherer, K. R., & Wallbott, H. G. (1994). Evidence for universality and cultural variation of
differential emotion response patterning. Journal of Personality and Social Psychology,
66(2), 310.
Schimmack, U. (2001). Pleasure, displeasure, and mixed feelings: Are semantic opposites
mutually exclusive? Cognition & emotion, 15(1), 81-97.
Schimmack, U. (2007). That's disgusting! …, but very amusing: Mixed feelings of amusement
and disgust AU - Hemenover, Scott H. Cognition and Emotion, 21(5), 1102-1113.
https://doi.org/10.1080/02699930601057037
Schmaelzle, R., & Huskey, R. (2023). Integrating Media Content Analysis, Reception Analysis,
and Media Effects Studies. Frontiers in Neuroscience, 17, 612.
Schneider, I. K., & Schwarz, N. (2017). Mixed feelings: the case of ambivalence. Current
Opinion in Behavioral Sciences, 15, 39-45.
https://doi.org/https://doi.org/10.1016/j.cobeha.2017.05.012
Schulte, F. P., Maderwald, S., Kraemer, N. C., & Brand, M. (2012). Bittersweet-neither happy
nor sad. An experimental comparison of the neural effects of bittersweet, negative, and
positive film clips using 7T fMRI. Proc. Intl. Soc. Mag. Reson. Med., 20, 2136.
Schultz, D., Izard, C. E., & Abe, J. (2005). The emotion systems and the development of
emotional intelligence. Emotional Intelligence, 51-67.
Schutte, N. S., Malouff, J. M., Hall, L. E., Haggerty, D. J., Cooper, J. T., Golden, C. J., &
Dornheim, L. (1998). Development and validation of a measure of emotional
intelligence. Personality and individual differences, 25(2), 167-177.
Sebastian, C. L., Fontaine, N. M., Bird, G., Blakemore, S.-J., De Brito, S. A., McCrory, E. J., &
Viding, E. (2012). Neural processing associated with cognitive and affective Theory of
Mind in adolescents and adults. Social cognitive and affective neuroscience, 7(1), 53-63.
https://doi.org/10.1093/scan/nsr023
Sedikides, C., Wildschut, T., Arndt, J., & Routledge, C. (2008). Nostalgia: Past, Present, and
Future. Current Directions in Psychological Science, 17(5), 304-307.
https://doi.org/10.1111/j.1467-8721.2008.00595.x
Sedikides, C., Wildschut, T., Cheung, W.-Y., Routledge, C., Hepper, E. G., Arndt, J., Vail, K.,
Zhou, X., Brackstone, K., & Vingerhoets, A. J. (2016). Nostalgia fosters self-continuity:
Uncovering the mechanism (social connectedness) and consequence (eudaimonic well-
being). Emotion, 16(4), 524.
179

Seeley, W. W. (2019). The salience network: a neural system for perceiving and responding to
homeostatic demands. Journal of neuroscience, 39(50), 9878-9882.
Seïler, N. R., & Craig, P. (2016). Empathetic technology. In Emotions, technology, and design
(pp. 55-81). Elsevier.
Seitz, R. J., Nickel, J., & Azari, N. P. (2006). Functional modularity of the medial prefrontal
cortex: involvement in human empathy. Neuropsychology, 20(6), 743.
https://doi.org/10.1037/0894-4105.20.6.743
Seth, A. K., & Friston, K. J. (2016). Active interoceptive inference and the emotional brain.
Philosophical Transactions of the Royal Society B: Biological Sciences, 371(1708),
20160007.
Seth, A. K., Suzuki, K., & Critchley, H. D. (2012). An interoceptive predictive coding model of
conscious presence. Frontiers in Psychology, 2, 395-395.
https://doi.org/10.3389/fpsyg.2011.00395
Sharpe, M. J., Wikenheiser, A. M., Niv, Y., & Schoenbaum, G. (2015). The State of the
Orbitofrontal Cortex. Neuron, 88(6), 1075-1077.
https://doi.org/10.1016/j.neuron.2015.12.004
Shirai, M., & Kimura, T. (2022). Degree of Meaningfulness of an Event’s Ending Can Modulate
Mixed Emotional Experiences Among Japanese Undergraduates. Perceptual and Motor
Skills, 129(4), 1137-1150.
Simmons, A., Stein, M. B., Matthews, S. C., Feinstein, J. S., & Paulus, M. P. (2006). Affective
ambiguity for a group recruits ventromedial prefrontal cortex. Neuroimage, 29(2), 655-
661. https://doi.org/10.1016/j.neuroimage.2005.07.040
Simons, J. J. P., Schneider, I. K., & Sanchez-Burks, J. (2018). Ambivalence, the person and the
attitude object: Individual differences in the experience of ambivalence.
Sincoff, J. B. (1990). The psychological characteristics of ambivalent people. Clinical
Psychology Review, 10(1), 43-68.
Singer, T., Seymour, B., O'doherty, J., Kaube, H., Dolan, R. J., & Frith, C. D. (2004). Empathy
for pain involves the affective but not sensory components of pain. Science, 303(5661),
1157-1162. https://doi.org/10.1126/science.1093535
Skerry, A. E., & Saxe, R. (2014). A common neural code for perceived and inferred emotion.
Journal of neuroscience, 34(48), 15997-16008.
Slater, M. D., Oliver, M. B., & Appel, M. (2019). Poignancy and mediated wisdom of
experience: Narrative impacts on willingness to accept delayed rewards. Communication
Research, 46(3), 333-354.
180

Smith, J. P., Glass, D. J., & Fireman, G. (2015). The understanding and experience of mixed
emotions in 3-5-year-old children. J Genet Psychol, 176(1-2), 65-81.
https://doi.org/10.1080/00221325.2014.1002750
Smith, R., & Lane, R. D. (2015). The neural basis of one's own conscious and unconscious
emotional states. Neuroscience & Biobehavioral Reviews, 57, 1-29.
https://doi.org/10.1016/j.neubiorev.2015.08.003
Song, H., Zou, Z., Kou, J., Liu, Y., Yang, L., Zilverstand, A., d'Oleire Uquillas, F., & Zhang, X.
(2015). Love-related changes in the brain: a resting-state functional magnetic resonance
imaging study. Front Hum Neurosci, 9, 71. https://doi.org/10.3389/fnhum.2015.00071
Steele, H., Steele, M., Croft, C., & Fonagy, P. (1999). Infant-Mother Attachment at One Year
Predicts Children’s Understanding of Mixed Emotions at Six Years. Social Development,
8(2), 161-178. https://doi.org/doi:10.1111/1467-9507.00089
Stephens, C. L., Christie, I. C., & Friedman, B. H. (2010). Autonomic specificity of basic
emotions: evidence from pattern classification and cluster analysis. Biol Psychol, 84(3),
463-473. https://doi.org/10.1016/j.biopsycho.2010.03.014
Stetson, C., Fiesta, M. P., & Eagleman, D. M. (2007). Does Time Really Slow Down during a
Frightening Event? PLoS One, 2(12), e1295.
https://doi.org/10.1371/journal.pone.0001295
Stolyarova, A. (2018). Solving the Credit Assignment Problem With the Prefrontal Cortex. Front
Neurosci, 12, 182. https://doi.org/10.3389/fnins.2018.00182
Strigo, I. A., & Craig, A. (2016). Interoception, homeostatic emotions and sympathovagal
balance. Philosophical Transactions of the Royal Society B: Biological Sciences,
371(1708), 20160010. https://doi.org/10.1098/rstb.2016.0010
Sukikara, M. H., Mota-Ortiz, S. R., Baldo, M. V., Felício, L. F., & Canteras, N. S. (2006). A
Role for the Periaqueductal Gray in Switching Adaptive Behavioral Responses. The
Journal of Neuroscience, 26(9), 2583-2589. https://doi.org/10.1523/jneurosci.4279-
05.2006
Sun, S., Zhen, S., Fu, Z., Wu, D.-A., Shimojo, S., Adolphs, R., Yu, R., & Wang, S. (2017).
Decision ambiguity is mediated by a late positive potential originating from cingulate
cortex. Neuroimage, 157, 400-414.
Sutherland, M. T., Carroll, A. J., Salmeron, B. J., Ross, T. J., & Stein, E. A. (2013). Insula's
functional connectivity with ventromedial prefrontal cortex mediates the impact of trait
alexithymia on state tobacco craving. Psychopharmacology (Berl), 228(1), 143-155.
https://doi.org/10.1007/s00213-013-3018-8
Szabó, Á. G., Farkas, K., Marosi, C., Kozák, L. R., Rudas, G., Réthelyi, J., & Csukly, G. (2017).
Impaired mixed emotion processing in the right ventrolateral prefrontal cortex in
181

schizophrenia: an fMRI study. BMC Psychiatry, 17(1), 391.
https://doi.org/10.1186/s12888-017-1558-x
Takano, R., & Nomura, M. (2022). Neural representations of awe: Distinguishing common and
distinct neural mechanisms. Emotion, 22(4), 669.
Tay, L., & Kuykendall, L. (2017). Why self-reports of happiness and sadness may not
necessarily contradict bipolarity: a psychometric review and proposal. Emotion Review,
9(2), 146-154.
Tettamanti, M., Rognoni, E., Cafiero, R., Costa, T., Galati, D., & Perani, D. (2012). Distinct
pathways of neural coupling for different basic emotions. Neuroimage, 59(2), 1804-1817.
https://doi.org/10.1016/j.neuroimage.2011.08.018
Thagard, P., Larocque, L., & Kajić, I. (2023). Emotional change: Neural mechanisms based on
semantic pointers. Emotion, 23(1), 182.
Thompson, M. M., Zanna, M. P., & Griffin, D. W. (1995). Let’s not be indifferent about
(attitudinal) ambivalence. Attitude strength: Antecedents and consequences, 4, 361-386.
Thompson, R. J., & Boden, M. T. (2019). State emotional clarity and attention to emotion: A
naturalistic examination of their associations with each other, affect, and context.
Cognition and Emotion, 33(7), 1514-1522.
Todd, R. M., Ehlers, M. R., Müller, D. J., Robertson, A., Palombo, D. J., Freeman, N., Levine,
B., & Anderson, A. K. (2015). Neurogenetic variations in norepinephrine availability
enhance perceptual vividness. Journal of neuroscience, 35(16), 6506-6516.
Todd, R. M., Miskovic, V., Chikazoe, J., & Anderson, A. K. (2020). Emotional objectivity:
Neural representations of emotions and their interaction with cognition. Annual review of
psychology, 71, 25-48.
Tohidifar, P., Plutz, M. J., Ordal, G. W., & Rao, C. V. (2020). The mechanism of bidirectional
pH taxis in Bacillus subtilis. Journal of bacteriology, 202(4), e00491-00419.
Tormala, Z. L., & Rucker, D. D. (2007). Attitude certainty: A review of past findings and
emerging perspectives. Social and Personality Psychology Compass, 1(1), 469-492.
Torres, E. P., Torres, E. A., Hernández-Álvarez, M., & Yoo, S. G. (2020). EEG-based BCI
emotion recognition: A survey. Sensors, 20(18), 5083.
Tottenham, N. (2014). The importance of early experiences for neuro-affective development.
Curr Top Behav Neurosci, 16, 109-129. https://doi.org/10.1007/7854_2013_254
Trost, W., Ethofer, T., Zentner, M., & Vuilleumier, P. (2012). Mapping aesthetic musical
emotions in the brain. Cereb Cortex, 22(12), 2769-2783.
https://doi.org/10.1093/cercor/bhr353
182

Tsai, J. L., Chentsova-Dutton, Y., Freire-Bebeau, L., & Przymus, D. E. (2002). Emotional
expression and physiology in European Americans and Hmong Americans. Emotion,
2(4), 380-397.
Tye, K. M. (2018). Neural Circuit Motifs in Valence Processing. Neuron, 100(2), 436-452.
https://doi.org/https://doi.org/10.1016/j.neuron.2018.10.001
Uddin, L. Q., Supekar, K. S., Ryali, S., & Menon, V. (2011). Dynamic reconfiguration of
structural and functional connectivity across core neurocognitive brain networks with
development. J Neurosci, 31(50), 18578-18589. https://doi.org/10.1523/jneurosci.4465-
11.2011
Vaccaro, A. G., & Fleming, S. M. (2018). Thinking about thinking: A coordinate-based meta-
analysis of neuroimaging studies of metacognitive judgements. Brain and Neuroscience
Advances, 2, 2398212818810591. https://doi.org/10.1177/2398212818810591
Vaccaro, A. G., Heydari, P., Christov-Moore, L., Damasio, A., & Kaplan, J. T. (2022).
Perspective-taking is associated with increased discriminability of affective states in the
ventromedial prefrontal cortex. Social cognitive and affective neuroscience, 17(12),
1082-1090.
Vaccaro, A. G., Kaplan, J. T., & Damasio, A. (2020). Bittersweet: The Neuroscience of
Ambivalent Affect. Perspectives on Psychological Science, 15(5), 1187-1199.
https://doi.org/10.1177/1745691620927708
van der Heiden, L., Scherpiet, S., Konicar, L., Birbaumer, N., & Veit, R. (2013). Inter-individual
differences in successful perspective taking during pain perception mediates emotional
responsiveness in self and others: An fMRI study. Neuroimage, 65, 387-394.
https://doi.org/https://doi.org/10.1016/j.neuroimage.2012.10.003
van Harreveld, F., Nohlen, H. U., & Schneider, I. K. (2015). The ABC of ambivalence:
Affective, behavioral, and cognitive consequences of attitudinal conflict. In Advances in
experimental social psychology (Vol. 52, pp. 285-324). Elsevier.
Van Harreveld, F., Rutjens, B. T., Rotteveel, M., Nordgren, L. F., & Van Der Pligt, J. (2009).
Ambivalence and decisional conflict as a cause of psychological discomfort: Feeling
tense before jumping off the fence. Journal of Experimental Social Psychology, 45(1),
167-173.
van Tilburg, W. A. P., Bruder, M., Wildschut, T., Sedikides, C., & Goritz, A. S. (2018). An
appraisal profile of nostalgia. Emotion. https://doi.org/10.1037/emo0000417
Vandaele, Y., & Daeppen, J.-B. (2022). From concepts to treatment: a dialog between a
preclinical researcher and a clinician in addiction medicine. Translational Psychiatry,
12(1), 401.
183

Vandekerckhove, M., & Panksepp, J. (2011). A neurocognitive theory of higher mental
emergence: From anoetic affective experiences to noetic knowledge and autonoetic
awareness. Neuroscience & Biobehavioral Reviews, 35(9), 2017-2025.
Vazard, J. (2022). Feeling the Unknown: Emotions of Uncertainty and Their Valence.
Erkenntnis, 1-20.
Veilleux, J. C., Conrad, M., & Kassel, J. D. (2013). Cue-induced cigarette craving and mixed
emotions: a role for positive affect in the craving process. Addict Behav, 38(4), 1881-
1889. https://doi.org/10.1016/j.addbeh.2012.12.006
Venkatraman, A., Edlow, B. L., & Immordino-Yang, M. H. (2017). The Brainstem in Emotion:
A Review. Front Neuroanat, 11, 15. https://doi.org/10.3389/fnana.2017.00015
Verplanken, B. (2012). When bittersweet turns sour: Adverse effects of nostalgia on habitual
worriers [doi:10.1002/ejsp.1852].
Viinikainen, M., Jääskeläinen, I. P., Alexandrov, Y., Balk, M. H., Autti, T., & Sams, M. (2010).
Nonlinear relationship between emotional valence and brain activity: evidence of
separate negative and positive valence dimensions. Human brain mapping, 31(7), 1030-
1040.
Viinikainen, M., Kätsyri, J., & Sams, M. (2012). Representation of perceived sound valence in
the human brain. Human brain mapping, 33(10), 2295-2305.
https://doi.org/doi:10.1002/hbm.21362
Voutilainen, L., Henttonen, P., Kahri, M., Kivioja, M., Ravaja, N., Sams, M., & Peräkylä, A.
(2014). Affective stance, ambivalence, and psychophysiological responses during
conversational storytelling. Journal of Pragmatics, 68, 1-24.
Wahl-Jorgensen, K. (2019). Emotions, media and politics. John Wiley & Sons.
Wang, Y., Song, W., Tao, W., Liotta, A., Yang, D., Li, X., Gao, S., Sun, Y., Ge, W., & Zhang,
W. (2022). A systematic review on affective computing: Emotion models, databases, and
recent advances. Information Fusion.
Weissman, D. G., Guyer, A. E., Ferrer, E., Robins, R. W., & Hastings, P. D. (2018). Adolescents'
brain-autonomic coupling during emotion processing. Neuroimage, 183, 818-827.
https://doi.org/10.1016/j.neuroimage.2018.08.069
White, T. P., Joseph, V., Francis, S. T., & Liddle, P. F. (2010). Aberrant salience network
(bilateral insula and anterior cingulate cortex) connectivity during information processing
in schizophrenia. Schizophrenia research, 123(2-3), 105-115.
Whitesell, N. R., & Harter, S. (1989). Children's Reports of Conflict between Simultaneous
Opposite-Valence Emotions. Child Development, 60(3), 673-682.
https://doi.org/10.2307/1130732
184

Wicker, B., Perrett, D. I., Baron-Cohen, S., & Decety, J. (2003). Being the target of another’s
emotion: a PET study. Neuropsychologia, 41(2), 139-146. https://doi.org/10.1016/s0028-
3932(02)00144-6
Wildschut, T., Sedikides, C., Arndt, J., & Routledge, C. (2006). Nostalgia: content, triggers,
functions. J Pers Soc Psychol, 91(5), 975-993. https://doi.org/10.1037/0022-
3514.91.5.975
Wilson, S. J., Creswell, K. G., Sayette, M. A., & Fiez, J. A. (2013). Ambivalence about smoking
and cue-elicited neural activity in quitting-motivated smokers faced with an opportunity
to smoke. Addict Behav, 38(2), 1541-1549. https://doi.org/10.1016/j.addbeh.2012.03.020
Wilson-Mendenhall, C. D., Barrett, L. F., & Barsalou, L. W. (2014). Variety in emotional life:
within-category typicality of emotional experiences is associated with neural activity in
large-scale brain networks. Social cognitive and affective neuroscience, 10(1), 62-71.
https://doi.org/10.1093/scan/nsu037
Wiltshire, T. J., & Fiore, S. M. (2014). Social cognitive and affective neuroscience in human–
machine systems: A roadmap for improving training, human–robot interaction, and team
performance. IEEE Transactions on Human-Machine Systems, 44(6), 779-787.
Winecoff, A., Clithero, J. A., Carter, R. M., Bergman, S. R., Wang, L., & Huettel, S. A. (2013).
Ventromedial prefrontal cortex encodes emotional value. Journal of neuroscience,
33(27), 11032-11039. https://doi.org/https://doi.org/10.1523/JNEUROSCI.4317-12.2013
Winkler, A. M., Ridgway, G. R., Webster, M. A., Smith, S. M., & Nichols, T. E. (2014).
Permutation inference for the general linear model. Neuroimage, 92, 381-397.
Wintre, M. G., & Vallance, D. D. (1994). A developmental sequence in the comprehension of
emotions: Intensity, multiple emotions, and valence. Developmental Psychology, 30(4),
509-514. https://doi.org/10.1037/0012-1649.30.4.509
Wu, H., Vaccaro, A. G., & Kaplan, J. (2023). Neural event segmentation during naturalistic story
listening. In Prep.
Wulf, T., Bowman, N. D., Velez, J. A., & Breuer, J. (2020). Once upon a game: Exploring video
game nostalgia and its impact on well-being. Psychology of Popular Media, 9(1), 83.
Wulf, T., Breuer, J. S., & Schmitt, J. B. (2022). Escaping the pandemic present: The relationship
between nostalgic media use, escapism, and well-being during the COVID-19 pandemic.
Psychology of Popular Media, 11(3), 258.
Wulf, T., Rieger, D., & Schmitt, J. B. (2018). Blissed by the past: Theorizing media-induced
nostalgia as an audience response factor for entertainment and well-being. Poetics, 69,
70-80.
185

Yan, Q., Chen, Y., Jiang, Y., & Chen, H. (2022). Exploring the impact of envy and admiration
on social media fatigue: Social media loneliness and anxiety as mediators. Current
Psychology, 1-14.
Yang, Y., & Sourjik, V. (2012). Opposite responses by different chemoreceptors set a tunable
preference point in E scherichia coli pH taxis. Molecular microbiology, 86(6), 1482-
1489.
Yang, Z., Izuma, K., & Cai, H. (2022). Nostalgia in the Brain. Current Opinion in Psychology,
101523.
Yang, Z., Wildschut, T., Izuma, K., Gu, R., Luo, Y. L., Cai, H., & Sedikides, C. (2022). Patterns
of brain activity associated with nostalgia: a social-cognitive neuroscience perspective.
Social cognitive and affective neuroscience, 17(12), 1131-1144.
Yates, T. S., Skalaban, L. J., Ellis, C. T., Bracher, A. J., Baldassano, C., & Turk-Browne, N. B.
(2022). Neural event segmentation of continuous experience in human infants.
Proceedings of the National Academy of Sciences, 119(43), e2200257119.
Young, P. T. (1918). An experimental study of mixed feelings. American Journal of Psychology,
29, 237-271.
Zajdel, R. T., Bloom, J. M., Fireman, G., & Larsen, J. T. (2013). Children's understanding and
experience of mixed emotions: the roles of age, gender, and empathy. J Genet Psychol,
174(5-6), 582-603. https://doi.org/10.1080/00221325.2012.732125
Zhou, X., Sedikides, C., Wildschut, T., & Gao, D. G. (2008). Counteracting loneliness: on the
restorative function of nostalgia. Psychol Sci, 19(10), 1023-1029.
https://doi.org/10.1111/j.1467-9280.2008.02194.x
Zhou, X., Wildschut, T., Sedikides, C., Chen, X., & Vingerhoets, A. J. (2012). Heartwarming
memories: Nostalgia maintains physiological comfort. Emotion, 12(4), 678-684.
https://doi.org/10.1037/a0027236

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Spatial and temporal patterns of brain activity associated with emotions in music

PDF

Individual differences in heart rate response and expressive behavior during social emotions: effects of resting cardiac vagal tone and culture, and relation to the default mode network

PDF

The brain's virtuous cycle: an investigation of gratitude and good human conduct

PDF

Biological and behavioral correlates of emotional flexibility and associations with exposure to family aggression

PDF

Sociocognitive and neurophysiological contributors to effective secondary teaching

PDF

Behavioral and neural correlates of core-self and autobiographical-self processes

PDF

Decoding the neural basis of valence and arousal across affective states

PDF

Inhibitory spillover in adolescence: a fMRI study

PDF

Development of error monitoring in preschool to 12th-grade students and relations in late childhood and adolescence to social-affective processing and emotions about math

PDF

Self-reported and physiological responses to hate speech and criticisms of systemic social inequality: an investigation of response patterns and their mediation…

PDF

Mapping the neural architecture of concepts

PDF

Decoding information about human-agent negotiations from brain patterns

PDF

Decoding situational perspective: incorporating contextual influences into facial expression perception modeling

PDF

Neural and behavioral correlates of fear processing in first-time fathers

PDF

Behavioral and neural influences of interoception and alexithymia on emotional empathy in autism spectrum disorder

PDF

The interpersonal effect of emotion in decision-making and social dilemmas

PDF

Difficulty-as-importance and difficulty-as-impossibility: unpacking the context-sensitivity and consequences of identity-based inferences from difficulty

PDF

Associations between longitudinal loneliness, epigenetic age, and dementia risk

PDF

Core, social and moral disgust processing in youth with autism spectrum disorder (ASD)

PDF

The impact of emotion, valence, and arousal on differential memory processes in younger and older people

Asset Metadata

Creator Vaccaro, Anthony Gianni (author)

Core Title The neuroscience of ambivalent and ambiguous feelings

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Degree Conferral Date 2023-08

Publication Date 06/07/2023

Defense Date 05/25/2023

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

Tag affective neuroscience,bittersweet,fMRI,mixed emotions,OAI-PMH Harvest,uncertainty

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Damasio, Antonio (committee chair), Kaplan, Jonas (committee chair), Immordino-Yang, Mary Helen (committee member), Stange, Jonathan (committee member), Sweeney, Mary (committee member)

Creator Email agvaccar@usc.edu,tonygvaccaro@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113169246

Unique identifier UC113169246

Identifier etd-VaccaroAnt-11932.pdf (filename)

Legacy Identifier etd-VaccaroAnt-11932

Document Type Dissertation

Format theses (aat)

Rights Vaccaro, Anthony Gianni

Internet Media Type application/pdf

Type texts

Source 20230609-usctheses-batch-1053 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

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

Repository Email cisadmin@lib.usc.edu