Aesthetic Appreciation: The View From Neuroimaging

Abstract

Our understanding of aesthetic appreciation has undergone a profound change during the past 20 years, as a result of the ability to study the human brain through neuroimaging. A number of findings cast into doubt important tenets of previous theories and models. Specifically, neuroscientific evidence suggests that aesthetic appreciation is not a distinct neurobiological process assessing certain objects, but a general system, centered on the mesolimbic reward circuit, for assessing the hedonic value of any sensory object. Furthermore, neuroscientific research also makes it clear that hedonic values are not determined solely by object properties, but subject to a range of object-extrinsic modulatory factors. This article reviews these findings and discusses how they demand a new experimental approach to aesthetic appreciation.

Keywords

aesthetic appreciation reward sensory valuation preference aesthetic judgment neuroaesthetics

The introduction, in the 1990s, of noninvasive neuroimaging methods such as positron emission tomography (PET), functional magnetic resonance imaging (fMRI), and magnetoencephalography (MEG), has provoked what can only be called a revolution within the field of empirical aesthetics. The access to in vivo neural activity that these methods provide has allowed researchers to probe neurobiological processes associated with aesthetic appreciation, and, thereby, gain access to the mechanisms that compute and implement aesthetic valuation. It is not hyperbolic, in my view, to claim that, with regard to our understanding of how aesthetic appreciation works, there is a before and an after neuroimaging became part of empirical aesthetics.¹

The neuroimaging revolution has been twofold. First, imaging research has allowed empirical aesthetics to identify the neural systems that, in humans, compute value signals involved in determining liking or disliking, making it possible to explore how various external and internal factors—such as perceptual information, task demands, or interoceptive body states—modulate these processes and affect liking. As a result, we have begun, for the first time, to understand the basic neurobiological principles of how aesthetic appreciation works.

Second, with its underlying neurobiological mechanisms becoming clearer, the way we conceptualize what aesthetic appreciation is has also begun to undergo radical change. More than anything, the emerging neuroscientific evidence has compelled us to realize that aesthetic appreciation is a much more complex phenomenon than previously thought. Where before neuroimaging aesthetic appreciation predominantly was viewed as more or less a type of “judgment,” applied to some stimulus (I like this painting; I dislike this dress), neuroscientific evidence has revealed aesthetic appreciation events to also be involved in, inter alia, predicting and motivating choice and behavior, influencing how other perceptual and cognitive processes operate, and modulating the brain’s various physiological systems. The fact that we humans consciously experience liking something is, in reality, just the top of the iceberg of a much larger number of unconscious processes that unfolds during the neurobiological event we call aesthetic appreciation.

In this article, I describe the central findings that have triggered this revolution in our conception of aesthetic appreciation. Together, they can be said to have opened a black box that remained stubbornly closed to the methods used previously in aesthetic research, especially the psychophysical method (Figure 1). Where before neuroimaging, exploration of aesthetic appreciation was limited to relating object properties to measurements of preference, with no possibility of probing the actual mechanisms relaying one into the other, neuroimaging has made these mechanisms accessible. As a consequence, questions that had long remained obscure have slowly found convincing answers. For example, as I will discuss in further detail later, for most of its history, empirical aesthetics has debated intensely whether emotional or cognitive processes underwrite aesthetic appreciation. As late as the early 2000s, this fundamental question was still hotly debated by scholars such as Cupchik and Martindale (e.g., Martindale, 2001).² Neuroimaging studies have finally solved this mystery for us, although the answer turned out to be more complex than imagined.

Figure 1.

Where the psychophysical method employed by empirical aesthetics before 2000 only allows empirical aesthetics to relate aesthetic judgments to object properties, neuroimaging gives it access to the actual neural mechanisms that computes hedonic value for a sensory stimulus.

Before neuroimaging, aesthetic appreciation was widely assumed to be a specific neural process, assessing an object’s “aesthetic” qualities. Making this determination was thought to be an end to itself and to follow certain rules: stimulus-judgment laws that determine how aesthetic appreciation unfolds when confronted with a specific object property. It has been jolting to discover that none of these ideas is backed up by neuroscientific evidence. Instead, experimental findings support a model of aesthetic appreciation where sensory stimuli are assessed for their relevance to internal adaptive needs and possible behavioral outcomes. According to this emerging conception, aesthetic appreciation does not come down to computing aesthetic judgments to a perceptual input. It comes down to assessing what value a stimulus has for the organism, in its current context, relative to previous experiences, its homeostatic state, and behavioral options. I will try to outline how a new theory of aesthetic appreciation must take these parameters into account, and what kind of questions they raise for the study of aesthetic appreciation going forward.

Inventing Aesthetic Appreciation

What were the main tenets of earlier theories of aesthetic appreciation, and where did they come from?

Aesthetic appreciation as a distinct psychological function, and a scientific problem to be investigated, was first invented by philosophers in the 18th century. This new idea, however, was not birthed ex nihilo, but firmly rooted in the way the human mind had been understood to work since the Middle Ages. The key assumption in this ancient theory of mind, dating back at least to Galen, was a functional separation between the senses and human reason (Kemp, 1996). The senses were thought to passively receive inputs from the surrounding world, upon which a number of “rational faculties” operated. Crudely put, emotional responses to sense data were thought to be passive reflections of these impressions (James, 1997), while cognitive responses were thought to be active. Specifically, reason was imagined to take sensory impressions and turn them into ideas, beliefs, and knowledge. The main question facing philosophers of the day was how this happens. From Leibniz and Locke in the 17th century, over Reid and Hume to Kant in the 18th, this was the central question animating philosophical thinking (Reed, 1997), and it was also this question that gave rise to aesthetics as a philosophical discipline.

Since the ancient Greeks, it had been assumed that human reason comprises a specific number of mental operations, “faculties,” that allow us to make meaningful sense of the onslaught of sense data. For instance, it was posited that when sense data are subsumed under a concept or an idea, we acquire knowledge about what is being perceived. Such mental operations were known as judgments, and the study of how judgments lead to truthful knowledge was called logic. But what about other forms of sense knowledge that does not fall under the auspices of logic? For instance, how can we know that a rose is beautiful from simple sense impressions such as color or smell? Certainly, experiencing a flower as beautiful is not akin to knowing it is red, or any other verisimilar judgment. The genius of philosophers such as Hutcheson, Hume, and Reid in Britain (Dickie, 1996; Kivy, 2003), or Baumgarten and Kant in Germany (Åhlberg, 2003; Wessel Jr., 1972), was to recognize that we need to think of this kind of mental judgments as different in nature from logical judgments and that the reason why such nonlogical judgments exist must be because the human mind is endowed with a specific faculty making them possible. Baumgarten christened the study of this new, putative faculty “aesthetics,” from the Greek word aisthanomai (“I perceive with my senses”). As he wrote in his dissertation, where he first used the word:

The Greek philosophers and the Church fathers have already carefully distinguished between things perceived and things known. It is entirely evident that they did not equate things known with things of sense, since they honored with this name things also removed from sense (therefore, images). Therefore, things known are to be known by the superior faculty as the object of logic; things perceived [are to be known by the inferior faculty, as the object] of the science of perception, or aesthetic. (Baumgarten, 1735, p. 75)

Thus, the study of aesthetics was originally envisioned of as a companion to logic, as the study of a specific way the human mind can respond to sensory impressions (Figure 2). Logical judgments assessed if something is true. Aesthetic judgments assessed if something is likable or beautiful (or something akin to this).³ In line with the prevailing way of understanding mental judgments, aesthetic judgments were conceived of as active mental responses to sense data. As a consequence, the two main questions that came to confront the new philosophical enterprise were the following: (a) First, what are the psychological mechanisms underlying the novel, putative faculty of aesthetic judgment? Can we identify its constitutive processes? (b) Second, what are the rules governing it? Why is a rose beautiful? Can we explain why specific sensory properties lead to specific aesthetic judgments?

Figure 2.

Schematic model of how 18th century philosophers viewed aesthetic appreciation as a specific kind of mental judgment the human mind can apply to sensory input.

Figure 3.

Components making up aesthetic appreciation, as envisioned by philosophers of aesthetics in the 18th century. Objects in our environments impinge on our senses, and a putative mental mechanism called aesthetic judgment assesses the value of this sense impression.

Figure 4.

The main ambition guiding the way empirical aesthetics has operated since Fechner: unearthing stimulus-judgment laws assumed to determine aesthetic appreciation. This is done by manipulating object properties and plotting how variance in object features (z, w) changes aesthetic judgments.

To sum up this short historical genealogy, aesthetic appreciation was developed as a concept by philosophers in the 18th century as an attempt to understand the fact that humans, as one of the ways they respond to sense data, assess sensory objects for their subjective—nonlogical—qualities, especially objects’ hedonic value. They proposed that the human mind contains a dedicated psychological faculty that performs this assessment by wielding an aesthetic judgment on received sense data. Importantly, for the way it was conceived, this theory was rooted in a medieval model of the human mind that has turned out to be wildly misleading, but still furnished the notion of aesthetic appreciation with a number of assumptions that have continued to inform later thinking:

First and foremost, the idea that aesthetic judgments are distinct mental operations. Because of this assumption, to this day, aesthetic appreciation is often thought to tap into distinct cognitive or emotional processes, such as unique aesthetic emotions (Schindler et al., 2017).

Second, the intuition that aesthetic judgments follows from the perceptual representation of a stimulus. This assumption not only suggests that perception precedes aesthetic appreciation but that the way the aesthetic judgment faculty responds is primarily conditioned by inputs from the perceptual system, and thus ultimately by the kind of properties an object is endowed with.

Third, the idea, following from Assumption 2 (that aesthetic appreciation, psychologically speaking, forms a relationship between perception and aesthetic judgment), that the key question confronting the study of aesthetics is how the latter follows from the first (Figure 3). There must be some kind of law that dictates how perception leads to judgment. This question is further compounded by the fact that people very obviously does not agree on how they judge identical percepts. How is this possible? What explains an individual person’s assessment? The story of later empirical aesthetics is essentially a story about the struggle to reconcile the received need for universal laws with observations of individual variance, manipulating object properties and other possible factors to establish the stimulus-judgment “laws” thought to drive aesthetic appreciation (Figure 4).

Finally, the classical notion of “judgment” casts aesthetic appreciation as an end to itself, a machinery for making “decisions” about how to assess the sensory input at hand. This assumption has gone on to impede the ability of empirical aesthetic to countenance aesthetic appreciation as the handmaiden to other neural process. I will return to this point.

Psychological Processes Involved in Aesthetic Appreciation

Obviously, the philosophers who invented aesthetics and the notion of aesthetic appreciation had little idea how the human brain actually is constructed and functions. Consequently, attempts to explain what psychological mechanisms might constitute the envisioned aesthetic judgment faculty remained speculative until the end of the 20th century. Two principal hypotheses dominated such theories: Aesthetic judgments could be implemented by “cognitive” processes, or they could be implemented by “emotional” processes (Figure 5). Thus, percepts could conceivably acquire their aesthetic value by having an emotional impact, or through being subject to specific cognitive operations. As already mentioned, both accounts had their proponents well into the 21th century.

Figure 5.

Possible psychological processes speculated to underlie aesthetic judgments before empirical neuroscience.

The first evidence-based models of how “perception,” “emotion,” or “cognition” really work only came to life in the 1960s. Two major experimental breakthroughs paved the way for this change and also profoundly influenced how empirical aesthetics would go on to translate the notion of aesthetic appreciation into psychology and neuroscience. One was the observation that perceptual systems are organized in a hierarchical system of different functional units that analyze different aspects of the sensory input (Felleman & van Essen, 1991; Livingstone & Hubel, 1988). This body of work suggested that information flows through a number of functional nodes that first extract certain important features of the sensory stimulus, then combine these features into Gestalts, and finally projects the outcome to “association” networks where the percept is infused with knowledge and memory. This model of perceptual information processing was extremely important in shaping psychological theories of how aesthetic appreciation works (Gregory, Harris, Heard, & Rose, 1995; O’Hare, 1981; Solso, 1994). It set up an account of aesthetic processing where bottom-up processes activate, first, an early stage of perceptual processing, the output of which is projected to a later stage molding this bottom-up input. Together, the two stages make sense of what is being perceived and evaluate the resulting percept for its perceived aesthetic qualities.

The pinnacle of the use of information-processing research to model the psychological mechanisms involved in aesthetic judgments came with the publication of Helmut Leder’s influential 2004 model of aesthetic appreciation (Figure 6).

Figure 6.

Leder’s model detailing information-processing steps leading to an aesthetic judgment (Leder et al., 2004).

In Leder’s model, processes thought to specifically underwrite aesthetic assessment of sensory objects are mapped onto known perceptual nodes of the visual system. For instance, sensory properties thought to be of importance for assessing visual stimuli, such as complexity or symmetry, are dealt with by early “perceptual analysis” nodes, while later knowledge-related nodes treat things like familiarity or style (Leder, Belke, Oeberst, & Augustin, 2004). Aesthetic judgments are proposed to result from the interaction between such earlier mechanisms and later “mastering” and “evaluation” processes.

It is interesting to note that even a modern model such as Leder’s retain the classical view of aesthetic judgments as “acts” on sense impressions. The Leder model suggests that early perceptual processes are “automatic,” while the later processes involved in “mastering” the percept are “deliberate.” Aesthetic judgments are viewed as outputs from this perceptual-cognitive information-processing stream, suggesting that a stimulus must travel through all computational steps before acquiring its aesthetic assessment—although it is not in fact specified in the 2004 model exactly which process, or perhaps processes, in the processing chain that “do” the actual judging.

While Leder’s model is mainly focused on perceptual-cognitive processes, giving short shrift to emotional processes—albeit they appear to be thought of as “by-products” of the perceptual-cognitive processing—a second major breakthrough in the 1960s put emotions front and center in another attempt to explain the psychology of aesthetic appreciation. This breakthrough came with animal studies relating regulatory systems, controlling food intake and metabolism, to specific neurobiological systems. A key idea to emerge from this line of research was the notion that emotions play a central role in regulating behavior to conform to physiological needs. Specifically, behavior that solves homeostatic needs is “rewarded” by a pleasurable feeling, while behavior that does not is “punished” by negative hedonic feelings (displeasure, disgust, etc.). As a consequence, such positive and negative affective states help promote adaptive behavior by “motivating” the organism to engage in advantageous, and avoid problematic, behavior (Berridge, 2004). Importantly, in animal models, it was possible to locate these regulatory functions to specific anatomical regions—such as the brainstem, hypothalamus, striatum—and physiological processes, paving the way for a possible neurobiological theory of emotions (Swanson, 2000).

One of the systems found to play a functional role in the regulation of metabolic activity, and thus possibly underlie the affective implementation of motivational states, was the so-called ascending reticular activating system (Fuller, Sherman, Pedersen, Saper, & Lu, 2011). The reticular activating system appeared to be involved in regulating changes to the body—heart rate, blood pressure, muscle tension—through the modulation of various physiological processes, especially arousal. Arousal, therefore, could be thought of as a candidate physiological component driving motivated behavior (Bradley, 2000), mediating between emotional states and behavior.

The first researcher to integrate this emerging body of work into models of aesthetic appreciation was Daniel Berlyne, who suggested that the concepts of arousal and reward not only help explain regulation of metabolic behavior, but why organisms “explore” its sensory surroundings to begin with (Berlyne, 1960, 1966). More specifically, Berlyne’s theory proposed that changes in arousal level could be a possible physiological mechanism determining why we like some objects we encounter, and dislike others: Sensory properties that succeed in elevating arousal produce a positive hedonic effect, rewarding the organism with an experience of pleasure (Berlyne, 1971). In contrast, stimuli that decrease arousal yield a negative hedonic value (Figure 7).

Figure 7.

Berlyne’s model of aesthetic appreciation. Aesthetic values for sensory impressions are thought to be mediated by the level of arousal the stimulus elicits.

Berlyne thus explains aesthetic judgments as an affective response to the sensory stimulus: We like stimuli that are able to enhance autonomic arousal and dislike stimuli that decrease arousal. His theory also presents us with a possible reason why organisms need an aesthetic system for assessing the hedonic value of sensory objects in the first place.

Unfortunately, while attractive in a number of ways, Berlyne’s theory turned out to suffer from several flaws. One was the fact that, while his theory does suggests a specific mechanism driving aesthetic judgments, it is far from concrete in explaining how arousal in fact interfaces with perception and other systems implementing the affective response. Furthermore, predictions following from Berlyne’s model with regard to which stimuli should enhance arousal, and hence be judged as pleasing, failed to pan out experimentally (e.g., Martindale, Moore, & Borkum, 1990). Finally, as neuroscience progressed, it also soon became clear that the role assigned to arousal as the causal factor driving affective states was far too simplistic (Silvia, 2005).

Neuroimaging Aesthetic Valuation

A major disadvantage to earlier theories of aesthetic appreciation based on information processing or arousal research was the lack of opportunity to test them directly on humans. Most of the neuroscientific work on perceptual-cognitive systems and arousal came from animal studies and was then inferred to also hold for aesthetic appreciation in humans. Only when neuroimaging techniques became available such a test became possible. Especially PET and fMRI, with their capacity for precise location of neural activity, promised to provide answers to the long-standing questions plaguing empirical aesthetics. I have already mentioned (in Note 2) how the first fMRI experiments imaging subjects while engaged in aesthetic judgments explicitly hoped to resolve the question of whether aesthetic appreciation relies on emotional or perceptual-cognitive structures. As Oshin Vartanian and Marcos Nadal wrote in 2007, the case for empirical aesthetics adopting neuroimaging as a method very much rested on the opportunity it offered to test the validity of existing models (Vartanian & Nadal, 2007).

It is in fact quite striking how the first wave of neuroaesthetics research embraced almost all of the assumptions inherited from the original inventors of aesthetic appreciation as a concept. To wit: (a) Most early studies explicitly hypothesized that aesthetic appreciation would rely on a distinct neurobiological system. This could be seen in the way neuroaesthetics studies only used (or indeed allowed) a select set of objects to count as relevant “aesthetic” stimuli—primarily art, faces, and certain visual and auditory object properties (e.g., Aharon et al., 2001; Blood, Zatorre, Bermudez, & Evans, 1999; Cela-Conde et al., 2004; Kawabata & Zeki, 2004; Vartanian & Goel, 2004a)—or specifically tried to contrast aesthetic judgment with other forms of judgment (Jacobsen, Schubotz, Höfel, & von Cramon, 2006). (b) Almost all early neuroaesthetics studies were designed with the stimulus as the independent variable, thus upholding the age-old assumption that aesthetic appreciation is caused by, and forms a response to, variations in object features. (c) Their main theoretical objective was to find the neural correlates of the aesthetic judgment “faculty,” specifically the neural processes underlying positive and negative judgments. (Consequently, most analyses contrasted liking and disliking states with either each other or some control state.) This experimental aim helped sustain the view of aesthetic appreciation as an end to itself, unrelated to other neural processes. Just as strikingly, all of these assumptions turned out to be in conflict with the actual findings produced by the and the studies that followed.

At first, it was somewhat difficult for neuroaesthetics to make sense of the results it produced. Contrast analyses found elevated activity in diffuse networks for positive and negative judgments. As Marcos Nadal lamented in a review, there was “a complete lack of coincidence among the results” (Nadal, Munar, Capó, Rosselló, & Cela-Conde, 2008, p. 385). A clearer picture first emerged with the publication of a number of statistical meta-analyses in the early 2010s (see Table 1). These meta-analyses calculated which neural loci are commonly activated across a large number of different evaluation experiments, identifying the neural structures found to be recurrently involved in valuation events irrespective of specific experimental designs or choice of analysis. Repeatedly in these meta-analyses, the key anatomical structures found to be involved in sensory valuation were structures located in the reward circuit: striatal structures (especially nucleus accumbens [NAcc] and pallidum), orbitofrontal cortex (OFC), the anterior cingulate cortex, insula, and the amygdala (Figure 8).

Figure 8.

Depiction of the reward circuit, outlining the location of its central anatomical nodes. (a) Medial view showing striatum, OFC, ACC, and amygdala and (b) lateral view showing insula.

Table 1.

Overview of Statistical Meta-Analyses Analyzing Neural Structures Found Commonly Activated in Imaging Studies Investigating Sensory Valuation in Humans.

Paper	Stimuli	OFC	ACC	Striatum	Insula	Amygdala
Bzdok et al., 2010	Faces	∗				∗
Brown, Gao, Tisdelle, Eickhoff, & Liotti, 2011	Visual	∗	∗	∗	∗	∗
	Auditory	∗			∗
	Olfactory		∗	∗	∗
	Gustatory	*	*		∗
Kühn & Gallinat, 2012	Faces, odor, drink, food, IAPS, paintings, music, photos	∗	∗
Sescousse, Caldú, Segura, & Dreher, 2013	Taste		∗	∗	∗
	Money	∗	∗	∗	∗	∗
	Erotic	∗	∗	∗	∗	∗
Bartra, McGuire & Kable, 2013	Monetary payoffs, liquids, arousing pictures, social feedback	∗	∗	∗	∗
Zou, van Hartvelt, Kringelbach, Cheung, & Chan, 2016	Odors	∗	∗		∗	∗

Note. Asterisks indicate that one of the reward circuit’s key structures was found to be involved. ACC = anterior cingulate cortex; OFC = orbitofrontal cortex; IAPS = International Affective Picture System.

When viewed in this amalgamated way, considering data from thousands of human subjects, scanned in many hundreds of studies, the neuroimaging evidence suggests that aesthetic appreciation is rooted, primarily, in neurobiological processes located in the reward circuit. But what does this finding entail for our understanding of aesthetic valuation? To start with, locating the neural machinery involved in aesthetic valuation to the reward circuit endorse the hypothesis that we like sensory objects when they trigger a positive affective response. In other words, it suggests that aesthetic judgments rely on hedonic valuation computations to assess whether an object is likable. However, these hedonic valuation computations may not function exactly like previous aesthetic theories imagined they did.

First and foremost, analyzing the common activation pattern for aesthetic appreciation studies not only reveal this pattern to be centered on the reward circuit, it also makes it clear that the human brain only has evolved one common neural network to compute hedonic value. Regardless of sensory modality or what kind of object is being assessed the hedonic value is computed by the same nuclei in striatum, OFC, striatum, anterior cingulate cortex, insula, and amygdala. Thus, objects traditionally thought to be distinctively “aesthetic,” such as art objects or faces, are appreciated by the brain using the same neurobiological value mechanisms it uses to assess liking for food, drinks, odors, landscapes, table, chairs, or computer products—even money (Table 1). Neuroscientists refer to this idea as the “common currency” hypothesis (Berridge & Kringelbach, 2015; Levy & Glimcher, 2012; Montague & King-Casas, 2007).

Can we know with certainty that it is in fact neurons in the reward circuit that encode hedonic values? One problem with relying on imaging studies is that most types of functional imaging data are, as a matter of principle, correlational in nature. Another problem is that it can often be very difficult to distinguish “true” value computations from auxiliary computations that are associated with value computations, but not actual value signals themselves (O’Doherty, 2014). For instance, it can be very hard to disentangle neural processes that encode consequences of valuation from processes that encode value as such. How can we be sure that the reward nodes found activated in imaging studies belong to the latter category and not to the first?

The main line of evidence supporting reward structures as the locus of hedonic valuation computations has come from observations of where in the brain activity is modulated when valuation is manipulated through some kind of intervention. This kind of manipulation can in fact be accomplished in neuroimaging studies as well. For example, in studies where subjects continue to ingest chocolate, their hedonic response will change over time from liking to disliking as they become increasingly sated. This variation in liking can be modeled as a parametric regressor, which can be used to locate neural activity that reflects this change. Many studies have located this type of neural activity to the reward circuit, especially OFC (e.g., Kringelbach, O’Doherty, Rolls, & Andrews, 2003; Small, Zatorre, Dagher, Evans, & Jones-Gotman, 2001). In animal models, it is also possible to directly monitor firing patterns in single cells while liking responses changes. For instance, if rats are depleted of salt, they start liking even high concentrations of sodium, a change reflected in the firing of NAcc neurons that do not respond to salt in nondepleted states (Tindell, Smith, Peciña, Berridge, & Aldridge, 2006).

Another line of research that demonstrates a causal relation between reward activity and change in hedonic value comes from electrophysiological stimulation where the stimulus remain unchanged, but activity in parts of the reward circuit is manipulated, with an increase or decrease in liking as a result (e.g., Peciña, Smith, & Berridge, 2006; Peng et al., 2015). A similar effect can be achieved by manipulating neurochemicals involved in reward processing. Injecting a μ-opioid agonist in specific parts of NAcc or ventral pallidum selectively enhances liking responses in rats (Smith & Berridge, 2005, 2007), while administering a μ-opioid antagonist to human subjects strongly attenuates liking responses for sensory inputs, including music (Mallik, Chanda, & Levitin, 2017).

Finally, a very interesting clinical phenomenon called specific musical anhedonia (SMA) strongly implicates the reward circuit as the central neural hub for value computation. SMA subjects experience a reduced pleasure response to music, despite having an intact ability to represent music perceptually, and a preserved ability to feel pleasure for nonmusical input such as visual art and monetary rewards (Mas-Herrero, Marcos-Pallares, Zatorre, & Rodriguez-Fornells, 2018; Mas-Herrero, Zatorre, Rodriguez-Fornells, & Marco-Pallaes, 2014). A recent fMRI study by Martínez-Molina, Mas-Herrero, Rodríguez-Fornells, Zatorre, and Marco-Pallares (2016) showed that this difference in hedonic response to music and other stimuli, for SMA subjects, can be explained by a difference in NAcc activity, caused by diminished white matter connectivity between the auditory areas responsible for representing music and this part of the reward circuit (Sachs, Ellis, Schlaug, & Loui, 2016). In fact, individual variability in connectivity between auditory cortex and the reward circuit also predicts individual differences in experience of musical pleasure in nonanhedonic subjects (Loui et al., 2017), suggesting that perceptual representations of stimuli must gain access to the reward circuit for the brain to be able to attach a hedonic value to it.

Together with the abundant evidence found in functional imaging studies, intervention studies makes a strong case that neurons in such structures as NAcc, pallidum, OFC, and so forth subserve value computations that allow the brain to assess sensory objects for their perceived pleasantness or unpleasantness. This, in turn, raises the question of how the reward circuit comes to make these value “decisions.”

The Functional Purpose of Aesthetic Appreciation

If the early wave of neuroaesthetic research can be described as a quest to identify the neurobiological processes involved in generating liking, later years have been dominated by an interest in understanding the computational principles guiding these processes. It is one thing to know which neural systems underlie hedonic valuation; it is another to understand why some stimuli get imbued with positive hedonic values, and other with negative value. This question has turned out only to be answerable when we take into account the general functional principles of the reward circuit.

As already noted in the earlier discussion of Berlyne’s arousal thesis, the moniker “reward” refers to the functional role hedonic valuation plays. From a functional point of view rewarding values can be thought of as affective states that motivate behavior by “nudging” the organism to pursue actions that are accompanied by feelings of pleasure and to avoid actions that give rise to displeasure or disgust. In this sense, hedonic values help reward or punish behavioral decisions (Rangel, Camerer, & Montague, 2008).

From a neurobiological point of view, research shows that motivated behavior can be broken down into three interrelated, but dissociable, phases (Figure 9): an appetitive phase, a consummatory phase, and a satiety phase (Berridge & Kringelbach, 2015; Kringelbach & Berridge, 2017; Swanson, 2000). The appetitive phase establishes a need and compels the organism to initiate behavioral acts that can alleviate this need. To take a simple example, thirst arises as receptors in the blood detects changes in fluid balance (plasma osmolarity and angiotensin II) and signals this information to neurons in the lamina terminalis (Zimmerman, Leib, & Knight, 2017). A network of neurons in the lamina terminalis triggers a suite of responses the purpose of which is the restoration of fluid balance, including thirst and salt appetite. These processes initiate search for and procurement of a fluid that contains the physiological components needed by the organism (e.g., sodium). When such a fluid is found, other neural processes control its intake and ingestion; together they form the consummation phase. As the ingested molecules restore the homeostatic balance, changes to plasma osmolarity, volume, and pressure are signaled back to the reward circuit, inhibiting or decreasing neural activity—the satiety phase.

Figure 9.

Schematic depiction of the three functional phases characterizing reward processing (Kringelbach & Berridge, 2017).AMY = amygdala; aMCC = anterior middle cingulate cortex; Claus = claustrum; dlPFC = dorsolateral prefrontal cortex; dmPFC = dorsomedial prefrontal cortex; FO = frontal operculum; HT = hypothalamus; Ins = insula; IPL = intraparietal lobule; med-temp = medial temporal conrtex; OFC = orbitofrontal cortex; pACC = pregenual anterior cingulate cortex; sACC = subgenual anterior cingulate cortex; SPL = superior parietal lobule; temp pole = temporal pole; vlOT = ventrolateral occipitotemporal; vlTG = ventrolateral temporal gyrus; vmPFC = ventromedial proefrontal cortex; VS = ventral striatum.

The reward circuit computes different value signals that help implement these three phases. For this reason, in reality, the reward circuit contains several different value mechanisms that reflect different aspects of value-based behavior regulation. Some value mechanisms help the organism forage and make sense of the surrounding sensory world as it relates to the homeostatic problem at hand. For instance, the reward circuit computes value signals that predict the likely reward value of a sensory input, informing and promoting behavior. Kent Berridge has dubbed such value signals “wanting” mechanisms, because they promote approach toward, and consumption of, rewards (Berridge, Robinson, & Aldridge, 2009). Value predictions not only motivate the organism to want a specific part of its surroundings (by predicting it will be rewarding) but also bias perception toward salient features and modulate neural activity in perceptual and cognitive systems. A second type of value signals helps integrate approach and avoidance drives with the brain’s executive and motor systems, to help the organism make decisions about what to consummate. Extensive research has demonstrated how neurons in the ventromedial prefrontal cortex, including OFC, compare values of available options and signal to premotor cortex and other parts of the executive system to select and implement behavioral choices (Padoa-Schioppa, 2011; Rushworth, Mars, & Summerfield, 2009). Finally, a third set of value signals helps compute the outcome of the consummatory phase, that is to say how rewarding the chosen behavior turned out to be. Berridge calls these value mechanisms “liking” mechanisms.

Because the purpose of hedonic values is to help regulate motivated behavior, we can only understand their nature—and, crucially, how a specific value is computed in a given context—if we understand value signals as a function of the way the reward circuit help regulate motivated behavior. We like, to put it crudely, what is good for the organism. We dislike what is bad for us. Specifically, to understand how value signals behave in individual situations, we need to understand how neural activity intersects with, and is modulated by, neural systems involved in regulating homeostatic physiology. For example, the way neurons in NAcc behave while computing value signals for glycemic replenishment is wholly dependent on a complex network of signaling between neurons in the hypothalamus and ventral tegmental area, modulated by hormones such as leptin or ghrelin that reflect interoceptive energy levels (Palmiter, 2007). Hedonic values for sensory inputs can be thought of as affective states that relate information about the exteroceptive world to the internal physiology of homeostatic regulation. It allows us to classify the auditory, visual, and so forth objects we encounter with regard to their potential positive or negative effect on our survival. It does so by integrating perceptual information with information from interoceptive systems, signaling homeostatic states, as well as information reflecting relevant behavioral task demands.

Factors Modulating Hedonic Valuation

In wake of the first wave of imaging studies investigating hedonic valuation, we have witnessed a second wave of experimental findings that overwhelmingly supports this revised model of aesthetic appreciation. We can summarize the tenets of this revised model in the following way. Aesthetic appreciation computes hedonic values for sensory input that informs the organism about whether the object is of positive or negative relevance to its regulatory needs and behavioral options. These hedonic values are not computed by distinct systems for different object categories, but by one common currency system, the reward circuit. (This finding very obviously entails that objects that have been proposed to elicit “disinterested” aesthetic judgments in fact also acquire their hedonic value through the activation of motivational value signals.) Furthermore, the value mechanisms computing how likable or dislikable an object is are not driven solely by projections from the perceptual system representing the object, but equally by projections from interoceptive and executive top-down systems. Figure 10 provides a schematic representation of the revised model of aesthetic appreciation.

Figure 10.

According to recent evidence, hedonic values are computed by nuclei in the reward circuit for sensory inputs. Value computations are, however, strongly modulated by inputs from interoceptive and executive systems, reflecting contextual task demands and homeostatic state.

I will limit myself to a few examples of this extensive empirical literature (see also Skov, in press). I have already mentioned how imaging studies have revealed neurons associated with value assessment to be modulated by satiation (Kringelbach et al., 2003; Small et al., 2001). As just discussed, the reason why the reward system is set up to valuate the same stimulus differently relative to satiation level is that having values that reflect homeostatic states helps calibrate ingestion of food and drink to match the physiological needs of the organism. As a consequence, food objects—even cues indicating the presence of possible food items (Beaver et al., 2006; Schur et al., 2009)—are deemed more valuable when satiation is low, and less valuable when high, and neural activity in structures such as NAcc, pallidum, or OFC, is found to vary as a consequence, even though the percept remains constant (Führer, Zysset, & Stumvoll, 2008; Katsuura, Hexkmann, & Taha, 2011; Siep et al., 2009; Thomas et al., 2015). Satiation, though, is only one example of the ways endogenous processes regulate reward activity, and as consequence modulate hedonic valuation for sensory objects. Another example would be the way estrus modifies hormone levels across women’s menstrual cycle, modulating both activity in the reward circuit (Caldú & Dreher, 2009; Yoest, Quigley, & Becker, 2018) and women’s hedonic values for various objects (Dreher et al., 2007; Frank, Kim, Krzemien, & Van Vugt, 2010; Ossewaarde et al., 2010; Rupp et al., 2009).

Where the organism’s endogenous state is communicated to the reward circuit via projections from interoceptive and neuroendocrinological systems, task demands related to ongoing behavior are primarily relayed through top-down signals originating in the prefrontal cortex. Perhaps most importantly, such top-down signals help modulate expectations for the sensory object being valued. An experimental manipulation often used to investigate changes in expectations, and how they modulate hedonic valuation, is the so-called framing effect (Okamoto & Dan, 2013). In framing studies, object-extrinsic cues are presented before a sensory object is processed and valued, influencing how these processes unfold. A large number of such studies have found that semantic labels, names, titles, or other forms of information modulate hedonic valuation of an object, even though there is no change in the sensory input (Fernqvist & Ekelund, 2014; Krishna, 2012; Okamoto & Dan, 2013; Piqueras-Fiszman & Spence, 2018). Neuroimaging experiments demonstrate that this modulation of hedonic value is associated with changes in reward activity (e.g., Kirk, Skov, Hulme, Christensen, & Zeki, 2009; McClure et al., 2004; Plassmann, O'Doherty, Shiv, & Rangel, 2008). Intriguingly, people who are less susceptible to the framing effect exhibit increased regulatory activity in the executive network, including dorsolateral prefrontal cortex (Aydogan et al., 2018; Schmidt, Skvortsova, Kullen, Weber, & Plassmann, 2017).

Together, research into variation in satiation and expectations demonstrate that both homeostatic states and task demands influence how aesthetic appreciation values sensory objects. More specifically, they make clear that hedonic values are not determined solely by the information the reward circuit receives from the perceptual system. Rather, the neurons computing how likable or dislikable an object is are strongly modulated by a host of contextual factors reflecting regulatory needs and situational task demands.

Consequences of Hedonic Valuation

Because hedonic values are computed by the brain to motivate behavior, they are not in fact self-contained neural states. Rather, value signals are projected to a range of other systems, influencing decision-making (Calhoun & Hayden, 2015; Koechlin, 2016; O’Doherty, Cockburn, & Pauli, 2017), and physiological processes important to the implementation of behavior. Like other emotional systems, hedonic valuation triggers what Damasio and Carvalho (2013) call action programs: changes in autonomic, endocrinological, and motor function processes relevant to the implementation of behavior. A well-known example hereof is the way musical chills elevate heat rate, skin conductance, or respiration (Craig, 2005; Salimpoor, Benovoy, Longo, Cooperstock, & Zatorre, 2009).

What is perhaps more surprising is that hedonic valuation processes also project to, and influence, perception. The hedonic value assigned to a sensory object modulates the way we attend to it and helps determine how the object is perceptually represented by the brain. For example, numerous studies have found that objects deemed more pleasurable draw rapid and preferred attention (e.g., di Pelligrino, Margarelli, & Mengarelli, 2011; Valuch, Pflüger, Wallner, Laeng, & Ansorge, 2015), even when experienced subliminally (Hung, Nieh, & Hsieh, 2016; Sui & Liu, 2009). Similarly, hedonic valuation influences which parts of a crowded sensory environment the brain expends valuable computational resources on exploring (Chen, Liu, & Nakabayashi, 2012; Liu & Chen, 2012; Li, Oksama, & Hyönä, 2016). Being pleasing or displeasing might even help a stimulus gain faster access to consciousness (Nakamura, Arai, & Kawabata, 2018; Rams⊘y & Skov, 2014).

Value expectations, based on the reward history of an object also heavily modulate neural activity, even in the earliest parts of the perceptual system (e.g., Hikosaka, Nakamura, & Nakahara, 2006; Serences, 2008; Summerfield & de Lange, 2014). For example, electrophysiological studies of neural activity in V1 and V4 have found that, as a visual cue becomes associated with reward value, activity in a subset of neurons change firing rates to reflect value attribution (Baruni, Lau, & Salzman, 2015; Gavornik, Shuler, Loewenstein, Bear, & Shuval, 2009; Goltstein, Meijer, & Pennartz, 2018; Shuler & Bear, 2006; Zold & Shuler, 2015). This modulation of spiking patterns may be regulated by specific projections from forebrain structures to V1 (e.g., Chubykin, Roach, Bear, & Shuler, 2013), although the exact mechanisms controlling the way reward value influence perceptual activity remain elusive.

Findings like these suggest that the brain’s perceptual systems are not passive conduits of stimulus information, to which the reward system simply responds with a positive or negative affective response when it has been determined what object features are present. Rather perception integrates information from valuation computation into the way a stimulus is being processed by the brain’s sensory systems. This may seem surprising if we assume that aesthetic appreciation is a type of judgment applied to perception, but not if we understand it to be a part of behavior regulation. On the latter view, it makes sense that the brain uses an object’s assigned value to establish to how salient and important it is relative to the behavioral decision-making the organism is engaged in.

Toward a New “New” Aesthetics

When Berlyne, in the 1970s, called for a new experimental aesthetics, he made a fairly scathing assessment of the old experimental aesthetics preceding him. “Experimental aesthetics has a long but not particular distinguished history,” he wrote (Berlyne, 1974, p. 5). “For most of the ensuing century [following Fechner], it made some progress (. . .), but in comparison with other branches of experimental psychology, its products were relative sparse and, on the whole, not profoundly enlightening” (Berlyne, 1974, p. 5). I am not going to judge his contribution, or others preceding the neuroimaging revolution, as equally unenlightening. But as I hope to have shown, our understanding of what aesthetic appreciation is, and how hedonic values are computed by the human brain, have been profoundly changed by the past two decades of neuroscientific research, upending most of the assumptions first established by 18th century philosophers. We can summarize the implications for research into the phenomenon of aesthetic appreciation under three headings:

Aesthetic appreciation is not a distinct mechanism. Empirical evidence amassed by neuroimaging and other neuroscientific methods strongly implicates the brain’s mesolimbic reward circuit as the central computational hub for assessing sensory objects’ hedonic value. Indeed, the reward circuit appears to compute hedonic values for all categories of sensory objects. Importantly, because the functional purpose of the reward circuit is to regulate motivated behavior, computation of hedonic values is subject to a number of important modulatory factors, including information about the homeostatic state of the organism, and task demands dictated by ongoing behavior (Figure 10). This account questions the traditional interpretation of aesthetic appreciation as a distinct psychological or neurobiological process, eliciting special aesthetic emotions for special aesthetic objects (Nadal & Skov, 2018; Skov & Nadal, 2018). There is no good reason to focus research in aesthetic appreciation studies on specific object categories because it is assumed that certain objects, especially art works, are treated differently by the reward circuit, nor for experiments to hypothesize that such objects evokes a specialized set of affective responses (Skov & Nadal, 2019). Rather, aesthetic appreciation must be thought of as a fundamental neurobiological phenomenon, yielding elementary hedonic values for cultural objects as well as food, sex, social behavior, and economic transactions.

Aesthetic appreciation is not driven only by object properties. From the original philosophical idea that aesthetic judgments respond to sense impressions to later psychophysical research in empirical aesthetics, the dominant hypothesis regarding aesthetic appreciation has been that aesthetic values are determined by object properties. Neuroimaging has largely falsified this hypothesis. Instead, evidence suggests that identical stimuli can give rise to diverse computational representations, both in perceptual networks and in the reward circuit, as well as different hedonic values, when contextual circumstances vary. This flexible nature of aesthetic appreciation reflects its functional purpose, as outlined in the earlier sections. For example, chocolate is valued higher when blood sugar levels are low, and lower when these are high, because this variance in hedonic value helps decide if it is advantageous or not to consume chocolate. Experimental work must reflect this fundamental fact.

Aesthetic appreciation is not an end to itself. In most previous theories of aesthetic appreciation, aesthetic judgments are, explicitly or tacitly, assumed to be an end point, the judgment output our sole interest. In empirical research, we see this assumption play out very strikingly in the way almost all experiments focus on how a particular manipulation affects ratings. In reality, the conscious feeling we introspectively attend to and report in such experiments is only one of many neural consequences of a hedonic valuation event. Information from the reward circuit is projected to a number of other neural systems, affecting physiological, endocrinological, executive, and perceptual processes. Although experiments in empirical aesthetics do routinely collect data about physiological, behavioral, or cognitive changes associated with hedonic valuation, we have not as a field been very good at conceptualizing these as “outputs” in their own right, or at integrating them into models of aesthetic appreciation. Certain neural consequences of aesthetic appreciation events, such as endocrinological changes, remain woefully understudied.

Even if this radical reevaluation of central tenets in our understanding of aesthetic appreciation does not completely invalidate Berlyne’s old “new” experimental aesthetics, it may suggest that we do need a new “new” experimental aesthetics. At least, it can be argued that we are now at a juncture where the main question of scientific aesthetics is not which object properties drive aesthetic appreciation, but how neural mechanisms compute hedonic values to assess the biological relevance of a sensory object to motivated behavior. How, and why, does neural information from sensory processing gain access to the reward circuit? How are value signals computed by different nuclei in the reward system, and how is neural activity associated with these processes regulated by other neural systems? What are their molecular constituents? While the contours of what aesthetic appreciation is are becoming clearer, its precise neurobiological architecture and computational principles remain obscure. The next step in the scientific exploration of aesthetic appreciation must make it its mission to tackle these profound mysteries.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Martin Skov is a senior researcher at the Danish Research Centre for Magnetic Resonance (Copenhagen University Hospital) and Center for Decision Neuroscience (Copenhagen Business School). His main research interests include: Modeling neurobiological mechanisms involved in sensory valuation. His publications include several edited volumes, here of two special issues on neuroaesthetics (2005, 2013), and the book Neuroaesthetics (Baywood, 2009), co-edited with Oshin Vartanian. Follow his research on Twitter: @mskov01.

References

Aharon

Etcoff

Ariely

Chabris

C. F.

O’Connor

Breiter

H. C.

(2001). Beautiful faces have variable reward value. Neuron, 32, 537–551.

Åhlberg

L.-O.

(2003). The invention of modern aesthetics: From Leibniz to Kant. Historicni Seminar, 4, 133–153.

Aydogan

Flaig

Ravi

S. N.

Large

E. W.

McClure

S. M.

Margulis

E. H.

(2018). Overcoming bias: Cognitive control reduces susceptibility to framing effects in evaluating musical performance. Scientific Report, 8(1), 6229.

Bartra

McGuire

J. T.

Kable

J. W.

(2013). The valuation system: A coordinate-based meta-analysis of BOLD fMRI experiments examining neural correlates of subjective value. Neuroimage, 76, 412–427.

Baruni

J. K.

Lau

Salzman

C. D.

(2015). Reward expectation differentially modulates attentional behavior and activity in visual area V4. Nature Neuroscience, 18, 1656–1663.

Baumgarten

A. G.

(1735). Reflections on poetry: Alexander Gottlieb Baumgarten's Meditationes philosophicae de nonnullis ad poema pertinentibus. Berkeley and Los Angeles, CA: University of California Press.

Beaver

J. D.

Lawrence

A. D.

van Ditzhuijzen

Davis

M. H.

Woods

Calder

A. J.

(2006). Individual differences in reward drive predict neural responses to images of food. Journal of Neuroscience, 26, 5160–5166.

Berlyne

D. E.

(1960). Conflict, arousal, and curiosity. New York, NY: McGraw-Hill.

Berlyne, D. E. (1966). Curiosity and exploration. Science, 153, 25--33.

10.

Berlyne

D. E.

(1971). Aesthetics and psychobiology. New York, NY: Appleton-Century-Crofts.

11.

Berlyne

D. E.

(1974). The new experimental aesthetics. In D. Berlyne (Ed.), Studies in the new experimental aesthetics. Steps toward an objective psychology of aesthetic appreciation (pp. 1--25). New York, NY: Taylor & Francis.

12.

Berridge

K. C.

(2004). Motivational concepts in behavioral neuroscience. Physiology & Behavior, 81, 179–209.

13.

Berridge

K. C.

Kringelbach

M. L.

(2015). Pleasure systems in the brain. Neuron, 86, 646–664.

14.

Berridge

K. C.

Robinson

T. E.

Aldridge

J. W.

(2009). Dissecting components of reward: “Liking”, “wanting”, and learning. Current Opinion in Pharmacology, 9, 65–73.

15.

Blood

A. J.

Zatorre

R. J.

Bermudez

Evans

A. C.

(1999). Emotional responses to pleasant and unpleasant music correlate with activity in paralimbic brain regions. Nature Neuroscience, 2, 382–387.

16.

Bradley

M. M.

(2000). Emotion and motivation. In J. T. Caioppo, L. G. Tassinary, & G. G. Berntson (Eds.), Handbook of psychophysiology (2nd ed.) (pp. 602--642). Cambridge, MA: Cambridge University Press.

17.

Brown

Gao

Tisdelle

Eickhoff

S. B.

Liotti

(2011). Naturalizing aesthetics: Brain areas for aesthetic appraisal across sensory modalities. Neuroimage, 58, 250–258.

18.

Bzdok

Langner

Caspers

Kurth

Habel

Zilles

, . . . Eickhoff

S. B.

(2010). ALE meta-analysis on facial judgments of trustworthiness and attractiveness. Brain Structure and Function, 215, 209–223.

19.

Caldú

Dreher

J.-C.

(2009). Gonadal steroid hormones’ influence on decision making processes. In J.-C. Dreher & L. Tremblay (Eds.), Handbook of reward and decision making (pp. 309--334). Amsterdam, the Netherlands: Academic Press.

20.

Calhoun

A. J.

Hayden

B. Y.

(2015). The foraging brain. Current Opinion in Behavioral Sciences, 5, 24–31.

21.

Cela-Conde

C. J.

Marty

Maestú

Ortiz

Munar

Fernández

, . . . Quesney

(2004). Activation of the prefrontal cortex in the human visual aesthetic perception. Proceedings of the National Academy of Sciences of the United States of America, 101, 6321–6325.

22.

Chen

Liu

C. H.

Nakabayashi

(2012). Beauty hinders attention switch in change detection: The role of facial attractiveness and distinctiveness. PLoS One, 7(2), e32897.

23.

Chubykin

A. A.

Roach

E. B.

Bear

M. F.

Shuler

M. G.

(2013). A cholinergic mechanism for reward timing within primary visual cortex. Neuron, 77, 723–735.

24.

Craig

D. G.

(2005). An exploratory study of physiological changes during “chills” induced by music. Musicae Scientae, 9, 273–282.

25.

Damasio

Carvalho

G. B.

(2013). The nature of feelings: Evolutionary and neurobiological origins. Nature Reviews Neuroscience, 14, 143–152.

26.

Dickie

(1996). A century of taste. Oxford, England: Oxford University Press.

27.

Di Pelligrino

Margarelli

Mengarelli

(2011). Food pleasantness affects visual selective attention. Quarterly Journal of Experimental Psychology, 64, 560–571.

28.

Dreher

J. C.

Schmidt

P. J.

Kohn

Furman

Rubinow

Berman

K. F.

(2007). Menstrual cycle phase modulates reward-related neural function in women. Proceedings of the National Academy of Sciences of the United States of America, 104, 2465–2470.

29.

Felleman

van Essen

(1991). Distributed hierarchical processing in the primate cerebral cortex. Cerebral Cortex, 1, 1–47.

30.

Fernqvist

Ekelund

(2014). Credence and the effect on consumer liking on food – A review. Food Quality and Preference, 32, 340–353.

31.

Frank

T. C.

Kim

G. L.

Krzemien

Van Vugt

D. A.

(2010). Effect of menstrual cycle phase on corticolimbic brain activation by visual food cues. Brain Research, 1363, 81–92.

32.

Führer

Zysset

Stumvoll

(2008). Brain activity in hunger and satiety: An exploratory visually stimulated fMRI study. Obesity, 16, 945–950.

33.

Fuller

Sherman

Pedersen

N. P.

Saper

C. B.

(2011). Reassessment of the structural basis of the ascending arousal system. Journal of Comparative Neurology, 519, 933–956.

34.

Gavornik

J. P.

Shuler

M. G.

Loewenstein

Bear

M. F.

Shuval

H. Z.

(2009). Learning reward timing in cortex through reward dependent expression of syntactic plasticity. PNAS, 106, 6826–6831.

35.

Goldstein

(1980). Thrills in response to music and other stimuli. Physiological Psychology, 8, 126–129.

36.

Goltstein, P.M., Meijer, G.T., & Pennartz, C.M. (2018). Conditioning sharpens the spatial representation of rewarded stimuli in the mouse primary visual cortex. eLife, 7, e37683.

37.

Gregory

Harris

Heard

Rose

(1995). The artful eye. Oxford, England: Oxford University Press.

38.

Hikosaka

Nakamura

Nakahara

(2006). Basal ganglia orient eyes to reward. Journal of Neurophysiology, 95, 567–584.

39.

Hung

S. M.

Nieh

C. H.

Hsieh

P. J.

(2016). Unconscious processing of facial attractiveness: Invisible attractive faces orient visual attention. Scientific Reports, 6, 37117.

40.

Jacobsen

Schubotz

R. I.

Höfel

von Cramon

D. Y.

(2006). Brain correlates of aesthetic judgment of beauty. NeuroImage, 29, 276–285.

41.

James

(1997). Passion and action. The emotions in seventeenth-century philosophy. Oxford, England: Clarendon Press.

42.

Katsuura

Hexkmann

J. A.

Taha

S. A.

(2011). μ-Opioid receptor stimulation in nucleus accumbens elevates fatty tastant intake by increasing palatability and suppressing satiety signals. American Journal of Physiology, 301, R244–R254.

43.

Kawabata

Zeki

(2004). Neural correlates of beauty. Journal of Neurophysiology, 91, 1699–1705.

44.

Kemp

(1996). Cognitive psychology in the middle ages. Westport, CT and London, England: Greenwood Press.

45.

Kirk

Skov

Hulme

Christensen

M. S.

Zeki

(2009). Modulation of aesthetic value by semantic context: An fMRI study. Neuroimage, 44, 1125–1132.

46.

Kivy

(2003). The seventh sense (2nd ed.). Oxford, England: Clarendon Press.

47.

Koechlin

(2016). Prefrontal executive function and adaptive behavior in complex environments. Current Opinion in Neurobiology, 37, 1–6.

48.

Kringelbach

M. L.

Berridge

K. C.

(2017). The affective core of emotion: Linking pleasure, subjective well-being, and optimal metastability in the brain. Emotion Review, 9, 191–199.

49.

Kringelbach

M. L.

O’Doherty

Rolls

E. T.

Andrews

(2003). Activation of the human orbitofrontal cortex to a liquid food stimulus is correlated with subjective pleasantness. Cerebral Cortex, 13, 1064–1071.

50.

Krishna

(2012). An integrative review of sensory marketing: Engaging the senses to affect perception, judgment, and behavior. Journal of Consumer Psychology, 22, 332–351.

51.

Kühn

Gallinat

(2012). The neural correlates of subjective pleasantness. Neuroimage, 61, 289–294.

52.

Leder

Belke

Oeberst

Augustin

(2004). A model of aesthetic appreciation and aesthetic judgements. British Journal of Psychology, 95, 489–508.

53.

Levy

D. J.

Glimcher

P. W.

(2012). The root of all value: A neural common currency for choice. Current Opinion in Neurobiology, 22, 1027–1038.

54.

Oksama

Hyönä

(2016). How facial attractiveness affects sustained attention. Scandinavian Journal of Psychology, 57, 383–392.

55.

Liu, C. H., & Chen, W. (2012). Beauty is better pursued: Effects of attractiveness in multiple-face tracking. The Quarterly Journal of Experimental Psychology, 65, 553–564.

56.

Livingstone

Hubel

(1988). Segregation of form, color, movement, and depth: Anatomy, physiology, and perception. Science, 240, 740–749.

57.

Loui

Patterson

Sachs

M. E.

Leung

Zeng

Przysinda

(2017). White matter correlates of musical anhedonia: Implications for evolution of music. Frontiers in Psychology, 8, 1–10.

58.

Mallik

Chanda

M. L.

Levitin

D. J.

(2017). Anhedonia to music and mu-opioids: Evidence from the administration of naltrexone. Scientific Reports, 7, 41952.

59.

Martindale

(2001). How does the brain compute aesthetic preference? The General Psychologist, 36, 25–35.

60.

Martindale

Moore

Borkum

(1990). Aesthetic preference: Anomalous findings for Berlyne’s psychobiological theory. American Journal of Psychology, 103, 53–80.

61.

Martínez-Molina

Mas-Herrero

Rodríguez-Fornells

Zatorre

R. J.

Marco-Pallares

(2016). Neural correlates of specific musical anhedonia. Proceedings of the National Academy of Sciences of the United States of America, 113, E7337–E7345.

62.

Mas-Herrero

Zatorre

R. J.

Rodriguez-Fornells

Marco-Pallaes

(2014). Dissociation between musical and monetary reward responses in specific musical anhedonia. Current Biology, 24, 699–704.

63.

Mas-Herrero

E. K.

Marcos-Pallares

Zatorre

R. J.

Rodriguez-Fornells

(2018). The impact of visual art and emotional sounds in specific musical anhedonia. Progress in Brain Research, 237, 399--413

64.

McClure

S. M.

Tomlin

Cypert

K. S.

Montague

L. M.

Montague

P. R.

(2004). Neural correlates of behavioral preference for culturally familiar drinks. Neuron, 44, 379–387.

65.

Montague

P. R.

King-Casas

(2007). Efficient statistics, common currencies and the problem of reward-harvesting. Trends in Cognitive Sciences, 11, 514–519.

66.

Nadal

Munar

Capó

M. A.

Rosselló

Cela-Conde

C. J.

(2008). Towards a framework for the study of the neural correlates of aesthetic preference. Spatial Vision, 21, 379–396.

67.

Nadal

Skov

(2018). The pleasure of art as a matter of fact. Proceedings of the Royal Society B: Biological Sciences, 285, 20172252.

68.

Nakamura

Arai

Kawabata

(2018). Prioritized identification of attractive and romantic partner faces in rapid serial visual presentation. Archives of Sexual Behavior, 46, 2327–2338.

69.

O’Doherty

(2014). The problem with value. Neuroscience and Biobehavioral Reviews, 43, 259–268.

70.

O’Doherty

Cockburn

Pauli

W. M.

(2017). Learning, reward, and decision making. Annual Review in Psychology, 68, 73–100.

71.

O’Hare

(1981). Psychology and the arts. Sussex, England: The Harvester Press.

72.

Okamoto

Dan

(2013). Extrinsic information influence taste and flavor perception: A review from psychological and neuroimaging perspectives. Seminars in Cell & Developmental Biology, 24, 247–255.

73.

Ossewaarde, L., van Wingen, G. A., Kooijman, S. C., Bäckström, T., Fernández, G., & Hermans, E. J. (2010). Changes in functioning of mesolimbic incentive processing circuits during the premenstrual phase. Social Cognitive and Affective Neuroscience, 6, 612–620.

74.

Padoa-Schioppa

(2011). Neurobiology of economic choice: A good-based model. Annual Review in Neuroscience, 34, 333–359.

75.

Palmiter

R. D.

(2007). Is dopamine a physiologically relevant mediator of feeding behavior? Trends in Neuroscience, 30, 375–381.

76.

Peciña

Smith

K. S.

Berridge

K. C.

(2006). Hedonic hot spots in the brain. The Neuroscientist, 12, 500–511.

77.

Peng

Gillis-Smith

Jin

Tränkner

Ryba

N. J.

Zuker

C. S.

(2015). Sweet and bitter taste in the brain of awake behaving animals. Nature, 527, 512–515.

78.

Piqueras-Fiszman

Spence

(2018). Sensory expectations based on product-extrinsic food cues: An interdisciplinary review of the empirical evidence and theoretical accounts. Food Quality and Preference, 40, 165–179.

79.

Plassmann

O'Doherty

Shiv

Rangel

(2008). Marketing actions can modulate neural representations of experienced pleasantness. PNAS, 105, 1050–1054.

80.

Ramsøy

T. Z.

Skov

(2014). Brand preference affects the threshold for perceptual awareness. Journal of Consumer Behavior, 13, 1–8.

81.

Rangel

Camerer

Montague

P. R.

(2008). A framework for studying the neurobiology of value-based decision making. Nature Reviews Neuroscience, 9, 545–556.

82.

Reed

E. S.

(1997). From soul to mind. The emergence of psychology from Erasmus Darwin to William James. New Haven, CT and London, England: Yale University Press.

83.

Rupp

H. A.

James

T. W.

Ketterson

E. D.

Sengelaub

D. R.

Janssen

Heiman

J. R.

(2009). Neural activation in the orbitofrontal cortex in response to male faces increases during the follicular phase. Hormones and Behavior, 56, 66–72.

84.

Rushworth

F. S.

Mars

R. B.

Summerfield

(2009). General mechanisms for decision making? Current Opinion in Neurobiology, 19, 1–9.

85.

Sachs

M. E.

Ellis

R. J.

Schlaug

Loui

(2016). Brain connectivity reflects human aesthetic responses to music. Social Cognitive and Affective Neuroscience, 6, 884–891.

86.

Salimpoor

V. N.

Benovoy

Longo

Cooperstock

J. R.

Zatorre

R. J.

(2009). The rewarding aspects of music listening are related to degree of emotional arousal. PLoS One, 4, e7487.

87.

Schindler

Hosoya

Menninghaus

Beermann

Wagner

Eid

Scherer

K. R.

(2017). Measuring aesthetic emotions: A review of the literature and a new assessment tool. PLoS One, 12, e0178899.

88.

Schmidt

Skvortsova

Kullen

Weber

Plassmann

(2017). How context alters value: The brain’s valuation and affective regulation system link price cues to experienced taste pleasantness. Scientific Reports, 7, 8098.

89.

Schur

E. A.

Kleinhans

N. M.

Goldberg

Buchwald

Schwartz

M. W.

Maravilla

(2009). Activation in brain energy regulation and reward centers by food cues varies with choice of visual stimulus. International Journal of Obesity, 33, 653–661.

90.

Serences

J. T.

(2008). Value-based modulations in human visual cortex. Neuron, 60, 1169–1181.

91.

Sescousse

Caldú

Segura

Dreher

J.-C.

(2013). Processing of primary and secondary rewards: A quantitative meta-analysis and review of human functional neuroimaging studies. Neuroscience and Biobehavioral Reviews, 37, 681–696.

92.

Shuler

M. G.

Bear

M. F.

(2006). Reward timing in the primary visual cortex. Science, 311, 1606–1609.

93.

Siep, N., Roefs, A., Roebroeck, A., Havermans, R., Bonte, M. L., & Jansen, A. (2009). Hunger is best spice: An fMRI study of the effects of attention, hunger and calorie content on food reward processing in the amygdala and the orbitofrontal cortex. Behavioural Brain Research, 198, 149–158.

94.

Silvia

P. J.

(2005). Emotional responses to art: From collation and arousal to cognition and emotion. Review of General Psychology, 9, 342–357.

95.

Skov

( in press). The neurobiology of sensory valuation. In M. Nadal & O. Vartanian (Eds.), The Oxford handbook of empirical aesthetics. Oxford, England: Oxford University Press.

96.

Skov

Nadal

(2018). Art is not special: An assault on the last lines of defense against the naturalization of the human mind. Reviews in the Neurosciences, 29, 699–702.

97.

Skov

Nadal

(2019). What is the nature of the perceptional and emotional processes involved in aesthetic experience? A response to Makin’s challenge. Manuscript submitted for publication.

98.

Small

D. M.

Zatorre

R. J.

Dagher

Evans

A. C.

Jones-Gotman

(2001). Changes in brain activity related to eating chocolate. From pleasure to aversion. Brain, 124, 1720–1733.

99.

Smith

K. S.

Berridge

K. C.

(2005). The ventral pallidum and hedonic reward: Neurochemical maps of sucrose “liking” and food intake. Journal of Neuroscience, 25, 8637–8649.

100.

Smith

K. S.

Berridge

K. C.

(2007). Opioid limbic circuit for reward: Interaction between hedonic hotspots of nucleus accumbens and ventral pallidum. Journal of Neuroscience, 27, 1594–1605.

101.

Solso

R. L.

(1994). Cognition and the visual arts. Cambridge, MA and London, England: The MIT Press.

102.

Sui

Liu

C. H.

(2009). Can beauty be ignored? Effects of facial attractiveness on covert attention. Psychonomic Bulletin and Review, 16(2), 276–281.

103.

Summerfield, C., & de Lange, F. P. (2014). Expectation in perceptual decision making: Neural and computational mechanisms. Nature Reviews Neuroscience, 15, 745–756.

104.

Swanson

L. W.

(2000). Cerebral hemisphere regulation of motivated behavior. Brain Research, 886, 113–164.

105.

Thomas

J. M.

Higgs

Dourish

C. T.

Hansen

P. C.

Harmer

C. J.

McCabe

(2015). Satiation attenuates BOLD activity in brain regions involved in reward and increases activity in dorsolateral prefrontal cortex: An fMRI study in healthy volunteers. American Journal of Clinical Nutrition, 101, 697–704.

106.

Tindell

A. J.

Smith

K. S.

Peciña

Berridge

K. C.

Aldridge

J. W.

(2006). Ventral pallidum firing codes hedonic reward when a bad taste turns good. Journal of Neurophysiology, 96, 2399–2409.

107.

Valuch

Pflüger

L. S.

Wallner

Laeng

Ansorge

(2015). Using eye tracking to test for individual differences in attention to attractive faces. Frontiers in Psychology, 6, 42.

108.

Vartanian

Goel

(2004a). Neuroanatomical correlates of aesthetic preference for paintings. NeuroReport, 15, 893–897.

109.

Vartanian

Goel

(2004b). Emotion pathways in the brain mediate aesthetic preference. Bulletin of Psychology and the Arts, 5(1), 37–42.

110.

Vartanian

Nadal

(2007). A biological approach to a model of aesthetic experience. In L. Dorfman, C. Martindale, & V. Petrov (Eds.), Aesthetics and innovation (pp. 429--444). Cambridge, England: Cambridge Scholars Press.

111.

Wessel

L. P.

Jr. (1972). Alexander Baumgarten’s contribution to the development of aesthetics. The Journal of Aesthetics and Art Criticism, 30, 333–342.

112.

Zimmerman

C. A.

Leib

D. E.

Knight

Z. A.

(2017). Neual circuits underlying thirst and fluid homeostasis. Nature Reviews Neuroscience, 18, 459–469.

113.

Yoest

K. E.

Quigley

J. A.

Becker

J. B.

(2018). Rapid effects of ovarian hormones in dorsal striatum and nucleus accumbens. Hormones and Behavior, 98, 210–218.

114.

Zold

C. L.

Shuler

M. G.

(2015). Theta oscillations in visual cortex emerge with experience to convey expected reward time and experienced reward rate. Journal of Neuroscience, 35, 9603–9614.

115.

Zou

L. Q.

van Hartvelt

T. J.

Kringelbach

M. L.

Cheung

E. F.

Chan

R. C.

(2016). The neural mechanism of hedonic processing and judgment of pleasant odors: An activation likelihood estimation meta-analysis. Neuropsychology, 30, 970–979.