Smelling Shapes: Crossmodal Correspondences Between Odors and Shapes
Advance Access publication October
Smelling Shapes: Crossmodal Correspondences Between Odors and Shapes
Grant Hanson-Vaux 0 1
Anne-Sylvie Crisinel 0 1
Charles Spence 0 1
0 Oxford , OX1 3UD , UK.
1 Department of Experimental Psychology, Crossmodal Research Laboratory, University of Oxford , UK
Crossmodal correspondences between odors and visual stimuli-particularly colors-are well-established in the literature, but there is a paucity of research involving visual shape correspondences. Crossmodal associations between 20 odors (a selection of those commonly found in wine) and visual shape stimuli (“kiki”/“bouba” forms-Köhler W. 1929. Gestalt psychology. New York: Liveright.) were investigated in a sample of 25 participants (mean age of 21 years). The odors were rated along a form scale anchored by 2 shapes, as well as several descriptive adjective scales. Two of the odors were found to be significantly associated with an angular shape (lemon and pepper) and two others with a rounded shape (raspberry and vanilla). Principal component analysis indicated that the hedonic value and intensity of odors are important in this crossmodal association, with more unpleasant and intense smells associated with more angular forms. These results are discussed in terms of their practical applications, such as in the use of bottle, logo, or label shape by marketers of perfume and wine to convey the prominent notes through congruent odor-shape pairing. In conclusion, these results support the existence of widespread crossmodal associations (or correspondences) between odors and visual shape stimuli.
crossmodal associations; crossmodal correspondences; odors; shapes
The environments in which we live are multisensory; often
even the simplest of everyday activities entail the interaction
of multiple sources of sensory information. The task of
eating a meal, for instance, involves a barrage of inputs from the
different sensory modalities, some of which are combined in
the perception of flavor. Crossmodal correspondences (or
associations) may thus reflect a strategy by which the human
brain (not to mention the brains of other species; see Ludwig
et al. 2011) can quickly and effectively deal with complex
sensory inputs; as Spence and Zampini (2006, p. 1009) put
it, a “range of implicit perceptual and cognitive processes
are continually engaged in blending and segregating
information from diverse sources”. Crossmodal correspondences,
then, may help an organism to solve the crossmodal
binding problem, that is, to know which of several stimuli that
happen to be presented in different modalities at any given
time are likely to belong together (see Ernst 2007; Spence
2011). Further, by using a preferential looking methodology,
visual-auditory correspondences have been shown to be
present as early as 3- to 4-months of age (Walker et al. 2010).
The crossmodal correspondence between visual lightness
and auditory pitch has recently been demonstrated in
chimpanzees as well (Ludwig et al. 2011), thus confirming their
widespread existence during development and across species.
In terms of consumer research, certain expectations may be
borne out of crossmodal associations. Consumers will rate
food and drink products more highly if they meet their
expectations than if they do not (Spence 2012; Yeomans
et al. 2008). Therefore, studying crossmodal associations is
a worthwhile endeavor, not only for its theoretical value but
for its practical implications for fields such as advertising and
food marketing. Perhaps it is no surprise, then, that there has
been a growth in interest in studying crossmodal interactions
in recent years—especially in relation to flavor perception.
For a small minority of individuals, crossmodal
experiences manifest themselves in a more extreme form—
namely, synesthesia. Synesthesia refers to the phenomenon
whereby stimulation in one sensory modality automatically
triggers a conscious perception, often in a second modality
(Cytowic, 1993). Although some researchers have argued
that crossmodal correspondences should be considered as a
weak form of synesthesia (e.g., Martino and Marks 2001),
there are a number of key differences between the condition
and findings of crossmodal correspondences—the meaning
of synesthesia may be distorted if the term is applied too
readily or broadly (see Deroy and Spence [in preparation];
Spence 2011). Not least of these differences, the perception
of, for example, a certain shape is “involuntary” in response
to tastes or flavors in taste/flavor-shape synesthetes. This
contrasts with the propensity for certain attributes to be linked
crossmodally, but, crucially, not perceived concurrently nor
involuntarily in neurologically normal individuals.
Research on the topic of crossmodal associations (or
correspondences) arguably finds its roots in the classic research
reported by Köhler (1929; see also Sapir 1929), where it was
demonstrated that there was a nonarbitrary association
between the auditory and visual domains under certain
conditions. The speech sounds “baluba” and “takete” were to be
attributed to 2 visual shapes (see Figure 1). It was found that
these stimuli were very often paired by participants in a
nonrandom (i.e., specific) manner. There was a strong preference
for “baluba” to be associated with a curvy, rounded shape,
and for “takete” to be paired with a sharp, angular one. The
same result has now been demonstrated in a wide variety of
cultures (Bremner et al. [in press]).
This phenomenon has been explained in terms of a matching
of the sharp directional changes of the lines of the angular shape
with the sharp phonemic inflections of the sound “takete” (or
“kiki”), as well as the sharp inflection of the tongue on the
palate (see Ramachandran and Hubbard 2001). This simple result
turns out to be surprisingly robust, with replications
obtaining the same general finding with an array of different word
stimuli being used (e.g., Köhler 1947—“takete” and “maluma”;
Ramachandran and Hubbard 2001—“kiki” and “bouba”).
Crossmodal correspondences between speech sounds and
visual shapes are thus well-documented in the literature, and
as such, more attention has been directed toward research
into perceptual interactions involving the other senses. Of
particular interest in terms of the present study is research
involving the sense of smell. Crossmodal associations and
interactions between odors and auditory
stimuli—specifically pitch—have been reported in a number of studies.
Participants display a consensus in ordering certain smells
by ascending pitch (Belkin et al. 1997; see also Crisinel and
Spence 2012), and congruent auditory information can
modulate the perceived pleasantness of an odor (see, e.g.,
Seo and Hummel 2011). Properties such as hedonic value
and complexity have been found to be important factors
Figure 1 Typical angular and rounded shapes similar to those used in
Köhler’s (1929) original study.
in terms of the matching of odor and pitch (Crisinel and
Spence 2012; cf., Schifferstein and Tanudjaja 2004).
Research into the links between odors and visual stimuli
is arguably a more piecemeal domain. Different visual
stimuli have been utilized in different studies. Predominantly
though, color has been used; for instance, nonrandom
linking across participants between odor and color hue has been
widely established (e.g., Demattè et al. 2006; Gilbert et al.
1996), and the robustness of such correspondences has been
demonstrated both explicitly (i.e., correspondences have
been documented to remain stable after 2 years—Gilbert
et al. 1996) and implicitly (Demattè et al. 2006). More intense
smells are associated with darker colors (Kemp and Gilbert
1997). Using a different type of visual stimulus, Gottfried
and Dolan (2003) reported that semantically congruent, as
opposed to incongruent, pictorial images resulted in faster
and more accurate olfactory detection. However,
considerably less research has been conducted into the possible
existence of crossmodal matches between odors and shapes.
Seo et al. (2010) recently conducted a study in which they
utilized abstract symbols. The study demonstrated that certain
symbols were consistently paired with specific odors across
participants, and that subsequent odor-symbol congruency
modulated odor pleasantness, as well as increased the amplitude of
olfactory event-related potentials in the human brain. However,
despite this scarcity of research on the topic, the shape of
perfume bottles is considered as a key element in determining the
success of a new perfume (e.g., Caldwell and Flammia 1991).
The shape of the packaging and/or label of many food and
nonfood fragrant products is also strongly associated with
specific brands (see Spence and Piqueras-Fiszman forthcoming).
The present study was designed to investigate any
associations that exist between odors commonly found in
wines (using the same stimulus set as in Crisinel and Spence
2012) and visual shape stimuli, namely the angular and rounded
shapes classically used in studies of word-shape association.
The odors used represent a wide range of categories (e.g.,
flowers, fruits, or spices), including stimuli with a trigeminal
component, such as pepper. The use of odors related to wine
seems interesting considering the propensity for wine writers
to describe wines with reference to shape. For instance, a wine
may be described as sharp or well-rounded (e.g., see Lehrer
2009; Peynaud 1996). Such descriptors are used to describe
the flavor of the wine as a whole, but given the important role
played by retronasal olfaction in flavor perception (see, e.g.,
Shepherd 2006), one can hypothesize that similar crossmodal
correspondences exist for wine-related odors smelled
orthonasally (although note that here single components of
wine odors were used). The present study goes beyond the
seminal work of Seo et al. (2010) in that we used a different
selection of olfactory stimuli together with a more controlled
shape dimension (here, a scale from more angular in shape
to more rounded in shape, see Figure 2). Our hypothesis was
that crossmodal associations would be obtained between the
wine odor stimuli and the angular and rounded shapes.
Figure 2 Scale used to rate the association of each odor with the angular
or rounded shape.
Materials and Methods
Twenty-five individuals participated in the experiment (15
female and 10 male; mean = 21.2 years old, range = 20–30 years),
from a population at the University of Oxford. None of the
participants reported a cold or any other impairment of their
sense of smell prior to taking part in the study.
The experiment complied with the standards set in the
Declaration of Helsinki for Medical Research involving
Human Subjects and was approved by the Central University
Research Ethics Committee of Oxford University.
Twenty odors (out of 54) from the “Nez du Vin” set (see
www.nezduvin.co.uk/nezduvin_54.htm)—a kit designed to
train amateur wine tasters to recognize and describe the
aromas found in wine—were used as olfactory stimuli (almond,
apple, apricot, blackberry, caramel, cedar, dark chocolate,
cut hay, green pepper, honey, lemon, liquorice, mushroom,
musk, pepper, pineapple, raspberry, smoked, vanilla, and
violet). These particular samples were used in order to cover
a broad range of aromas, and further, to facilitate
comparisons with findings from past research, which selected exactly
the same odors (Crisinel and Spence 2012). The odors were
contained in small transparent numbered glass bottles. Thus,
without a covering, the color of the liquid inside was visible.
The color of the liquids ranged from transparent to dark
brown. In order to check any effect the color of the liquid
might have on the various ratings, for a subset of the
participants (N = 10), the bottles were covered with white stickers.
For the visual stimuli, angular and rounded shapes were
used as the anchors on a 9-point scale (with the angular
“kiki” shape on the left side and the rounded “bouba” shape
on the right—see Figure 2). Fifteen adjective scales were also
presented, with the descriptors: complex, intense, pleasant,
familiar, floral, fruity, acrid, earthy, nutty, spicy, woody, bitter,
salty, sour, and sweet. These were also rated on a 9-point scale
(from “not at all” on the left side, to “extremely so” on the
right) and text accompanied each scale asking participants
to what extent the current odor was, for example, “complex.”
The experimenter obtained informed consent before
participants took part in the study. The participants were seated
in a darkened booth and were then handed the first bottle.
The order of odors was randomized for each participant.
They were instructed that they could sniff the odor as much
as they wished. Each odor was to be rated on a series of
dimensions, with the experiment programmed in E-Prime
(Version 1.2). This included a dimension of shape, with a
scale anchored on one end by an angular shape and on the
other by a rounded shape. Following the shape dimension,
15 adjective scales were presented, one at a time, in a
random order. After all of the ratings had been made for the
first odor, that bottle was taken away by the experimenter
and the next one was provided, until all 20 odors had been
rated by the participant. The experiment typically lasted for
approximately 40 min. The participants were told that they
could withdraw from the study at any time should they so
desire (though none did).
In order to test for an unwanted effect of the visibility of
the color of the odor liquid, two-way (20 × 2) mixed analyses
of variance were conducted for all ratings, with the
withinsubjects variable of odor and the between-subjects variable
of bottle covering. It was found that although there were
significant main effects of odor on all ratings (P < 0.001), there
were no significant main effects of the bottles being covered,
except for the familiarity rating (P = 0.008), nor were there
significant interactions between odors and the bottles being
covered, suggesting that visual information from the
transparent odor bottles did not affect the participants’ ratings.
One-way t-tests were carried out in order to assess which
of the individual odors were significantly different from the
scale’s midpoint (thus, the test value was set to 5). Bonferroni
adjusted alpha levels of 0.0025 were used (0.05/20). The
significantly angular aromas were lemon—t(24) = −3.73,
P = 0.001, and pepper—t(24) = −3.54, P = 0.002. Odors that
were judged to be significantly rounded were raspberry—
t(24) = 4.23, P < 0.001, and vanilla—t(24) = 3.94, P = 0.001
(see Figure 3).
Factor analysis was conducted on the data. Principal
component analysis (PCA) was used on the 15 adjective
ratings as well as the rating of shape, using SPSS version
19. The Kaiser–Meyer–Oklin measure of sampling
adequacy (Kaiser 1970, 1974) was 0.65 and Bartlett’s test of
sphericity (Bartlett 1954) was significant, so the data were
of sufficient quality to proceed. From the component matrix
3 components were extracted, with eigenvalues of at least
1.0. These accounted for 83.5% of the total variance; 58.6%,
18.5%, and 6.4% for the 3 components, respectively. Based
on inspection of the scree plot, 2 components were retained,
and Varimax rotation was performed. Consequently, the
first component now accounted for 54.0% of the variance,
and the second component for 23.1%. The first factor had
strong (>0.05) positive loadings of acrid, bitter, complex,
Figure 3 Mean shape score of each of the 20 odor samples, ordered
according to their value; error bars indicate the standard errors, and
asterisks denote the odors that were significantly different from the scale’s
midpoint of 5.
earthy, nutty, salty, spicy, and woody, while it had strong
negative loadings of floral, fruity, pleasant, and sweet (see
Figure 4A). This could be argued to characterize the hedonic
value, but also seems to represent complexity. The second
factor had strong positive loadings of acrid, bitter, familiar,
intense, and sour. This, then, seems to be linked to intensity,
in particular the intensity of sourness. The lemon odor had
the highest score on this component. The shape ratings had
negative loadings on both the first (−0.44) and second (−0.80)
component. Scores for each odor are shown in Figure 4B.
The results of the experiment reported here further our
understanding of the crossmodal correspondences that exist
between odors and visual stimuli; the hypothesis that
crossmodal associations between visual shape and odor would be
obtained was supported. These associations are shared, but
further research will be needed to establish their stability over
time in individuals (cf., Gilbert et al. 1996). This result builds
on the existing findings of links between odor and color
(Demattè et al. 2006; Gilbert et al. 1996; Kemp and Gilbert
1997; Schifferstein and Tanudjaja 2004) and helps to confirm
and extend—in a more controlled, dimensional way—an
underrepresented area in the literature: namely crossmodal
associations between odor and abstract symbols/shapes (Seo
et al. 2010). The results indicated that the lemon and pepper
odors were significantly associated with the angular shape,
whereas the raspberry and vanilla odors were significantly
associated with the rounded one instead. The other odors
tested might be associated with intermediate shapes (partly
rounded, partly angular). However, the scale used does not
allow one to differentiate between an absence of any
association with shapes and an association with an intermediate
shape for ratings around the midpoint of the scale. Further
studies could disambiguate these results by using
intermediate shapes along the scales, and/or having a separate scale to
rate the strength of the matching.
Interestingly, the results of the present study are similar to
those of studies that have used the sense of taste, rather than
the sense of smell, in their crossmodal research. For instance,
sweet tastes have been found to be associated with
roundedness, and bitterness is associated with angularity. Ngo et al.
(2011), for example, found that sweet milk chocolate was
rated as more rounded, and bitter dark chocolate was
associated with sharper shapes. Elsewhere, participants in a study
by Deroy and Valentin (2011) linked sweet beers (as opposed
to acidic ones) with more rounded shapes. Considering that
odors easily acquire taste properties when they have been
experienced together (Stevenson et al. 1995), the finding
that vanilla and raspberry are rated as significantly rounded
seems to accord well with this propensity for sweetness to
be linked with roundedness. Likewise, it is interesting to
note that lemon, a typically sour/bitter odor (note that
people sometimes confuse sour and bitter—see Hettinger et al.
1999) was rated as significantly angular. One can also
wonder whether the trigeminal system plays a role in the
crossmodal associations with shapes, as the pepper odor, which
has a strong trigeminal component, was associated with the
The results of the PCA suggest that perceived intensity
and sourness are important factors in crossmodal
correspondences: More intense and sour odors tend to
be associated with a more angular shape, whereas more
subtle ones are associated with the rounded shape. Hedonic
value was also found to be a factor in the crossmodal
correspondence. That is, the strong negative loading of
shape onto the hedonic component suggests that unpleasant
(i.e., intensely bitter, sour, acrid) odors are associated with
a more sharp, angular form. The finding that hedonic value
is involved accords well with previous results showing that
pleasantness is the most salient or dominant underlying
dimension in odorants (see, e.g., Zarzo 2008), although the
crossmodal correspondences observed here are not entirely
explained by the hedonic value of the stimuli. Further, other
studies of crossmodal associations have demonstrated that
hedonic value is involved in correspondences, for example,
between odor and musical pitch (Crisinel and Spence 2012).
Moreover, the results of the present study are consistent with
previous evidence that angularity is unpleasant (see Bar and
Neta 2006). However, other factors such as odor intensity
and complexity, as well as the presence of a trigeminal
component, should certainly be explored to provide a more
comprehensive understanding of the correspondences.
An idea for future work stemming from these findings is the
addition of an implicit association test to assess the
robustness of the study’s cross-modal associations. Prior work has
Figure 4 Loading plot from the PCA conducted on the various ratings
of the odors (A). Note that a higher rating on the shape scale corresponds
to a stronger association with the rounded shape. Score plot of the odors
(B). Odors that were significantly associated with the angular shape are
represented by a star, whereas those that were associated with the rounded
shape are represented by an open circle.
already utilized this methodology; for example, Demattè
et al. (2006) reported that participants responded faster, and
more accurately, to odor-color pairings that were strongly
associated with each other as opposed to when there was a
weak association or else no association. Here, the most
angular odors (such as lemon and pepper) and the most rounded
ones (such as raspberry and vanilla) could be paired with
the shapes to test for facilitatory effects and to test for the
robustness of the associations. Other research has found that
congruent pictorial shapes (e.g., the shape of a strawberry
when smelling that aroma), as well as color information, can
aid olfactory discrimination (Demattè et al. 2009). Colors
and pictorial cues seem to lead to the access of an object’s
representation at different levels of information processing,
with color activating a perceptual level, whereas pictorial
involves a semantic level (see Naor-Raz et al. 2003). With
the present study, however, it could be predicted that shapes
would influence the perceptual features of odors as, unlike
the shapes utilized in Demattè et al. (2009), those used here
were abstract, and not pictorial (i.e., they were not explicitly
linked to, nor did they signify, a certain odor). Indeed,
previous research on correspondences between odors and abstract
symbols has demonstrated effects of shape congruency on
odor intensity and pleasantness ratings (Seo et al. 2011).
The practical implications of the results expand into the
marketing of products such as perfumes, beverages, and
foods (see Spence and Piqueras-Fiszman forthcoming). It is
important to note that synesthetic, crossmodal language is
often used in advertising and marketing. It has been shown
that adverts using this kind of phrasing are rated as being
more pleasant and novel than those using literal language,
especially if they can stimulate a sense that cannot be
stimulated directly by the object in question; for instance, using
language relating to sounds when advertising perfume (see
Nelson and Hitchon 1999). One reason, therefore, why
certain fragrances are successful could be due to the
appropriate use of congruent shape-smell pairings—both perhaps
in terms of geometric shape images on the packaging and
the form of the bottle itself. Indeed, there are many options
and opportunities for perfume marketers in the area of
innovative bottle design. Two successful perfumes—DKNY’s
“Golden Apple” and “Nina” by Nina Ricci—utilize rounded
bottles in the form of an apple. There is a semantic
congruency here already, as these perfumes contain notes of apple,
but further, some degree of a link to hedonic value might be
hypothesized given the present results.
In addition, given that the present study used wine odors,
bottle, logo, and label shapes could be manipulated in the
marketing of wines in order to capture the prominent note.
For instance, for a fruity wine with strong notes of
raspberry, then the results reported here indicate that more
rounded shapes would be better than in the case of a wine
that had prominent lemon or pepper notes. One should note,
though, that a wine’s aroma typically consists of a
complex combination of olfactory notes. Further research will
therefore be needed to determine, for example, what shape is
considered congruent for a wine that combines notes
associated with angularity and roundedness. The use of shape in
wine marketing could aim to be congruent with, for instance,
whether it is an overall “sharp” or “well-rounded” wine (see
Lehrer 2009; Peynaud 1996). In doing so, this could lead to
more positive perceptions of the product from consumers,
with its key sensory features underscored through
marketers’ attention to crossmodal correspondences (e.g., see Gal
et al. 2007).
who tasted shapes. New
Bar M , Neta M. 2006 . Humans prefer curved visual objects . Psychol Sci . 17 ( 8 ): 645 - 648 .
Bartlett MS . 1954 . A note on multiplying factors for various chi square approximations . J Roy Stat Soc . 16 : 296 - 298 .
Belkin K , Martin R , Kemp SE , Gilbert AN . 1997 . Auditory pitch as a perceptual analogue to odor quality . Psychol Sci . 8 : 340 - 342 .
Caldwell HM , Flammia D. 1991 . The development of American dominance in perfume marketing . In Marketing history - its many dimensions: Proceedings of the Fifth Conference on Historical Research in Marketing and Marketing Thought; 1991 April 19 -21; Kellogg Center of Michigan State University. Taylor CR, editor. Michigan State University . pp. 352 - 367 .
Crisinel AS , Spence C. 2012 . A fruity note: crossmodal associations between odors and musical notes . Chem Senses . 37 ( 2 ): 151 - 158 .
Cytowic RE . 1993 . The G. P. Putnam's Sons.
Demattè ML , Sanabria D , Spence C. 2006 . Cross-modal associations between odors and colors . Chem Senses . 31 ( 6 ): 531 - 538 .
Demattè ML , Sanabria D , Spence C. 2009 . Olfactory discrimination: when vision matters? Chem Senses . 34 ( 2 ): 103 - 109 .
Deroy O , Valentin D. 2011 . Tasting liquid shapes: investigating cross-modal correspondences . Chemosens Percept . 4 : 80 - 90 .
Ernst MO . 2007 . Learning to integrate arbitrary signals from vision and touch . J Vis . 7 ( 5 ):7. 1 - 714 .
Gal D , Wheeler SC , Shiv B. 2007 . Cross-modal influences on gustatory perception . J Exp Psychol Gen. Abstract available at SSRN: http://ssrn. com/abstract=1030197
Gilbert AN , Martin R , Kemp SE . 1996 . Cross-modal correspondence between vision and olfaction: the color of smells . Am J Psychol . 109 ( 3 ): 335 - 351 .
Gottfried JA , Dolan RJ . 2003 . The nose smells what the eye sees: crossmodal visual facilitation of human olfactory perception . Neuron . 39 ( 2 ): 375 - 386 .
Hettinger TP , Gent JF , Marks LE , Frank ME . 1999 . A confusion matrix for the study of taste perception . Percept Psychophys . 61 ( 8 ): 1510 - 1521 .
Kaiser H. 1970 . A second generation Little Jiffy . Psychometrika. 35 : 401 - 415 .
Kaiser H. 1974 . An index of factorial simplicity . Psychometrika . 39 : 31 - 36 .
Kemp SE , Gilbert AN . 1997 . Odor intensity and color lightness are correlated sensory dimensions . Am J Psychol . 110 ( 1 ): 35 - 46 .
Köhler W. 1929 . Gestalt psychology . New York : Liveright.
Köhler W. 1947 . Gestalt psychology . 2nd ed . New York : Liveright.
Lehrer A. 2009 . Wine & conversation. 2nd ed. Oxford: Oxford University Press.
Naor-Raz G , Tarr MJ , Kersten D. 2003 . Is color an intrinsic property of object representation? Perception . 32 ( 6 ): 667 - 680 .
Nelson MR , Hitchon JC . 1999 . Loud tastes, colored fragrances, and scented sounds: how and when to mix the senses in persuasive communications . Journalism Mass Comm Q . 76 : 354 - 372 .
Ngo M , Misra R , Spence C. 2011 . Assessing the shapes and speech sounds that people associate with chocolate samples varying in cocoa content . Food Qual Prefer . 22 : 567 - 572 .
Peynaud E. 1996 . The taste of wine: the art and science of wine appreciation . New York : John Wiley & Sons.
Ramachandran VS , Hubbard EM . 2001 . Synaesthesia-a window into perception, thought and language . J Consciousness Stud . 8 : 3 - 34 .
Sapir E. 1929 . A study in phonetic symbolism . J Exp Psychol . 12 : 225 - 239 .
Schifferstein HN , Tanudjaja I. 2004 . Visualising fragrances through colours: the mediating role of emotions . Perception . 33 ( 10 ): 1249 - 1266 .
Seo HS , Arshamian A , Schemmer K , Scheer I , Sander T , Ritter G , Hummel T. 2010 . Cross-modal integration between odors and abstract symbols . Neurosci Lett . 478 ( 3 ): 175 - 178 .
Seo HS , Hummel T. 2011 . Auditory-olfactory integration: congruent or pleasant sounds amplify odor pleasantness . Chem Senses . 36 ( 3 ): 301 - 309 .
Shepherd GM . 2006 . Smell images and the flavour system in the human brain . Nature . 444 ( 7117 ): 316 - 321 .
Spence C. 2011 . Crossmodal correspondences: a tutorial review . Atten Percept Psychophys . 73 ( 4 ): 971 - 995 .
Spence C. 2012 . Managing sensory expectations concerning products and brands: capitalizing on the potential of sound and shape symbolism . J Consum Psychol . 22 : 37 - 54 .
Spence C , Zampini M. 2006 . Auditory contributions to multisensory product perception . Acta Acustica united with Acustica . 92 : 1009 - 1025 .
Stevenson et al 1995 . The acquisition of taste properties by odors . Learn Motiv . 26 : 433 - 455 .
Walker P , Bremner JG , Mason U , Spring J , Mattock K , Slater A , Johnson SP . 2010 . Preverbal infants' sensitivity to synaesthetic cross-modality correspondences . Psychol Sci . 21 ( 1 ): 21 - 25 .
Ludwig VU , Adachi I , Matsuzawa T. 2011 . Visuoauditory mappings between high luminance and high pitch are shared by chimpanzees (Pan troglodytes) and humans . Proc Natl Acad Sci USA . 108 ( 51 ): 20661 - 20665 .
Yeomans MR , Chambers L , Blumenthal H , Blake A. 2008 . The role of expectancy in sensory and hedonic evaluation: the case of smoked salmon ice-cream . Food Qual Prefer . 19 : 565 - 573 .
Martino G , Marks LE . 2001 . Synesthesia: strong and weak . Curr Dir Psychol Sci . 10 : 61 - 65 .
Zarzo M. 2008 . Psychologic dimensions in the perception of everyday odors: pleasantness and edibility . J Sens Stud . 23 : 354 - 376 .