Modulation of aesthetic value by semantic context: An fMRI study
Ulrich Kirka,b,⁎, Martin Skovc,1, Oliver Hulmea,b,1, Mark S. Christensenc,d, Semir Zekia,b
aWellcome Laboratory of Neurobiology, Anatomy Department, University College London, Darwin Building, Gower Street, London WC1E 6BT, UK
bWellcome Department of Imaging Neuroscience, University College London, Darwin Building, Gower Street, London WC1E 6BT, UK
cDanish Research Centre for Magnetic Resonance, Copenhagen University Hospital, Hvidovre, DK-2650, Denmark
dDepartment of Exercise and Sport Sciences, University of Copenhagen, The Panum Institute, DK-2200, Denmark
a b s t r a c ta r t i c l ei n f o
Received 14 April 2008
Revised 1 August 2008
Accepted 13 October 2008
Available online 29 October 2008
Aesthetic judgments, like most judgments, depend on context. Whether an object or image is seen in daily
life or in an art gallery can significantly modulate the aesthetic value humans attach to it. We investigated the
neural system supporting this modulation by presenting human subjects with artworks under different
contexts whilst acquiring fMRI data. Using the same database of artworks, we randomly labelled images as
being either sourced from a gallery or computer generated. Subjects' aesthetic ratings were significantly
higher for stimuli viewed in the ‘gallery’ than ‘computer’ contexts. This contextual modulation correlated
with activity in the medial orbitofrontal cortex and prefrontal cortex, whereas the context, independent of
aesthetic value, correlated with bilateral activations of temporal pole and bilateral entorhinal cortex. This
shows that prefrontal and orbitofrontal cortices recruited by aesthetic judgments are significantly biased by
subjects' prior expectations about the likely hedonic value of stimuli according to their source.
© 2008 Elsevier Inc. All rights reserved.
It is known, both anecdotally and experimentally, that aesthetic
value is influenced by context. Behavioural studies have shown that
presenting artworks accompanied with titles, text, and other forms of
cognitive information can significantly influence an observer's
reported evaluation of the artwork (e.g., Cupchik et al., 1994; Leder
et al., 2006; Russell, 2003). However, the neural sources of this context
specific modulation remain uncharted.
Recently, neuroimaging studies have investigated the neural
correlates of aesthetic evaluation. Experiments using musical
sequences as stimuli (Blood et al., 1999; Blood and Zatorre, 2001;
Brown et al., 2004; Menon and Levitin, 2005; Koelsch et al., 2006)
have found that post-hoc ratings of subjective pleasantness correlate
with activity in the striatum (dorsal and ventral), amygdala, para-
hippocampal gyrus, insula, orbitofrontal cortex (OFC), and the anterior
cingulate cortex (ACC). A similar finding was reported in two studies
using paintings as stimuli (Vartanian and Goel, 2004; Kawabata and
Zeki, 2004). The main effect of the subjects' aesthetic appreciation of
the images (as determined bysubjective ratings) resulted in activity in
the caudate nucleus, OFC and ACC. The network recruited during an
aesthetic judgment (when compared to a symmetry judgment)
includes enhanced activity in frontomedian cortex, lateral OFC,
inferior frontal gyrus, posterior cingulate, temporal pole, and
temporoparietal junction (Jacobsen et al., 2006). Together, these
studies suggest that subjective assessment of the aesthetic value of
works of art, engage a network of brain structures known to be
involved in the processing of reward, perceptual processing, and
The question we address here is whether explicit contextual in-
formation influences activity in specific part of this extended network
of areas. Since previous studies have demonstrated that activity in
the medial OFC is modulated by contextual information in experi-
mental conditions where subjects evaluate their preference for liquids
or odours, we hypothesized that the medial OFC would be the most
likely target of such contextual modulation. de Araujo et al. (2005)
demonstrated that human subjects will rate a test odour as sig-
nificantly more pleasant when it is paired with a pleasant visual word
as when paired with an unpleasant one. They observed a neural
correlate of this behavioural modulation in medial OFC. This shows
that high-level cognitive input such as word labels influence brain
activity in medial OFC. Similarly, McClure et al. (2004) investigated the
neural systems involved in generating preferences produced by two
different brands of soft drinks (Coca-Cola and Pepsi) and found that
the rated preference of unlabelled drinks (i.e. without cognitive
influences) was reflected in activations of the ventromedial part of the
prefrontal cortex (VMPFC), whereas brand knowledge modulated
dorsolateral prefrontal cortex, visual cortex, midbrain and hippocam-
pus. This result was recently extended by Plassmann et al. (2008) who
showed that knowledge of the monetary value of a wine increases
subjects' reports of preference, and reported a neural correlate of this
effect in medial OFC. However, no neuroimaging study has charted the
NeuroImage 44 (2009) 1125–1132
⁎ Corresponding author. Wellcome Laboratory of Neurobiology, Anatomy Depart-
ment, University College London, Darwin Building, Gower Street, London WC1E 6BT, UK
E-mail address: email@example.com (U. Kirk).
1These authors contributed equally to this work.
1053-8119/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/ynimg
neural correlates of contextual modulation of subjective preference
for visual stimuli, including visual works of art.
To test this hypothesis subjects were asked to aesthetically rate a
series of abstract paintings while in the scanner. We manipulated the
context under which subjects viewed the artworks by labelling 50% as
being from a prestigious gallery (‘gallery’ label) and 50% as being
generated by the experimenters using a computer program (‘compu-
ter’ label) (see Fig. 1). This prevented any systematic difference in
visual stimulation between the two conditions and thus controlled
only the prior expectation by changing the context. Our aim was to
test which areas were significantly modulated by context alone, and
which were modulated by both context and aesthetics ratings. Pre-
screeningof subjects and post-hoc testing confirmedthat subjects had
no prior knowledge or familiarity of the paintings.
We hypothesized that the context in which subjects viewed
artworks would significantly affect both their aesthetic evaluation in
terms of their behavioural rating as well as the neural processing
supporting this change. We base our hypothesis on the assumption
that the context associated with the ‘gallery’ label would induce a
higher expectation of reward than the context associated with the
‘computer’ label, and that this difference in reward expectation would
be reflected in the ratings and in the neural activity of the reward
Materials and methods
Fourteen subjects (5 females; mean age 26.3 years, age range 23–
29 years, 4 left-handed) participated in the study. They were all
undergraduate or graduate students and all had normal or corrected-
to-normal vision, and none had a historyof neurological or psychiatric
disorders. We excluded subjects with a formal education in any art-
related field in order to reduce the familiarity effect of the stimulus
class. Written informed consent from all subjects and ethical approval
(KF-01-131/03) were obtained before the experiment.
Visual chromatic reproductions of original abstract paintings
served as stimuli. In total 200 stimuli were selected from online
sources (see Supplementary material for a complete list of stimulus
material). Labels accompanied stimulus-presentation and were pre-
sented below each visual stimulus (see Fig. 1). We manipulated the
context in which the stimuli were presented by applying one of two
labels in different blocks: Half the stimuli were labelled as belonging
to the Louisiana Museum of Modern Art, Denmark, corresponding to
‘gallery’ conditions, while the other half were labelled as ‘computer’
(implying that the experimenters generated the stimuli through a
computer program, Photoshop). Prior to scanning, subjects were
instructed in the use of the aesthetic rating scale. No reference was
made to the labelling manipulation of the stimulus material. In
support of the experimental manipulation subjects were given the
following instruction: “Inside the scanner you will be presented with
200 abstract paintings. There are two types of paintings. 50% of the
paintings have been borrowed from the Louisiana Museum of Modern Art
(Copenhagen, Denmark). The other 50% have been generated by the
experimenter by means of a computer program (Photoshop)”. All
subjects had prior knowledge of the existence and prestige of
Louisiana Museum of Modern Art. All labels were counterbalanced
across subjects, such that visual parameters were balanced across
label conditions. Since systematic differences in visual stimulation
between the two conditions were controlled only the prior expecta-
tions of the subjects varied by changing the context.
Prior to scanning, stimuli were rated according to an aesthetic
rating scale in a behavioural pilot study by a separate cohort of
subjects (9 subjects; 3 females; mean age 25.3 years; age range 24–
28 years). Level of appeal was rated using a scale from 1 to 5, where 1:
“very unappealing” and5:“veryappealing”. Basedonthese results, we
found that the reports conformed to an approximately uniform
balanced distribution across the aesthetic rating scale.
The experimental protocol consisted of a block-design with 5
presentations in each block with identical labels making a total of 20
blocks for each context. A block design was used in order to minimize
the number of task switches required of subjects. On each trial, a
fixation cross appeared for 1000 ms on a grey background followed by
a stimulus presentation for 5000 ms. Following the appearance of
each stimulus, subjects were instructed, within the 5000 ms stimulus
Fig. 1. Displays experimental paradigm where a fixation cross was shown for 1 s
followed by a block of 5 presentations of ‘gallery’ labelled conditions (A) or ‘computer’
labelled conditions (B), which was presented for 5 s in which the subjects were
instructed to indicate the level of aesthetic appeal by means of button-press on a scale
from 5 (highest appeal) to 1 (lowest appeal). The inter-trial interval was 1 s (fixation
cross) between the presented images. (C) The experimental protocol consisted of a
block-design with 5 presentations in each block. The 5 trials within every block had
duration of 6 s (fixation+stimulus). The TR was 2400 ms. Labels were counterbalanced
across subjects. The stimuli were all non-canonical chromatic abstract artworks.
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
duration, to rate the stimulus (using the same scale as mentioned
above) by pressing one of five buttons on a key pad with their right
hand (Fig.1). Total scanning-time per subject was 20 min per session.
Post-scanning, subjects werepresented (outside the scanner) with the
stimuli again and asked to rate each stimulus on a 5-step familiarity
rating scale (1; not familiar – 5; very familiar). Familiarity ratings were
entered intothe design matrix as a parametric regressor of no interest.
The stimuli were presented at a screen resolution of 1024×
768 pixels displayed at a visual angle of 24×18°, and centred in a
500×500 pixel resolution surrounded by a grey background. Stimuli
were presented and responses collected using E-prime (Psychology
Software Tools, Inc.). The stimuli were back-projected via an LCD
projector onto a transparent screen positioned over the subjects'
head and viewed through a tilted mirror fixed to the head coil.
fMRI data acquisition
The functional imaging was performed with a 3 Tscanner (Siemens,
Magnetom Trio, Erlangen, Germany) to acquire gradient T2⁎ weighted
(BOLD) contrast (echo time, TE=30 ms; repetition time, TR=2400 ms;
flip angle, FA=90°). The EPI sequence was optimized in order to reduce
signal drop-out in OFC (Deichmann et al., 2003). Each functional image
was acquired in an interleaved way, beginning with 2nd slice (slice no.
2,4…40, 1,3…39) when counted from the bottom, comprising 40 axial
slices each 3.0 mm thick, consisting of 64×64 voxels with an inplane
resolution of 3×3 mm. The slices were acquired in the transverse plane
and tilted around the x-axis in order to give near whole-brain coverage,
excluding the cerebellum. Each session consisted of 500 volumes. The
subjects' pulse and respiration were recorded using an MRI-compatible
functional scan, a T1 weighted MPRAGE structural sequence was
acquired, using a phased array head coil to provide high-resolution
fMRI data analysis
Image pre-processing and data analysis was performed using
SPM2 (Wellcome Department of Imaging Neuroscience, London, UK).
The EPI images were realigned spatially (Friston et al.,1995). This was
followed by temporal realignment, which corrected for slice-time
differences using the middle slice as reference slice. Images were then
normalized to the Montreal Neurological Institute (MNI) template
provided in SPM2. Finally a spatial filtering was performed by
applying a Gaussian smoothing kernel of 8 mm FWHM (full width at
Following pre-processing a general linear model was applied to the
fMRI time-series where stimulus onset was modelled as single impulse
response functions including stimulus duration (5000 ms) and then
convolved with the canonical haemodynamic response function (HRF)
including its temporal and dispersion derivatives in order to capture
variations in the onset and width of the BOLD responses.
A parametric regression analysis was used (Buchel et al.,1998) that
allowed us tomodel 0thorderand1storderhaemodynamic responses
using orthogonalized polynomial expansion functions. This was
performed for each of the two conditions (gallery and computer
labels) using subject-specific aesthetic ratings in order to model
parametric modulations of aesthetic ratings for the 1st order
expansions. The 0th order parametric regression analysis allowed us
to model the two conditions (gallery and computer) independently of
the aesthetic ratings. This gave the overall design character of a mixed
design incorporating the condition (gallery and computer labels) in
the 0th order expansion and the aesthetic evaluation of the individual
stimuli into the 1st order expansion and. Because the HRF is a
smoothing filter in the temporal domain, the 0th order expansion,
which was constructed as a convolution of the box-car (‘on’ when the
stimulus was presented and ‘off’ in the 1 s in between) with the HRF,
was verysimilar to, although not identical, a traditional box-car which
for the ‘gallery’ labelled images would be ‘on’ for the full 30 s and ‘off’
when the computer label was ‘on’. The 1st order expansion on the
other hand incorporates the individual aesthetic ratings in such a way
that the underlying model (i.e. onset time+5 s duration) is scaled with
the aesthetic rating and then convolved with the HRF. Thereby a
regressor was constructed where the expected haemodynamic signal
varied for the individual images according to the aesthetic rating.
First-level analysis was performed on each subject to generate a single
mean image corresponding to each term of the polynomial expansion.
As the behavioural data demonstrated a statistically significant
difference in aesthetic ratings (see Fig. 2) between gallery and
computer conditions, further regressors were applied to separate
the data and control for these behavioural differences between
conditions. For each subject the frequency of each of the five rating
between gallery and computer conditions since such a difference
would confound the 0th order parametric regression analysis with
Fig. 2. Subjects' aesthetic ratingsgrouped accordingtothe labelled conditions. (A) Mean
aesthetic ratings across subjects for the two conditions. The SEM across subjects is
shown. (B) Distribution of the frequencies of aesthetic responses across all subjects for
the two conditions (5=high appeal; 1=low appeal).
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
aesthetic ratings. For each subject the residual rating bins from this
process were put in separate regressors that were applied in the
subsequent 1st order parametric modulation. The mean images from
the first level analysis were entered into a second-level, random
effects (RFX) analysis accounting for the between subject variance. An
ANOVA model using the beta-estimates of the two conditions for the
0th and 1st order expansions was applied. Equal variance was not
assumed, thus SPM2's options for non-sphericity correction was
applied (Glaser and Friston, 2004).
cardiac pulsation we included RETROICOR (RETROspective Image based
CORrection method) nuisancecovariates in thedesign matrix (Glover et
and respiratory oscillations. We included six regressors for respiration
and ten regressors for cardiac pulsation. We also included twenty-four
regressors that remove residual movement artefacts with spin history
(Friston et al., 1996). This set of nuisance regressors have also been
shown to reduce inter and intra subject variation significantly (Lund et
al., 2005). Having all four types of nuisance regressors in the design
improves the assumption of independently and identically distributed
errors (Lundetal., 2006).Fortheanalysis ahighpass filterwithacut-off
frequency at 1/128 Hz was applied.
Using t-contrasts allowed us to test for correlations of the fMRI
BOLD signal and the parameters of interest performed respectively as
0th order and 1st order parametric modulations. Reported p-values
(pb0.05) were controlled for false discovery rate (FDR) correction
(Genovese et al., 2002). Additionally for regions which failed to reach
FDR significance but for which we had a priori hypotheses, we report
uncorrected p-values. Unless otherwise stated reported p-values are
uncorrected for multiple comparisons. The co-ordinates of all
activations are reported in MNI space.
The aesthetic ratings collected during scanning for the two
stimulus conditions are shown in Fig. 2. The mean aesthetic rating
for ‘gallery’ labelled conditions was 2.86 (SEM=0.05), and for
‘computer’ labelled conditions 2.57 (SEM=0.07). As hypothesized,
statistical analysis revealed that the stimuli were rated as being
significantly more aesthetically pleasing when labelled as ‘gallery’
than when labelled as ‘computer’ (ANOVA, F(1,26)=10.49; pb0.003).
Mean reaction times (RT) were 1224 ms (SEM=98.5) for gallery
conditions and 1187 ms (SEM=96.5) for computer conditions (Fig. 3).
The labels did not give rise to a significant difference in RT (ANOVA, F
(1,26)=0.07; pN0.78) between conditions.
Correlation between the BOLD signal and aesthetic ratings
To test whether aesthetic ratings differentially modulate brain
activity due to the labelling of the two conditions, we employed a
parametric regression analysis (Buchel et al., 1998) using the mean
Fig. 3. Behavioural data collected during scanning. The figure shows the mean reaction-
times (RT) across all subjects. Bars indicate SEM.
Fig. 4. Upper panel: The figure shows activation in right medial OFC where the BOLD signal correlates with the 1st order linear term for the contrast [G–C]. The activation is overlaid
on saggital, coronal and axial sections of the canonical SPM structural image. Lower panel: Parameter estimates for voxels in medial OFC for the two conditions Gallery (G) and
Computer (C) where the x-axis reflects the two stimulus conditions, and the y-axis shows BOLD signal changes. Error bars indicate 90% confidence interval.
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
centred subject-specific behavioural responses (aesthetic judgments)
as separate regressors for each label.
We identified brain areas where the 1st order parametric regression
to the computer (C) context. This contrast revealed significant effects in
right medial OFC (12, 48, −20; z=4.37; pb0.05, corrected for multiple
comparisons using false discovery rate, FDR) (Fig. 4). Although the peak
voxels were located in the medial OFC of the right hemisphere (Fig. 4),
significant activations were found in a corresponding region of the left
OFC at a lower statistical threshold of pb0.001 (−14, 46, −23; z=3.51)
(figure not shown). Furthermore, the left frontal pole on the ventral
aspect of the medial prefrontal cortex (VMPFC) (−10, 60, 2; z=3.49;
pb0.001) was active (Fig. 5). The converse comparison, which should
show areas whose linear response to aesthetic ratings is greater for the
regions (pN0.001) (figure not shown).
A conjunction analysis was performed to identify common brain
regions that are linearly responsive to aesthetic ratings regardless of
context. This was performed as a conjunction between the 1st order
parametric regressors for both G and C conditions. This analysis did
not reveal significant voxels (pN0.0012).
Correlation between the BOLD signal and context irrespective of
For the main effect [G–C], which reflects brain areas which are
more active in the gallery vs. computer condition irrespective of the
actual aesthetic rating given, activity was found in bilateral entorhinal
cortex, BA 28/34 (see Table 1 and Fig. 6A). This region, located on the
parahippocampal gyrus adjacent to the hippocampus formation, is
heavily interconnected with the hippocampus. Furthermore, we
observed activity in bilateral temporal pole/medial temporal pole
(see Table 1 and Fig. 6B) and a large cluster of voxels in bilateral visual
cortex corresponding to BA 17 (see Table 1). The main effects show an
effect of context even when there was no change in the visual
stimulation across the two conditions.
For the converse main effect of context [C–G] we did not observe
significant activations (pN0.001) (figure not shown).
Finally, we performed a conjunction analysis to identify areas
involved in contextual processing irrespective of the two conditions.
This was performed as a conjunction between G and C conditions. It
did not result in any significant activity (pN0.0012), indicating that
contextual processing are recruited differentially by the two labels in
brain areas shown in Fig. 6.
Correlation between recognition effects and aesthetic ratings
To control whether recognition effects induced by the labels might
have contributed to the results, such that the cortical differences
between conditions reflected recognition effects rather than aesthetic
judgments in medial OFC (Frey and Petrides, 2002), we regressed onto
familiarity data (collected post-scanning for all subjects) in order to
search for brain areas that showed a correlation between familiarity
and aesthetic ratings. Small volume corrections (SVC) were applied
constraining our analysis tothe medial OFC activation (12, 48, −20 and
−14, 46, −23). This analysis did not produce any supra-threshold
voxels at pN0.001 (not shown), indicating that recognition effects did
not contribute to the results in medial OFC.
The aim of this study was to investigate the neural system
supporting contextual modulation of aesthetic ratings. We presented
images to subjects in two different contexts, which we hypothesized
Fig. 5. Activation in left polar frontal cortex where the BOLD signal correlates with the 1st orderlinear term of the aesthetic ratings for the contrast [G–C]. The activation is overlaid on
saggital, coronal and axial sections of the canonical SPM structural image.
Location of brain regions that respond preferentially to [Gallery–Computer] irrespective
of aesthetic ratings
Activations are shown at (pb0.001, uncorrected) and (pb0.005, uncorrected, in
parentheses). ⁎pb0.003, corrected; cluster-level. L, left hemisphere; R, right hemisphere.
Fig. 6. The main effect of context [G–C]. (A) Significant voxels (pb0.001) were found in
right entorhinal cortex. Activations were found in a corresponding region on the left at a
lower statistical threshold (pb0.005). Panel A displayed at pb0.005 to show the extent
of the activation in entorhinal cortex. (B) Activations were also found in bilateral
temporal pole (pb0.001). The activations are overlaid on axial and coronal sections of
the canonical SPM structural image.
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
would induce different prior expectations of the hedonic values of
each image, with gallery contexts inducing a greater expectation of
hedonic value than the same image presented in a computer-
generated context. The behavioural results showed this hypothesis
tobe correct, since images under the gallerylabelwere ratedashaving
a significantly higher mean aesthetic value than those carrying the
computer label, even in the absence of systematic differences in visual
stimulation across conditions.
The fMRI data showed that activity in the medial OFC exhibited a
stronger correlation with aesthetic ratings under the gallery context
compared to the computer context, thus demonstrating that the
aesthetic response profile in medial OFC is sensitive to context. In
contrast, bilateral activations in the entorhinal cortex, adjoining the
hippocampus, temporal pole and primary visual cortex showed a
greaterresponse tothe galleryvs.computercontext irrespectiveof the
actual aesthetic ratings. Note that for this contrast, we controlled for
the distribution of aesthetic ratings, so that the conditions used were
selected to ensure that the distributions of aesthetic ratings were
identical for both galleryand computer conditions. Thus differences in
mean aesthetic rating cannot account for the activations reported.
The activation of medial OFC is consistent with current evidence
which suggests that this region represents attributed hedonic values
fromavarietyof sensorymodalities, includinggustatory (O'Dohertyet
al., 2001a; Small et al., 2001, 2003), olfactory (Anderson et al., 2003;
Gottfried et al., 2002; Rolls et al., 2003a), somatosensory (Rolls et al.,
2003b), correlated with modulated activity (Blood et al., 1999; Blood
and Zatorre, 2001; Brown et al., 2004; Menon and Levitin, 2005;
Koelsch et al., 2006) and vision (Aharon et al., 2001; Kawabata and
Zeki, 2004; O'Doherty et al., 2003b; Vartanian and Goel, 2004) as well
as for more abstract rewards such as money (Elliott et al., 2003;
Knutson et al., 2001; O'Doherty et al., 2001b). As an extension of these
previous findings, the present study provides evidence that the
relative to cognitive and semantic input even when there are no
systematic image-wise differences across the two conditions. Further-
more, the results of the present study confirm and extend earlier
studies demonstrating a relation between contextual modulation of
subjects' preference for perceived stimuli and medial OFC activity.
Studiesusing brandinformationto influencepreferencefor softdrinks
(McClure et al., 2004), visual word descriptors to influence preference
for odours (de Araujo et al., 2005), and price information to influence
preference for wine (Plassmann et al., 2008) all found that this
behavioural effect modulated activity in medial OFC or VMPFC. The
present study extends this effect to the domain of visual processing.
This generality across gustatory, olfactory and visual modalities
strongly suggests that medial OFC is a common centre for the
integration of different sources of information pertaining to the
processing, and high-level cognitive inputs.
An interesting question is how the OFC result should be interpreted
in relation to models of the neural mechanisms underlying aesthetic
evaluation. The medial OFC might be involved in two functional sub-
processes: (1) the evaluative categorization processes associated with
making an aesthetic judgment, i.e. the subjective report; or (2) or the
processes correlated withthe codingofsubjective pleasure, i.e. stimulus
hedonic value. Results indicate a neural difference between making an
active aesthetic judgment of the aesthetic value of some stimulus and
attending to the same stimulus in other ways (passive viewing;
symmetrical judgments) (Höfel and Jacobsen, 2007a,b; Jacobsen and
Höfel, 2003, Jacobsen et al., 2006). However, no involvement of medial
part of OFC was found in a neuroimaging study investigating the
processes involved in making aesthetic judgments compared to making
ity in aesthetic judgments were found in the lateral aspects of the OFC
(Jacobsen et al., 2006). The parametric modulation of medial OFC in the
present studyare thus more likely to be related to the subjective coding
of hedonic valence, than mechanisms associated with the active act of
making an overt judgment.
In the study by McClure et al. (2004) images of Coca-Cola vs. Pepsi
cans influenced activations in areas that are traditionally considered as
more cognitive than flavour related areas, including hippocampus,
midbrain, primary visual cortex and dorsolateral prefrontal cortex
(McClureetal., 2004).Weobservedbilateral activationintheentorhinal
cortex in the main effect [G–C] irrespective of the actual aesthetic
ratings. The entorhinal cortex adjoins and is interconnected with the
hippocampus, a region that has been consistently related to processing
of episodic memories (Brown and Aggleton, 2001; Eichenbaum et al.,
1996, 2007). Hippocampus activation is associated with trials in which
subjects correctly recollect contextual information compared to ones in
which they do not (Cansino et al., 2002). Other findings suggest that
midbrain dopaminergic systems involved in reward expectation could
directly modulate declarative memory formation in the hippocampus
(Adcock et al., 2006; Bernabeu et al.,1997; Wittmann et al., 2005). This
evidence is in line with our initial hypothesis that it is the subjects'
conception of the image, rather than its sensory properties, that
primarily determines its hedonic value. Similarly, recent results have
shown that neurons in primary visual cortex encode associations
between visual stimuli and subsequent prediction of reward timing
(Shuler and Bear, 2006). These data demonstrate that reward-timing
activity can occur very early in the sensory-processing paths. One
possible mechanism responsible for this effect might be due to subjects
stimulus labels. The different expectations of hedonic value could be
determined in different ways. The difference in context in this
experiment straddles at least two possible factors. The first might be
loosely described as a difference in prestige; the art gallery is more
prestigious than the computer as a source of artworks. Accordingly the
more prestigioustheartgallery,themorecompetitionthereis forartists
have a greater monetary value than those which are not (on average).
Neglecting economic factors of supply and demand, one might have the
simple prior that the more expensive the artwork the more likely it is to
have a higher hedonic value for a given observer. The context induced
changes in aesthetic modulation could be due to either or both of these
factors and dissociating these two will be the focus of future work.
In contrast to an event-related design in which the labelling would
allowed us to look at responses to cognitive labels that were fully
predictable within blocks. It has been suggested that block designs is
confounded by reward expectation (O'Doherty et al., 2003a; Winston
et al., 2007). Accordingly activity observed in the present study could
possibly be due to reward expectation induced by the block design
rather thanbeingdriven by theassessmentof aesthetic value.However,
we argue against this possibility in that evidence suggests that ventral
et al., 1992). OFC, on the other hand, is known to be involved in
representing the hedonic value of a stimulus (O'Doherty et al., 2003a;
themselves and it is likely that responses were not observed in ventral
striatum because there was no confound of reward expectation.
Although the entorhinal cortex and temporal pole (BA 38) are
associated with different roles in declarative memory (Kroll et al.1997;
Nakamura and Kubota 1995; Sargolini et al. 2006), our data suggest a
connection between them. The temporal pole is thought to be
implicated in semantic memory retrieval (Damasio et al., 1996;
Mummery et al., 2000). This region is involved in the storage and recall
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
of contextual information, particularly when affectively salient (Lane et
to be active in tasks that require explicit evaluation judgments such as
emotional intensity (Cunninghamet al.,2004), and aesthetic judgments
(Jacobsen et al., 2006). Our data fits with these prior findings and raise
the possibility that the entorhinal cortex and temporal pole may be
engaged during recollection of art-related and cultural information that
influence aesthetic judgments during gallery conditions, while VMPFC
and medial OFC are more involved in attaching hedonic properties to
them. These two systems do not appear to function independently of
each other but are modulated and cooperate to influence aesthetic
judgments induced by semantic context.
We thank Dr. T. Lund for helpful discussions. T. Ramsoy provided
useful comments on the manuscript. U. Kirk was supported by a PhD
scholarship from the Danish Medical Research Council; M. Skov was
supported by Hvidovre Hospital’s research foundation; O. Hulme was
supported by the Medical Research Council, United Kingdom; M.S.
Christensen was supported by the Faculty of Science, University of
Copenhagen and the Danish Medical Research Council; S. Zeki was
supported by a grant from the Wellcome Trust, London. The MR-
scanner was donated by the Simon Spies Foundation.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.neuroimage.2008.10.009.
Adcock, R.A., Thangavel, A., Gabrieli, A.W., Knutson, B., Gabrieli, J.D.E., 2006. Reward-
motivated learning: mesolimbic activation precedes memory formation. Neuron
Aharon, I., Etcoff, N., Ariely, D., Chabris, C.F., O'Connor, E., Breiter, H.C., 2001. Beautiful faces
have variable reward value: fMRI and behavioral evidence. Neuron 32, 537–551.
Anderson, A.K., Christoff, K., Stappen, I., Panitz,D., Ghahremani, D.G., Glover, G., Gabrieli,
J.D.E., Sobel, N., 2003. Dissociated neural representations of intensityand valence in
human olfaction. Nat. Neurosci. 6, 196–202.
Medina, J.H., 1997. Involvement of hippocampal cAMP/cAMP-dependent protein
kinase signalling pathways in a late memory consolidation phase of aversively
motivated learning in rats. Proc. Natl. Acad. Sci. U. S.A. 94, 7041–7046.
Berridge, K.C., 1996. Food reward: brain substrates of wanting and liking. Neurosci.
Biobehav. Rev. 20, 1–25.
Blood, A.J., Zatorre, R.J., 2001. Intensely pleasurable responses to music correlate with
Blood, A.J., Zatorre, R.J., Bermudez, P., Evans, A.C.,1999. Emotional responses to pleasant
and unpleasant music correlate with activity in paralimbic brain regions. Nat.
Neurosci. 2, 382–387.
Brown, M.W., Aggleton, J.P., 2001. Recognition memory: what are the roles of the
perirhinal cortex and hippocampus? Nat. Rev., Neurosci. 2, 51–61.
Brown, S., Martinez, M.J., Parsons, L.M., 2004. Passive music listening spontaneously
engages limbic and paralimbic systems. NeuroReport. 15, 2033–2037.
Buchel, C., Holmes, A.P., Rees, G., Friston, K.J., 1998. Characterizing stimulus response
functions using nonlinear regressors in parametric fMRI experiments. NeuroImage
Cansino, S., Maquet, P., Dolan, R.J., Rugg, M.D., 2002. Brain activity underlying encoding
and retrieval of source memory. Cereb. Cortex 12, 1048–1056.
Cunningham, W.A., Raye, C.L., Johnson, M.K., 2004. Implicit and explicit evaluation:
fMRI correlates of valence, emotional intensity, and control in the processing of
attitudes. J. Cogn. Neurosci. 16 (10), 1717–1729.
Cupchik, G.C., Shereck, L., Spiegel, S.,1994. The effects of textual information on artistic
communication. Visual Arts Res. 20, 62–78.
Damasio, H., Grabowski, T.J., Tranel, D., Hichwa, R.D., Damasio, A.R.,1996. A neural basis
for lexical retrieval. Nature 380, 499–505.
de Araujo, I.E., Rolls,E.T., Velazco, M.I., Margot, C., Cayeux, I., 2005. Cognitive modulation
of olfactory processing. Neuron 46, 671–679.
Deichmann, R., Gottfried, J.A., Hutton, C., Turner, R., 2003. Optimized EPI for fMRI
studies of the orbitofrontal cortex. NeuroImage 19, 430–441.
Eichenbaum, H., Schoenbaum, G., Young, B., Bunsey, M.,1996. Functional organization of
the hippocampal memory system. Proc. Natl. Acad. Sci. U. S. A. 93, 13500–13507.
Eichenbaum, H., Yonelinas, A.P., Ranganath, C., 2007. The medial temporal lobe and
recognition memory. Annu. Rev. Neurosci. 21, 123–152.
Elliott, R., Newman, J.L., Longe, O.A., Deakin, J.F., 2003. Differential response patterns in
the striatum and orbitofrontal cortex to financial reward in humans: a parametric
functional magnetic resonance imaging study. J. Neurosci. 23, 303–307.
Friston, K.J., Ashburner, J., Frith, C.D., Poline, J.B., Heather, J.D., Frackowiak, R.S.J., 1995.
Spatial registration and normalization of images. Hum. Brain Mapp. 2, 165–189.
Friston, K.J., Williams, S., Howard, R., Frackowiak, R.S.J., Turner, R., 1996. Movement-
related effects in fMRI time-series. Magn. Reson. Med. 35, 346–355.
Genovese, C.R., Lazar, N.A., Nichols, T., 2002. Thresholding of statistical maps in
functional neuroimaging using the false discovery rate. NeuroImage 15, 870–878.
Glaser, D., Friston, K.J., 2004. Variance components. In: Frackowiak, R.S.J., Friston, K.J.,
Frith, C.D., Dolan, R.J., Price, C.J., Zeki, S., Ashburner, J., Penny, W. (Eds.), Human Brain
Function. Academic Press, Elsewier, pp. 781–792.
Glover, G.H., Li, T.Q., Ress, D., 2000. Image-based method for retrospective correction of
physiological motion effects in fMRI: RETROICOR. Magn. Reson. Med. 44, 162–167.
Gottfried, J.A., Deichmann, R., Winston, J.S., Dolan, R.J., 2002. Functional heterogeneity
in human olfactory cortex: an event-related functional magnetic resonance
imaging study. J. Neurosci. 22, 10819–10828.
Höfel, L., Jacobsen, T., 2007a. Electrophysiological indices of processing aesthetics:
spontaneous or intentional processes? Int. J. Psychophysiol. 65, 20–31.
Höfel, L., Jacobsen, T., 2007b. Electrophysiological indices of processing symmetry and
aesthetics. A resultof judgment cetgorization or judgment report? J. Psychophysiol.
Jacobsen, T., Höfel, L., 2003. Descriptive and evaluative judgment processes: behavioral
and electrophysiological indices of processing symmetry and aesthetics. Cogn.
Affect. Behav. Neurosci. 3, 289–299.
Jacobsen, T., Schubotz, R.I., Höfel, L., Von., Cramon, D.Y., 2006. Brain correlates of
aesthetic judgment of beauty. NeuroImage 29, 276–285.
Kawabata, H., Zeki, S., 2004. Neural correlates of beauty. J. Neurophysiol. 91,1699–1705.
Knutson, B., Adams, C.M., Fong, G.W., Hommer, D., 2001. Anticipation of increasing
monetary reward selectivity recruits nucleus accumbens. J. Neurosci. 21, RC159.
Koelsch, S., Fritz, T., Von., Cramon, D.Y., Muller, K., Friederici, A.D., 2006. Investigating
emotion with music: an fMRI study. Hum. Brain Mapp. 27, 239–250.
Kroll, N.E., Markowitsch, H.J., Knight, R.T., von Cramon, D.Y., 1997. Retrieval of old
memories: the temporofrontal hypothesis. Brain 120, 1377–1399.
Lane, R.D., Fink, G.R., Chau, P.M.L., Dolan, R.J., 1997. Neural activation during selective
attention to subjective emotional responses. NeuroReport 8, 3969–3972.
Leder, H., Carbon, C.C., Ripsas, A.L., 2006. Entitling art: Influence of title information on
understanding and appreciation of paintings. Acta Psychol. (Amst.) 121, 176–198.
Lund, T.E., Nørgaard, M.D., Rostrup, E., Rowe, J.B., Paulson, O.B., 2005. Motion or activity:
their role in intra- and inter-subject variation in fMRI. NeuroImage 26, 960–964.
Lund, T.E., Madsen, K.H., Sidaros, K., Luo, W.L., Nichols, T.E., 2006. Non-white noise in
fMRI: does modeling have an impact? NeuroImage 29, 54–66.
McClure, S.M., Li, J., Tomlin, D., Cypert, K.S., Montague, L.M., Montague, P.R., 2004.
Neural correlates of behavioral preference for culturally familiar drinks. Neuron 44,
Menon, V., Levitin, D.J., 2005. The rewards of music listening: response and
physiological connectivity of the mesolimbic system. NeuroImage 28, 175–184.
Mummery, C.J., Patterson, K., Price, C.J., Ashburner, J., Frackowiak, R.S., Hodges, J.R.,
2000. A voxel-based morphometry study of semantic dementia: relationship
between temporal lobe atrophy and semantic memory. Ann. Neurol. 47, 36–45.
Nakamura, K., Kubota, K., 1995. Mnemonic firing of neurons in the monkey temporal
pole during a visual recognition memory task. J. Neurophysiol. 74, 162–178.
O’Doherty, J., Rolls, E.T., Francis, S., Bowtell, R., McGlone, F., 2001a. Representation of
pleasant and aversive taste in the human brain. J. Neurophysiol. 85, 1315–1321.
O’Doherty, J.P., Kringelbach, M.L., Rolls, E.T., Hornak, J., Andrews, C., 2001b. Abstract
reward and punishment representation in the human orbitofrontal cortex. Nat.
Neurosci. 4, 95–102.
O’Doherty, J., Winston, J., Critchley, H., Perrett, D., Burt, D.M., Dolan, R.J., 2003a. Beauty
in a smile; the role of medial orbitiofrontal cortex in facial attractiveness.
Neuropsychologia 41, 147–155.
O’Doherty, J.P., Critchley, H., Deichmann, R., Dolan, R.J., 2003b. Dissociating valence of
outcome from behavioral control in human orbital and ventral prefrontal cortices.
J. Neurosci. 23, 7931–7939.
Plassmann, H., O, Doherty, J., Shiv, B., Rangel, A., 2008. Marketing actions can modulate
neural representations of experienced pleasantness. Proc. Natl. Acad. Sci. U. S. A.
Rolls, E.T., Kringelbach, M.L., de Araujo, I.E., 2003a. Different representations of pleasant
and unpleasant odours in the human brain. Eur. J. Neurosci. 18, 695–703.
Rolls, E.T., O, Doherty, J., Kringelbach, M.L., Francis, S., Bowtell, R., McGlone, F., 2003b.
Representations of pleasant and painful touch in the human orbitofrontal and
cingulate cortices. Cereb. Cortex 13, 308–317.
Russell,P.A., 2003. Effort after meaning and the hedonic value of paintings. Br. J. Psychol.
Sargolini, F., Fyhn, M., Haftin, T., McNaughton, B.L., Witter, M.P., Moser, M.B., Moser, E.I.,
2006. Conjunctive representation of position, direction, and velocity in entorhinal
cortex. Science 312, 758–762.
Schultz, W., Apicella, P., Scarnati, E., Ljungberg, T., 1992. Neuronal activity in monkey
ventral striatum related to the expectation of reward. J. Neurosci. 12, 4595–4610.
Shuler, M.G., Bear, M.F., 2006. Reward timing in the primary visual cortex. Science 331,
Small, D.M., Zatorre, R.J., Dagher, A., Evans, A.C., Jones-Gotman, M., 2001. Changes in
brain activity related to eating chocolate. Brain 124, 1720–1733.
Small, D.M., Gregory, M.D., Mak, Y.E., Gitelman, D., Mesulam, M.M., Parrish, T., 2003.
Dissociation of neural representation of intensity and affective valuation in human
gustation. Neuron 39, 701–711.
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132
Smith, A.P., Henson, R.N., Dolan, R.J., Rugg, M.D., 2004. fMRI correlates of the episodic Download full-text
retrieval of emotional contexts. NeuroImage 22, 868–878.
Tremblay, L., Schultz, W., 1999. Relative reward preference in primate orbitofrontal
cortex. Nature 398, 704–708.
Vartanian, O., Goel, V., 2004. Neuroanatomical correlate of aesthetic preference for
paintings. NeuroReport 15, 893–897.
Winston, J.S., O, Doherty, J., Kilner, J.M., Perrett, D.I., Dolan, R.J., 2007. Brain systems for
assessing facial attractiveness. Neuropsychologia 45, 195–206.
Wittmann, B.C., Schott, B.H., Guderian, S., Frey, J.U., Heinze, H.J., Duzel, E., 2005.
Reward-related FMRI activation of dopaminergic midbrain is associated with
enhanced hippocampus-dependent long-term memory formation. Neuron 45,
U. Kirk et al. / NeuroImage 44 (2009) 1125–1132