Behavioural Brain Research 278 (2015) 147–154

Contents lists available at ScienceDirect

Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr

Research report

Do you like Arcimboldo’s? Esthetic appreciation modulates brain activity in solving perceptual ambiguity M. Boccia a,b,∗ , F. Nemmi c , E. Tizzani a , C. Guariglia a,b , F. Ferlazzo a , G. Galati a,b , A.M. Giannini a a

Department of Psychology, Sapienza University of Rome, Italy Neuropsychology Unit, IRCCS Fondazione Santa Lucia of Rome, Italy c Klingberg Lab, Neuroscience Department, Karolinska Institute, Sweden b

h i g h l i g h t s • Activity in the fusiform gyrus contributes to the esthetic experience of ambiguous portraits. • State of mind interacts with perceptual features in the superior parietal lobe. • Study confirms the interaction between top-down and bottom-up processes.

a r t i c l e

i n f o

Article history: Received 26 May 2014 Received in revised form 22 September 2014 Accepted 24 September 2014 Available online 5 October 2014 Keywords: Neuroaesthetic fMRI Ambiguity Esthetic judgment

a b s t r a c t Esthetic experience is a unique, affectively colored, self-transcending subject–object relationship in which cognitive processing is felt to flow differently than during everyday experiences. Notwithstanding previous multidisciplinary investigations, how esthetic experience modulates perception is still obscure. We used Arcimboldo’s ambiguous portraits to assess how the esthetic context organizes ambiguous percepts. The study was carried out using functional magnetic resonance imaging (fMRI) in healthy young volunteers (mean age 25.45; S.D. 4.51; 9 females), during both an explicit esthetic judgment task and an artwork/non-artwork classification task. We show that a distinct neural mechanism in the fusiform gyrus contributes to the esthetic experience of ambiguous portraits, according to the valence of the esthetic experience. Ambiguous artworks eliciting a negative esthetic experience lead to more pronounced activation of the fusiform face areas than ambiguous artworks eliciting a positive esthetic experience. We also found an interaction between task and ambiguity in the right superior parietal lobule. Taken together, our results demonstrate that a neural mechanism in the content-dependent brain regions of face processing underlies the esthetic experience of ambiguous portraits. Furthermore, they suggest that esthetic experience interacts with perceptual qualities of stimuli in the right superior parietal lobe, supporting the idea that esthetic experience arises from the interaction between top-down orienting of attention and bottom-up perceptual facilitation. © 2014 Elsevier B.V. All rights reserved.

1. Introduction In the past decade several studies have assessed neural underpinnings of esthetic experience [1–7], giving rise to the field of Neuroaesthetics [8]. Actually, earlier theories about the mind processes subtending esthetic experience can be found both in Kant

∗ Corresponding author at: Dipartimento di Psicologia, Università Sapienza di Roma, via dei Marsi 78, 00100 Roma, Italy. Tel.: +39 06 49917527; fax: +39 06 49917711. E-mail address: [email protected] (M. Boccia). http://dx.doi.org/10.1016/j.bbr.2014.09.041 0166-4328/© 2014 Elsevier B.V. All rights reserved.

[9] and Schopenhauer [10]. Especially with Schopenhauer’s theory, the notion of esthetic attitude began to emerge. If an esthetic attitude does exist, then it has to be supposed that esthetic experience requires an intentional shift from an automatic visuo-perceptual processing to an esthetic state of mind, more explicitly directed to the sensory experience [11,12]. Indeed, esthetic experience allows for the objects processing in a unique, emotionally colored, selftranscending subject–object relationship [13,14]. Interestingly, the neural bases of esthetic experience were recently studied as a function of the interaction between topdown orienting of attention and bottom-up perceptual facilitation [3]. These authors found that different neural networks subtend

148

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

a pragmatic and an esthetic orientation toward the artworks and that different brain areas underpin the esthetic orientation in relation to different perceptual features. Specifically, esthetic orientation, compared with pragmatic orientation, activated the left lateral prefrontal cortex, while paintings that facilitated visuospatial exploration activated the left superior parietal lobule. Esthetic experience has been seen as the result of the interaction between individual predisposition toward the artworks (i.e., top-down processes) and perceptual facilitation (i.e., bottom-up processes) [3]. Capitalizing on visual neuroscience studies, Chatterjee [15] developed a theoretical model of the cognitive and affective processes involved in visual esthetic preference. A series of information-processing stages would occur during the esthetic experience of artworks: first, all of the elementary visual features of artworks are processed like all other visual objects; second, attentional processes redirect information processing to salient visual properties, such as color, shape, and composition, by means of the fronto-parietal attentional network; third, the attentional network modulates processing within the ventral visual stream. Specifically, the content of artworks (i.e., landscapes, portraits, etc.) is processed by the attributional areas of the ventral visual stream, such as the parahippocampal place area (for landscapes) or the fusiform face areas (for portraits). Fourth, feed-back and feed-forward processes, linking attentional and attributional circuits, enhance the experience of the visual stimuli. Finally, in most cases emotional systems are also involved. The notion that content-dependent brain regions are involved in esthetic experience has repeatedly been proved [2,16], and recently these regions have been hypothesized to serve as a neural trigger for pervasive effects of esthetic experience [16], which is processed by a wider neural network, including also regions involved in judgment and decision making and emotional processing [17]. Cela-Conde et al. [17] proposed that the whole esthetic experience mainly consists of two distinct cognitive events, which take place at different time spans: (a) an initial general appraisal of the esthetic qualities (i.e., the perception of a visual stimulus as beautiful or not), which the authors called “esthetic appreciation sensu stricto”; and (b) a delayed appraisal of detailed aspects of the esthetic experience (i.e., whether it is interesting or original), which the authors called “esthetic appreciation sensu lato”. The esthetic appreciation sensu stricto was mainly related to a network of areas encompassing occipital and frontal regions, while the esthetic experience sensu lato mainly corresponded to the activation of the default mode network [17]. These results, together with others [18], support the idea that the whole esthetic experience is the results of the coordination of different cognitive processes, which involve different brain regions [18]. In line with these data, fMRI evidence suggests that the whole esthetic experience is associated with the activation of interconnected regions (i.e., frontal cortex, precuneus and posterior cingulate cortex), which mainly match the default mode network [19]. Zeki [20] hypothesized that studying ambiguity allows for highlighting different brain stations accountable for the microconsciousness for an attribute and their interactions. Perceptual ambiguity, that is, the quality of perceptual stimuli of being open to more than one interpretation, is one of the tricks artists use to evoke an esthetic experience in observers and it is commonly a hallmark of great artworks [20]. It is well known that ambiguity leads to a perceptual oscillation [21] from one percept to another [22,23], thus resulting in distributed micro-consciousness, both in time and space [20]. Indeed, neuroimaging studies have repeatedly demonstrated a transient brain activity change during perceptual reversals of non-artistic ambiguous stimuli [22,23]. In particular studies assessing perceptual reversal in ambiguous figures, such as the Rubin’s face-vase illusion, found higher activation

in the fusiform face area (FFA) during the spontaneous vase-tofaces reversal [23]. Jakesch et al. [24] found that ambiguity is an essential ingredient in esthetic appreciation of artworks. These authors using Magritte’s ambiguous paintings found that ambiguous paintings were preferred over non-ambiguous ones, even if they were found to be harder to process. They also found that, in watching abstract paintings, a moderate level of ambiguity is associated with a higher level of experienced pleasure and interest [25]. The authors concluded that this effect is mainly due to the involvement of higher esthetic judgment. Using Arcimboldo’s ambiguous portraits, Boccia et al. [26] recently found that perceptual ambiguity interacted with the individual predisposition toward the artworks to give rise to the esthetic experience. In perceiving ambiguity, under strong constraints, one interpretation may stand out and determine the whole percept [21]. In the present study, we hypothesized that the esthetic frame in perceiving art, acting as a strong constraint, biases cognitive processing by means of top-down mechanisms and thus affects the perceptual organization of ambiguous artworks, resulting in a different pattern of activity in content-dependent brain regions [3], as previously demonstrated for non-artistic ambiguous illusions [22,23]. Corollary to this hypothesis is the possibility that a specific neural mechanism underlies the interaction between the esthetic state of mind and ambiguous percepts. Starting from Zeki’s hypothesis [20], which posited that ambiguity allows for highlighting different brain stations accountable for the micro-consciousness for an attribute and their interactions, we used in our study Arcimboldo’s ambiguous portraits (also used in our recent study [26]), which are characterized by the presence of part-whole ambiguity. In Arcimboldo’s artworks more than one interpretation of the work is possible, since he disguised faces assembling them from objects. Indeed, each of his paintings could be interpreted in two ways, as an array of objects or as a face. In the part-whole ambiguity, each object has a perceptual meaning per se and is combined with others to create a different pattern with a different meaning (i.e., a human face). Attentional shifts might take place between the perception and interpretation of the parts of the painting and the whole portrait [22]. We used functional magnetic resonance imaging (fMRI) to measure local cerebral activation while participants performed either a like/dislike esthetic judgment (AeJ) or an artistic/non-artistic classification task (ClT) on a set of artistic and non-artistic pictures depicting ambiguous or non-ambiguous portraits (Fig. 1A and B). Due to the use of portraits, we expected to find an activation of the occipito-temporal network of face processing, which includes the medial portion of the inferior occipital lobe, the fusiform gyrus and the inferior temporal cortex [27]. Furthermore, based on the results of previous studies [22,23], we expected that ambiguous pictures would lead to higher activation of these brain areas [22,23,27] as well as, due to the presence of whole-part ambiguity, of areas related to visuo-constructive abilities and feature binding, such as the inferior parietal lobule [28–30]. In particular we searched for interactions between the task (AeJ vs. ClT) and the presence of ambiguity, and between esthetic valence (liked vs. disliked pictures) and the type of pictures (artistic vs. non artistic, ambiguous vs. non ambiguous).

2. Materials and methods 2.1. Subjects Twenty healthy right-handed subjects (mean age 25.45 and S.D. 4.513; 9 females) without a history of neurological and psychiatric disorders took part in this study. None had artistic expertise, as shown by their responses to the art questionnaire [31]. All subjects

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

149

Fig. 1. Stimuli and task. (A) Exemplars of the four categories of pictures used in the experiment: ambiguous artworks (AA), non-ambiguous artworks (nAA), ambiguous nonartworks (AnA), and non-ambiguous non-artworks (nAnA). (B) Timeline of the experiment. Each block starts with a written instruction about the next task to be performed (esthetic judgment or classification task), followed by ten trials, two from each of the four categories plus two scrambled images.

gave their written informed consent. This study was approved by the local ethical committee of Santa Lucia Foundation in Rome, in agreement with the Declaration of Helsinki. 2.2. Stimuli The study had a 2 × 2 × 2 factorial design, with two tasks and four categories of stimuli organized along the two orthogonal dimensions of Art and Ambiguity (Fig. 1A and B). Thirty-two ambiguous artworks (AA) were chosen from Arcimboldo’s collection. These stimuli presented a whole-part kind of perceptual ambiguity. In these portraits, the artist disguised the faces by assembling them from objects (e.g. books, fruit, etc.) (see Fig. 1A and B). To resolve the ambiguity in Arcimboldo’s paintings requires perceptually shifting from objects to faces, thus from parts to the whole, from details to global shape. Thirty-two Renaissance portraits of the same period of Arcimboldo paintings, matched for gender and face position with Arcimboldo’s stimuli, were chosen for the category of non-ambiguous artworks (AnA). Unlike Arcimboldo’s works, these portraits have no ambiguous features. In addition to these artistic stimuli, we included two categories of non-artistic stimuli: 32 photos of people, matched for gender and face position, without any ambiguous characteristics (non-artistic, non-ambiguous stimuli – nAnA), and 32 hidden face pictures with the same kind of ambiguity (whole-part) as Arcimboldo’s portraits (non-artistic, ambiguous stimuli – nAA). Stimuli were 500 per 750 pixels (width and height), subtending 10 degrees of visual angle horizontally, and were presented on a black background. Starting from this set of stimuli, we also created 32 scrambled images (S) to control for early stages of visual processing. 2.3. Procedure Each participant underwent four fMRI scans, each consisting of eight blocks of ten pictures. Stimuli were presented in a pseudorandomized sequence across blocks and scans in the two tasks,

and the same stimulus was never repeated in the following scan. Tasks were alternated in each session, that is, between four blocks of esthetic judgment (AeJ) and four blocks of classification (ClT). In the AeJ, we asked subjects whether or not they liked each picture, based on their esthetic preferences. In the ClT, we asked subjects to judge (using their common sense) whether each picture could or not be shown in an art gallery. Each block began with written instructions about the next task to be performed (Fig. 1B): the string “like?” indicated the AeJ, and the string “gallery?” indicated the ClT. The instructions were followed by ten trials, two from each of the four categories, plus two scrambled images, arranged in a pseudo-random sequence. Subjects answered by using an fMRI compatible keyboard with two buttons (for “yes” or “no”). For the AeJ, we computed the proportion of positive preferences (i.e., “yes” responses) made by the subjects for each stimulus category. For the ClT, we computed the proportion of correct responses, that is, when subjects correctly accepted artworks or correctly rejected non-artworks.

2.4. Image acquisition A Siemens Allegra scanner (Siemens Medical System, Erlangen, Germany), operating at 3 T and equipped for echo planar imaging (EPI), was used to acquire functional magnetic resonance images. Head movements were minimized by mild restraint and cushioning. Functional MRI images (EPI) were acquired for the entire cortex using BOLD signal imaging (38 slices, in plane resolution = 3 × 3 mm, slice thickness = 3.75 mm, repetition time [TR] = 2.47 s, echo time [TE] = 30 ms, flip angle = 70 degree). For each of the four fMRI scans, we acquired 198 fMR volumes. We also acquired a three-dimensional high-resolution T1-weighted structural image for each subject (Siemens MPRAGE, 176 slices, in-plane resolution = 0.5 × 0.5 mm, slice thickness = 1 mm, TR = 2 s, TE = 4.38 ms, flip angle = 8 degree). Image analysis was performed using SPM8 (http://www.fil.ion.ucl.ac.uk/spm).

150

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

Table 1 Activation clusters (PFDR < 0.001; cluster size >10 voxels), hemisphere, MNI coordinates and cluster size (mm3 ) and effects. Notes: H = hemisphere; task = esthetic judgment > classification task; ambiguity = ambiguous > non-ambiguous stimuli; 2nd level interaction = task per ambiguity interaction, with regions with higher activation for esthetic judgment of ambiguous stimuli; like = like > dislike stimuli; 3rd level interaction = preference per art per ambiguity interaction, with regions with different pattern of activation for artistic and non artistic ambiguous stimuli in relation to the esthetic preference (IFG = inferior frontal gyrus; SMA = supplementary motor area; MFG = middle frontal gyrus; SFG = superior frontal gyrus; FG = fusiform gyrus; LG = lingual gyrus; ITG = inferior temporal gyrus; PHG = parahippocampal gyrus; IOG = inferior occipital gyrus; MOG = middle occipital gyrus; SPL = superior parietal lobule; IPL = inferior parietal lobule). Volume (mm3 )

Region

H

MNI coordinates x

y

z

Lateral MFG Insula SFG IPL MFG, anterior cingulum FG, PHG SMA, SFG SFG, anterior cingulum IFG IFG, orbitofrontal cortex SPL MOG IOG, MOG FG, LG, ITG Anterior IFG Middle cingulum PHG SMA, SFG Orbitofrontal SFG, anterior cingulum IFG Supramarginal gyrus, IPL IOG, FG, MOG SPL MOG FG, LG, ITG

L L L L L L L L L L L L L L R R R R R R R R R R R R

−33 −27 −21 −51 −3 −33 −6 −3 −45 −45 −27 −30 −42 −27 54 3 21 6 39 3 51 48 45 24 36 30

23 14 44 −34 44 −31 14 59 8 26 −58 −85 −79 −49 41 23 −31 20 32 59 11 −31 −79 −67 −82 −46

49 −17 43 49 −17 −23 64 22 28 −5 58 22 −8 −11 10 34 1 61 −17 16 28 49 −8 55 16 −11

Significant effects (p < 0.05) Task

243 702 2835 3591 4050 4320 4644 7452 7803 10,125 12,744 14,634 15,930 20,925 594 1404 1431 1620 2079 5103 6696 7128 12,069 13,230 13,581 21,951

2.5. Image analysis The first four volumes were discarded to allow for T1 equilibration. All images were corrected for head movements using the first volume as a reference. Participants’ EPI images were co-registered onto their T1 image, normalized to the standard MNI-152 EPI template using the main realigned image as a source, and spatially smoothed using an 8-mm-FWHM isotropic Gaussian kernel. Functional images were analyzed in each subject separately on a voxel-by-voxel basis according to the general linear model. In a first model, separate regressors were included for each combination of task (AeJ, ClT) and picture type (AA, AnA, nAA, nAnA, S). In a second model, we also modeled trials within AeJ as a function of the esthetic judgments (like/dislike) given in individual trials. Group analysis was performed on parameter estimate images that resulted from the individual models, treating subject as a random factor. We first computed an omnibus F-contrast by comparing each picture type (within each task) vs. the respective scrambled condition to subtract the earliest stages of visual processing. The resulting statistical parametric map was thresholded at p < 0.001 (corrected for multiple comparisons using false discovery rate [32], and cluster size >10 voxels). The voxels that emerged from the omnibus F-test were grouped into regions, that is, clusters of adjacent significant voxels, according to their anatomical localization (Table 1). For each subject and region, we computed a regional estimate of the amplitude of the hemodynamic response in each experimental condition by entering a spatial average (across all voxels in the region) of the pre-processed time series into the individual general linear models. The rationale for this choice was to take in account in the following analyses all the regions that were related to observing our stimuli, in the different tasks, over and above the effect of mere visual stimulation. These regional hemodynamic response estimates were then analyzed with two repeated measures 2 × 2 × 2 analyses of variance. The first analysis aimed to explore the combined effect of

Ambiguity

2nd level interaction

Like

3rd level interaction

+ + +

+ +

+ + + + + + + + + +

+ + + + + + + + + +

+ +

+ + + + + +

+ +

+ + +

+ + + + + +

+

+ +

Art and Ambiguity on esthetic judgments; it was framed as Task (AeJ, ClT) by Art (art, non-art) by Ambiguity (ambiguous, nonambiguous). The second analysis focused on AeJ trials and aimed to explore the effect of subjective esthetic preferences; it was framed as esthetic judgment (like, dislike) by Art by Ambiguity. 3. Results Participants liked on average 41.25% (S.D. 32.04%) of AA and 40.16% (S.D. 28.19) of AnA; non-artistic stimuli were judged positively on average 58.91% (S.D. 25.95%) of nAA and 43.28% (23.37%) of nAnA. In ClT, subjects correctly identified as artworks on average 78.91% (S.D. 33.93%) of AA and 88.44% (S.D. 16.94%) of AnA. They correctly rejected non-artistic stimuli on average 85.31% (S.D. 11.87%) and 93.44% (S.D. 13.79%) of nAA and nAnA, respectively. The comparison between pictures (within each task) vs. the respective scrambled condition showed activation in a widespread network of areas spanning from the occipital to the frontal lobe (see Table 1). Interestingly, this network includes areas of face processing on the medial part of the occipito-temporal cortex, such as fusiform gyrus, inferior occipital gyrus, as well as frontal and parietal subregions (Table 1). We first explored effects due to Task, Art and Ambiguity within this network of areas. We found that, compared to the classification task, esthetic judgments activated the orbito-frontal cortex, insula, supplementary motor area, superior and inferior frontal gyrus in the left and the middle cingulum in the right hemisphere, as well as the bilateral anterior middle cingulum (Fig. 2A). Furthermore, we found that the network of areas of face processing [27], which encompasses the inferior and middle occipital gyrus, fusiform gyrus, lingual gyrus, and inferior temporal gyrus in both hemispheres, were more activated by AeJ than ClT, suggesting a content-dependent activation for the esthetic evaluation of faces. When we explored the effect of ambiguity, we found that

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

151

Fig. 2. Effects of task, artistic nature and ambiguity on brain activation. (A) Areas with higher activation for esthetic judgment compared with the classification task are shown in red-to-yellow patches. (B) Areas with higher activation for ambiguous (vs. non-ambiguous) artistic and non artistic pictures are shown in blue-to-light-blue patches. (C) Areas showing a task by ambiguity interaction are shown in red-to-light blue patches. (D) Percent BOLD signal changes in the eight conditions in the right superior parietal lobule (SPL-RH). Notes: LH = left hemisphere; RH = right hemisphere. (For interpretation of the references to color in this figure legend, please see the web version of this article).

processing ambiguous vs. non-ambiguous stimuli (Fig. 2B) also yielded to higher activation of the areas of face processing, as well as parietal and frontal areas. Indeed we found that the occipitotemporal network of face processing [27] was more activated by ambiguous than by non ambiguous stimuli, both artistic and non-artistic, suggesting that these areas are more involved when individuals processed an ambiguous face. Also the right inferior frontal gyrus and the bilateral superior and inferior parietal lobule were more activated by ambiguous than non-ambiguous stimuli. More interestingly, we found a second level task by ambiguity interaction in the right superior parietal lobule (Fig. 2C). Post hoc analysis (Bonferroni test) of this interaction revealed that the right superior parietal was modulated by the task (being more activated by AeJ than by ClT) but only for ambiguous pictures (Fig. 2D). We then analyzed brain responses as a function of the esthetic valence (liked or disliked), as resulting from the esthetic judgments given online by each individual to each picture during the AeJ task. This analysis had the advantage of evaluating esthetic valence within the same category of pictures, thus avoiding confounding factors such as differences in complexity and heterogeneity between stimuli of different categories (i.e., ambiguous and non ambiguous artworks and ambiguous and non ambiguous non-artistic pictures). Positive esthetic experience, compared with negative esthetic experience (i.e., liked vs. disliked pictures), yielded neural activation in a set of frontal areas, such as the orbitofrontal cortex, insula, supplementary motor area and anterior cingulum, as well as in the occipito-temporal areas of face processing (i.e., the fusiform gyrus; [27]) and the parahippocampal cortex. We did not find any effect for the opposite contrast (i.e., dislike > like pictures).

Interestingly, we found a third-level Art by Ambiguity by Judgment in the medial occipito-temporal areas of face processing, i.e., the inferior temporal gyrus, the fusiform and the parahippocampal gyri (Fig. 3). These regions evidenced higher activation when subjects disliked (vs. liked) Arcimboldo’s ambiguous artworks and when subjects liked (vs. disliked) non-ambiguous artworks or liked both ambiguous and non-ambiguous non-artworks. Thus, this network of areas showed an intriguing neural mechanism that appears to subtend the valence of esthetic experience in watching artistic and non-artistic ambiguous portraits (i.e., Arcimboldo’s Artworks). 4. Discussion Starting from previous findings in the field of neuroaesthetics [1–7], in the current study we hypothesized the existence of a neural mechanism underlying the interaction between the esthetic state of mind and ambiguous percepts, and that the esthetic experience modulates the neural network involved in processing ambiguous portraits (i.e. Arcimboldo’s artworks). We found that esthetic experience is related to activation in a network of areas including, other than the “face-processing” occipito-temporal network, also bilateral frontal areas, which have already been hypothesized to be involved in esthetic appreciation [33–36]. The role of these areas in esthetic experience can be fully appreciated considering their involvement in emotional labeling of sensory information [37]. Evaluating sensory information, such as making an esthetic judgment about visual stimuli, requires the interaction between cognitive and emotional processes [15,37]. Anterior cingulate cortex and orbitofrontal cortex are examples of brain sites where cognition and emotion interact to allow

152

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

Fig. 3. Effects of subjective judgment, artistic nature and ambiguity on brain activation. (A) Areas showing third-level interaction (Art by Ambiguity by Preferences) are displayed in red-to-yellow patches. (B) Percent BOLD signal changes in the eight conditions within the regions showing third-level interaction (Art by Ambiguity by Preferences). The fusiform gyri (FG), parahippocampal gyri (PHG), lingual gyrus (LG) and inferior temporal gyrus (ITG) were less activated when subjects liked (vs. disliked) Arcimboldo’s artworks and were more activated when subjects liked (vs. disliked) non-ambiguous portraits, ambiguous and non-ambiguous non-artworks. Notes: LH = left hemisphere; RH = right hemisphere. (For interpretation of the references to color in this figure legend, please see the web version of this article).

evaluating sensory information [37,38]. In particular, orbitofrontal cortex directly receives input from late visual areas (such as that along the inferior temporal cortex). Thus, orbitofrontal is extremely well positioned to tune perceptual processing in visual areas, and allow emotional evaluation of stimuli [37]. In our study, it is possible that the frontal regions we found to be involved in esthetic judgment, such as the anterior cingulate cortex and orbitofrontal cortex, interact with FFA in the esthetic evaluation of portraits, mainly processing the general aspects of esthetic evaluation. And very interestingly, we also found that activation in areas related to esthetic judgment was enhanced when the perceived stimulus received a positive esthetic judgment. It is noteworthy that the occipito-temporal areas involved in face processing resulted more activated during esthetic judgments than during classification task. This result supports the hypothesis of a contentdependent effect of esthetic appreciation, already advanced by Kawabata and Zeki [2], who found that esthetic appreciation of different categories of paintings was associated with distinct and specialized visual areas of the brain. For example, the esthetic evaluation of landscapes [2] is connected with the modulation of brain areas involved in place processing [39,40], while the disruption of extrastriate body area by means of repetitive transcranial magnetic stimulation reduces observers’ esthetic sensitivity of body stimuli [41]. The regions of face processing were also more activated in relation to ambiguous portraits than to non-ambiguous ones, possibly due to the perceptual shift occurring between the parts and the faces in ambiguous stimuli [22]. Furthermore, we found that neural activity in the right superior parietal lobe showed a Task by Ambiguity interaction, being more activated when individuals performed the esthetic judgment of ambiguous portraits. This result strongly confirms the existence of a neural mechanism underlying the interaction between the esthetic state of mind and ambiguous percepts, as previously suggested by Cupchik et al. [3], supporting the hypothesis that esthetic experience arises from the interaction between an individual predisposition toward the artworks (i.e., esthetic or classification

tasks) and perceptual facilitation (i.e., ambiguous or non ambiguous portraits). When we analyzed the main effect of the valence of esthetic experience (like vs. dislike) we found that the neural network of esthetic experience, revealed by the comparison between AeJ and ClT, was more activated by positive esthetic experience (like > dislike), while the opposite contrast (dislike > like) failed to reach significance. This is in line with results of previous studies. For example, Kawabata and Zeki [2] found that stimuli judged as beautiful or ugly modulated the neural activity in the same cortical areas and that the modulation of activity within those areas correlated with the judgment of a stimulus as being beautiful or not [2]. They also suggested that neural activity in each of these areas reflects the judgmental category. Surprisingly, the comparison between the esthetic appreciation (like vs. dislike) of artistic and of non-artistic ambiguous portraits (Preference by Art by Ambiguity interaction) showed that the same neural substrates process different types of esthetic appreciation in different categories of stimuli. Namely, the fusiform and parahippocampal gyrus were not only more active for disliked than for liked ambiguous artistic stimuli (i.e. Arcimboldo’s paintings), but also for liked than for disliked pictures of all other categories. In other words, ambiguous artworks eliciting a negative esthetic experience lead to more pronounced activation of regions of face processing [27], while ambiguous artworks eliciting a positive esthetic experience lead to reduced activation of these areas. This result is intriguing for at least three reasons: first, it showed that a specific neural mechanism located in content-dependent areas, namely the fusiform gyrus and inferior occipito-temporal cortex, devoted to face processing, contributes to the esthetic appreciation of artistic and non-artistic ambiguous portraits. Second, it indicates that the same cortical areas subtend both positive and negative esthetic experiences, but with different patterns, so that the pattern of activity in these regions predicts the category of pictures. Indeed, this is particularly true for the ambiguous artworks which yield a pattern of activity opposite to that yielded by all of the other categories. Similarly, Kawabata and Zeki [2] found

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

that the perception of different categories of paintings was associated with distinct and specialized visual areas (content-dependent areas) and that the modulation of activity within the same areas correlated with the judgment of a stimulus as being beautiful or not. Third, we found within these content-dependent areas (i.e. fusiform gyrus and inferior occipito-temporal areas) an interaction between individuals’ esthetic preference, Art and Ambiguity, suggesting that the content-dependent areas activity contribute to the arising of esthetic experience, together with a broader neural network, including areas generally involved in esthetic experience (as discussed before) [3]. Indeed, if the content-dependent regions serve as a neural trigger for pervasive effects of esthetic experience [16], it is possible that they are only the first step toward the whole esthetic experience of both ambiguous and non-ambiguous artworks. Thus, the whole esthetic experience results to be associated with the activation of a wider neural network, which includes also regions that process the more general aspect of esthetic experience, i.e., the anterior cingulate cortex and orbitofrontal cortex. This observation is in line with the twofold model of esthetic experience, proposed by Cela-Conde and colleagues [17]: indeed, it is possible that the distinct pattern we found in the fusiform gyrus and the inferior occipital-temporal cortex mainly correspond to the esthetic appreciation sensu stricto, as proposed by Cela-Conde et al. [17], while the general pattern of activation we found in frontal regions corresponds to the esthetic appreciation sensu lato. Overall, our results suggest that while specific aspects of esthetic experience (for example, the type of artworks) are processed within content-dependent brain regions (such as FFA), general aspects of esthetic evaluation are processed within a frontal network of areas, which includes orbitofrontal and anterior cingulate cortices, both of which are strictly involved in emotional labeling of sensory inputs [37]. The neural mechanism we found in the bilateral fusiform and parahippocampal gyrus (i.e. Preference per Art per Ambiguity interaction) deserves some more speculative considerations. We found that neural activity in each of these areas reflects the judgmental category, being these areas more activated when participants disliked Arcimboldo’s portraits and when they appreciate all the other categories. This result may suggest that in appreciating an artistic ambiguous stimulus (i.e. Arcimboldo’s) a weaker involvement of regions of face processing is required. We suggest that this pattern of activation means that esthetic appreciation leads to holistic perception of the artworks, so that solving ambiguity (i.e., to see the face in place of the array of objects) becomes unnecessary because of the simultaneous acceptance of part and whole. In this light, esthetic experience of artworks can act as a strong constraint in perceiving ambiguity and can lead to a completely coherent percept only when a positive esthetic experience is elicited, making a unique interpretation of the percept unnecessary and leading to simultaneous acceptance of part and whole. On the other hand, a constellation of features not requiring or forcing a unique interpretation could have provoked the positive feelings toward Arcimboldo’s work; indeed, positive feelings can arise from the simultaneous acceptance of two different interpretations. When effort is required to solve the ambiguity by accepting only one of the two possible interpretations (i.e. the face), a negative feeling is elicited. The different pattern of activations observed in these areas for the other stimuli can be interpreted as follows. Indeed, we found that the esthetic judgment (like vs. dislike) modulates the activity of the fusiform and parahippocampal areas according to the valence of elicited feelings also during watching artistic non-ambiguous as well as non-artistic ambiguous and non-artistic non-ambiguous stimuli, with an opposite pattern of activity to that elicited by watching Arcimboldo’s. Specifically, positive feelings lead to higher activation of these areas when non ambiguous artistic and non

153

artistic portraits are presented; the same effect is observed when an ambiguous non artistic face is presented, showing that the positive appreciation of the “surprising” discovery of a portrait in a fac¸ade is processed when the perceptual oscillation is solved by recognizing a face in it (i.e., activating the fusiform face area). Some caution is required in interpreting and discussing our results. In particular, one could argue that our stimulus set was complex and highly heterogeneous, so that differences in activations induced by different picture categories might be due to factors other than those we were interested in. The problem of complexity is a classical methodological issue in experimental psychology, whose unique and definite solution is still a matter of debate. This issue is critical in the field of neuroaesthetics, since artworks are not elementary stimuli and matching stimulus features to study specific aspects of esthetic pleasure is very difficult. Thus, different studies have adopted different strategies to deal with this issue. For example, Di Dio et al. [1] compared canonical and modified sculptures to investigate whether there is an objective, biological basis for the experience of beauty in art. They modified just one feature in their modified sculptures, i.e., the proportion, which is their independent variable. In the context of present study, it is impossible to modify just one feature to control for whole-part ambiguity in Arcimboldo’s. Thus, in order to investigate whether esthetic pleasure raised from viewing an artwork affects the perception of the whole-part ambiguity in Arcimboldo’s portraits, we included non-artistic portraits with the same kind of whole-part ambiguity in our set of stimuli. We also included Renaissance portraits and non-artistic photos to control for both the esthetic pleasure in absence of ambiguity and the perception of a face. We matched the most relevant features of the stimuli across categories (e.g., gender and face position) and, at the same time, included pictures with considerable variability also within each category. This assured that any observed difference was related to the dimensions of interest. Second the core of our results is based on comparisons within each stimulus category. In particular, we compared esthetic judgments vs. non-esthetic classifications of the same artworks and non-artistic pictures. Such a comparison implicitly controls for stimulus differences. Similarly, we compared liked vs. disliked items within each category. Thus, we were reasonably sure that the observed differences were due to the investigated variables and not to confounding factors, such as different degree of complexity and heterogeneity between stimuli. 5. Conclusion In sum, our results demonstrate that a neural mechanism in the content-dependent brain regions of face processing contributes to the esthetic experience of ambiguous portraits (i.e., Arcimboldo’s works). Furthermore, they suggest that esthetic experience interacts with perceptual qualities of stimuli in the right superior parietal lobe, supporting previous conclusions, which posited that esthetic experience arises from the interaction between topdown orienting of attention (i.e. esthetic or pragmatic state of mind) and bottom-up perceptual facilitation (perceptual qualities of the stimuli). The complexity of our paradigm, as well as the uniqueness of the stimuli used in our study, makes it impossible to draw definitive conclusions about the role of esthetic pleasure in ambiguous artworks, and prevent any possible generalization of our results to other aspects of esthetic experience and other categories of paintings. Furthermore, it has to be highlighted that different types of ambiguity can be found in artworks. One of the higher levels of ambiguity in art is represented by the multiple narrative interpretations that can be given to a masterpiece, even if in presence of stable picture (see for example The Pearl Earring, by Johannes Vermeer) [20]. Anyway, some artists (see for example works by Giuseppe

154

M. Boccia et al. / Behavioural Brain Research 278 (2015) 147–154

Arcimboldo or Salvador Dali) deliberately made of ambiguity an artistic form. In our study we restricted our observations to a specific type of ambiguity, that is the perceptual part-whole ambiguity of Arcimboldo’s artworks. This allowed us to test for our experimental hypothesis without spurious perceptual effects, but at the same time it restricts any possible generalization of our results to other types of perceptual complexity and ambiguity in art. Anyway, the present study, by comparing esthetic experiences elicited by part-whole ambiguity with that elicited by non-ambiguous stimuli, not only allowed us to begin to disentangle the role played in the neuroaesthetic networks by different areas, but also opened novel routes toward new areas of interest in the field of neuroaesthetics. For example, it could be of interest to assess whether other content-dependent regions (i.e., parahippocampal place area) show for ambiguous landscapes the same pattern of activity we found for ambiguous portraits. Thus, further investigations are necessary to shed more light upon this topic. Finally, whether the contentdependent brain regions serve as a neural trigger for pervasive effects of esthetic experience, it could be of interest to better understand what happens in the following processing stages of the esthetic experience when their activity is biased by the perceptual uncertainty. Acknowledgment We thank M.B. who gave written informed consent for publication of his photograph. References [1] Di Dio C, Macaluso E, Rizzolatti G. The golden beauty: brain response to classical and renaissance sculptures. PLoS ONE 2007;11:e1201. [2] Kawabata H, Zeki S. Neural correlates of beauty. J Neurophysiol 2004;91:1699–705, http://dx.doi.org/10.1152/jn.00696.2003. [3] Cupchik GC, Vartanian O, Crawley A, Mikulis DJ. Viewing artworks: contributions of cognitive control and perceptual facilitation to aesthetic experience. Brain Cogn 2009;70:84–91. [4] Nadal M. The experience of art: insights from Proc. R. Soc. neuroimaging. Progr Brain Res 2013;204:135–58. [5] Nadal M, Pearce MT. The Copenhagen Neuroaesthetic Conference: perspectives and pitfalls for an emerging field. Brain Cogn 2011;76:172–83. [6] Vartarian O, Skov M. Neural correlates of viewing paintings: evidence from a quantitative meta-analysis of functional magnetic resonance imaging data. Brain Cogn 2014;87:52–6. [7] Cela-Conde CJ, Agnati L, Huston JP, Mora F, Nadal M. The neural foundations of aesthetic appreciation. Progr Neurobiol 2011;94:39–48. [8] Di Dio C, Gallese V. Neuroaesthetics: a review. Curr Opin Neurobiol 2009;19:682–7. [9] Kant I (J.H. Bernard, Trans.) Critique of judgment. London: MacMillan; 1914 (original work published, 1790). [10] Schopenhauer A (E.J. Payne, Trans.) The world as will and representation (Vol. 1) (E.J. Payne). New York: Dover; 1969 (original published, 1818). [11] Cupchik GC. From perception to production: a multilevel analysis of the aesthetic process. In: Cupchik GC, Laszlo J, editors. Emerging visions of the aesthetic process: psychology, semiology, philosophy. New York: Cambridge University Press; 1992. p. 83–99. [12] Cupchik GC, Winston AS. Confluence and divergence in empirical aesthetics, philosophy, and mainstream psychology. In: Carterette EC, Friedman MP, editors. Handbook of perception & cognition, cognitive ecology. San Diego, CA: Academic Press; 1996. p. 62–85.

[13] Markovic S. Components of aesthetic experience: aesthetic fascination, aesthetic appraisal, and aesthetic emotion. i-Perception 2012;3:1–17. [14] Mamassian P. Ambiguities and conventions in the perception of visual art. Vis Res 2008;48:2143–53. [15] Chatterjee A. Prospects for a cognitive neuroscience of visual aesthetics. Bull Psychol Arts 2004;4:55–60. [16] Chatterjee A, Thomas A, Smith SE, Aguirre GK. The neural response to facial attractiveness. Neuropsychology 2009;23:135–43. [17] Cela-Conde CJ, García-Prieto J, Ramasco JJ, Mirasso CR, Bajo R, Munar E, et al. Dynamics of brain networks in the aesthetic appreciation. Proc Natl Acad Sci U S A 2013;110:10454–61. [18] Munar E, Nadal M, Castellanos NP, Flexas A, Maestú F, Mirasso C, et al. Aesthetic appreciation: event-related field and time-frequency analyses. Front Hum Neurosci 2012;5:185, http://dx.doi.org/10.3389/fnhum.2011.00185. [19] Vessel EA, Starr GG, Rubin N. The brain on art: intense aesthetic experience activates the default mode network. Front Hum Neurosci 2012;6:66, http://dx.doi.org/10.3389/fnhum.2012.00066. [20] Zeki S. The neurology of ambiguity. Conscious Cogn 2004;13:173–96. [21] Yevin I. Ambiguity in art. Complexus 2006;3:74–82. [22] Kleinschmidt A, Bi?chel C, Zeki S, Frackowiak RSJ. Human brain activity during spontaneously reversing perception of ambiguous figures. Proc R Soc Lond 1998;265:2427–33. [23] Andrews TG, Schluppeck D, Homfray D, Matthews P, Blakemore C. Activity in the fusiform gyrus predicts conscious perception of Rubin’s vase–face illusion. NeuroImage 2002;17:890–901. [24] Jakesch M, Forster M, Leder H. Image ambiguity and fluency. PLOS ONE 2013;8(9):e74084. [25] Jakesch M, Leder H. Finding meaning in art: preferred levels of ambiguity in art appreciation. Quart J Exp Psychol 2009;62:2105–12. [26] Boccia M, Barbetti S, Margiotta R, Guariglia C, Ferlazzo C, Giannini AM. Why do you like Arcimboldo’s portraits? Effect of perceptual style on aesthetic appreciation of ambiguous artworks. Atten Percept Psychophys 2014, http://dx.doi.org/10.3758/s13414-014-0739-7. [27] Haxby JV, Gobbini MI, Furey ML, Ishai A, Schouten JL, Pietrini P. Distributed and overlapping representations of faces and objects in ventral temporal cortex. Science 2001;293(5539):2425–30, http://dx.doi.org/10.1126/science.1063736. [28] Watson CE, Chatterjee AA. A bilateral frontoparietal network underlies visuospatial analogical reasoning. Neuroimage 2012;59(3):2831–8. [29] Amorapanth PX, Widick P, Chatterjee A. The neural basis for spatial relations. J Cogn Neurosci 2010;22(8):1739–53. [30] Shafritz KM, Gore JC, Marois R. The role of the parietal cortex in visual feature binding. PNAS 2002;99(16):10917–22. [31] Chamorro-Premuzic T. Personality, intelligence, and art. Personal Individ Diff 2004;36:705–15. [32] Genovese R, Lazar NA, Nichols T. Thresholding of statistical maps in functional neuroimaging using the false discovery rate. NeuroImage 2002;15:870–8. [33] Aharon I, Etcoff N, Ariely D, Chabris CF, O’Connor E, Breiter HC. Beautiful faces have variable reward value: fMRI and behavioral evidence. Neuron 2001;32(3):537–51. [34] O’Doherty J, Winston J, Critchley H, Perrett D, Burt DM, Dolan RJP. Beauty in a smile: the role of medial orbitofrontal cortex in facial attractiveness. Neuropsychologia 2003;41(2):147–55. [35] Fusar-Poli P, Placentino A, Carletti F, Landi P, Allen P, Surguladze S, et al. Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J Psychiatry Neurosci 2009;34(6):418–32. [36] Jacobsen T, Schubotz RI, Lea Höfel L, Cramonb DY. Brain correlates of aesthetic judgment of beauty. NeuroImage 2006;29:276–85. [37] Pessoa L. On the relationship between emotion and cognition. Nature Rev Neurosci 2008;9:148–58. [38] Rolls ET, Grabenhorst F. The orbitofrontal cortex and beyond: from affect to decision-making. Progr Neurobiol 2008;86(21):6–244. [39] Epstein R, Kanwisher N. A cortical representation of the local visual environment. Nature 1998;392:598–601. [40] Boccia M, Nemmi F, Guariglia C. Neuropsychology of environmental navigation in humans: review and meta-analysis of fMRI studies in healthy participants. Neuropsychol Rev 2014, http://dx.doi.org/10.1007/s11065-014-9247-8. [41] Calvo-Merino B, Urgesi C, Orgs G, Aglioti SM, Haggard P. Extrastriate body area underlies aesthetic evaluation of body stimuli. Exp Brain Res 2010;204:447–56, http://dx.doi.org/10.1007/s00221-010-2283-6.