The Subject(s) at the Centre of Aesthetic Experiences

by Giacomo Bignardi, Kirren Chana, MacKenzie Trupp, & Sasha Koushk-Jalali

Reflections on Edward Vessel’s Introduction to Visual Neuroaesthetics

On 15th November 2018, MSc Music, Mind and Brain and MSc Psychology of the Arts, Neuroaesthetics and Creativity students had the pleasure of having Edward Vessel discuss the nature of visual aesthetic experiences and their neural correlates.

Edward Vessel Neuroaesthetics

Figure 1. Edward Vessel is a neuroscientist studying the neural basis of aesthetic experiences and the neurobiology of information foraging at the Max Planck Institute for Empirical Aesthetics.

Just as philosopher G. Santayana (1995) defines ‘beauty’ as the pleasure evoked by an object, and not the object itself, Vessel studies aesthetic subject matter not based on the particularities of the stimuli but rather on subjective responses to them. Contrary to some philosophers, Vessel believes aesthetics should also be studied from a scientific perspective. According to him, ‘aesthetic appreciation represents a fundamental way of interacting with the world.’

Vessel has not set out to use brain imaging to define what is beautiful nor what art is, but instead stresses a distinction between external objective stimuli and internal subjective factors. Features of our visual world have less importance than the perceiver’s reaction to them. This focus of low-level (visual) vs higher-level (emotionally driven, or semantic) factors presents difficulties for aesthetic scientific research. To address this, Vessel stresses the importance of individual differences in aesthetic experiences, and rejects the use of a simple average as a tool to measure aesthetic preference. By definition, when one averages ratings of liking, one loses information about subjective preferences.

Pairwise Correlation, Mean Minus One, and the Central Role of Meaning

The pairwise cross-correlation distribution is a tool used to measure agreement. A pairwise cross-correlation is calculated using the correlations of ratings of all possible pairs.

cross-observer pairwise correlation
Figure 2. Vertical axis represents the number of pairs (people!). Horizontal axis represents the pairwise correlation values. The vertical line represents the average value for different distributions. The agreement bar on the left is a visual representation of how agreement changes as a function of the distribution. Images were computed in R-studio Version 1.1.419. The images were then post edited.

If the distribution is centred around 0, people had low agreement (i.e. low pairwise correlations), while if its centre is shifted toward higher numbers, we can say that people agree more.

What did Vessel do with this tool? He showed agreement on perceived beauty of images of abstract compositions is lower than those for images of real-world scenes. That is, people tend to agree when rating the perceived beauty of images of sceneries, but disagree for images that are abstract (Vessel & Rubin, 2010).

Individual preferences (Vessel & Rubin, 2010)
Figure 3. Agreement on ratings of abstract images (left) is lower than for images of real-world scenes (right). Source: Vessel & Rubin (2010).

This led Vessel to hypothesise that there is a central role of meaning (semantic) in aesthetic experiences: our experiences are generalisable by the degree to which we have shared semantic interpretations. If people interpret an object in the same way, their aesthetic reaction to it will probably be similar. That is, according to Vessel:

“Shared semantics leads to shared preferences”

Another powerful statistical tool that Vessel utilised to measure agreement, and once more the role of meaning, is Mean Minus One (MM1). This specifically measures shared and unique variance amongst participants’ ratings, and quantifies this to a value between 0 and 1. This is indicative of how much people agree between themselves, with 1 being a perfect match (100% of the variance is shared) and 0 being no agreement (0% of the variance is shared). Imagine collecting n x m ratings, where n = number of subjects and m = number of objects. This is roughly how MM1 is computed:

Mean Minus One (MM1)
Figure 4. Every column represents one participant, while every row represents one object. The numbers represent the measure of interests (ranging from 1 to 7 on a Likert scale measurement). Steps to compute MM1: 1) Compute the mean of the scores for object 1(first row) for every subject except for subject one (in first column) and iterate the process for m objects (row); 2) Compute the correlation between the ratings of subject 1 and the computed means; 3) Iterate the process, but excluding one new subject every time (e.g. Subject 2, 3… n); 4) Convert the correlation scores to z scores; 5) Compute the average of the z scores; 6) Convert z – to – r again.

Neuroaesthetics: from measure to brain correlates of subjective Aesthetic experiences

Vessel has stated that, broadly speaking, aesthetic experiences have higher-level (semantic) and lower-level (purely visual) aspects, with one level perhaps contributing more towards people’s overall reactions than the other. If this is the case, then are there measurable neural correlates associated with one or both of these states?

Functional Magnetic Resonance Imaging (fMRI) is a technique widely used to research neural correlates in neuroscience. This methodology allows neuroscientists to establish a neurological correlation of activated brain area and a specified task (such as aesthetic rating). fMRI studies suggest beauty, regardless of its modality source, is processed by the medial prefrontal cortex (mPFC) (Kawabata & Zeki, 2004; Isizhu & Zeki, 2011). However, Vessel does not seem convinced. Firstly, the mPFC processes several experiences beyond beauty, such as subjective value (Kable & Glimcher, 2007) and social cognition (Amodio & Frith, 2006), suggesting the mPFC may not be specific to processing beauty. Moreover, looking specifically at perceived beauty might narrow the broader question regarding aesthetic experiences.

Thus, Vessel shifted the focus from ‘how beautiful‘ to ‘how moving‘ images were with the following instruction:

“…Respond on the basis of how much this images ‘moves’ you… what works you find powerful, pleasing, or profound”

(Vessel, Starr & Rubin, 2012, p.3)
Figure 5. The Nightmare, 1781. Henry Fuseli. Hindola Raga, c. (1790–1800), Pahari Hills, Kangra school. An Ecclesiastic, c. 1874. Mariano José Maria Bernardo Fortuny y Carbo Constant, c. 1988. Valerie Jaudon.

Participants were instructed to translate the degree to which they were “moved” on a scale from 1 – 4 while in the fMRI scanner. Stimuli were photographs of paintings from a variety of artists, allowing individual preferences to emerge (we know this because MM1 results show low agreement across participants).

Occipito-temporal regions of the brain were found to be linearly related to ratings (more ‘moved by’ = more brain activation). Additionally, a network of anterior brain regions were activated only by images considered most ‘moving’ (rated as 4). Some of these specific locations are important hubs in the Default Mode Network (DMN) (Vessel, Starr & Rubin, 2012; 2013).

Art reaches within: aesthetic experience, the self and the default mode network (Vessel, E., Star, G., G & Rubin, N., 2013)
Figure 6. “Distinct patterns of response to artworks as a function of their ratings in a distributed network of brain regions” (Vessel, Starr, & Rubin, 2013). Image links to the paper.

Occipito-temporal regions in the brain process external stimuli, such as visual stimuli. Anterior regions process higher cognitive function, such as internally generated thoughts and emotions. The DMN is more active during non-task periods, and it is one of the brain networks associated with internally-oriented cognition (click here to read more about the DMN). When we are ‘moved by’ art, the network that processes external stimuli (Occipito-temporal) is engaged simultaneously with the network that processes internally oriented events (anterior/DMN). Vessel suggests that intense “aesthetic experience involve(s) the integration of sensory and emotional reactions in a manner linked with … personal relevance” (Vessel et al., 2012, p.1). Thus, it seems that ‘being moved’, which is an access point to aesthetic experiences, might be categorically different from other experiences.

For Vessel, this is only the beginning, as studying the aesthetic experiences of the subject rather than the subject matter will give a central role to meaning. Aesthetic experiences, aside from beauty alone, seem to constitute moments where the external world is meaningfully integrated by the internal one, leaving the individual moved.

35148f_8691ccb672274ecc85fdbd9f38b14700~mv2.gifFigure 7. Aesthetic Experiences = Boom. It seems that after a certain threshold is reached, aesthetic magic happens. Image: fuse* – Multiverse”. Courtesy of FUSE.


An Ecclesiastic, c. 1874. Mariano José Maria Bernardo Fortuny y Carbo Retrieved from

Amodio, D. M., & Frith, C. D. (2006). Meeting of minds: the medial frontal cortex and social cognition. Nature reviews neuroscience, 7(4), 268.

Hindola Raga, c. (1790–1800), Pahari Hills, Kangra school Retrieved from

Ishizu, T., & Zeki, S. (2011). Toward a brain-based theory of beauty. PloS one, 6(7), e21852.

Kable, J. W., & Glimcher, P. W. (2007). The neural correlates of subjective value during intertemporal choice. Nature neuroscience, 10(12), 1625.

Kawabata, H., & Zeki, S. (2004). Neural correlates of beauty. Journal of neurophysiology, 91(4), 1699-1705.

Mariano José Maria Bernardo Fortuny y Carbo Constant, c. 1988. Valerie Jaudon. Retrieved from

Multiverse. (2018). [Gif]. Retrieved from

Santayana, G. (1995). The sense of beauty: Being the Outlines of an Aesthetic Theory. New York: Dover (Original work published 1896).

The Nightmare, 1781. Henry Fuseli Retrieved from

Vessel, E. A. (2018). [Photograph]. Retrieved from

Vessel, E. A., & Rubin, N. (2010). Beauty and the beholder: highly individual taste for abstract, but not real-world images. Journal of vision, 10(2), 18-18.

Vessel, E. A., Starr, G. G., & Rubin, N. (2012). The brain on art: intense aesthetic experience activates the default mode network. Frontiers in human neuroscience, 6, 66.

Vessel, E. A., Starr, G. G., & Rubin, N. (2013). Art reaches within: aesthetic experience, the self and the default mode network. Frontiers in Neuroscience, 7, 258.

Zabelina, D. L., & Andrews-Hanna, J. R. (2016). Dynamic network interactions supporting internally-oriented cognition. Current opinion in neurobiology, 40, 86-93.


This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s