For at least several centuries, in the West, the artistic phenomenon has presented itself as follows: a person (the artist), signs a particular object or message (the work), which other persons (the recipients, the public, the critics) perceive, taste, read, interpret, and evaluate […] The techno-cultural environment that is emerging, however, gives rise to new art forms, ignoring the distinction between emission and reception, creation, and interpretation […] This new art form allows what is precisely no longer an audience to experience other methods of communication and creation. (Levy, 1996, p. 336).

Emotions, Cognition, and Aesthetic Experience

The exploration of aesthetic and emotional reactions to artworks spans an extensive historical journey, from Plato and Aristotle’s theories to Arthur Schopenhauer’s (1788–1860) and Immanuel Kant’s (1724–1804) analyses on the perception of beauty. An interesting contemporary approach to this topic is represented by neuroaesthetic studies aiming to identify the neural correlates associated with aesthetic experiences and the cognitive processes activated when individuals engage with art, music, literature, or other aesthetic stimuli. These studies usually consider how visual perception is influenced by various factors, including bottom-up and top-down processes. Bottom-up processing refers to the analysis of sensory information starting from basic features and building up to more complex interpretations. In the context of art, this involves the initial analysis of visual elements, such as lines, shapes, and colors, before forming a holistic perception of the artwork. Top-down processing, on the other hand, involves using prior knowledge, expectations, and cognitive factors to guide and influence perception. When viewing art, our past experiences, cultural background, and acquired knowledge about artistic styles and conventions can shape how we interpret and appreciate the artwork.

By studying how the brain processes visual stimuli and how this processing relates to subjective aesthetic experiences, neuroaesthetics researchers aim to gain insights into the complex interplay between neural mechanisms, perception, and art appreciation. In this essay, I will focus in particular on neuroaesthetics studies focused on how the brain perceives and processes aesthetic stimuli in relation to visual artworks, and in particular to CGI-generated creative content, such as images and videos.

A relevant starting point is the Neural Correlates of Beauty theory, which investigates the cognitive and neural mechanisms involved in the perception of beauty and other aesthetic qualities. This theory proposes that certain aesthetic qualities or features, such as symmetry, complexity, harmony, and novelty, activate particular brain areas or networks, leading to the subjective experience of beauty.Footnote 1 However, it’s important to note that this theory is still a subject of ongoing research and debate, and there is no universally agreed-upon set of neural correlates that define beauty. Different individuals and cultures may have different perceptions of beauty, and there may be variations in brain responses accordingly.

Neurobiologists such as Anjan Chatterjee, Edward Vessel, and Semir Zeki have conducted extensive research on this topic, investigating how specific brain regions, particularly in the visual cortex and the reward system, are involved in aesthetic judgments and preferences (Chatterjee & Cardillo, 2022; Kawabata & Zeki, 2004; Vessel et al., 2022). Leading tests combining neuroimaging techniques and functional magnetic resonance imaging (fMRI), they studied the brain’s response to stimuli perceived as beautiful, observing that when individuals view aesthetically pleasing stimuli, such as artworks, landscapes, or faces, specific brain regions related to reward, emotion, and visual processing are often activated.

However, visual stimuli work differently for different individuals, and recent neuroscientic studies have considered the role of aesthetic preferences in the appreciation of artworks, exploring how each individual appreciates different visual artworks and the emotional responses generated by the encounter with these artworks. V.S. Ramachandran has explored how mirror neurons (which are active both when we perform an action and when we observe someone else performing the same action) play a role in empathy and aesthetic appreciation (Ramachandran, 2012). Along the same lines, cognitive neuroscientist Oshin Vartanian has conducted research on the neural mechanisms underlying aesthetic experiences and emotional responses to art in order to analyze how the brain processes and evaluates different aesthetic qualities, such as beauty and novelty (Jung & Vartania, 2018). Helmut Leder et al. (2004) have examined emotional responses to aesthetic appreciation and aesthetic judgments of specific artworks, exploring the reasons why modern art, with its multitude of individualized styles, innovation, and conceptual nature, offers enriching aesthetic experiences.

These studies indicate that the judgement of different aesthetic features of artworks by the human brain involves a complex interplay of cognitive processes, emotional responses, and sensory perception. While the precise mechanisms are still being investigated, several factors have been identified as key contributors to how the brain evaluates aesthetic features in art, in particular visual elements such as color, shape, texture, and composition. Sensory processing areas, like the visual cortex, extract and process this information, allowing the brain to perceive and differentiate artworks’ various aesthetic qualities.

The Beauty of Art—Paintings and Visual Composition

The evaluation of aesthetic features in art is highly subjective and can vary among individuals due to personal preferences, cultural influences, and individual differences in neural processing. As we have mentioned, an important element in the aesthetics evaluation of an artwork are emotions. The brain’s emotional centers, such as the amygdala and insula, play a role in assessing the emotional valence and intensity of an artwork. Positive emotional experiences, such as feelings of beauty, awe, or joy, can enhance the evaluation of aesthetic features and contribute to a positive aesthetic experience, and vice versa.

Researchers such as Kent Berridge (1947) and Antonio Damasio (2006) have established how brain’s reward and pleasure systems—including the ventral striatum and orbitofrontal cortex—are engaged during the evaluation of aesthetic features. Activation in these regions indicates the brain’s reward response to visually pleasing stimuli.

Cognitive processes, including attention, memory, and categorization, influence the evaluation of aesthetic features. Attentional mechanisms direct focus to specific visual elements or features, while memory and categorization processes allow the brain to compare and contrast artworks based on previous experiences and learned aesthetic norms. Aesthetic features that are perceived as novel, complex, or harmonious often elicit greater reward responses, leading to a more positive evaluation. Expertise, gained through artistic training or exposure to art, can enhance these cognitive processes and refine the evaluation of aesthetic features.

If we consider a specific case study, i.e., the visual perception of paintings, several key elements related to the aesthetic judgment of an artwork come into play, in particular:

  • Composition: The arrangement of visual elements within a painting can influence how our brain perceives and processes the artwork. Certain compositional principles, such as the rule of thirds or the golden ratio, can create a sense of balance and harmony in paintings, leading to positive aesthetic experiences.

  • Color and emotion: Colors can have a direct impact on our perception of a painting. Different colors activate specific areas of the brain associated with emotions. For example, warm colors like red and orange may elicit feelings of excitement or energy, while cool colors like blue and green can evoke a sense of calmness or serenity.

  • Visual attention: Paintings can guide our visual attention and direct our gaze to specific areas of interest. Artists strategically use visual cues, such as lines, contrast, and focal points, to attract attention and engage viewers. These techniques can influence the way our brain processes and interprets the artwork.

Recent studies using brain imaging techniques have identified neural correlates associated with subjective aesthetic judgments. Cupchik et al. (2009) investigated the neural responses to different art styles by examining brain activity using fMRI. Participants were shown paintings from various artistic movements, such as Impressionism, Cubism, and Abstract Expressionism. The findings revealed distinct patterns of brain activation associated with the perception of each art style, suggesting that different styles of painting elicit specific neural responses. For example, representational paintings with realistic depictions of objects tended to activate brain regions involved in object recognition, while abstract paintings with ambiguous or non-representational forms engaged areas associated with visual abstraction and interpretation.

Another interesting example is represented by “Dynamics of brain networks in the aesthetic appreciation” by Cela-Conde et al. (2013). In this study, participants were exposed to classical artworks and their brain activity was measured. The results revealed that viewing paintings activated brain areas related to visual processing, emotion, reward, and memory. The researchers found that lines and edges in the paintings activated the brain regions responsible for processing visual information and spatial attention. Focal points, or areas of the painting with high contrast or salience, were associated with increased activity in the visual cortex and prefrontal cortex, indicating that these areas drew viewers’ attention and engaged their cognitive processing.

Overall, these studies provide evidence that viewing paintings perceived as beautiful led to increased activity in the medial orbitofrontal cortex, a region associated with reward and aesthetic pleasure, suggesting a neural basis for the perception of beauty in paintings. But what happens when we apply this neuroscientific approach to the study of the cognitive and emotional perception of CGI-generated visual content?

Perceptual Responses to CGI Artworks

Let’s consider Susanne Langer’s theory of aesthetic experience based on symbolic forms, emphasizing the role of emotions in aesthetic responses (Langer, 1949). This theory revolves around the idea that human beings have a fundamental need to create and engage with symbolic forms, i.e., expressive structures or patterns that embody human feelings, experiences, and ideas. They include not only traditional art forms like music, visual art, and literature but also more everyday symbolic activities such as rituals, gestures, and language itself. According to Langer, when we engage with symbolic forms, we enter the aesthetic experience—a state of heightened awareness—and undergo an aesthetic transformation. I propose to combine Langer’s theory with Dominic McIver Lopes’ views of aesthetic perception, specifically regarding computer-generated imagery (CGI) (2010). In his book, A Philosophy of Computer Art, Lopes explores the aesthetic potential of computer-generated art, analyzing in particular how the nature of CGI, its use of algorithms, and its capacity for creating complex and imaginative visuals can lead to new aesthetic experiences. He considers the role of perception, interpretation, and engagement with CGI artworks, challenging conventional notions of authenticity, originality, and representation in art. According to his point of view, CGI should be approached with a new set of aesthetic criteria that are suited to its unique characteristics, rather than evaluated solely on the basis of traditional artistic standards:

Computer art works invite and indeed prescribe repeat encounters. Users expect something new with each interaction and are attuned to the differences between the displays they generate. Through many interactions and displays, they come to see the possibilities the work holds for them. (Lopes, 2010, p. 60).

Building on Lopes’s analysis, let’s consider how CGI’s ability to generate virtual worlds and characters leads to questions about the ontological status of the represented subjects and how our perceptual engagement with these images impacts our emotional and cognitive responses. Probably the most famous example of artwork made by an AI in recent years is “Portrait of Edmond de Belamy” by Paris-based arts collective Obvious (Alleyne, 2018), the first AI-generated artwork to be auctioned at a major art house.

Created using a Generative Adversarial Network (GAN), the piece depicts a fictional eighteenth-century portrait and sparked debates about the role of AI in the art world. Another famous example is “The Next Rembrandt” by ING and Microsoft. In this project, AI algorithms analyzed data from Rembrandt’s paintings to create a new artwork in the style of the renowned Dutch artist. The resulting piece demonstrates the ability of AI to emulate artistic styles and create new works that resonate with historical art (see www.nextrembrandt.com). More recently, the production of AI-generated images by software like DeepDream and StyleGAN sparked debate. Using deep neural networks, these softwares generate pictures characterized by surreal and psychedelic aesthetic, with viewers interpreting them in diverse ways, ranging from mesmerizing and captivating to disorienting or unsettling (Mordvintsev et al., 2015).

Academics and theorists are focusing on the psychological and cognitive outcomes of the perceptive encounters between the viewers and these artworks. For example, The Computational Creativity Research Group at Goldsmiths, University of London, explores various aspects of aesthetics, perception, and evaluation of computer-generated artwork. By integrating AI techniques and computational approaches, this research group aims to understand and simulate human-like creativity in machines, to explore how acts of creativity, traditionally considered uniquely human, are perceived when produced by artificial intelligence. Analyzing in detail the observers’ ability to distinguish between computer-generated and human-made art, and exploring how categorizing artworks influence their perceived aesthetic value (Chamberlain et al., 2017), this research not only emphasizes the technical aspects of generating art but also examines the philosophical and ethical implications of AI’s role in creative practices.

Perception of Hyperrealistic Digital CGI

Hyperrealistic digital avatars are computer-generated characters that closely resemble real individuals. They are created using advanced computer graphics techniques (such as Dall-E or Metahuman softwares) and aim to achieve a high level of realism in their appearance and behavior. Neuroaesthetics studies have recently delved into the perception of hyperrealistic digital avatars by the human brain, and as we analyze the realm of realistic-looking humanoids, it’s imperative to initially confront the phenomenon of the Uncanny Valley. This perceptual response arises when humanoid objects, like robots or computer-generated characters, bear an uncanny resemblance to human beings but fall just short of being convincingly human.

The concept suggests that as the appearance of a humanoid object becomes more human-like, there is a corresponding increase in our affinity or positive emotional response towards it. However, there is a critical point at which the object’s resemblance to a human reaches a certain level of similarity, but not close enough to be indistinguishable. At this point, instead of evoking positive emotions, the object triggers negative emotions, leading to a dip in the emotional response curve. This dip represents the uncanny valley.

The concept was first theorized by Freud in 1919, in the framework of a study on the breaking of boundaries between fact and fiction (Freud, 2018). He described the Uncanny as the feeling that arises when someone has an intellectual uncertainty about the livingness of a certain being. The concept was further explored by Japanese robotics professor Masahiro Mori in the 1970s, with regard to the relation between an object’s degree of resemblance to a human being and the human emotional response to that same object (Mori, 2005).

Notable examples of realistic looking humanoids that often generate such an uneasy feeling in the individuals with whom they interact are Sophia—a humanoid robot Developed by Hanson Robotics known for her realistic appearance and human-like expressions—and the Geminoids, a series of humanoid robots developed by Japanese roboticist Hiroshi Ishiguro designed to resemble specific individuals and exhibit lifelike facial movements and expressions. When an object falls into the uncanny valley, people often experience a sense of discomfort or revulsion, or a feeling of eeriness and uneasiness. The object’s subtle deviations from human appearance, such as unnatural facial features, strange movement patterns, or lack of appropriate emotional expressions, contribute to this effect.

One possible explanation for the uncanny valley effect is that humans have a heightened sensitivity to detecting subtle deviations from the norm in human appearance and behavior. Our brains are wired to recognize and respond to familiar human characteristics, but when confronted with an entity that closely resembles a human but exhibits uncanny deviations, our perceptual system detects the mismatch, triggering a negative emotional response. Overcoming this effect is a significant challenge in robotics and animation. While there is no definitive solution to this day, several approaches are being tested by engineers, in particular emotional design. Focusing on creating robots or characters that evoke positive emotions and empathy can help bridge the gap. By emphasizing traits such as kindness, humor, or cuteness, designers can make the robot more relatable and endearing. Furthermore, recognizing that different people have varying thresholds for the Uncanny Valley effect, designers can offer customization options. Allowing users to personalize the appearance or behavior of the humanoid robot can help increase acceptance and reduce discomfort. As technology progresses, we can expect further improvements in creating humanoid robots and characters that are more believable and comfortable to interact with.

The Outcomes of Realistic-Looking Avatars

In recent years human interaction with digital avatars has become increasingly prevalent in various domains, ranging from entertainment to education, from customer service to health and therapy. Sherry Turkle, a professor of social studies of science and technology, has written extensively on the impact of such a technology on individual cognitive and emotional spheres (Turkle, 2012). According to her “second self” theory,      digital avatars serve as a representation or extension of an individual’s identity in virtual environments. When people interact with them, they often project parts of their personalities, desires, and aspirations onto these virtual representations. Digital avatars become a ‘second self’ that individuals use to explore and express different aspects of their identity, sometimes in ways that differ from their offline persona.

Turkle suggests that the relationship between individuals and their digital avatars is a form of self-presentation and self-exploration. Through these avatars, individuals can experiment with different identities and engage in activities that they may not feel comfortable doing in their real lives. For some, digital avatars become an opportunity to embody idealized versions of themselves or explore fantasies and experiences that are not readily available offline. Turkle’s theory also addresses the potential impact of these interactions on social relationships. She argues that while digital avatars provide opportunities for self-expression and connection, they can also lead to a sense of disconnection and superficiality. In virtual environments, individuals may find it easier to present themselves in curated ways, hiding certain aspects of their real identity or adopting idealized versions of themselves.

Further research in this field has shown that when people embody attractive avatars, they tend to display more confident and outgoing behavior compared to when they embody less attractive avatars. This theory, known as the Proteus Effect, suggests that people’s cognition and behavior can be influenced by the visual representation of their digital avatars. This cognitive identification can lead to increased self-esteem, self-efficacy, and feelings of ownership over the avatar. In particular, Nick Yee and Jeremy Bailenson examined the influence of avatar customization on users’ self-perception and behavior in virtual environments (Yee & Bailenson, 2007). During a test study participants were assigned avatars with either attractive or unattractive features and those who had attractive avatars displayed more confident and extroverted behaviors, indicating a stronger identification with their avatars.

Such an approach to the digital domain might lead to the potential consequences of relying heavily on digital avatars for social interactions, suggesting that prolonged engagement with digital worlds may have as a consequence a diminished ability to engage in face-to-face interactions and develop empathy with others. In particular, since hyperrealistic avatars often aim to evoke emotional engagement from viewers, neuroaesthetics researchers are now investigating the neural mechanisms underlying our ability to attribute mental states, emotions, and intentions to these virtual characters.

Dr. Robin R. Murphy, a leading figure in the field of human–robot interaction and artificial intelligence, has explored what is known as the simulation theory, suggesting that humans tend to project human-like characteristics onto AI systems, attributing intentions, emotions, and social behaviors to them (Murphy, 2004). According to Murphy,  such an anthropomorphic tendency allows us to project our own emotional responses onto AI systems. We may empathize, feel attachment, or even develop a sense of social connection with AI based on our simulations of their internal experiences.

In short, when viewing aesthetically pleasing avatars, the brain’s reward and pleasure systems may be activated, leading to feelings of enjoyment, admiration, or attraction. When we see a realistic avatar expressing specific emotions, our brains have a tendency to mimic or mirror those emotions. For example, if an avatar displays happiness or sadness, it may evoke corresponding emotional responses in the viewer. Additionally, when we interact with an avatar that closely resembles a real person, we may develop a sense of familiarity, connection, or even attachment to it.

Conclusion

In recent years, neuroscientific studies have uncovered fascinating findings regarding the cognitive and emotional processes involved in art perception, highlighting the complex interplay between bottom-up sensory processing and top-down cognitive mechanisms and revealing how our visual system combines perceptual information with prior knowledge and expectations to construct meaning from artistic stimuli.

If we consider AI-generated art, we discover how CGI often presents novel visual aesthetics and unique artistic styles that may challenge traditional cognitive and sensorial engagement with art. The perceptual responses to CGI artworks have demonstrated a complex interplay of pros and cons, offering both exciting possibilities and potential drawbacks. On the positive side, CGI allows for unparalleled creativity, enabling artists and filmmakers to bring their visions to life in ways previously unimaginable. The ability to generate hyper-realistic scenes and characters has enriched the realms of entertainment, design, and scientific visualization. The perception of these artworks can evoke a sense of curiosity, fascination, or even surprise, as viewers encounter new and unexpected visual forms. Additionally, the emergence of realistic avatars has proven to be a powerful tool for communication and self-expression.However, alongside these remarkable advancements come inherent challenges. Users who become closely identified with their avatars risk to dissociate from reality, and this in turn can lead to detrimental emotional and cognitive effects, such as addiction and social isolation. The uncanny valley effect remains a significant concern, as the quest for perfect realism in CGI can sometimes lead to eerie and unsettling perceptual responses. This can create emotional and cognitive dissonance, hindering the immersive experience that CGI seeks to achieve. As technology continues to evolve, addressing these challenges and striking the right balance between realism and likability will be crucial to ensure that users can fully harness the benefits of CGI and realistic avatars while mitigating their negative impacts.