Towards a sensorimotor aesthetics of performing art

Abstract

The field of neuroaesthetics attempts to identify the brain processes underlying aesthetic experience, including but not limited to beauty. Previous neuroaesthetic studies have focussed largely on paintings and music, while performing arts such as dance have been less studied. Nevertheless, increasing knowledge of the neural mechanisms that represent the bodies and actions of others, and which contribute to empathy, make a neuroaesthetics of dance timely. Here, we present the first neuroscientific study of aesthetic perception in the context of the performing arts. We investigated brain areas whose activity during passive viewing of dance stimuli was related to later, independent aesthetic evaluation of the same stimuli. Brain activity of six naïve male subjects was measured using fMRI, while they watched 24 dance movements, and performed an irrelevant task. In a later session, participants rated each movement along a set of established aesthetic dimensions. The ratings were used to identify brain regions that were more active when viewing moves that received high average ratings than moves that received low average ratings. This contrast revealed bilateral activity in the occipital cortices and in right premotor cortex. Our results suggest a possible role of visual and sensorimotor brain areas in an automatic aesthetic response to dance. This sensorimotor response may explain why dance is widely appreciated in so many human cultures.

Introduction

Dance is a dynamic visual form of artistic expression that has accompanied mankind since ancient times. Many different dance styles and traditions exist, covering most human cultures. In dance performance, an observer watches the body movements of the dancer. The spatial and temporal features of the dancer’s movement can induce aesthetic experience in the observer. Dance choreographers and performers exploit these influences to communicate their intentions to the viewer. Dance typically also involves several other elements, including but not limited to narrative, music, costume, and interaction between dancers. No single one of these additional elements is necessary, though the synergy between them undoubtedly increases the aesthetic impact of dance performance. How is this impact coded in the observer’s mind and brain? Here, we present a study that focuses on the neural mechanism underpinning the aesthetic experience of observers watching dance.

The term aesthetic is derived from the Greek word Aesthesis, and was re-defined by Baumgarten in the eighteen century as the gratification of the senses or sensuous delights (Goldman, 2001). A particular type of sensory stimulus, often but not exclusively a work or art, produces a psychological state usually termed aesthetic experience. The neural mechanisms that correlate with these internal processes associated to aesthetic experience have been a recent focus of several studies, and created a new field of ‘neuroaesthetics’ (Cela-Conde et al., 2004, Kawabata and Zeki, 2004). Neuroaesthetics addresses the brain processes underlying aesthetic judgement, evaluation and interpretation. Evaluation is a particularly important element of aesthetic experience. Aesthetic experience can involve two kinds of evaluation: one attributes intrinsic perceptual properties to the stimulus (e.g., ‘it is beautiful’), and the second characterises the observer’s response attitude to the stimulus (e.g., ‘I like it’).

Two classical perspectives on aesthetics have led researchers to focus on different components of aesthetics. On the one hand, objectivist theories treat beauty and other aesthetic properties as attributes of stimuli, processed or perceived by the observer like other attributes. These theories often relied on psychophysical studies aiming to identify particular stimulus properties that induce aesthetic experience. Studies of the so-called ‘golden section’ in abstract geometrical figures provide one example (McManus & Weatherby, 1997), but similar approaches have been used with paintings (Jacobsen, 2004, Jacobsen et al., 2004, Jacobsen and Hofel, 2002, McManus, 1980, Whissell, 1980). Such studies generally suggest that aesthetic experience depends on compositional arrangement between parts of the stimulus, and between individual parts and the whole. Thus, symmetry, balance, complexity, and order of an image all contribute to its aesthetic impact (Leder, Belke, Oeberst, & Augustin, 2004). Objectivist theories often claim some level of generality for these psychophysical regularities across individuals, since most individuals’ perceptual systems process stimulus patterns in the same way.

On the other hand, subjectivist theories maintain that ‘beauty is in the eye of the beholder’, and is largely a matter of attitudes, such as individual taste and preference. This theory receives support from the strong effect of familiarity on aesthetic judgement. We tend to like what we know (Zajonc, 1968). Hence, individual differences in aesthetic judgement may arise from individual differences in prior experience, notably due to cultural environment. Because subjectivist theories focus on individuals, rather than on the generality of aesthetic experience, they often make rather few predictions that can be tested in traditional laboratory experiments. However, they have recently been reinforced by several neuroaesthetic studies identifying brain areas associated with specific aesthetic attitudes such as like/dislike, even though participants may have differed among themselves in which stimuli evoked these attitudes (Cela-Conde et al., 2004, Jacobsen et al., 2006, Kawabata and Zeki, 2004, Vartanian and Goel, 2004).

In general, subjects in these studies view several paintings, and give a rating of how much they like each one. Previous neuroaesthetic studies suggest that at least two broad brain networks may be involved in this process. One is primarily perceptual, centred on sensory and attentional regions. The second, centred on the prefrontal cortex, seems primarily cognitive and/or hedonic. More specifically, brain activity associated with viewing paintings that subjects liked, as opposed to disliked, was found in both visual and prefrontal brain areas (Cela-Conde et al., 2004, Kawabata and Zeki, 2004, Vartanian and Goel, 2004). These included the occipital gyri and fusiform gyri bilaterally in the visual cortex. The same areas were shown to respond to the emotional significance of faces and pictures in other studies that did not focus on aesthetic experience (Iidaka et al., 2002, Paradiso et al., 1999, Vartanian and Goel, 2004). A prefrontal correlate of aesthetic evaluation has also been reported. This included the orbitofrontal cortex, which often reflects the reward value of a stimulus (Kawabata and Zeki, 2004, Rolls, 2000a, Rolls, 2000b), and the prefrontal dorsolateral cortex, which plays a role as a centre of the perception–action interface and is critical for the monitoring and comparison of multiple events in working memory (Cela-Conde et al., 2004, Petrides, 2000). Such studies can reveal the neural correlates of liking or disliking aesthetically-relevant stimuli. However, they provide little information about the physical properties responsible for the aesthetic experience. In other words, they cannot project their conclusions back into stimulus space, to identify which parameters of a stimulus lead to its being liked or disliked. This limitation is a consequence of comparing different sets of stimuli according to the attitudes expressed by each subject (Cela-Conde et al., 2004, Kawabata and Zeki, 2004).

Jacobsen et al. (2006) recently used a subjectivist approach to investigate the neural correlates of aesthetic judgments of abstract geometric forms, and found essentially similar results. Beauty judgments for these stimuli activated both temporoparietal regions, but also fronto-median and prefrontal regions. They then correlated aesthetic preference with symmetry judgments for the same stimuli, and found significant overlap. Psychophysical studies reached similar conclusions (Leder et al., 2004, McManus, 1980). Interestingly, some of these areas were previously shown to underlie other types of human judgement (i.e., social and moral judgement) (Cunningham, Johnson, Gatenby, Gore, & Banaji, 2003).

Overall, previous neuroaesthetic studies demonstrate an important role of prefrontal areas in aesthetic evaluation. These areas are also involved in representing the reward value of stimuli, which would not be considered aesthetic (Aharon et al., 2001, Rolls, 2000a, Rolls, 2000b, Small et al., 2001). Moreover, the aesthetic evaluation network partially overlaps with brain regions processing social and moral behaviour (Jacobsen et al., 2006). Interestingly, these frontal areas have also increased disproportionately in size and complexity during recent human evolution. They may contribute both to the large difference in sophistication of the social interactions between humans and other mammals, and to creation of art by humans, but not other species (Cela-Conde et al., 2004).

In summary, three main issues emerge from these studies. First, most studies have used static images such as paintings, photographs, or abstract forms. Second, most studies have investigated brain activity during explicit aesthetic evaluation or appreciation. Few studies have considered automatic effects of viewing stimuli that have aesthetic potential. Finally, studies have focused on aesthetic judgements of individuals, rather than the potential of stimuli to elicit similar aesthetic responses in a group of individuals. Here, we extend previous neuroaesthetic work by using a new class of stimuli (dance), by using an implicit task during stimulus presentation, and by studying group-average rather than individual responses.

Dance is a dynamic, fluent and fugitive visual art form. To simplify our approach to the aesthetic appreciation of human dance, we have reduced dance to its core motor elements, as the key to its aesthetic value. For this, we studied the kinematics of dance movements while keeping all other features such as visual background constant. We therefore removed narrative, costume, and musical elements. We reasoned that any variation in neural activity related to aesthetic valuation could then be explained only by differences in the movements’ kinematics. A disadvantage of this simplifying approach, however, is that the stimulus itself and the subject’s aesthetic response to it may fall far short of the typical dance performance situation. On the other hand, simplified experiments allow any positive findings to be interpreted more clearly. For example, an experimenter can more confidently identify specific stimulus parameters that lead to particular aesthetic responses.

In addition, simplified dance stimuli allow aesthetic responses to be directly related to studies on selectivity of relevant brain areas. Two well-established brain systems appear to be specifically relevant to processing of dance kinematics. These are brain regions responsible for processing motion stimuli such as random dots or moving objects. These activate area MT/V5 (Sunaert, Van Hecke, Marchal, & Orban, 1999). A second system is concerned with processing a special class of stimulus motions, caused by the actions of other individuals. This system comprises areas in the premotor, and parietal cortices (so-called ‘mirror neuron areas’), and also the superior temporal sulcus. To give one example, this latter area responds to observation of biological but not non-biological motion (Grossman and Blake, 2002, Puce and Perrett, 2003, Vaina et al., 2001). The human mirror system is also known to be involved in action observation in general (Decety and Grezes, 1999, Grafton et al., 1996, Rizzolatti et al., 1996), and dance observation in particular (Calvo-Merino et al., 2005, Calvo-Merino et al., 2006, Cross et al., 2006). We hypothesised that variation in aesthetic responses to a set of dance movements might correlate with the extent to which the movements activated these specialised visual and motor brain systems.

We therefore recorded brain activity in naïve subjects with no dance experience while they watched video clips of two dance styles (classical ballet and capoeira). No explicit aesthetic or evaluative questions were asked at the time of viewing inside the scanner. A range of dance styles and kinematic parameters were chosen to probe general brain responses to dance movements, rather than recognition or association with individual culturally familiar stimuli. We then showed the same stimuli to the subjects again in a later testing session. This time, the subjects gave ratings to evaluate each stimulus along a number of standard aesthetic dimensions. The group average of all subjects’ ratings was used to identify brain areas coding these aesthetic dimensions. Finally, we evaluated the characteristics of the dance movements that lead to most and least activation of any brain area thus identified. This allowed us to identify kinematic features encoded by that brain area, and related to its aesthetic coding.