The Sussex Baby Lab, in the School of Psychology at the University of Sussex, recently conducted a study where the research team showed Dutch post-impressionist, Vincent van Gogh’s landscape paintings to baby and adult participants. Recording how long the babies engaged with the paintings, and asking the adults which painting they found the most pleasing, the researchers were able to gain a deeper understanding of how our visual preferences develop from an early age. In this interview, Philip McAdams, Doctoral Researcher at the University of Sussex, explains the project and it’s outcomes.
Above image: The Starry Night, 1889, Vincent van Gogh, Oil on canvas, 73.7 × 92.1 cm | 29.01 × 36.26 in, Museum of Modern Art, New York
How Early Do We Develop Visual Preferences in Art?
An interview with Philip McAdams, Doctoral Researcher at the University of Sussex
The following images show the paintings preferred by adults on the left, and infants results on the right, in order of most favoured, 1 – to least favoured, 40. All images are oil on canvas by Vincent van Gogh, and have been cropped as they were for the subjects in the research experiment.
Clare: What were the main lines of inquiry at the start of this research project?
Philip: We wanted to investigate the extent to which adults’ aesthetic preferences could be traced back to babies’ visual preferences. Investigating babies’ visual preferences to art, can also provide insight into baby vision and visual development.
Clare: What led the team to decide to work with van Gogh’s paintings?
Philip: van Gogh is considered by many to be the most famous and liked artists of all time, so a good artist to study for preference. Furthermore, van Gogh is known for his use of vibrant colours which are saturated enough for babies to see.
Clare: Why are all the paintings cropped to the same dimensions?
Philip: To carefully control the experiment we centre cropped all the paintings, given that they all had various dimensions. That way there is no variation in size which may affect preference. Furthermore, to analyse the images for their visual features, some of the algorithms we use require an image to be square.
Clare: Can you tell us about the participants you recruited for the study? How many were there, how old were the babies? What age range were the adults?
Philip: We recruited 25 babies’ (4-8 months old) and 25 adults (18-40). We used statistical techniques to ensure that we had a powerful enough study to make inferences.
Clare: How long did the baby participants engage with the most preferred painting, in comparison to the least?
Philip: On average, babies engaged 57% more with their most preferred paintings in comparison to their least preferred paintings.
Clare: From the study, what colour would you say babies enjoy the most? And what colour do adults enjoy the most?
Philip: Babies showed a visual preference for a highly saturated green-yellow (chartreuse), adults found paintings with green more pleasant whereas they found paintings with yellow the least pleasant.
Clare: Is there a theory as to why babies enjoy the colour green? Many studies show that blue is the most popular colour among adults.
Philip: It is currently unknown. One theory for adults suggest that adults like colours that are associated with objects that elicit a positive response, e.g. blues with a clear sky and clean water, and green with fruit and trees. It is possible that this association is evolved over generations and present even in young babies, and/or this preference came about via life experience in their own particular cultural context and environment.
Clare: What can the study tell us about how our visual preferences evolve over time, rather than just staying the same? Why do adults and babies prefer different colours and forms?
Philip: Despite remarkable similarities in preference, the differences we discovered in the way the visual features and complexity of the artwork contribute to infants’ and adults’ responses suggest that infants have their own responses to art that are distinct from adult aesthetics. Infants’ responses to art is likely a result of both their visual immaturity, such as poor visual acuity and contrast sensitivity, and reduced visual experience as well as their relatively immature emotional and cognitive processing. Adults’ aesthetic responses to artwork are affected by cognitive and emotional factors such as context, experience, memory, and culture, whereas, these are less of a factor for infants.
Clare: Is there any correlation between the most popular pictures and the least cropped works – can we read any insight into what kinds of compositions/shapes are most engaging?
Philip: That’s an interesting question, however only four of the paintings were portrait orientation, with the rest being landscape orientation. Which wouldn’t be enough cropped information to make a difference overall. Furthermore, although most of the paintings were of different dimensions, the dimensions were all similar enough for the landscape oriented paintings, so that there was a similar amount of cropped information.
In terms of what compositions are most engaging, we investigated this in an objective way by analysing each painting for its chromatic (colour) and spatial (form) ‘image statistics’. Image statistics are a way of measuring the visual information in an image. For example, we can calculate the average colour in a painting, giving us a single numeric value, or calculate the variation in brightness, or how many edges there are, or how these edges are aligned (e.g. straight, curved), or how random the elements of the painting are. We then investigate any relationships between these image statistics and infant and adult responses, e.g. calculating whether infants look longer when there are more edges in a painting.
So, in terms of what we found, the main similarity was that adults found art more pleasant, and infants looked longer the greater the contrast in colours and brightness of the art. The amount of randomness of the elements and the edges in the paintings also contributed to both infant looking and adult liking. However, adults preferred more randomness while infants preferred less – speaking again to infants immature visual systems.
Adults also found art more pleasant when there were fewer straight edges, whereas infants looked longer at the art the greater the amount of straight and curved edges. Other compositional aspects were that adults found art more pleasant that had a mid-range fractal dimension – this is a painting with repeating patterns at smaller and smaller scales that match those found in natural scenes. Infants looked longer at art with greater ‘gappiness’ – e.g. a range of gaps giving a less complex and more perceivable image.
Clare: What does this study show about the significance of visual art in general?
Philip: Our study shows that visual art is a stimulus that can provoke a response even in young babies. This could be because visual art shares many visual features including image statistics with natural scenes to which the visual system has evolved to process. Our study also shows that babies have sophisticated responses to visual art, and that visual art is a stimulus that can be shown to young babies.
Clare: What are some of the most surprising outcomes of the research project?
Philip: Firstly, it was surprising that babies looked longer at the paintings which adults rated higher for pleasantness. Previously it was thought that baby vision was quite basic and that babies prefer simple patterns but this research shows that babies can perceive much more complex images and prefer to look at some paintings over others, similar to adults. Secondly, it was surprising that we could explain two-thirds of baby and adult preferences using a combination of various chromatic and spatial image statistics. That is, that the visual features of the artwork was driving adult liking and infant looking to such a large extent.
Clare: How do you plan to use the information you have learned from the project?
Philip: We will use the findings from this study to add to the knowledge of the role of image statistics in visual development. This is important since image statistics have a major role in mature vision, and in object and scene perception, recognition, and categorisation in adults. The role of image statistics in visual development can also tell us about when vision is optimised for natural scenes, with implications for understanding the role of experience in the optimisation of vision to natural scenes.
It would be interesting to study infants’ and adults’ responses to abstract art, so that we could maybe more effectively reveal similarity between infants’ visual preferences and adults’ aesthetic responses to different image statistics. Other artists’ work and other genres would be an interesting angle to investigate too. We also wish to investigate aesthetic responses to natural scenes, which as I mentioned share similar chromatic and spatial images statistics as art. We also want to further explore infants’ sensitivity and preference for particular images statistics that we found to be important in this present study using more controlled stimuli.
Find out more at The Sussex Baby Lab