Reward processing in sexual desire

Thesis

Reference

Reward processing in sexual desire

SENNWALD, Vanessa

Abstract

The main purpose of the present thesis was to explore reward processing in sexual desire.

More specifically, we aimed to apply the incentive salience hypothesis to sexual desire. We investigated the influence of sexual reward-related Pavlovian cues on instrumental actions to potentially elucidate variations of sexual desire. Accordingly, we showed evidence that inter-individual differences are key in determining effort mobilization for sexual rewards and importantly, that there is a relationship between intensity of the participants' perceived sexual desire and the intensity of the effort they are willing to mobilize, which was not the case for the reward's hedonic impact. Moreover, we also validated a paradigm that is suitable to measure both outcome specific and general Pavlovian-to-instrumental transfers involving sexual rewards within the same participants. Altogether, these empirical studies are an initial step in showing that the incentive salience hypothesis could be a promising theory to explain variations of sexual desire.

SENNWALD, Vanessa. Reward processing in sexual desire. Thèse de doctorat : Univ.

Genève, 2017, no. FPSE 693

DOI : 10.13097/archive-ouverte/unige:101069 URN : urn:nbn:ch:unige-1010696

Available at:

http://archive-ouverte.unige.ch/unige:101069

Disclaimer: layout of this document may differ from the published version.

1 / 1

Section de Psychologie

Sous la direction de David SANDER

REWARD PROCESSING IN SEXUAL DESIRE

THESE Présentée à la

Faculté de psychologie et des sciences de l’éducation de l’Université de Genève

pour obtenir le grade de Docteur en Psychologie

par

Vanessa SENNWALD De

La Chaux-du-Milieu (NE) Thèse No 693

GENEVE, novembre 2017

No étudiant: 06-217-814

Acknowledgements

I would like to express my deepest gratitude to the many amazing people that have helped me throughout my thesis. The stimulating and understanding environment they created made this intense period of my life infinitely more pleasant.

The first person I would like to thank is my thesis advisor, Professor David Sander. You are one of the kindest and most caring people I have ever encountered. I truly could not have hoped for a better supervisor. Aside from being an amazing person, I have learned a lot from you about being a rigorous researcher and owe you my gratitude for taking a chance on me and being so encouraging.

I would also like to express my deepest gratitude to Dr. Sylvain Delplanque and Professor Tobias Brosch. You have both helped me tremendously throughout this thesis. I am deeply thankful to Sylvain for helping make these last few years so enjoyable. Not only have you shared your scientific knowledge with me, you have also always had my back and I sincerely appreciate it. I also owe my gratitude to Tobias for our valuable discussions about this project. You have always been supportive of my work and have helped improve it tremendously with your valuable and as quick as lightning feedback.

Thank you to Dr. Francesco Bianchi-Demicheli, for always being very enthusiastic and positive about this research project as well as for sharing his insights based on his clinical experience.

Thank you to the Professor Brian Knutson for accepting to be a member of my jury.

Thank you to my colleagues and friends at the CISA for helping me as well as for being supportive and wonderful: Ryan Murray, Matthieu Ischer, Géraldine Coppin, Aline Pichon, Catherine Audrin, Patricia Cernadas Curotto, François Jaquet, Florian Cova, Céline Tarditi Joz-Roland, Aude Ferrero, Ulf Hahnel, Sylvain Tailamée, Gilles Chatelain, Daniela Sauge, and Marion Gumy.

A very special thank you goes to Eva Pool, my work wife and the other half of Evan. I definitely could not have done this without you. Over the last decade of our friendship, you have helped me in more ways than I think you are even aware of.

Aside from being a constant source of optimism, you have brought so much humor and happiness to this period of my life. I would also like to express my deepest gratitude to Yoann Stussi for his endless support. I truly value your opinion and I cherish our friendship. I would also like to say a big thank you to Ben Meuleman and Leo Ceravolo for their continued support. More specifically, thank you to Ben for our many hilarious discussions, no matter how absurd they were and to Leo for being so comforting and cheerful.

A big thanks goes to Charlotte Harding-Price, my non-researcher friend. You have been one of my biggest cheerleaders and have always been there for me, which is definitely not lost on me.

Finally, I would like to express my deepest gratitude to my family and husband. I am so lucky for being surrounded by such loving and supportive people. I appreciate all that you have done for me and this thesis would not have been possible if it was not for you.

Abstract

The main purpose of the present thesis was to explore reward processing in sexual desire. More specifically, we aimed to apply the incentive salience hypothesis to sexual desire. We investigated the influence of sexual reward-related Pavlovian cues on instrumental actions to potentially elucidate variations of sexual desire.

Accordingly, in the first two studies, we showed evidence that inter-individual differences are key in determining effort mobilization for sexual rewards and importantly, that there is a relationship between intensity of the participants’

perceived sexual desire and the intensity of the effort they are willing to mobilize, which is not the case for the hedonic impact of the reward. Moreover, in a third empirical study, we validated a paradigm that is suitable to measure both outcome- specific and general Pavlovian-to-instrumental transfers involving sexual rewards within the same participants. Altogether, these empirical studies are an initial step in showing that the incentive salience hypothesis could be a promising theory to explain variations of sexual desire. Moreover, they provide paradigms that may help further understand sexual desire and the motivational biases influencing its variations, which could potentially help better understand sexual desire disorders.

Table of Contents

Preliminary Note ... 1

1. Introduction and overview ... 1

1.1 Sexual desire ... 3

1.2 Neurophysiological mechanisms underlying rewards ... 5

1.3 Psychological mechanisms underlying rewards ... 6

1.3.1 Learning component ... 8

1.3.2 Motivational and affect components ... 14

1.4 The incentive salience hypothesis and sexual desire ... 16

2. Empirical part ... 23

2.1 Inter-individual differences underlie cue-triggered ‘wanting’ for sexual reward . 25 2.1.1 Introduction ... 27

2.1.2 Experiment 1 ... 30

2.1.2.1 Method ... 31

2.1.2.2 Results ... 36

2.1.2.3 Discussion ... 38

2.1.3 Experiment 2 ... 39

2.1.3.1 Method ... 40

2.1.3.2 Results ... 40

2.1.3.3 Discussion ... 49

2.1.4 General Discussion ... 50

2.2 Evidence for the existence of outcome-specific and general Pavlovian-to- Instrumental transfers involving sexual rewards ... 53

2.2.1 Introduction ... 55

2.2.2 Method ... 56

2.2.3 Results ... 61

2.2.4 Discussion ... 64

3. General discussion and conclusion ... 67

3.1 Implication for problematic sexual reward seeking behaviors ... 72

3.2 Limitations and future perspectives ... 74

3.2.1 Attention ... 74

3.2.2. Brain mechanisms ... 86

3.2.3 Reward-seeking learning styles ... 88

3.3 Conclusions ... 88

4. References ... 91

5. Résumé en français ... 111

i

Preliminary Note

A commentary and a review article are integrated in the introduction and the discussion, respectively. The paragraphs included from each of these works are indicated in the footnotes of the text.

i

1

1. Introduction and overview

2

Chapter 1 | introduction & Overview

3

Sexuality is a crucial aspect of human life. On a biological level, it is important for reproduction and the survival of the species (e.g., Panksepp, 1998), however, on a more individual level, it highly influences one’s perceived well-being (Flynn et al., 2016; Schmiedeberg, Huyer-May, Castiglioni, & Johnson, 2017).

Indeed, studies using self-reports have shown a positive link between perceived sexual satisfaction and quality of life (Flynn et al., 2016) as well as life satisfaction (Schmiedeberg et al., 2017). Overall, individuals that are satisfied with their sexual lives also tend to be more satisfied with their quality of life. Moreover, it has been suggested that the pleasure gained from sexuality can even potentially improve general health (Jannini, Fisher, Bitzer, & McMahon, 2009). Given its importance and benefits, it is key to better understand the psychological mechanisms underlying sexual behavior.

1.1 Sexual desire

Sexual and other appetitive behaviors, such as eating and drinking, have neurobiological and psychological mechanisms in common (Georgiadis &

Kringelbach, 2012; Toates, 2014). They are activities typically associated with pleasure, which encourages their repetition, particularly in a bodily state of deficiency (Toates, 2014). Appetitive behaviors occur in a cycle of expectation, consumption, and satiety (Figure 1.1; Kringelbach, Stein, & Van Hartevelt, 2012; Georgiadis &

Kringelbach, 2012). In the case of sexual behavior, sexual desire, which is the subjective experience of being attracted toward a person with potential rewarding effects (Both, Everaerd, & Laan, 2007), needs to be triggered to enter the cycle and generate sexual action (Georgiadis & Kringelback, 2012).

Interestingly, sexual desire was not considered a phase of the human sexual response cycle until the late 1970s when Helen Kaplan, a sexual therapist, observed that her patients presented a lack of interest in undertaking sexual activity (Kaplan, 1995). Based on her clinical experiences and the work of Kupferman (1991), she posited that sexual desire is a drive modulated by the brain, and more specifically regulated by the central nervous system’ control mechanisms (Kaplan, 1995). In her view, normal variations of sexual desire are controlled and depend on a good balance between excitatory and inhibitory systems. These systems involve physiological and psychological sexual inciters (e.g., testosterone, aphrodisiac or an attractive partner)

4

and suppressors (e.g., hormone disorders, depression, or unattractive partner) that influence hypothalamic and limbic regions. She proposes that there are sensors in the brain detecting and signaling whether one has had enough sexual activity to maintain homeostasis. The excitatory system activates sexual motivation, for example, if one’s sexual activity has been low for a while, one will likely feel intense sexual desire upon the perception of an attractive sexual opportunity. However, the inhibitory system will keep the individual from taking risks and override sexual motivation in case of emergency such hazards in the environment. A similar view of sexual desire is still being echoed by researchers such as Bancroft, Graham, Janssen and Sanders (2009), who have proposed that a dual control model, involving inhibition and excitation, could more broadly explain human sexual responses, including the sexual desire phase.

Figure 1.1. Illustration of the sexual pleasure cycle. It consists of three repeating phases: expectation, consummation, and satiety. First, a perceived stimulus has to elicit enough interest in an individual for them enter the pleasure cycle. Relevant sexual stimuli will usually elicit desire, leading to arousal. If the conditions are appropriate, consummation may take place, which could lead to an orgasm, followed by a state of satiety. Though learning can occur throughout the whole cycle, orgasms are a driving force in the learning process. Additionally, even in the case of incomplete sexual response cycles (e.g., an orgasm was not achieved), learning may occur. Reprinted from Georgiadis and Kringelbach (2012).

Chapter 1 | introduction & Overview

5

It is important to note that Kaplan’s conceptualization of sexual desire has not been based on scientific evidence but rather on her clinical observations (Levin, 2009). However, clinical observations have played part in the growing interest to study sexual desire and its variations using empirical data. Indeed, there have since been various approaches to elucidate how sexual desire works, such as functional magnetic resonance imaging (fMRI) studies investigating the structures involved (e.g., Stoléru & Mouras, 2007). Additionally, it has recently been postulated that incentive motivation models could potentially explain variations of sexual desire, which propose that stimuli (or incentives) in the environment are key in influencing sexual motivation (Both et al., 2007; Toates, 2009, 2014). More specifically, incentive motivation models suggest that reward-seeking behavior is due to an interaction between the needs of an organism and the incentive stimuli perceived in their environment (e.g., Berridge, 2004; Bindra, 1974; Bolles, 1972; Toates, 1997). Thus, motivational and affective processes involved in reward processing are fundamental in better understanding how sexual desire is triggered and what determines its intensity. Consequently, an understanding of the neurophysiological and psychological mechanisms underlying rewards more generally is necessary to explain variations of the intensity of sexual desire.

1.2 Neurophysiological mechanisms underlying rewards

Rewards are incentives eliciting motivation and goal pursuit (Kelley &

Berridge, 2002). The brain evolved to appropriately respond to natural rewards such as food and sex to ensure survival and reproduction. However, this neural network, which responds to incentives, is also stimulated by other rewards such as drugs (Kelley & Berridge, 2002). Moreover, researchers have found ways to use electrical stimulations in animals to determine which brain sites are involved in reward processing (e.g., Olds & Milner 1954). In turn and over time, self-stimulation, pharmacological, physiological, and behavioral studies in animals have helped elucidate which brain regions are responsible for reward-seeking behaviors (e.g., Haber & Knutson, 2010; Hikosaka, Bromberg-Marin, Hong, & Matsumoto, 2008;

Kelley & Berridge, 2002; Rolls, 2000; Schultz, 2000; Stefani & Moghaddam, 2006;

Wise, 2002) and as such, it has been demonstrated that the cortical-basal ganglia circuit is at the center of the reward system (Haber & Knutson, 2010). The key

6

structures in this circuit are the anterior cingulate cortex, the orbital prefrontal cortex, the ventral striatum, the ventral pallidum, and the midbrain dopamine neurons.

Though the nucleus accumbens and the ventral tegmental area dopamine neurons seem to play the largest part in this circuit, there are other brain areas which regulate the reward system such the amygdala and the hippocampus (Haber & Knutson, 2010).

Importantly, dopamine and the ventral striatum, and more specifically, the nucleus accumbens, have been demonstrated to be fundamental in reward-seeking behaviors (Berridge & Robinson, 2003). Neurobiological manipulations of these brain areas have been crucial in determining the psychological components involved in these types of behaviors and in potentially explaining maladaptive behaviors such as addictions (Berridge & Robinson, 2003). Based on these findings, a key model was developed suggesting that reward processing is not made up of a single unit but actually three distinct components.

1.3 Psychological mechanisms underlying rewards

Over the last few decades, researchers in affective neuroscience have developed a model of reward processing, the incentive salience hypothesis, which proposed that reward can be distinguished into three distinct components: learning, motivation, and affect (Figure 1.2; Berridge & Robinson, 2003). Each category is represented by two psychological components, which can be processed at either an explicit or implicit level. These components can operate under different psychological and neural mechanisms. Cognitive incentive or wanting, reward expectation, and the conscious experience of pleasure are conscious experiences, which individuals have access to and are aware of. On the other hand, incentive salience or ‘wanting’, associative learning, and objective affective reactions are processed at an unconscious level, which means individuals are not always aware these are taking place. Though sometimes implicit processing’s outcome can be transformed into conscious representations, explicit awareness is not necessary for these implicit psychological components to have great influence on behavior. Furthermore, although the implicit and explicit levels of motivation serve different purposes, they usually function concurrently to motivate behavior. Each variation of the components can be behaviorally measured in various ways, explicit processes such as cognitive incentives are typically measured using self-reports, whereas implicit processes such

Chapter 1 | introduction & Overview

7

as incentive salience can be measured through Pavlovian-to-instrumental transfers. It should be noted that these reward components interact with one another.

Importantly, the present thesis focuses on the more implicit aspects of this model, specifically the motivational component using paradigms developed in the animal literature. More precisely, the incentive salience hypothesis, as formulated by Berridge and Robinson (2003), is based on the relationships organisms build between previously neutral stimuli that have acquired a positive value through learning and the way they behave upon their perception, depending on their physiological state. This model aims to explain how rewards or cue-paired rewards influence our actions when they are relevant to us depending on our needs and how they can explain maladaptive reward-seeking behaviors such as addictions.

Figure 1.2. Reward components and their descriptions. Reward processing consists of three major components: motivation, learning, and affect. The components can be broken down into two levels of psychological components, a more explicit level and a more implicit level. There are various ways of measuring each psychological mechanism. It is important to note that the focus of the present thesis is on the more

8

implicit level of the motivational component, incentive salience (highlighted in a thicker black border). Reprinted and modified from Berridge and Robinson (2003).

1.3.1 Learning component

The learning component involves any knowledge acquired about the associations between stimuli and actions related to the reward. Learning is essential to make predictions about the reward, prepare any anticipatory response, for cues in the environment to be influential and for goal-directed actions (Berridge & Robinson, 2003).

Learning and behavior: Goal-directed, habitual, and Pavlovian systems

An important research tradition has led to the proposal that there exists at least three parallel and competitive learning systems that influence behavior: the goal- directed, the habitual, and the Pavlovian systems (e.g., Daw & O’Doherty, 2013;

Rangel, Carmer, & Montague, 2008). First, the goal-directed system is in charge of the learning of an action in relation to its valued outcome, for example one learns that if a coin is inserted in the vending machine, then one will receive a treat. This system involves learning an association between a specific action (e.g., inserting a coin) and valuable outcome (e.g., the treat; action-outcome learning). The ability to select goal- directed actions requires a representation of the states, actions, and goals available, and involves a flexible computation of action-plans. Second, the habitual system is in charge of learned associations between stimuli and actions (e.g., stimulus-action learning), which will result in a certain outcome; this system is related to the stimulus- driven mechanism. Indeed, unlike goal-directed behaviors, habitual behaviors are reflexively elicited by the stimulus without considering the current goals of the individual. This implies that actions can be selected even if the outcome of the action is no longer relevant to or valued by the individual. Thus, under the control of the habitual system, the action is not performed with the intention of obtaining or

1Adapted from Sennwald,V., Pool, E., & Sander, D. (2017). Considering the influence of the Pavlovian system on behavior: Appraisal and value representation.

Psychological Inquiry, 28(1), 52-55.

Chapter 1 | introduction & Overview

9

avoiding the outcome, whereas in the goal-directed system, the representation of the outcome is critical. Finally, a third distinct system has been proposed: the Pavlovian system. The Pavlovian system has an associative nature consisting in the learning between a stimulus and its associated outcome (e.g., stimulus-outcome learning); for example, learning the association between a vending machine sign and the reception of a treat. For a long time, the Pavlovian system was presumed to be stimulus driven, by learning simple stimulus-value associations (Doll, Simon, & Daw, 2012), however, recent evidence suggests that the Pavlovian system is capable of building a more complex representation of the stimulus and its associated outcome (Dayan &

Berridge, 2014; Prévost, McNamee, Jessup, Bossaerts, & O’Doherty, 2013; Robinson

& Berridge, 2013;). Furthermore, while the neural basis of these systems is not fully established (Rangel et al., 2008), it has been proposed that these systems potentially rely on separate neural networks (Balleine & O’Doherty, 2010). More precisely, the goal-directed system has been suggested to rely on a neural circuit involving brain areas such as the medial prefrontal cortex, the dorsomedial striatum, and the mediodorsal thalamus, whereas the habitual system has been suggested to rely on a network involving the sensory-motor cortex, the dorsolateral striatum, and the posterior thalamus. Finally, the Pavlovian system related to rewards has been suggested to rely on brain areas such as the orbitofrontal cortex and the ventral striatum.

Interestingly, these systems can interact but they also are in competition with each other to determine behavior. The pull for power between these systems can sometimes lead to maladaptive behavior, in which an inappropriate action such as a habit is selected instead of a goal-directed behavior; for example, a Belgian driver driving on the wrong side of the road in the United Kingdom. Importantly, there can be an interaction between the Pavlovian and the other instrumental systems that results in the learning of an action that is invigorated by the perception of a Pavlovian cue (Campese et al., 2014; Daw & O’Doherty, 2013), this phenomenon is known as Pavlovian-to-instrumental transfer (PIT). Over the last couple of decades, it has been brought to the forefront that this interaction between the Pavlovian system and the other instrumental systems might provide useful insights in the understanding of maladaptive behaviors (e.g., Corbit & Balleine, 2015).

While the Pavlovian system shares similarities with both the habitual and the goal-directed systems, it is in fact different from them. We will develop these points

10

further hereafter by using appetitive learning literature to compare the different systems. It is important to note that these learning systems are conceived to be similarly involved in both appetitive and aversive learning (Rangel, Camerer, &

Montague, 2008). However, in the present thesis, we will focus on appetitive learning, given that the focus is on sexual desire.

The Pavlovian and the habitual systems: the role of the outcome value representation

As previously mentioned, the Pavlovian system has an associative nature and is driven by the perception of stimuli, just like the habitual system. However, the association learned is between a stimulus and its associated outcome (e.g., stimulus- outcome learning) rather than a stimulus and an action (e.g., stimulus-action learning).

Even though the Pavlovian system is in charge of learning associations between stimuli and outcomes, Pavlovian learning still produces behaviors (e.g., Dayan, Niv, Seymour, & Daw, 2006). Once the associative learning is completed, the Pavlovian stimulus acquires the ability to evoke behaviors originally triggered by the outcome.

Therefore, both habitual and Pavlovian behaviors are triggered by the stimulus perception, and for a long time these two systems were assumed to make similar computations by simply storing caches of stimulus-value or action-value associations (Doll, Simon, & Daw, 2012). However, there is a fundamental difference between these two kinds of behaviors: whereas habitual behaviors are suggested to be independent from representation of the original outcome associated with actions, Pavlovian behavior has been proposed to be highly sensitive to the value of the outcome. A clear illustration of this phenomenon can be found in the outcome devaluation literature. Devaluation procedures consist in decreasing the value of an outcome or a stimulus-associated outcome through satiation (procedure used on humans and animals) or by associating the outcome with an illness (procedure used on animals), resulting in the modification of the behavior related to the outcome (Balleine & O’Doherty, 2010). For instance, in a study by Hogarth and Chase (2011), young adult smokers learned to press keys that would either allot them tobacco points or chocolate points. The participants subsequently underwent a devaluation procedure either for the tobacco or for the chocolate, where they had to smoke a cigarette or eat

Chapter 1 | introduction & Overview

11

chocolate to satiation. Finally, in a choice task, the young adult smokers’ were less likely to press on the keys associated with the outcome that was specifically devalued.

A large corpus of literature has consistently shown that the influence of Pavlovian stimuli on behavior is highly sensitive to outcome devaluation procedures (e.g., Holland & Staub, 1979; Lex & Huber, 2010). For example, Pool, Brosch, Delplanque and Sander (2014) showed that chocolate lovers’ attention was involuntarily oriented towards a cue-associated to a chocolate odor, resulting in the participants being quicker to discriminate the position of a line when it was preceded by the cue compared to a neutral stimulus. However, after a devaluation procedure, in which the participants were satiated with chocolate, the cue associated to the chocolate odor no longer involuntarily oriented participants attention and in turn no longer affected the participants’ behavior. This study shows that the cue no longer being relevant to the chocolate lovers was crucial in determining their behavior.

In contrast, the influence of the habitual system on behavior is insensitive to changes in the outcome value (e.g., Adams, 1982; de Wit et al., 2012; Tricomi, Balleine, & O’Doherty, 2010). In a famous experiment, Adams showed that depending on the amount of training a rat receives to press on a lever to receive a reward, it will either stop reproducing the action to receive the outcome or it will continue the action after an outcome devaluation procedure. Indeed, two groups of rats were trained to press on a lever to receive sucrose. The first group received a small amount of training, inducing goal-directed behavior, whereas the second group was over-trained, inducing habits. After the sucrose was associated with an illness in both groups, the first group of rats stopped pressing on the lever to receive the now devalued sucrose, however, the second group of over-trained rats continued to press on the lever to obtain the devalued sucrose. These findings demonstrate that, unlike the goal-directed mechanism and the Pavlovian system, habits do not seem to be dependent on the representation of the outcome value.

As shown above, the Pavlovian system and the habitual system are quite different even if they share a similarity in that they both involve the learning of an association of a stimulus and an outcome or action respectively. Already, the type of association is different, the former is in charge of associations between a stimulus and an outcome, and the latter is in charge of associations between a stimulus and an action. Additionally, the behavior produced by the Pavlovian system relies on the outcome value, whereas the habitual system is insensitive to outcome value.

12

The Pavlovian and goal-directed systems: the roles of incentive salience and expected pleasantness

A major aspect in which the Pavlovian and the goal-directed systems are similar is that the representation of the outcome value, which can be subject to devaluation in both cases, plays a key role in the control of behavior. However, it has been suggested that changes in the representation of the outcome value do not affect the behavior produced by these systems in the same way. The different behavior productions of these systems are due to the fact that they rely on distinct incentive processes (e.g., Ostlund & Balleine, 2008).

The goal-directed mechanism is dependent on the hedonic experience related to the outcome to determine the performance of the goal-directed actions (Balleine &

O’Doherty, 2010). In contrast, according to the incentive salience hypothesis, the incentive processes underlying the Pavlovian system in appetitive behaviors do not depend on the pleasure experienced due to the outcome. The incentive salience hypothesis proposes that reward processing is composed of multiple components including ‘wanting’ and liking (Berridge & Robinson, 2003). Reward pursuit behaviors are said to be guided by the ‘wanting’ component, which consists of the effort one is willing to mobilize in order to obtain a reward, rather than the liking component, the pleasure experienced during the consumption of the reward (Berridge

& Robinson, 2003). In general, ‘wanting’ and liking are positively correlated, that is to say that if one likes a particular reward then one will put in a proportional amount of effort to obtain said reward. However, under certain circumstances such as addictions (Berridge, Robinson, & Aldridge, 2009) or stress (Pecina, Schulkin, &

Berridge 2006; Pool, Brosch, Delplanque, & Sander, 2015), ‘wanting’ and liking can be dissociated, meaning that an individual will work very hard to obtain a reward that is not necessarily liked any more. The incentive salience hypothesis proposes that the influence of Pavlovian stimuli does not depend on the hedonic properties of the associated outcome, but rather relies on a different mechanism consisting of the synergetic interaction between an individual’s brain (e.g., level of mesolimbic dopamine) or physiological state (e.g., hunger or satiety) and the Pavlovian stimulus (Zhang, Berridge, Tindell, Smith, & Aldridge, 2009). After the perception of the Pavlovian cue, behavior is not determined by the pleasantness expectations about the

Chapter 1 | introduction & Overview

13

associated outcome, but is directly modulated by the relevant physiological state.

While the incentive salience hypothesis only makes predictions about the physiological relevance of a stimulus to the organism, we proposed that this concept could be expanded, using appraisal theory, to the psychological construct of affective relevance (Pool, Sennwald, Delplanque, Brosch, & Sander, 2016). Affective relevance consists in the interaction between the perception of a stimulus and an individual’s current concerns (Sander, Grandjean, & Scherer, 2005). These concerns, which are affective representations, encompass both physiological and psychological concerns (e.g., needs, values). For a reward-associated stimulus to trigger a motivational state (i.e., ‘wanting’), it has to be relevant for the current concerns of the individual (Pool et al., 2016; Robinson & Berridge, 2013; Tindell, Smith, Pecina, Berridge, & Aldridge, 2006; Zhang et al., 2009). Essentially evidence suggests that the goal-directed system relies on the experienced pleasantness of the outcome (Balleine & O’Doherty, 2010) and the Pavlovian system depends on the affective relevance the outcome has for the individual’s current concerns (Pool et al., 2016;

Robinson & Berridge, 2013; Tindell, et al., 2006; Zhang et al., 2009).

Additionally, while the Pavlovian system’s conditioned behavior ceases once an outcome has been devalued (Robinson & Berridge, 2013), the goal-directed system requires explicit feedback that the outcome has been devalued through re-experience to reduce behavior, therefore updating the expectations of the outcome pleasantness (Balleine & Dickinson, 1991; Ostlund & Balleine, 2008; Pool et al., 2016). These differences between the systems can be illustrated with Balleine and Dickinson’s (1991) study. In this study, rats were taught to press on a lever in order to receive sucrose. Once the sucrose was associated with illness in a devaluation procedure, their behavior did not stop under extinction. The rats had to consume the devalued sucrose to stop pressing on the lever, revealing that to adapt their goal-directed actions, rats needed to update the value representation of the outcome through the direct consumption experience. Therefore, whereas the goal-directed mechanism relies on memory updates of the outcome pleasantness, the Pavlovian system relies on an online valuation of the outcome relevance for the current state of the individual, meaning as soon as the outcome has been devalued, the behavior produced by the Pavlovian stimuli will cease. Evidence of this phenomenon can be seen in Robinson and Berridge’s (2013) study, in which they taught rats to turn their repulsion towards a cue-associated to an unpleasant salty sensation into an attraction towards it.

14

Robinson and Berridge demonstrated that when the rats were put in a sodium depletion state, the cue associated to the unpleasant salty sensation became relevant, in turn making the rats want the cue-associated to the previously unpleasant salty sensation. This study shows that the relevance of the outcome associated with the Pavlovian stimulus is key in influencing behavior.

In brief, though outcome value is important for both the Pavlovian and the goal-direct system, they rely on different incentive processes. Unlike the latter, the former does not seem to rely on the hedonic properties of the outcome but rather on its relevance for the current concerns of the individual. Moreover, the behavior the Pavlovian system produces seems to be dependent on an online valuation of the outcome relevance, whereas the goal-directed system requires a memory update of the value of the outcome to modify behavior.

1.3.2 Motivational and affect components

As previously mentioned, the motivational component, which is also known as incentive salience or ‘wanting’, is an implicit process in which organisms mobilize effort in an action to obtain a reward once they have perceived an associated cue in their environment, it promotes approach toward rewards as well their consumption (Berridge, Robinson, & Aldridge, 2009; Berridge & Robinson, 2003). ‘Wanting’ is typically measured using Pavlovian-to-instrumental transfers tasks (e.g., Balleine, 1994; Corbit & Balleine, 2005, 2011; Wyvell & Berridge, 2000), during which organisms learn to associate a reward with a Pavlovian stimulus and an instrumental action. ‘Wanting’ is measured through the influence of Pavlovian stimuli on instrumental actions, which translates to an increase of effort that is mobilized to perform the instrumental action in the presence of the Pavlovian stimulus.

Furthermore, it has been hypothesized that there can be two types of Pavlovian influence on instrumental actions: outcome-specific and general (Dickinson &

Balleine, 2002; Holland, 2004). It has been posited that the outcome-specific PIT selects the instrumental action during the perception of the reward-associated cue, whereas the general PIT has been suggested to determine the vigor of the actions performed, independently of the outcome value (Corbit & Balleine, 2011; Prévost, Liljeholm, Tyszka, & O’Doherty, 2012). Importantly, the incentive salience hypothesis relies on the fact that the cue perceived which will trigger an organism’s

Chapter 1 | introduction & Overview

15

behavior has to be relevant for its physiological needs (Robinson & Berridge, 2013;

Zhang et al., 2009). Additionally, as previously mentioned, it has been proposed that incentive salience could also potentially rely on affective relevance (Pool et al., 2016).

On the other hand, the affect component, or liking, consists of the hedonic impact the reward has on the organism during its consumption (Berridge & Robinson, 2003). In animals and human infants it has been measured through facial expressions, however, for adult humans it is usually evaluated through ratings during the consumption of the reward.

Studies in affective neuroscience have suggested that reward-seeking behaviors depend on the motivational component rather than the hedonic impact of the reward affecting maladaptive behaviors such as addictions, pathological gambling, overeating (Berridge & Robinson, 2003; Finlayson, King, & Blundell, 2007;

Goldstein et al., 2010; Wölfling et al., 2011). Berridge and Robinson (1998) first proposed the existence of distinct components and the concept of incentive salience, when they depleted dopamine in the nucleus accumbens of rats and showed that only their motivated behavior was affected but not their liking of the reward or the learning of new incentives. They suggested that dopamine in the nucleus accumbens is necessary for ‘wanting’ incentives, unlike liking or the learning of new likes and dislikes. Though the dissociation between ‘wanting’ and liking has been established in rats using neurobiological manipulations, in humans it has been shown through physiological manipulations such as stress (Pool et al., 2015). Using a combination of a PIT task and a socially evaluated cold pressor task, Pool et al. (2015) showed that in comparison to a stress-free group, participants who had undergone the stress manipulation mobilized more effort in an instrumental action in the presence of a reward-related cue, without a parallel increase in liking in comparison. Moreover, studies have suggested that their neurochemical components differ and that they map onto distinct brain areas of the reward system (Berridge, Robinson, & Aldridge, 2009). Indeed, it has been shown that the dopamine neurotransmitter is crucial in modulating ‘wanting’ (Berridge & Robinson, 1998, 2003; Berridge, Robinson, &

Aldridge, 2009), whereas the opioid, endocannabinoid, and GABA-benzodaizepine neurotransmitter systems in particular hotspots of the limbic structures such as the nucleus accumbens shell and the ventral pallidum are important to generate liking (Berridge & Robinson, 2003; Berridge, Robinson, & Aldridge, 2009).

16

The dissociation between ‘wanting’ and liking has largely been suggested due to neurobiological manipulations in animals, however, there has been an interest in investigating these concepts on humans to further understand human motivational processes (Pool et al., 2016). Yet, results across studies have been inconsistent and consequently, it has been doubted that these components can be disentangled within human reward processing (Havermans, 2011, 2012). Notably, the operationalization of these components has not been coherent with the definitions established in the animal literature (Pool et al., 2016). Indeed, ‘wanting’ and liking have not always been measured at the right time in the experiments, that is to say that animal

‘wanting’ is typically measured during the presence of the cue, without the presence of the reward, and that animal liking is measured during the consumption of the reward. However, some studies have operationalized these components in the same way as each other. Pool et al. (2016) hypothesize that these contradictions are due to a major confound, in which the components are confused with expected pleasantness, a cognitive process that does not correspond to the hedonic experience of animal liking or ‘wanting’.

1.4 The incentive salience hypothesis and sexual desire

Research testing the incentive motivation models (e.g., Corbit & Balleine, 2005; Dickinson & Balleine, 1990; Robinson & Berridge, 2013) has mainly explored the psychological mechanisms underlying the relationship between reward-paired stimuli and the reward’s relevance for the physiological needs of the organism.

Though there has been a consensus among some scholars that the incentive salience hypothesis can be applied to multiple types of rewards, reward-seeking behaviors such as overeating (e.g., Finlayson et al., 2007), homeostatic needs (e.g., hunger, thirst, salt appetite) have been at the center of these investigations (e.g., Corbit &

Balleine, 2005; Robinson & Berridge, 2013). Organisms are in the pursuit of such rewards to will try to seek to correct homeostatic imbalances that could lead to death if not taken care of (Schultz, 2015; Toates, 2014). For instance, if an organism is parched, it will ingest any drink, even if it is distasteful, to try and maintain the energy levels necessary for survival (Toates, 2014). By contrast, though sex is necessary for reproduction and the survival of the species (Olsen, 2011), deprivation of sex will not lead an individual to their death and moreover, for sexual action to be generated,

Chapter 1 | introduction & Overview

17

certain conditions will still need to be met (Toates, 2014). Research on sexual preferences has shown that inter-individual differences are crucial in sexuality (Chivers, Seto, & Blanchard, 2007; Toates, 2009). The same stimulus will not elicit sexual desire and arousal in every individual. Importantly, how one identifies in regard to their sexual identity, including their sexual orientation, will severely influence their sexual behavior and sexuality in general (Chivers et al., 2007; Tally &

Stevens, 2017).

While sexual rewards differ in some ways to food and drink, they rely on comparable neurobiological and psychological processes (Georgiadis & Kringelbach, 2012; Toates, 2014). Crucially, for the incentive salience hypothesis to be applied to variations of the intensity of sexual desire, two fundamental aspects need to be met.

The incentive salience hypothesis relies on two tenets (e.g., Berridge & Kringelbach, 2015; Berridge & Robinson, 2003; Zhang et al., 2009): (1) motivational processes are triggered if the reward-paired cue perceived is relevant for the needs of the individual;

(2) ‘wanting’ and liking represent two distinct components of reward processing.

Therefore, sexual motivation should be triggered by appraisal mechanisms, which evaluate whether each stimulus perceived in the environment is relevant for an individual’s current concerns such as their needs, goals, and values (Frijda, 1986;

Sander, Grandjean, & Scherer, 2005; Scherer, 2013). The current concerns are not only limited to homeostatic needs, they also encompass more complex aspects related to self-concepts such as sexual orientation (Talley & Stevens, 2017), however, the essential notion is that these needs, goals, and values are pertinent for an individual at the moment the stimulus is perceived (Scherer, 2013). Intra- and inter-individual differences, therefore, highly modulate appraisal mechanisms (Sander, Grandjean, &

Scherer, 2005). Moreover, there has been evidence that reward-paired cues relevant for the individual’s current concerns trigger ‘wanting’ (e.g., Dayan & Berridge, 2014;

Pool et al., 2016; Robinson & Berridge, 2013; Tindell et al., 2006; Zhang et al., 2009). Thereby, sexual reward processing should also rely on inter-individual differences, where not all sexual stimuli will trigger ‘wanting’ in all individual. It will depend on their specific concerns, experiences, cognitive resources, and other psychological factors, which will modulate how they evaluate the stimuli in their environment.

Additionally, for the incentive salience hypothesis to be applied to sexual desire two distinct components, ‘wanting’ and liking, should playing different roles in

18

its variation. Though their distinctiveness has typically been demonstrated through dopamine manipulations in rats (e.g., Berridge & Robinson, 1998), subjective craving in drug addiction has also been linked to ‘wanting’ and not liking (Robinson &

Berridge 1993). Given that craving is considered to be a conscious subjective experience (Anselme & Robinson, 2016), ‘wanting’ is possibly better predicted by the level of intensity the object is desired than by how much the individual likes it. For sexual desire, this translates to a link between incentive salience and the subjective perception of the level of intensity an object or experience sexually arouses rather than with its hedonic impact. Recent observations suggest that this conceptual distinction can indeed be very important to understand typical and problematic variations of the intensity of sexual desire. Indeed, Krishnamurti and Loewenstein (2012) aimed to create and validate a self-reported measure to distinguish sexual

‘wanting’ from liking through a series of questions on individuals’ sexual behavior with their partners. They showed through three studies that individuals with typical variations of sexual desire intensity can report different levels of ‘wanting’ and liking toward their habitual sexual partner, suggesting these components are distinct constructs. Moreover, a recent study (Ferdenzi et al., 2015) revealed that women with a clinical form of sexual hypo-desire liked faces and men’s voices as much as a control group, but reported a significantly lower level of attraction toward the same men’s faces and voices than the control group. These findings suggest that, in a clinical form of sexual hypo-desire, the liking component could be unaltered, and that there could be a selective decrease of the ‘wanting’ component of sexual reward processing. They also highlight the relevance of having a valid hypothesis that differentiates the various components of sexual reward processing. In fact, if problematic variations of the intensity of sexual desire, such as hypoactive sexual desire disorder, are underlain by a selective variation of the ‘wanting’ component but not of the liking one, then the therapeutic interventions should target the psychological and neural processes involved in this specific component of reward processing.

In summary, it has been suggested that the conceptualization of sexual desire through an incentive salience hypothesis lens could be potentially powerful to predict and understand typical and problematic sexual desire (Toates, 2009, 2014). Moreover, studies using auto-reported measures (Krishnamurti & Loewenstein, 2012; Ferdenzi et al., 2015) point out the high relevance of distinguishing different components of

Chapter 1 | introduction & Overview

19

sexual reward processing, such as ‘wanting’ and liking. However, there has been a lack of studies conducted in an experimental setting that use operationalizations of these concepts that are close to the ones in animal studies, to test whether human sexual reward processing involves distinct components, and whether the psychological mechanisms underlying sexual reward components are the same as the ones postulated by the incentive salience hypothesis.

The main purpose of the present thesis was to apply the incentive salience hypothesis to sexual desire. More specifically, we aimed to investigate the influence of sexual reward-related Pavlovian cues on instrumental actions to elucidate the underlying mechanisms of the variations of sexual desire. We postulated that sexual reward processing is composed of two different components, one of which can be separated into two forms: (i) ‘wanting’, measured by general PIT, is triggered by the perception of a stimulus associated with a potential sexual partner appraised as relevant for the current sexual concerns (e.g., sexual orientation) of the individual; (ii)

‘wanting’, measured by specific PIT, is triggered by the perception of a stimulus associated with a potential sexual partner related to previous pleasurable interactions;

and (iii) liking is the experienced pleasure during the interactions with a potential sexual partner (Figure 1.3).

20

Figure 1.3. Illustration of the incentive salience hypothesis applied to sexual reward processing. ‘Wanting’ or incentive salience and liking are involved. First, ‘wanting’ is triggered by the perception of a stimulus associated with a potential sexual partner appraised as relevant for the current sexual concerns of the individual. The perception of the stimulus either triggers general motivation, in which any approach action is energized, or outcome-specific motivation, in which a specific action is selected and then energized, leading to the interaction with the potential sexual partner. Liking will then be experienced during the actual interaction with the potential sexual partner.

Past experiences of liking will play a role in outcome-specific motivation.

Additionally, through this research, we aimed to replicate the findings from human studies using protocols that measure the same concepts as animal ‘wanting’

and liking (e.g. Prévost, Liljeholm, Tyszka, & O’Doherty, 2012; Talmi, Seymour, Dayan, & Dolan, 2008) considering their potential to improve our understanding of motivational biases in typical and maladaptive reward-seeking behaviors. Recently, there has been an increase of studies adapting the Pavlovian-to-instrumental transfer paradigm to a human population (Cartoni, Balleine, & Baldassare, 2016), however, few have been successful in measuring both outcome-specific and general motivational biases within the same paradigm and participants (e.g., Morris, Quail, Griffiths, Green, & Balleine, 2015; Nadler, Delgado, & Delamater, 2011; Prévost et al., 2012). More specifically, the general motivational bias has been the more elusive of the two biases as findings have inconsistently demonstrated its existence (Jeffs &

Duka, 2017). Additionally, there has been replication crisis in psychology as a whole, in which replications of research findings have not been getting the same results as the original studies (e.g., Maxwell, Lau, & Howard, 2015) Consequently, replications of the existing findings in reward processing are crucial and even necessary in order to validate a paradigm appropriate to measure both outcome-specific and general motivation.

Additionally, we also aimed to build on the existing literature by extending our findings to motivational states dependent on inter-individual differences such as sexual orientation. To this end, we empirically tested whether the incentive salience hypothesis can be applied to sexual desire, using a Pavlovian-to-instrumental task and sexual rewards on a healthy human population (Chapter 2). Accordingly, in the first two studies, we tested whether individual-differences are a determining factor in

Chapter 1 | introduction & Overview

21

triggering sexual ‘wanting’. Participants of different sexual orientations (heterosexual and homosexual) were recruited and sexual rewards were selected correspondingly.

We predicted that the effort mobilized during the perception of a sexual stimulus- paired Pavlovian cue is dependent on the relevance the sexual stimulus has for the individual. That is to say that the cue associated with an erotic image of a woman would trigger greater effort mobilization in heterosexual men than in homosexual men and that a cue associated with an erotic image of a man would trigger greater effort mobilization in homosexual men than in heterosexual men. Furthermore, we wanted to investigate whether a link could be made between the subjective perception of an individual’s sexuality and their ‘wanting’ and liking measured using an experimental paradigm. In the third study, we aimed to disentangle outcome-specific and general Pavlovian-to-instrumental transfers in sexual motivation as well as validate a paradigm that is suitable to measure both biases.

In the final chapter of this thesis (Chapter 3), we discussed the contribution of our three empirical studies’ findings to the study of sexual desire and also discussed the limitations of our research as well the potential future perspective.

22

23

2. Empirical part

24

Chapter 2 | Empirical part

25

2.1 Inter-individual differences underlie cue-triggered ‘wanting’

for sexual reward ²

Abstract

Reward-seeking behaviors are necessary for our survival, they help us nourish ourselves and find mates. Through two experimental studies, we build on incentive motivation theories by investigating inter-individual differences in sexual reward- seeking behaviors. We demonstrated that Pavlovian-instrumental transfer effects for sexual rewards are strongly modulated by inter-individual differences. More specifically, this effect critically relied on the sexual orientation of the individual.

Furthermore, we also showed that, unlike the hedonic impact of the reward, the intensity of the Pavlovian-instrumental transfer effect predicts variations of sexual desire in everyday life. These findings suggest that incentive motivation models are well suited to explain variations of sexual desire, which could be key in future therapeutic interventions for sexual desire disorders.

2Reprint of: Sennwald, V., Pool, E., Delplanque, S., Brosch, T., Bianchi-Demicheli, F., & Sander, D. (2017). Inter-individual differences underlie cue-triggered ‘wanting’

for sexual reward. Manuscript submitted for publication.

26

Chapter 2 | Empirical part

27

2.1.1 Introduction

Rewards are an essential part of life; they enhance species’ chances at surviving and reproducing by influencing their eating, drinking, and mating behaviors (Schultz, 2015). Sexual stimuli exert strong effects on human behavior (e.g., Gola, Wordecha, Marchewka, & Sescousse, 2016; Pool, Brosch, Delplanque, & Sander, 2016; Sennwald et al., 2016), rendering them a powerful tool to investigate human reward processing. Additionally, a better understanding of reward mechanisms elicited by these kinds of stimuli may help shed light on inter-individual variations of sexual desire, which is a very important but under-investigated issue with implications for individual well-being and for society as a whole (Schmiedeberg, Huyer-May, Castiglioni, & Johnson, 2017).

In the last few decades, incentive motivation models (e.g., Berridge, 2004;

Bindra, 1974; Bolles, 1972; Toates, 1997) have brought forth major advancements in the understanding of the neurobiological and psychological mechanisms underlying reward-seeking behaviors. It should be noted that reward processing is not unitary but involves multiple components including ‘wanting’ and liking (Berridge & Robinson, 2003). The ‘wanting’ component consists in the effort an organism mobilizes to obtain a reward and is triggered by the perception of the reward and its associated cues. It is classically measured using the Pavlovian-instrumental transfer test (e.g., Balleine, 1994; Corbit & Balleine, 2005, 2011; Wyvell & Berridge, 2000), in which a reward is associated with a Pavlovian stimulus and an instrumental action. ‘Wanting’

is typically measured by the increase of the instrumental action in the presence of the Pavlovian stimulus. In contrast, the liking component consists of the pleasure experienced upon consumption of a reward (Berridge & Robinson, 2003); it is therefore measured during the consumption experience. Though they are usually positively correlated (i.e., the more an organism likes something the more they will want it), they can also be dissociated under certain circumstances such as addictions (Berridge & Robinson, 2003) or stress (Pool, Brosch, Delplanque, & Sander, 2015;

Pool, Delplanque, Coppin, & Sander, 2015), where an individual can work very hard to obtain a reward that they will not like any more once they have obtained it. These studies illustrate that ‘wanting’ and liking are distinct components and that they rely on different mechanisms. Indeed, it has been suggested that cue-triggered ‘wanting’

28

does not rely on the hedonic properties of the associated reward, but rather on the reward’s relevance for the organisms’ motivational states (e.g., Dickinson & Balleine, 1990; Pool, Sennwald, Brosch, Delplanque & Sander, 2016; Zhang, Berridge, Tindell, et al., 2009). For instance, Dickinson and Balleine (1990) showed that rats which were trained to press a lever to receive a sucrose solution or saline upon the perception of a cue, increased their pressing performance to obtain the sucrose solution under extinction when it was relevant to their motivational state (i.e., when they were hungry).

A strong line of research (e.g., Corbit & Balleine, 2005; Dickinson & Balleine, 1990; Robinson & Berridge, 2013) testing incentive motivation models has explored the mechanisms underlying the link between the perception of reward-associated stimuli and reward relevance for the current organisms’ motivational states. These studies have mainly manipulated homeostatic needs (e.g., hunger, thirst, salt appetite;

Balleine, 1994; Corbit & Balleine, 2005; Robinson & Berridge, 2013) in order to test how they interact with food reward seeking behaviors. Interactions between motivational states and other kinds of relevant rewards such as sexual rewards have been less investigated. Food and drink are types of rewards organisms will seek to try to correct homeostatic imbalances (Schultz, 2015) and are biologically imperative behaviors (Toates, 2014). If an organism is extremely hungry or thirsty, any food or drink ingested will suffice no matter how abhorrent it is to the organism in order to maintain chemical and energy levels within their body (Toates, 2014). However, while sex is a primary reward in the sense that it is important for the reproduction and survival of a species (Olsen, 2011), when organisms are deprived of sex certain conditions will still need to be met for them to engage in sexual behaviors (Toates, 2014). Indeed, sexuality is critically dependent on inter-individual differences (Chivers, Seto, & Blanchard, 2007; Toates, 2009), which means that the same stimulus will not elicit sexual desire and sexual arousal in everyone. More specifically, an individual’s sexual identity, which includes their sexual orientation, will influence their sexual behavior and sexuality in general (Chivers, Seto, &

Blanchard, 2007; Talley & Stevens, 2017).

Though sexual reward processing and food reward processing differ in some respects, they both rely on similar neurobiological and psychological mechanisms (Toates, 2014). Recently, it has been suggested that the incentive motivation models

Chapter 2 | Empirical part

29

can be used to predict and understand variations in the intensity of sexual desire (Toates, 2009, 2014). This suggestion relies on two fundamental aspects.

First, the motivational component of sexual reward processing is triggered by appraisal mechanisms, which determine whether a stimulus encountered in the environment is relevant for the organism’s current concerns such as their needs, goals, and values (Frijda, 1986; Sander, Grandjean, & Scherer, 2005; Scherer, 2013). These current concerns encompass multiple categories, from homeostatic needs to more complex aspects tied to self-concepts such as sexual orientation (Talley & Stevens, 2017); however, the important notion is that these concerns are pertinent for a given individual at a certain moment in time (Scherer, 2013). Appraisal mechanisms are therefore largely affected by inter-individual differences (Sander, Grandjean, &

Scherer, 2005). Importantly, it has been shown that reward-associated stimuli trigger

‘wanting’ due to their relevance for the current concerns of the individual (Dayan &

Berridge, 2014; Pool et al., 2016; Robinson & Berridge, 2013; Tindell, et al., 2006;

Zhang et al., 2009). This implies the existence of very strong inter-individual differences in reward processing: a stimulus having specific objective properties triggering ‘wanting’ in an individual but not in another, according to their concerns, experiences, cognitive resources, and other psychological factors influencing the way in which the individual appraises the stimulus.

Second, this hypothesis postulates that sexual reward processing involves two distinct components, ‘wanting’ and liking, which differentially contribute to variations in sexual desire. In Robinson and Berridge’s (1993) version of the incentive motivation model of addictions, they propose that subjective craving in drug addiction is related to ‘wanting’ and not the hedonic impact the reward has for the individual.

Considering that craving is a conscious subjective experience (Anselme & Robinson, 2016), this suggests that the subjective perception of the intensity with which an object or experience is desired can potentially be better predicted by ‘wanting’ than by liking. By extension, given the close relationship between sexual desire and sexual arousal (Toates, 2009), the subjective perception of the intensity with which an object or experience sexually arouses an individual could also potentially be linked to

‘wanting’ rather than liking.

In the present article, we aimed to build on the previous literature testing incentive motivation models, which have thus far focused on basic motivational states (e.g., hunger, thirst, salt appetite), by expanding their range to a larger variety of

30

motivational states that are dependent on inter-individual differences such as sexual orientation in the processing of human sexual reward. In two studies, we tested whether ‘wanting’ triggered by sexual reward associated stimuli is determined by the relevance the stimulus has for the sexual orientation of the individual. To this end, heterosexual and homosexual participants were recruited and sexual rewards were selected for each sexual orientation. Sexual stimuli of the opposite sex were considered more relevant for heterosexual participants than for homosexual participants; whereas sexual stimuli of the same sex were considered more relevant for the homosexual participants than for the heterosexual participants. We predicted that cue-triggered ‘wanting’ for a sexual reward is underlain by the relevance the sexual stimuli has for the sexual concerns of the individual. Furthermore, we tested whether the subjective perception of the intensity of ones’ sexuality can be better predicted by using experimental paradigms that distinctively measure the ‘wanting’

and liking components of sexual reward processing.

2.1.2 Experiment 1

In this experiment, we investigated to what extent a typical measure of cue- triggered ‘wanting’ in a classical Pavlovian instrumental transfer paradigm is determined by its relevance to the sexual orientation of an individual. As a starting point, the relevance of the sexual reward to the individual was manipulated by recruiting heterosexual men and using an erotic image of a man and of a woman.

First, participants rated their liking of neutral images, and erotic images of women and men. Subsequently, to measure ‘wanting’, participants underwent the Pavlovian- instrumental transfer test typically used on animals (Cartoni, Balleine, & Baldassarre;

2016; Holmes, Marchand, & Coutuerau, 2010), which has also been adapted to humans (e.g., Talmi, Seymour, Dayan, & Dolan, 2008; Van Steenbergen, Watson, Wiers, Hommel, & de Wit; 2017). Participants initially learned to associate sexual outcomes, their most liked image of a naked woman and of a naked man, with an instrumental action (i.e., squeezing a handgrip). Subsequently, they learned to associate geometric shapes with said image of the woman (CS1+), of the man (CS2+) and of a neutral image (CS3+). In a testing phase, which was administered under