Spatial Vision, Vol. 21, No. 3–5, pp. 379–396 (2008)  Koninklijke Brill NV, Leiden, 2008. Also available online - www.brill.nl/sv Towards a framework for the study of the neural correlates of aesthetic preference MARCOS NADAL 1,∗ ,ENRICMUNAR 1 , MIQUEL ÀNGEL CAPÓ 2 , JAUME ROSSELLÓ 1 and CAMILO JOSÉ CELA-CONDE 2 1 Department of Psychology, Universitat de les Illes Balears, Crta Valldermossa s/n, km 7,5, Palma de Mallorca 07122, Spain 2 Department of Philosophy, Universitat de les Illes Balears, Crta Valldermossa s/n, km 7,5, Palma de Mallorca 07122, Spain Received 21 March 2006; accepted 10 March 2007 Abstract—Aiming to provide a tentative framework for the study of the neural correlates of aesthetic preference, we review three recent neuroimaging studies carried out with the purpose of locating brain activity associated with decisions about the beauty of visual stimuli (Cela-Conde et al., 2004; Kawabata and Zeki, 2004; Vartanian and Goel, 2004). We find that the results of the three studies are not in line with previous neuropsychological data. Moreover, there are no coincidences among their results. However, when they are mapped on to Chatterjee’s (2003) neuropsychological model of aesthetic preference it becomes clear that neuroimaging data are not contradictory, but complementary, and their interpretation is enriched. The results of these studies suggest that affective processes have an important role in aesthetic preference, and that they are integrated with cognitive processes to reach a decision regarding the beauty of visual stimuli. Future studies must aim to clarify whether certain methodological procedures are better suited to study any of the particular cognitive operations involved in aesthetic preference, and ascertain the extent to which the proposed framework is compatible with the aesthetic appreciation of musical stimuli. Keywords: Brain; fMRI; MEG; aesthetic preference; beauty. INTRODUCTION The question of the neural correlates of artistic production, aesthetic preference, and similar phenomena, has not been addressed by means of functional neuroimaging techniques until quite recently, although it has been investigated from a neuropsy- chological perspective (see Chatterjee, 2004, for a review). Here we offer an in- tegrative perspective of neuroimaging studies of aesthetic preference. We will not ∗ To whom correspondence should be addressed. E-mail: marcos.nadal@uib.es 380 M. Nadal et al. try to offer a general overview of the neural correlates of appreciation for artistic and aesthetic objects. This is a very broad field that includes interesting studies ad- dressing the brain basis of different phenomena, such as artistic production, artistic appreciation, the effects of artistic education, visual arts, music, and so on. We have chosen to be selective in this paper, concentrating on visual art and aesthetics, and leaving aside several interesting studies carried out in relation to music (reviewed in Peretz and Zatorre, 2005). Thus, the present work focuses mainly on the neural correlates of visual aesthetic preference. Despite the limited scope of the results produced within this line of research and the current absence of a solid grounding framework, there are already several new studies being carried out, making an integrative framework much more urgent. Thus, the main objective of the present work is to begin the construction of this scaffolding by suggesting a series of testable hypotheses based on existing results. Our conclusions are not meant as closed, indisputable facts, but as a set of possible means to work on a solid framework for future studies in this field. The results of each of the reviewed neuroimaging studies will be discussed in reference to previous literature, and they will be brought together with the aid of a recent cognitive model of aesthetic preference. Cognitive processes involved in aesthetic preference Several views on aesthetic preference have been developed during the last century, including Psychoanalysis, Gestalt theory and Empirical Aesthetics. Within the last of these, aesthetic preference has been related with arousal (Berlyne, 1970, 1971), prototypicality (Martindale, 1988; Martindale et al., 1988), and appraisals (Silvia, 2005), among other factors. Potentially, any of these perspectives could serve to ground our interpretation of neuroimaging results. However, given that the present work aims to provide a tentative relation between neural activity and cognitive processes, we need a model that specifies the cognitive operations underlying aesthetic preference, as well as their interactions, and that, at the same time, can bridge the cognitive-neural levels. Leder et al. (2004) have recently proposed a comprehensive model of visual aes- thetic preference and perception, which includes five processing stages: perceptual analysis of the visual stimulus, implicit memory integration, explicit classification, cognitive mastering, and the emergence of a cognitive state, resulting from the pre- vious stages, and an affective state, that results from the continuous interactions between previous stages and affective systems in the brain. The cognitive state is the source of aesthetic preference, while aesthetic emotion is grounded on the af- fective state. The different stages suggested by Leder and colleagues (2004) include several operations, involving different variables known to affect aesthetic preference. How- ever, their model is formulated at a general psychological level, which makes it difficult to establish straightforward hypotheses about specific brain activity asso- ciated with those operations. In contrast, Chatterjee’s (2003) framework for the Neural correlates of aesthetic preference 381 neural correlates of aesthetic preference is directly grounded on visual neuroscience, which makes it an ideal candidate to bridge this gap. Chatterjee (2003) suggested that aesthetic preference involves three processing stages, common to the perception of any visual stimulus. Early visual processes extract and analyze simple compo- nents in different brain areas. Operations in the intermediate stage segregate some elements and group others, forming coherent representations. In late visual process- ing stages certain regions of the object are selected for further scrutiny, memories are activated, objects are recognized and associated with meanings. In the case of visual stimuli found to be aesthetically pleasing, these operations elicit emo- tional processes, which feedback into the system via attentional mechanisms. As in Leder and colleagues’ (2004) model, there is a second output, here represented by the decision-making processes required by most experimental designs. Chatter- jee (2003) suggested that processing aesthetic stimuli involves similar visual brain regions as processing any other kind of visual stimuli. What sets aesthetic prefer- ence apart from other cognitive processing of visual stimuli is precisely the engage- ment and interplay of additional non-perceptual processes, such as emotions and decision-making. Regarding the neural correlates of these operations, Chatterjee (2003) suggested that early visual processing of the basic features of artworks and other aesthetic stimuli takes place in occipital brain regions, like any other kind of stimuli. However, features processed in intermediate stages, such as shape or composition, can engage frontal-parietal attentional circuits, which enhance the processing of those attributes within the ventral visual stream (Chatterjee, 2003). He anticipated that the tasks of stating preferences and making decisions about objects would most likely be associated with activity in the dorsolateral frontal and medial frontal cortices. Pinpointing the neural correlates of the emotional facet of aesthetic experience is more difficult, given that its very nature is less clear. In spite of this, Chatterjee (2003) suggested that given that anterior medial temporal lobe, medial and orbital cortices in the frontal lobe, as well as subcortical structures, mediate emotions, they might also be involved in the affective component of the aesthetic experience. NEUROIMAGING STUDIES OF AESTHETIC PREFERENCE There are currently four published neuroimaging studies concerned with the neu- roanatomical correlates of aesthetic preference for visual stimuli. Three of them have very similar objectives. Cela-Conde et al. (2004) aimed to locate “brain ar- eas activated during the visual perception of aesthetic objects” (Cela-Conde et al., 2004, p. 6321). Kawabata and Zeki (2004) wanted to verify whether “there are brain areas that are consistently active across subjects when they perceive a painting as being beautiful and, conversely, whether there are brain areas that are specifically active when they view paintings that they consider to be ugly” (Kawabata and Zeki, 2004, p. 1699). Vartanian and Goel (2004) carried out their study to “determine 382 M. Nadal et al. the neuroanatomical correlates of aesthetic preference for paintings” (Vartanian and Goel, 2004, p. 893). Another similarity is that the three studies were designed to contrast participants’ brain activity associated with positively and negatively rated stimuli. Thus, while their brain activity was being recorded, participants created dif- ferent stimuli conditions varying in beauty or preference as a result of their aesthetic preference ratings. A fourth study (Jacobsen et al., 2006) may appear to address the same question as the aforementioned three. However, it differs in at least two very important issues. First, whereas the other three studies were concerned with the differences in brain activation when judging stimuli as beautiful or not beautiful, Jacobsen et al. (2006) compared the whole process of aesthetic decision making with another kind of decision, that of symmetry. Thus, the results of the studies by Kawabata and Zeki (2004), Vartanian and Goel (2004) and Cela-Conde et al. (2004) refer specifically to the neural correlates of judging stimuli as beautiful compared to those of judging them as ugly, whereas Jacobsen et al. (2006) designed their experiment to identify the neural correlates of judging the beauty of images compared to judging their symmetry. Thus, results obtained by Jacobsen et al. (2006) refer to the neural correlates of the judgment process itself. This difference is far from trivial; in fact it has important consequences for the comparison of the results of this study with the others. Jacobsen et al. (2006) acknowledged this in the first paragraph of their paper when referring to the other set of studies: “However, these approaches focus on the particular valences of preferences, e.g., by parametric manipulation of levels of attractiveness or by direct comparison of beautiful versus ugly or neutral pictures. In contrast, none of these studies aimed at identifying the network of aesthetic judgment per se” (Jacobsen et al., 2006, p. 276). In sum, while the question of the brain correlates of judging the beauty of images in contrast to other judgments, such as symmetry, is an interesting one, it is not the same as asking about the neural correlates of judging images as beautiful compared with judging them as ugly. The second difference between the study by Jacobsen et al. (2006) and the rest resides in the stimuli presented to the participants. Vartanian and Goel (2004) and Kawabata and Zeki (2004) asked their participants to express their aesthetic preferences only for artworks, and Cela-Conde et al. (2004) asked theirs to do so for artworks, decorative images and photographs. However, Jacobsen et al. (2006) asked participants to judge the beauty (and symmetry) of black and white abstract patterns created by the authors. These stimuli consisted of a black circle containing a centered white square, oriented like a rhombus, in which small black triangles were arranged to form a variety of graphic patterns. At present it is not clear whether decisions about the beauty of artistic and decorative stimuli involve the same cognitive processes as judging the beauty of simple geometric patterns. Berlyne (1971) considered the tradeoff related with the use of artistic vs simple visual materials in experimental aesthetics: “In the former case, there is the advantage of studying reactions to real art and the disadvantage that any two works of art differ from each other in several different respects, so that the actual factor responsible for Neural correlates of aesthetic preference 383 any differences in reactions to them is difficult to pin down. The use of artificially simple material overcomes this drawback but may be open to the criticism that it is a long way from anything that could be regarded as art and may thus prevent us from identifying essential components of real-life aesthetic behavior” (Berlyne, 1971, p. 12). We believe that the introduction of adequate control procedures reduces many of the disadvantages of using artistic and decorative materials, and that the use of simple visual patterns might engage different cognitive operations to those that enable aesthetic appreciation in natural conditions. Furthermore, given that symmetry is a very salient feature of the materials used by Jacobsen et al. (2006), their results might be difficult to generalize to other stimuli whose symmetry is less prominent. For these reasons, the remainder of the present work we will concentrate on the studies by Kawabata and Zeki (2004), Vartanian and Goel (2004) and Cela- Conde et al. (2004). Summary of the neuroimaging results Regarding the results of the three studies, we shall consider only the contrasts performed between the conditions of positively and negatively valued stimuli. Kawabata and Zeki (2004) and Vartanian and Goel (2004) obtained interesting results when comparing brain activity before different categories of stimuli, such as abstract vs representational, but such issues will not be commented on here, given that this review is primarily concerned with the neural basis of general aesthetic preference. The results of the three studies are illustrated in Fig. 1. Kawabata and Zeki (2004) registered participants’ brain activity with fMRI while rating the beauty of stimuli. They found that activity in the orbitofrontal cortex was greater for stimuli classified as beautiful, while activity in the motor cortex was greater for stimuli classified as ugly. Cela-Conde et al. (2004) used magnetoencephalography (MEG) to record brain activity during the aesthetic preference task. Their results showed that activity in the left dorsolateral prefrontal cortex increased in late latencies (400–1000 ms) when participants judged stimuli as beautiful, as compared to the non-beautiful condition. By means of fMRI Vartanian and Goel (2004) found that the activity in the right caudate nucleus decreased as preference ratings decreased, while activity in the left anterior cingulate gyrus and bilateral occipital gyri, increased with preference ratings. At least two issues merit comment. First, there seems to be a discontinuity between these three neuroimaging studies and those carried out using lesion and electroencephalographic methods. For instance, none of the neuroimaging studies found significant activity in the amygdala, which might have been expected based on Adolphs and Tranel’s (1999) results. Their study of patients with amygdalar lesions had suggested that this structure is involved in guiding preference for visual stimuli, Figure 1. (See color Plate XII) Neural correlates identified by the three neuroimaging studies. Abbreviations — ACC: Anterior cingulate cortex; CN: Caudate nucleus; DLPFC: Dorsolateral prefrontal cortex; MC: Motor cortex; OFC: Orbitofrontal cortex; OG: Occipital gyri. 384 M. Nadal et al. specifically those which are normally judged to be aversive. The neuroimaging technique used by Cela-Conde et al. (2004) does not allow the recording of activity in the amygdala, due to its spheroid structure, but Kawabata and Zeki’s (2004) and Vartanian and Goel’s (2004) studies were well suited to pick up any significant amygdalar activity, especially in the ugly or non-preferred condition. The only amygdalar activity was observed when Kawabata and Zeki (2004) compared the portrait and non-portrait conditions independently of their aesthetic rating. It is possible that regarding aesthetic preference the conclusions of lesion or degenerative studies do not easily extend to neuroimaging experiments of healthy participants. Alternatively, it might be the case that these three neuroimaging studies did not include unpleasant enough stimuli required to detect amygdalar activity. This possibility could be experimentally tested by offering participants a broader affective range of stimuli. Based on several electroencephalographic studies (Brattico et al., 2003; Jacobsen and Höfel, 2001, 2003), Jacobsen and Höfel (2003) had suggested a two-stage model of aesthetic preference. During the first stage, which takes place around 300 milliseconds after the stimulus has been presented, an initial impression is formed. This process is associated with anterior frontomedian activity, mainly when participants consider stimuli to lack aesthetic value. The second stage, a deeper aesthetic evaluation, begins close to 600 milliseconds after stimulus onset and is related with wide right hemisphere activity. Although the results of three studies could verify the involvement of frontomedian and right-hemisphere activity, only the study by Cela-Conde et al. (2004) was well suited, due to the temporal resolution of MEG, to test the suggested sequence of activity. But as it turned out, the frontomedian activity Jacobsen and Höfel (2003) found associated with negative ratings did not appear in any of the neuroimaging studies. In fact, frontal activity detected by Cela-Conde et al. (2004) (dorsolateral), Kawabata and Zeki (2004) (orbitofrontal) and Vartanian and Goel (2004) (anterior cingulate) was associated with positive ratings. Additionally, in the study by Cela-Conde et al. (2004) all brain activity correlating with aesthetic preference during the first second after stimuli onset was limited to the left prefrontal dorsolateral cortex. Activity in other areas, some of which were located in the right hemisphere, was also recorded, but revealed no significant differences between the beautiful and ugly conditions. These discrepancies may owe to the use of different kinds of stimuli. Jacobsen and Höfel (2003) based their model on studies using patterns composed by simple geometric forms, whereas Cela-Conde et al. (2004) used complex artworks and photographs. The comparison of the results from neuroimaging studies on the assessment of facial beauty (Aharon et al., 2001; O’Doherty et al., 2003; Senior, 2003) with those from the three studies attempting to identify the neural correlates of aesthetic preference reveals only one coinciding brain region. Kawabata and Zeki’s (2004) study, the only one out of the three to include portraits, revealed significant activity in the orbitofrontal cortex before stimuli classified as beautiful, just as was observed with beautiful faces with high reward value. Neural correlates of aesthetic preference 385 The second unexpected fact is the complete lack of coincidence among the results of the three studies when comparing their results regarding the difference in brain activity between the beautiful and ugly conditions (see Fig. 1). Although each of the reviewed studies leaves room for improvement, we believe that their limitations need not lead to an invalidation of their findings, at least until they are replicated or experimentally disproved. On the other hand, none of these studies asserted that the areas they had identified were the exclusive neural correlates aesthetic preference. In fact they all acknowledged that these areas influence and are influenced by the activity in other brain areas. It might be the case that the three studies captured only a subset of operations involved in the complex cognitive task of deciding about the aesthetics of visual stimuli, and that differences in the experimental designs and procedures lead them to reflect diverse aspects of the aesthetic experience. Thus, each study might be reporting a partial picture of the neural correlates of aesthetic preference. But before we attempt to sort out the cognitive operations associated with the neural correlates identified by each study, we need to address the reasons for the discrepancy among the results of the three studies. The suggestions commented below, and summarized in Table 1, are meant here as mere possibilities, given that at present there are no experimental data to demonstrate that each of these factors has a direct influence on the neural correlates of aesthetic preference. However, the revision of the possible reasons behind the divergence of neuroimaging results can be fruitful in suggesting new testable hypothesis. Table 1. Differences among the three neuroimaging studies Study Kawabata and Zeki Vartanian and Goel Cela-Conde et al. (2004) (2004) (2004) Results Greater activity in Activity in right caudate Greater activity in mOFC for beautiful nucleus decreases with DLPFC for beautiful stimuli. decreasing preference. stimuli. Greater activity in Activity in left cingulate motor cortex for gyrus and occipital gyri ugly stimuli. increases with increasing preference. Technique Event-related fMRI Event-related fMRI MEG Exposure 2 s 6 s 3 s Time 500 ms interstimuli int. No interstimuli int. 1–1.2 s interstimuli int. Task Indicate beautiful, Indicate (0–4) degree of Indicate beautiful or not neutral or ugly preference beautiful Participants 5 M 5 F 10 F 2 M 8 F Stimuli 16 ×(abstract, still (20 ×(representational, 40 × (abstract art, life, landscapes, abstract)) ×(original, classic, impressionist, portraits) × filtered, altered) (=120) postimpressionist) + (beautiful, neutral, 160 photos (=320) ugly) (=192) Procedure Pre-classification No No 386 M. Nadal et al. The most obvious difference among the three studies is the use of a different neu- roimaging technique (MEG) by Cela-Conde and colleagues (2004) with regards to the other two studies (event-related fMRI ). The involvement of the detected brain areas in aesthetic preference is inferred from different parameters. Whereas MEG detects magnetic fields generated by excitatory and inhibitory postsynaptic poten- tials in the dendrites of pyramidal neurons (Lounasmaa et al., 1996; Maestú et al., 2005), fMRI offers an indirect measure of neural activity related with hemodynamic and metabolic responses underlying neuronal events, probably reflecting input and intracortical processes (Logothetis et al., 2001). Furthermore, both techniques have different spatial and temporal resolutions, which require different exposure times to the stimuli and interstimuli intervals (see Table 1). Verifying that indeed neu- roimaging technique has an impact on the detected neural correlates of aesthetic preference could be achieved by using a single protocol for MEG and fMRI, or by the joint EEG and fMRI recording (Debener et al., 2006). Another major difference among the three studies is the task that participants were asked to perform. Kawabata and Zeki (2004) asked their participants to rate the beauty of the stimuli on a 3-point scale (beautiful, neutral, ugly), whereas Cela-Conde et al. (2004) used a dichotomous scale (beautiful, not beautiful). In contrast, Vartanian and Goel’s (2004) participants were asked to rate their degree of preference for the pictures on a 5-point scale. Leder and colleagues (2005) suggested that preference ratings are associated with a strong affective or reward component, and that the task of rating beauty might elicit a stronger cognitive component. Hence, both tasks might have partially different neural correlates, as suggested by studies of facial beauty (Aharon et al., 2001). Another important difference among the three studies that could lead to the discrepancies in their results is the composition of the groups of participants. Vartanian and Goel (2004) included 10 women and 2 men, Kawabata and Zeki (2004) included 5 men and 5 women, while Cela-Conde et al. (2004) included 8 women. This seems to be a relevant issue, in the light of studies that have found gender differences in aesthetic preference (Bernard, 1972; Burges Cruz, 2000; Eysenck and Castle, 1971; Furnham and Walker, 2001; Johnson and Knapp, 1963; Neperud, 1986; Polzella, 2000), and the increasing evidence of sex differences in the neural correlates of several cognitive (Bell et al., 2006; Boghi et al., 2006; Georgopoulos et al., 2001; Haier et al., 2006) and affective tasks (Azim et al., 2005; Kemp et al., 2004; Mackiewicz et al., 2006; Piefke et al., 2005; Tranel et al., 2005). Thus, it is not known how confounding it is to jointly analyze men and women’s results, or how limiting it is to only include participants from one of the sexes. The composition of the material might also turn out to have an important role in explaining the differences among the three studies being reviewed here. The only commonality in this respect is that the three studies included both abstract and representational stimuli, though in different proportions. Cela-Conde et al. (2004) were the only ones to include non-artistic in addition to artistic stimuli. Kawabata and Zeki’s (2004) study was the only one to include portraits. Finally, Neural correlates of aesthetic preference 387 Vartanian and Goel (2004) presented participants with two altered versions of each picture, in addition to the original form. There is an extensive literature showing differences in the aesthetic preference for visual stimuli according to their degree of abstraction, artistic qualities, and modification (Bernard, 1972; Cela-Conde et al., 2002; Furnham and Walker, 2001; Hekkert and van Wieringen, 1996a, 1996b; Johnson and Knapp, 1963; Lindauer, 1990; Neperud, 1986; Winston and Cupchik, 1992). It is possible that these behavioural differences may have an expression at the neural level, as suggested by Kettlewell and Lipscomb’s (1992) neuropsychological study, though at present we can only assume that neural networks related with object recognition contribute differently to the aesthetic appreciation of abstract and representational visual stimuli. Finally, there are profound differences among the three procedures, which come to reflect some of the divergences we have already pointed out. For instance, in contrast to the studies by Vartanian and Goel (2004) and Cela-Conde et al. (2004), Kawabata and Zeki (2004) included a stimuli pre-selection procedure. This might have inadvertently elicited recognition processes in the task participants performed while their brain activity was being recorded, as suggested by results from previous studies (Cela-Conde et al., 2002; Nadal et al., 2006) that revealed an association between mnemonic processes and aesthetic preference. The other procedural aspect that is different among the studies is the preparation of stimuli. There are several variables that have been shown to affect aesthetic preference of visual stimuli, such as psychophysical variables (luminance, contrast, predominant wavelength), complexity, novelty, prototypicality, and so on (e.g. Berlyne, 1970; Martindale and Moore, 1988). Furthermore, it has been shown that some of these variables have an impact on the neural correlates of visual processing (Daffner et al., 1998, 2000; Müller et al., 2003; Nicki and Gale, 1977; Sasaki et al., 2005). Whereas Cela-Conde et al. (2004) homogenized the complexity, novelty, color spectrum and luminance of the stimuli, in addition to their size and resolution, Vartanian and Goel (2004) and Kawabata and Zeki (2004) did not report whether they carried out any detailed procedure, beyond normalizing size and resolution, in order to control these variables. POSSIBLE RELATIONS BETWEEN BRAIN ACTIVITY AND COGNITIVE PROCESSES INVOLVED IN AESTHETIC PREFERENCE We have discussed the possible reasons for the lack of coincidence among the results of three neuroimaging studies addressing the neural correlates of aesthetic preference, concentrating on different procedural aspects. Given that aesthetic preference is a process that involves multiple cognitive operations, which take place in different brain areas and in different time frames, it is very possible that the results of the three neuroimaging studies reviewed here, conditioned by their respective experimental designs, refer to the neural correlates of only some of these processing operations. The fact that none of the three studies is grounded 388 M. Nadal et al. on a specified psychological model of the cognitive operations involved in aesthetic preference prevents the authors from designing adequate control procedures and baseline conditions. This would allow them to isolate more specific brain activity, or a subset of operations involved in aesthetic preference. Thus, for the moment, we must settle for what seems to be an incomplete and partial knowledge of the neural correlates of aesthetic or beauty preferences. Therefore, the relation between specific cognitive operations contributing to aesthetic preference and each of the brain areas identified by the neuroimaging studies remains to be determined. Vartanian and Goel (2004) found lower preference ratings associated with de- creased activity of the caudate nucleus, higher ratings associated with increased activity in the bilateral occipital giry and left cingulate sulcus. Caudate nucleus activity has been associated with the processing of primary rewarding stimuli by animals (Rolls, 1999; Schultz et al., 2000) as well as abstract rewards in humans, such as imaginary money (Delgado et al., 2000; Knutson et al., 2000). These results suggest that the caudate nucleus’ activity during aesthetic preference might reflect the subcortical processing of the emotional response to aesthetic stimuli. With regard to the increased activity in visual areas, Kaestner and Ungerleider (2000) have shown that attention modulates the processing of relevant visual stimuli by enhancing neuronal responses at different levels of visual processing in the brain. Poghosyan and colleagues (2005) showed that attention can modulate neuronal responses to certain locations of the visual field, whole visual objects, or specific visual features, such as color or shape. They noted that these modulatory effects were stronger in extrastriate visual areas, though different features of selective attention can also affect activity in striate cortex. However, it has been shown that the emotional valence of images can also modulate activity in visual areas (Lang et al., 1998; Shulman et al., 1997). So, it is not easy to determine whether the occipital activity associated with preferred images observed by Vartanian and Goel (2004) is associated with attentional or emotional modulation of visual processing. In fact, Lane and colleagues (1999) found that emotional valence, arousal and attention independently increased the activity in extrastriate cortex, suggesting that there is a common influence that converges on early visual processing. They also noted that the overlap of the activity patterns associated with emotion and attention could be a reflection of their overlap at a behavioral level, in the sense that attentional resources would automatically be recruited during emotional states. Maunsell (2004) suggested that the brain might not even have different neuronal signals related with attention and reward, broadly defined to include, in addition to primary reinforcers, other factors that can motivate behavior, such as the preference for places or stimuli. In this sense, the allocation of attention could be the representation of the subject’s actual assessment of reward. Thus, occipital activity identified by Vartanian and Goel (2004) might reflect enhanced processing of preferred stimuli, related with attentional or affective processes. Finally, the increase in cingulate activity could be a manifestation at an emotional level of this attentional engagement of preferred stimuli. Bush and colleagues [...]... emotional response, enhancement of early visual processes and decision-making Two aspects of the emotional response associated with aesthetic preference have been captured by two of the reviewed studies: the representation of reward value and the awareness of the emotional state The cortical processing of the magnitude of the reward of the stimuli to be aesthetically judged probably corresponds to activity... Cognitive appraisals and interest in visual art: Exploring an appraisal theory of aesthetic emotions, Empirical Studies of the Arts 23, 119–133 Tranel, D., Damasio, H., Denburg, N L and Bechara, A (2005) Does gender play a role in functional asymmetry of ventromedial prefrontal cortex? Brain 128, 2872–2881 Vartanian, O and Goel, V (2004) Neuroanatomical correlates of aesthetic preference for paintings,... just as Kawabata and Zeki (2004) observed for aesthetic stimuli Second, the medial orbitofrontal cortex, where Kawabata and Zeki (2004) found activity associated mostly with beautiful images, was most active after a reward and decreased its activity after punishment, whereas the activity in the lateral orbitofrontal cortex increased after punishments and decreased after reward O’Doherty and colleagues... role of an increased activity in the occipital cortex while rating visual stimuli as beautiful Fourth, the comparison of the neural correlates of rating beauty and rating preference of visual stimuli, as well as the use of affective and cognitive priming methods, can determine whether these two measures of aesthetic experience elicit cognitive and affective processes to the same measure Additionally,... activity in the medial orbitofrontal cortex identified by Kawabata and Zeki (2004): images rated as beautiful were associated with a higher reward value than those rated as ugly It remains to be studied whether extending the degree of ugliness of the images in the sample, including really unpleasant stimuli, would elicit activity in lateral orbitofrontal cortex in an aesthetic preference task, as observed... means of an affective priming experimental paradigm The same could be said about the relation of the motor cortex with the negative aesthetic experience It also remains to be tested whether the inclusion of stimuli which are more aversive elicits activity in the lateral orbitofrontal cortex and amygdala Third, controlling for attentional and affective processes will allow determining the true role of. .. Cela-Conde and colleagues (2004) Neural correlates of aesthetic preference 391 sual aesthetics This framework provides a general scaffolding for the results of the three studies, and suggests that their results are not contradictory, but complementary Mapping the results of the three neuroimaging studies onto Chatterjee’s (2003) model, briefly described above, clarifies the meaning of each of their results, allows... aesthetic preference The orbitofrontal cortex has often been related with the representation of the reward value of stimuli and with the integration of affective information relayed from limbic areas (Krawczyk, 2002) Orbitofrontal activity has been observed in association with the deliverance of primary tactile and gustatory reinforcers (Francis et al., 1999; O’Doherty et al., 2002) However, abstract reinforcers... bidirectional arrows, to show attentional processes as ubiquitous as opposed to a specific processing stage, and to change the original label ‘representational domain’ by ‘high vision’, thus avoiding the ambiguity of the term representation while maintaining the idea of a stage of higher-level visual processing Abbreviations — V & G: Vartanian and Goel (2004); K & Z: Kawabata and Zeki (2004); C-C et al: Cela-Conde.. .Neural correlates of aesthetic preference 389 (2000) suggested that the anterior cingulate cortex is part of an attentional system that regulates cognitive and emotional processing They noted that the anterior cingulate cortex can be functionally divided in two regions: a cognitive division and an affective division The latter encompasses the region identified by Vartanian and Goel (2004), and has . be addressed. E-mail: marcos.nadal@uib.es 380 M. Nadal et al. try to offer a general overview of the neural correlates of appreciation for artistic and aesthetic. those operations. In contrast, Chatterjee’s (2003) framework for the Neural correlates of aesthetic preference 381 neural correlates of aesthetic preference