35.1. INTRODUCTION

Writing in the early nineteenth century, the gastronomic pioneer, Brillat-Savarin was “tempted to believe that smell and taste are in fact but a single sense, whose laboratory is the mouth and whose chimney is the nose” (Brillat-Savarin 1825). Much of the subsequent history of perception research in the chemical senses has, in contrast, been characterized by a focus on discrete sensory channels, and their underlying anatomy and physiology. However, there has recently been renewed interest in examining flavor as a functional perceptual system. This has been borne to some extent out of a realization that in our everyday food experiences, we respond, perceptually and hedonically, not to discrete tastes, odors, and tactile sensations, but to flavors constructed from a synthesis of these sensory signals (Prescott 2004b).

This refocus regarding flavor is very much in line with the ecological approach to perception that had been advocated by Gibson (1966). Gibson argued that the primary purpose of perception is to seek out objects in our environment, particularly those that are biologically important. As such, the physiological origin of sensory information is less salient than that the information can be used in object identification. Effectively, then, the key to successful perception is that sensory information is interpreted as qualities that belong to the object itself. Within this context, flavor can be seen as a functionally distinct sense that is cognitively “constructed” from the integration of distinct physiologically defined sensory systems (such as olfaction and gustation) that are “functionally united when anatomically separated” (Gibson 1966, p. 137) in order to identify and respond to objects that are important to our survival, namely, foods.

35.2. CHEMOSENSORY INTERACTIONS AND INTEGRATION

Cross-modal sensory integration is frequently inferred from the influence of one modality on responses to another. Commonly, this is an enhanced (sometimes supra-additive) response to information from one sensory system due to concurrent input from another modality (Calvert et al. 1999). For example, in a noisy environment, speech comprehension is improved if we see the speaker's lip movements (Sumby and Polack 1954). Even information that is irrelevant to a task enhances neural response to task-relevant stimuli and augments behavioral performance (Stein et al. 1988). There is similarly evidence that tastes and odors, when encoded together as a flavor, interact to modify the perception of one another.

The most obvious expression of odor–taste interactions is the widely observed attribution of qualities that are more usually associated with basic taste qualities to odors (Burdach et al. 1984). When asked to describe the odor of caramel or vanilla, most people will use the term “sweet-smelling”; similarly, “sour” is used for the odor of vinegar (Stevenson and Boakes 2004). In one descriptive analysis of characteristics of a wide range of odors (Dravnieks 1985), 65% of assessors gave “sweetness” as an appropriate descriptor for the odor of vanillin, whereas 33% described the odor of hexanoic acid as being sour. These descriptions appear to have many of the qualities of synesthesia, in which a stimulus in one sensory modality reliably elicits a consistent corresponding stimulus in another modality (Martino and Marks 2001; Stevenson et al. 1998). Whereas in other modalities, synesthesia is a relatively uncommon event, the possession of taste properties by odors is almost universal, particularly in the case of commonly consumed foods. In fact, for some odors, taste qualities may represent the most consistent description used. Stevenson and Boakes (2004) reported data showing that, over repeat testing, ratings of taste descriptors for odors (e.g., sweetness of banana odor) were at least as reliable as ratings of the primary quality of the odor (i.e., banana).

This commonplace phenomenon could be dismissed as merely imprecise language (since highly specific odor descriptors are elusive) or even metaphor, given that the odor name is likely to refer to an object, which might also be sweet or sour. However, there are measurable consequences of such odor taste qualities, in that these odors, when added to tastants in solution, can modify the taste intensity. The most frequent finding is the ability of food odors such as strawberry or vanilla to enhance the sweetness of sucrose solutions (Frank and Byram 1988; Frank et al. 1989). This phenomenon is both taste- and odor-specific. For example, the sweet-smelling odor of strawberry will enhance a sweet taste, but the odor of bacon will not. Conversely, a nonsweet taste, for example, saltiness, will not be enhanced by strawberry (Frank and Byram 1988). Stevenson et al. (1999) showed that the smelled sweetness of an odorant was the best predictor of that odorant's ability to enhance a sweet taste when they were presented together in solution. Similarly, the ability of food odors to enhance saltiness in solution has been shown as highly correlated with the extent to which the foods themselves were judged to be salty (Lawrence et al. 2009).

Subsequently, these findings were extended by studies showing that odors added to tastants can also suppress taste intensity. Prescott (1999) found that odors judged to be low in smelled sweetness (peanut butter, oolong tea) suppressed sweetness when added to sucrose in solution, in contrast to raspberry odor, which enhanced it. Stevenson et al. (1999) reported that sweet-smelling caramel odor not only enhanced the sweetness of sucrose in solution but also suppressed the sourness of a citric acid solution. Importantly, this latter effect parallels the pattern of interactions seen with binary taste mixtures, in that the addition of sucrose would similarly suppress the sourness of citric acid. Such findings provide evidence that odor taste properties reflect a genuine perceptual phenomenon.

The ability of odors possessing smelled taste qualities to influence tastes has also been demonstrated in paradigms using measures other than intensity ratings. Dalton et al. (2000) assessed orthonasal (sniffed) detection thresholds for the odorant benzaldehyde, which has a cherry/almond quality, while subjects held a sweet taste (saccharin) in the mouth. Detection thresholds for the odor was significantly reduced compared with benzaldehyde alone, or in combination with either water or a nonsweet taste (monosodium glutamate, a savory quality). The most plausible interpretation of these findings is that the smelled sweetness of benzaldehyde and tasted sweetness of saccharin were being integrated at subthreshold levels. Similar odor threshold effects have also been found using a somewhat different experimental protocol in which both the odorant and tastant were presented together in solution (Delwiche and Heffelfinger 2005).

Reciprocal effects of odors on tastes are also found. These include increases in the detection accuracy of a sweet taste at around threshold in the presence of an orthonasally presented congruent odorant (strawberry) as compared to one that was not sweet (ham) (Djordjevic et al. 2004), as well as a similar effect using a priming procedure, in which the odorant preceded the taste presentation (Prescott 2004b), showing that a sweet-smelling odor produced a greater change in detectability, relative to no odor, than did another, nonsweet odorant. Similar priming effects at suprathreshold levels have been demonstrated behaviorally in a study in which subjects were asked to identify a taste quality sipped from a device that also simultaneously presented an odor—either congruent or incongruent—orthonasally. Speed of naming of tastes during presentation of congruent odor/taste pairs (sweet smelling cherry odor/sucrose; sour smelling grapefruit odor/citric acid) was faster relative to incongruent pairs (cherry odor/citric acid; grapefruit odor/sucrose), or neutral/control pairs (either butanol or no odor plus either sucrose or citric acid) (White and Prescott 2007).

35.3. ASSOCIATIVE LEARNING AND INTEGRATION

The importance of taste-related odor properties for understanding sensory integration in flavors derives principally from the fact that these effects are thought only to arise once the odor and taste have been repeatedly experienced together as a mixture in the mouth, most typically in the context of foods or beverages. This process has been repeatedly demonstrated experimentally. Novel odors that have little or no smelled sweetness, sourness, or bitterness when sniffed take on these qualities when repeatedly paired in solution with sweet, sour, or bitter tastes, respectively (Prescott 1999; Stevenson et al. 1995, 1998, 1999; Yeomans et al. 2006, 2009). Recent studies have expanded these findings beyond associative relationships of odors with tastes. Thus, odors paired with high fat milks themselves became fattier smelling and were able to increase the perceived fattiness of the milks, when added to the milks subsequent to conditioning (Sundquist et al. 2006).

Such acquired perceptual similarity has been seen as an example of a “learned synesthesia,” in which qualities in one sensory system (olfaction) are able to evoke qualities in another (taste) only as a result of their frequent co-occurrence (Stevenson et al. 1998). The nature of the change that the odor undergoes has been explained in terms of increasing congruency (similarity) with the taste, in that they possess qualities in common, as a result of coexposure (Frank et al. 1993; Schifferstein and Verlegh 1996). Hence, the sweetness of a taste such as sucrose is seen to be more congruent with the sweet-smelling odor of caramel than it is with the odor of bacon, which typically has no sweet smell. It is only after this coexposure that the odor enhances a (now) congruent taste (Prescott 1999; Stevenson et al. 1999). Thus, Frank et al. (1993) found that the degree of enhancement produced by an odor for a particular taste was significantly correlated with ratings of the perceived similarity or congruency of the odorant and tastant. This suggests, therefore, that whether an odor/taste combination is seen as congruent is dependent on prior association of the components as a combination.

Given the associative origin of these effects in the context of foods and beverages, we might expect cross-cultural differences in the extent to which particular odors and tastes are judged as congruent. For example, the odor of pumpkin is likely to smell sweeter in those cultures where it is incorporated into desserts (e.g., United States) as compared to cultures where it is savoury. Consistent with this, it has been reported that French and Vietnamese vary in their judgments of odor/taste harmony—that is, the extent to which an odor and taste are seen as congruent (Nguyen et al. 2002).

One explanatory model for these effects proposes that each experience of an odor always invokes a search of memory for prior encounters with that odor. If, in the initial experience of the odor, it was paired with a taste, a cross-modal configural stimulus—that is, a flavor—is encoded in memory. Subsequently sniffing the odor alone will evoke the most similar odor memory—the flavor— that will include both the odor and the taste component. Thus, for example, sniffing caramel odor activates memorial representations of caramel flavors, which includes a sweet taste component. This results either in perceptions of smelled taste properties such as sweetness or, in the case of a mixture, a perceptual combination of the memorial odor representation with the physically present taste in solution (Stevenson and Boakes 2004; Stevenson et al. 1998).

35.4. CROSS-MODAL CHEMOSENSORY BINDING

In vision, aspects of a scene or object include features such as form, color, or movement, combined to form a coherent perception. The neural processing of form can be shown to be independent of that of color, but our perception is always that the two visual phenomena are bound seamlessly together. To understand flavor perception, it is similarly crucial to know the mechanisms responsible for binding tastes, odors, and tactile sensations into one coherent, cross-modal percept.

Studies of interactions between visual, auditory, and somatosensory systems have demonstrated the importance of spatial and/or temporal contiguity in facilitating cross-modal sensory integration (Calvert et al. 1998; Driver and Spence 2000; Spence and Squire 2003). In flavors, the different stimulus elements are associated temporally. However, although both gustatory and somatosensory receptors are spatially located in the mouth, olfactory receptors are not. The question then arises of how odors become bound to taste and touch. Central to this process is the olfactory location illusion, in which the odor components of a food appear to originate in the mouth (Rozin 1982). Thus, we never have a sense that the oranginess of orange juice is being perceived within the nose, even if we are aware that it is an odor. This illusion is both strong and pervasive, despite the fact that we are frequently presented with evidence of the importance of the olfactory component in flavors, for example, through a blocked nose during a head cold. One common manifestation of this phenomenon is the interchangeability of chemosensory terms such as flavor and taste in common usage— that is, we routinely fail to make a distinction between olfactory and taste qualities within flavors.

The location illusion itself may depend on both the spatial and temporal contiguity of the discrete sensory inputs. von Bekesy (1964) illustrated the likely importance of temporal factors as potential determinants of odor/taste integration by showing that the perceived location of an odor (mouth vs. nose) and the extent to which an odor and taste were perceived as one sensation or two could be manipulated by varying the time delay between the presentation of the odor and taste. With a time delay of zero (simultaneous presentation), the apparent locus of the odor was the back of the mouth and the odor/taste mixture was perceived as a single entity. When the odor preceded the taste, the sensation was perceived as originating in the nose (see Figure 35.1). Although this report is consistent with models of binding across other sensory modalities, von Bekesy (1964) did not provide sufficient details to judge the reliability of his conclusions. The number of other studies addressing this issue is also very limited. A demonstration that odor-induced taste enhancement can occur whether the odor is presented orthonasally or retronasally, providing that the odor and taste are presented simultaneously (Sakai et al. 2001) does suggest a key role for temporal synchrony in facilitating integration. Pfieffer et al. (2005) manipulated both spatial and temporal contiguity for the odor and taste while assessing the threshold for benzaldehyde odor (almond/cherry) in the presence of a sub-threshold sweet taste, failing to find convincing evidence of their manipulations. However, a recent preliminary finding suggests that synchronicity judgments of odor and taste may be less sensitive to onset discrepancies than other multimodal stimulus pairs, including audiovisual stimuli, and odors and tastes, each paired with visual stimuli (Kobayakawa et al. 2009). One interpretation of such a finding, if confirmed, together with the data of Sakai et al. (2001), would be that odor–taste binding operates under less stringent requirements for spatiotemporal synchrony than multisensory integration within other sensory systems. In turn, binding under conditions in which there is a tolerance for asynchrony might reflect the high adaptive significance of chemosensory binding. Alternatively, at least in the case of temporal asynchrony, congruency between the odor and taste may be crucial. Hence, it has been demonstrated that judgments of audiovisual asynchrony are more difficult when the different modalities are bound by a common origin (Spence and Parise 2010).

FIGURE 35.1

Temporal and spatial determinants of odor/taste integration. Combination of smell and taste into a single sensation. A varying time difference between stimuli moves locus of sensation from tip of the nose back to the throat and forward again to tip of (more...)

The olfactory location illusion is effectively an equivalent phenomenon to the auditory/visual “ventriloquism effect” in that, like the ventriloquist's voice, the location of the odor is captured by other sensory inputs. The extent to which either concurrent taste or somatosensation, or both, is chiefly responsible for the capture and referral of olfactory information to the oral cavity is not known. However, the somatosensory system is more strongly implicated since it provides more detailed spatial information than does taste (Lim and Green 2008). Moreover, in neuroimaging studies, odors that are available to bind with tastes—that is, those presented retronasally (via the mouth) —have been shown to activate the mouth area of the primary somatosensory cortex, whereas the same odors presented via the nose do not (Small et al. 2005). This distinction, which occurs even when subjects are unaware of route of stimulation, suggests a likely neural correlate of the binding process, and supports the idea that somatosensory input is the underlying mechanism.

In fact, our tastes experiences may themselves be multimodal. Under most circumstances, taste and tactile sensations in the mouth are so well integrated that we cannot begin to disentangle them, and there is growing evidence that our everyday experiences of taste are themselves multisensory, in that they involve somatosensory input (Green 2003; Lim and Green 2008). Taste buds are innervated by somatosensory fibers (Whitehead et al. 1985) and various categories of somatosensory stimuli are also capable of inducing taste sensations. Thus, it has been noted that about 25% of fungiform papillae respond to tactile stimulation by fine wires with a taste quality (Cardello 1981). More recently, tastes have been shown to be elicited by heated and cooled probes placed on areas innervated by cranial nerves VII and IX, which subserve taste (Cruz and Green 2000), and by the application of the prototypical “pure” irritant, capsaicin, to circumvallate papillae (Green and Hayes 2003). Further evidence points to the ability of tactile stimulation to capture taste, presumably by providing superior spatial information and enhancing localization (Delwiche et al. 2000; Lim and Green 2008; Todrank and Bartoshuk 1991). Tactile information may therefore have an important role in binding tastes, perhaps together with odors, both to one another and to a physical stimulus such as a food.

The binding of odors to tastes and tactile stimuli may also rely on processing information about the origins of odor stimulation. Orthonasally presented odors are more readily identified and have lower thresholds than the same odors presented retronasally via the mouth (Pierce and Halpern 1996; Voirol and Daget 1986), and there is a strong suggestion that the two routes of stimulation are processed with some independence. Thus, neuroimaging studies show different activation patterns in cortical olfactory areas as a result of route of administration (Small et al. 2005). From an adaptive point of view, this makes sense. Olfaction has been described (Rozin 1982) as the only dual sense because it functions both to detect volatile chemicals in the air (orthonasal sniffing) and to classify objects in the mouth as foods or not, and each of these roles has unique adaptive significance. Since the mouth acts as the gateway to the gut, our chemical senses can be seen as part of a defense system to protect our internal environment—once something is placed in the mouth, there is high survival value in deciding whether consumption is appropriate. Sensory qualities (tastes, retronasal odors, tactile qualities) occurring together in the mouth are therefore bound into a single perception, which identifies a substance as a food (cf. Gibson 1966).

35.5. ATTENTIONAL PROCESSES IN BINDING

Even though an odor's sniffed “taste” qualities and its ability to enhance that taste in solution are highly correlated (Stevenson et al. 1999), demonstrating, for example, that a sweet-smelling odor can enhance the sweetness of sucrose in solution appears to operate under some constraints. This became evident from findings that whether an odor enhances taste was dependent on task requirements. Thus, Frank et al. (1993) found that although strawberry odor enhanced the sweetness of sucrose in solution when the subjects were asked to judge only sweetness, the enhancement was not evident when other sensory qualities of these mixtures, such as sourness and fruitiness, were rated as well. In addition, the sweetness of the strawberry/sucrose mixtures was suppressed when the subjects rated total intensity of the mixture and then partitioned their responses into sweet, salty, sour, bitter, and/or other tastes. Interestingly, these effects were also noted for some taste mixtures, in which the elements are often judged as similar (e.g., sour/bitter taste mixtures), but not others with dissimilar components (e.g., sweet/bitter mixtures; Frank et al. 1993). Similarly, significantly less sweetness enhancement was found when subjects rated the odor as well as taste intensity of flavors (sweetness plus strawberry or vanilla) than when they rated sweetness alone (Clark and Lawless 1994).

In attempting to explain such effects, Frank and colleagues (Frank 2003; Frank et al. 1993; van der Klaauw and Frank 1996) suggested that, given perceptual similarity between an odor and taste, the conceptual “boundaries” that the subject sets for a given complex stimulus will reflect the task requirements. In the case of an odor/taste mixture in which the elements share a similar quality, combining those elements is essentially optional. This explanation invokes the notion that integration of perceptually similar dimensions is determined by the attentional focus demanded in the task. These effects of instructional sets are analogous to those seen in studies of cross-modal integration of vision and hearing. For example, focusing on the overall similarity of visual or auditory stimulus pairs, representing different stimulus dimensions, versus focusing on their component dimensions, can influence whether the pairs are treated as interacting or separable dimensions (Melara et al. 1992). This suggests the possibility that the apparent influence of the number of rating scales on odor/taste interactions results from the impact of these scales on how attention is directed toward the odor and taste. In keeping with this view, van der Klaauw and Frank (1996) were able to eliminate taste enhancement by directing subjects' attention to the appropriate attributes in a taste/odor mixture, even when they were only required to rate sweetness.

35.6. ANALYSIS AND SYNTHESIS IN PERCEPTION OF FLAVOR

These attentional effects appear to correspond to the differing modes of interaction that occur within sensory modalities. The blending of odors to form entirely new odors is a commonplace occurrence in flavor chemistry and perfumery (at least for odor mixtures with greater than two components; see Laing and Willcox 1983), and hence is referred to as synthetic interaction (analogous to the blending of light wavelengths). By contrast, the mixing of tastes is typically seen as an analytic process, because individual taste qualities do not fuse to form new qualities and, like simultaneous auditory tones, can be distinguished from one another in mixtures. A further category of interaction, namely, fusion—the notion of sensations combined to form a single percept, rather than combining synthetically to form a new sensation—has also been proposed and applied to flavor perception (McBurney 1986).

The notion of fusion in flavor perception implies that the percept remains analyzable into its constituent elements even when otherwise perceived as a whole. Thus, although our initial response is to apple flavor—an effortless combining of all of its sensory qualities into a single percept—we can, if required, switch between a synthetic approach to flavor and an analysis of the flavor elements. Hence, apple flavor can be both a synthetic percept and, with minimal effort, a collection of tastes (sweet; sour), textures (crisp; juicy) and odor notes (lemony; acetone-like; honey) (see Figure 35.2). A more precise way of conceptualizing flavor therefore is that cross-modal sensory signals are combined to produce a percept, rather than combining synthetically—in the way that odors themselves do—to form a new sensation. During normal food consumption, we typically respond to flavors synthetically—an approach reinforced by the olfactory illusion and by the extent to which flavor components are congruent. As noted earlier, this implies a sharing of perceptual qualities, for example, sweetness of a taste and of an odor, derived from prior experience of these qualities together.

FIGURE 35.2

Synthetic and analytic views of a flavor. In each case, sensory signals are identical, but perception differs—whole flavor of apple versus a collection of sensory qualities on which different apples may vary.

Conversely, analytic approaches to complex food or other flavor stimuli (e.g., wines) are often used by trained assessors to provide a descriptive profile of discrete sensory qualities, as distinct from an assessment of the overall flavor. Asking assessors to become analytical appears to produce the same inhibitory effects on odor–taste interactions noted in studies by Frank et al. (1993) and others. In one study using both trained descriptive panelists and untrained consumers (Bingham et al. 1990), solutions of the sweet-smelling odorant maltol plus sucrose were rated as sweeter than a solution of sucrose alone by the untrained consumers. In contrast, no such enhancement was found in the ratings of those trained to adopt an analytical approach to the sensory properties of this mixture.

In experimental paradigms, whether an odor/taste mixture is perceived analytically or synthetically can be determined by the responses required of the subject. Multiple ratings of appropriate attributes force an analytical approach, whereas a single rating of a sensory quality that can apply to both congruent odors and tastes (e.g., the tasted sweetness of sucrose and the smelled sweetness of strawberry odor) encourages synthesis of the common quality from both sensory modalities. The components of these flavors may not be treated separately when judged in terms of sweetness or other single characteristics. When instructions require separation of the components, however, this can be done—the components of a flavor are evaluated individually, and sweetness enhancement is eliminated. In other words, rating requirements lead to different perceptual approaches (analytical or synthetic) that, in turn, influence the degree of perceptual integration that occurs. A recent study of odor mixtures has indicated that an analytical approach is similarly able to influence the integration of the individual mixture components, as shown in a reduction in the extent to which subjects perceived a unique quality distinct from those of the components (Le Berre et al. 2008).

35.7. INVESTIGATING COGNITIVE PROCESSES IN FLAVOR PERCEPTION

Thus, the concept of fusion suggests that flavor perception is highly dependent on both past experience with specific odor/taste combinations (the origin of congruence) and cognitive factors that influence whether the flavor elements are combined or not. The most influential model of visual binding proposes that individual visual features are only loosely associated during early stages of processing, most likely by a common spatial location, but are bound to form a coherent perception as a result of attention directed toward combining these features as aspects of the same object or scene (Treisman 1998, 2006). Similarly, the configural account of odor/taste perceptual learning (Stevenson and Boakes 2004) implies that when attention is directed toward a flavor, it is attended to as a single compound or configuration, rather than a collection of elements. A configural explanation for the ability of an odor to later summate with the taste to produce enhanced sweetness implies an attentional approach that combines the odor and taste, rather than identifying them as separate elements in the flavor. In other words, for a complete binding of flavor features via configural learning, synthesis of the elements via attending to the whole flavor is critical. The limited evidence that exists suggests that the binding and joint encoding of odors, tastes, and tactile sensations is automatic. This is indicated both by the finding that perceptual changes in odors after pairing with tastes appears not to require conscious awareness on the part of the subject of the particular odor-taste contingencies (Stevenson et al. 1998) and data suggesting that a single coexposure of an odor and taste can result in transfer of the taste properties to the odor (Prescott et al. 2004). Thus, such learning should be sensitive to manipulations in which attention is directed toward the identity of their constituent elements.

One approach to examining the role of these factors has been to force subjects to adopt contrasting attentional strategies (analytic vs. synthetic) while either experiencing or judging odor/taste mixtures. If it is the case that odor/taste interactions can be influenced by the extent to which an analytical or synthetic perceptual approach is taken during rating, then this suggests the possibility that the extent to which the odors and tastes become integrated (as shown by increased perceptual similarity) might similarly be determined by the way in which the components of the flavor are associated during their joint exposure. In turn, any influence of odors on those tastes in solution may similarly be modulated. Hence, an exposure strategy that emphasizes the distinctiveness of the elements in the odor/taste mixture (an analytical perceptual strategy) should inhibit increases in the taste properties of the odor, and the subsequent ability of the odor to influence tastes in solution. In contrast, treating the elements as a synthetic whole is likely to encourage the blurring of the perceptual boundaries, fostering subsequent odor/taste interactions.

Consistent with this, pre-exposure of the elements of the specific odor–taste flavor compounds that were later repeatedly associated—in Pavlovian terms, unconditional stimulus or conditional stimulus pre-exposure—eliminated any change in the odors' perceptual qualities following the pairing (Stevenson and Case 2003). Thus, pre-exposed odors later paired with sweet or sour tastes did not become sweeter or sourer smelling, whereas taste-paired odors that had not been pre-exposed did. In contrast, initial attempts to disrupt configural integrity by directing attention toward the elemental nature of the compound stimulus during associative pairing were unsuccessful. Neither training subjects to distinguish the individual odor and taste components of flavors prior to learning (Stevenson and Case 2003) nor predisposing subjects to adopt an analytical strategy by requiring intensity ratings of these odor and taste components separately during their joint exposure (Prescott et al. 2004) were initially successful in influencing whether odors paired with a sweet tastes became sweeter smelling. This is probably attributable to methodological reasons.

If it is the case that odors and tastes are automatically coencoded as a flavor in the absence of task demands that focus attention on the elements, then experimental designs in which odor and taste elements appear together without such an attentional strategy are likely to predispose toward synthesis. Hence, the analytical strategy used by Stevenson and Case (2003) was likely to be ineffective since they asked subjects during the exposure to rate overall liking for the odor–taste compound, an approach that may have encouraged integration of the elements. The analytical manipulation in Prescott et al.'s (2004) study may not have influenced the development of smelled sweetness because it took place after the initial pairing of the sweet taste and odor that occurred before the formal associative process—that is, as the preconditioning measure in the pre–post design. As noted earlier, a second study in Prescott et al.'s (2004) report demonstrated that a single odor–sweet taste coexposure can produce an odor that smells sweet.

More recently, it has been demonstrated that when such methodological considerations are addressed, prior analytical training in which attention is explicitly directed toward the individual elements in an odor and sweet taste mixture does inhibit the development of a sweet-smelling odor (Prescott and Murphy 2009; see Figure 35.3a). In this study, subjects only ever received a particular odor taste combination under conditions in which they had been trained to respond to the combination in explicitly synthetic or analytical ways. Moreover, the fact that the training used different odor/taste combinations than were later used in the conditioning procedure suggests that an attentional approach (analytical or synthetic) was being induced in the subjects during training that was then applied to new odor/taste combinations during conditioning.

FIGURE 35.3

Changes in perceptual characteristics of odors and flavors as a function of odor–taste coexposure and attentional strategy. (a) Mean ratings of smelled sweetness of odors increase after repeat paired with a sweet taste in solution, but only for a (more...)

The findings from this study have important theoretical implications, in that they are clearly consistent with configural accounts of perceptual odor–taste learning and flavor representation (Stevenson and Boakes 2004; Stevenson et al. 1998). Under conditions where attention is directed toward individual stimulus elements during conditioning, the separate representation of these elements may be incompatible with learning of a configural representation. This explanation is supported by the demonstration that an analytical approach also acted to inhibit a sweet-smelling odor's ability to enhance a sweet taste when the odor/taste combination were evaluated in solution after repeated pairing (Prescott et al. 2004; see Figure 35.3b). In other words, an analytical attentional strategy can be shown to interfere with either the development of a flavor configuration resulting from associative learning, or the subsequent ability of this configuration to combine with a physically present tastant.

35.8. HEDONIC IMPLICATIONS OF CHEMOSENSORY INTEGRATION

Likes and dislikes naturally arise from the integrated perception of flavor, since we are responding to substances that we have learned to recognize as foods and that are therefore biologically, culturally, and socially valued. Initial (“gut”) responses to foods are almost always hedonic and this naturally precedes accepting or rejecting the food. Hence, perhaps unique among multisensory interactions, multisensory integration in the chemical senses is, to greater and lesser extents, a process that has an inherently hedonic dimension.

As with perceptual changes in odors, hedonic properties of flavors arise from associative learning. Although our initial responses to odors may or may not be strongly positively or negatively valenced, tastes evoke emotions that are innate (Prescott 1998; Steiner et al. 2001). Because of this hedonic valence, repeated pairing odors with tastes not only produces a transfer of perceptual properties, leading to odor–taste properties, but also a change in the hedonic character of the odor, and hence also of the flavor. Thus, repeat pairing of a novel odor that is hedonically neutral with a liked sweet taste typically produces an increase in liking for that odor; conversely, pairing with a bitter taste produces a disliked odor (Baeyens et al. 1990; Zellner et al. 1983). This form of learning, known as evaluative conditioning (EC; Levey and Martin 1975), is procedurally identical to odor–taste perceptual learning. Nevertheless, evaluative and perceptual associative conditioning can be distinguished by the fact that conditioned increases in the taste properties of odors can occur without consistent changes in liking (Stevenson et al. 1995, 1998) and also by the reliance of odor–taste evaluative, but not perceptual, learning on the motivational state of the individual. Hence, EC is reduced or eliminated under conditions of satiation, whereas perceptual learning is unaffected (Yeomans and Mobini 2006). EC, but not perceptual learning, also relies on the relative hedonic value of the tastes. Although even relatively weak bitterness per se is universally negative (Steiner et al. 2001), in adults there is variation in the extent to which sweetness is hedonically positive (Pangborn 1970). However, when this is controlled for, by selecting “sweet likers”—commonly defined as those whose hedonic responses tend to increase with increasing sweetener concentration—odors paired with sweet tastes reliably become more liked (Yeomans et al. 2006, 2009).

A configural or holistic learning model of the type discussed earlier in relation to perceptual changes in odors paired with tastes, also accounts for odor–taste evaluative learning by proposing that the configuration includes a hedonic component “supplied” by the taste, which is evoked when the odor or flavor is experienced (De Houwer et al. 2001). This model is supported for EC by an identical finding for analytical versus synthetic attention as that shown with perceptual learning. That is, training to identify the elemental nature of the odor–taste compound during learning also eliminates the transfer of hedonic properties from the taste to the odor (Prescott and Murphy 2009), suggesting that the formation of an odor–taste configuration that includes hedonic values has been inhibited. Recent evidence also suggests that, even after learning, the hedonic value of a flavor can be altered by the extent to which an analytical approach is taken to the flavor. Comparisons between acceptability ratings alone and the same ratings followed by a series of analytical ratings of flavor sensory qualities found a reduction of liking in the latter condition (Prescott et al. 2011), suggesting that analytical approaches are inhibitory to liking even once that liking has been established. The explanation for this effect is that, as with the similar effects on perceptual learning reported by Prescott et al. (2004), an analytical attentional strategy is induced by knowledge that the flavour is to be perceptually analyzed, reducing the configuration process responsible for the transfer of hedonic properties. This finding joins a number of others indicating that analytical cognitions are antagonistic toward the expression of likes and dislikes (Nordgren and Dijksterhuis 2008).

An additional consequence of EC has been demonstrated in studies that have measured the behavioral consequences of pairing an odor with a tastant that may be valued metabolically. A considerable body of animal (Myers and Sclafani 2001) and human (Kern et al. 1993; Prescott 2004a; Yeomans et al. 2008b) literature exists showing that odor–taste pairing leading to learned preferences is highly effective when a tastant that provides valued nutrients is ingested. This process can be shown to be independent of the hedonic value of the tastant—for example, by comparing conditioning of odors using sweet tastants that provide energy (e.g., sucrose) with those that do not (Mobini et al. 2007). As with EC generally, this form of postingestive conditioning is sensitive to motivational state and is maximized when conditioning and evaluation of learning take place under relative hunger (Yeomans and Mobini 2006). It has also been recently demonstrated that a novel flavor paired with ingested monosodium glutamate (MSG) not only increased in rated liking, even when tested without added MSG, but also, relative to a non-MSG control, produced behavioral changes including increases in ad libitum food intake and rated hunger after an initial tasting of the flavor (Yeomans et al. 2008).

Finally, one interesting behavioral consequence of odor–taste perceptual integration has been a demonstration that a sweet-smelling odor significantly increased pain tolerance relative to a noodor control (Prescott and Wilkie 2007). Given that the effect was not seen in an equally pleasant, but not sweet-smelling, odor, the conclusion drawn was that the odor sweetness was acting in an equivalent manner to sweet tastes, which have been shown to have this same effect on pain (Blass and Hoffmeyer 1991). Although the presumption is that such effects are also the result of the same learned integration that produces the sweet smell and the ability to modify taste perceptions, the crucial demonstration of this has yet to be carried out. It does suggest, however, that the process of elicitation of a flavor representation by an odor may have broad behavioral as well as perceptual and hedonic consequences.

There have been some recent attempts to explore the practical implications of odor–taste learning, opening opportunities to perhaps exploit its consequences. It has been shown, for example, that the enhancement of tastes by congruent odors seen in model systems (i.e., solutions) also occurs in foods, with bitter- and sweet-smelling odors enhancing their respective congruent tastes in milk drinks (Labbe et al. 2006). Also consistent with data derived from model systems was a failure in these studies for a sweet-smelling odor to enhance the sweetness of an unfamiliar beverage. Most recently, an examination of the potential for odors from a range of salty foods to enhance saltiness in solution (Lawrence et al. 2009) raised the possibility that such odors could be used to effectively reduce the sodium content of foods, without the typical concurrent loss of acceptability that occurs (Girgis et al. 2003). Similarly, the finding that odors can take on fatlike properties after associative pairing with fats (Sundquist et al. 2006) might allow odors to partially substitute for actual fat content in foods. These studies therefore point to an exciting prospect, in which research aimed at understanding multisensory processes in flavor perception may lead to applications that ultimately have important public health consequences.

REFERENCES

1. Baeyens F, Eelen P, Van den Bergh O, Crombez G. Flavor-flavor and color-flavor conditioning in humans. Learning and Motivation. 1990;21:434–445.
2. Bingham A.F, Birch G.G, De Graaf C, Behan J.M, Perring K.D. Sensory studies with sucrose-maltol mixtures. Chemical Senses. 1990;15(4):447–456.
3. Blass E.M, Hoffmeyer L.B. Sucrose as an analgesic for newborn infants. Pediatrics. 1991;87(2):215–218. [PubMed: 1987534]
4. Brillat-Savarin J.-A. The physiology of taste. London: Penguin Books: 1825. 1994.
5. Burdach K.J, Kroeze J.H.A, Koster E.P. Nasal, retronasal, and gustatory perception: An experimental comparison. Perception & Psychophysics. 1984;36(3):205–208. [PubMed: 6522211]
6. Calvert G.A, Brammer M.J, Bullmore E.T, Campbell R, Iversen S.D, David A.S. Response amplification in sensory-specific cortices during crossmodal binding. NeuroReport. 1999;10:2619–2623. [PubMed: 10574380]
7. Calvert G.A, Brammer M.J, Iversen S.D. Crossmodal identification. Trends in Cognitive Sciences. 1998;2(7):247–253. [PubMed: 21244923]
8. Cardello A.V. Comparison of taste qualities elicited by tactile, electrical and chemical stimulation of single human taste papillae. Perception & Psychophysics. 1981;29:163–169. [PubMed: 7255095]
9. Clark C.C, Lawless H.T. Limiting response alternatives in time-intensity scaling: An examination of the halo-dumping effect. Chemical Senses. 1994;19(6):583–594. [PubMed: 7735838]
10. Cruz A, Green B.G. Thermal stimulation of taste. Nature. 2000;403:889–892. [PubMed: 10706285]
11. Dalton P, Doolittle N, Nagata H, Breslin P.A.S. The merging of the senses: Integration of sub-threshold taste and smell. Nature Neuroscience. 2000;3:431–432. [PubMed: 10769380]
12. De Houwer J, Thomas S, Baeyens F. Associative learning of likes and dislikes: A review of 25 years of research on human evaluative conditioning. Psychological Bulletin. 2001;127(6):853–869. [PubMed: 11726074]
13. Delwiche J.F, Heffelfinger A.L. Cross-modal additivity of taste and smell. Journal of Sensory Studies. 2005;20:512–525.
14. Delwiche J.F, Lera M.L, Breslin P.A.S. Selective removal of a target stimulus localized by taste in humans. Chemical Senses. 2000;25:181–187. [PubMed: 10781025]
15. Djordjevic J, Zatorre R.J, Jones-Gotman M. Effects of perceived and imagined odors on taste detection. Chemical Senses. 2004;29:199–208. [PubMed: 15047594]
16. Dravnieks A. Atlas of odor character profiles. Philadelphia, PA: American Society for Testing and Materials: 1985.
17. Driver J, Spence C. Multisensory perception: Beyond modularity and convergence. Current Biology. 2000;10::R731–R735. [PubMed: 11069095]
18. Frank R.A. Response context affects judgments of flavor components in foods and beverages. Food Quality and Preference. 2003;14:139–145.
19. Frank R.A, Byram J. Taste-smell interactions are tastant and odorant dependent. Chemical Senses. 1988;13(3):445–455.
20. Frank R.A, Ducheny K, Mize S.J.S. Strawberry odor, but not red color, enhances the sweetness of sucrose solutions. Chemical Senses. 1989;14(3):371–377.
21. Frank R.A, van der Klaauw N.J, Schifferstein H.N.J. Both perceptual and conceptual factors influence taste-odor and taste-taste interactions. Perception & Psychophysics. 1993;54(3):343–354. [PubMed: 8414893]
22. Gibson J.J. The senses considered as perceptual systems. Boston: Houghton Mifflin Company: 1966.
23. Girgis S, Neal B, Prescott J. A one-quarter reduction in the salt content of bread can be made without detection. European Journal of Clinical Nutrition. 2003;57(4):616–620. [PubMed: 12700625]
24. Green B.G. Studying taste as a cutaneous sense. Food Quality and Preference. 2003;14:99–109.
25. Green B.G, Hayes J.E. Capsaicin as a probe of the relationship between bitter taste and chemesthesis. Physiology & Behavior. 2003;79:811–821. [PubMed: 12954427]
26. Kern D.L, Mc Phee L, Fisher J, Johnson S, Birch L.L. The postingestive consequences of fat condition preferences for flavors associated with high dietary fat. Physiology & Behavior. 1993;54:71–76. [PubMed: 8327611]
27. Kobayakawa T, Toda H, de van Gotow N. 2009. Synchronicity judgement of gustation and olfaction. Paper presented at the Association for Chemoreception Sciences, Sarasota, FL.
28. Labbe D, Damevin L, Vaccher C, Morgenegg C, Martin N. Modulation of perceived taste by olfaction in familiar and unfamiliar beverages. Food Quality and Preference. 2006;17:582–589.
29. Laing D.G, Willcox M.E. Perception of components in binary odour mixtures. Chemical Senses. 1983;7(3–4):249–264.
30. Lawrence G, Salles C, Septier C, Busch J, Thomas-Danguin T. Odour-taste interactions: A way to enhance saltiness in low-salt content solutions. Food Quality and Preference. 2009;20(3):241–248.
31. Le Berre E, Thomas-Danguin T, Beno N, Coureaud G, Etievant P, Prescott J. Perceptual processing strategy and exposure influence the perception of odor mixtures. Chemical Senses. 2008;33:193–199. [PubMed: 18071197]
32. Levey A.B, Martin I. Classical conditioning of human ‘evaluative’ responses. Behavior Research & Therapy. 1975;13:221–226. [PubMed: 1191163]
33. Lim J, Green B.G. Tactile interaction with taste localization: Influence of gustatory quality and intensity. Chemical Senses. 2008;33:137–143. [PubMed: 17962229]
34. Martino G, Marks L.E. Synesthesia: Strong and weak. Current Directions in Psychological Science. 2001;10(2):61–65.
35. McBurney D.H. Taste, smell, and flavor terminology: Taking the confusion out of fusion. In: Meiselman H.L, Rivkin R.S, editors. Clinical measurement of taste and smell. New York: Macmillan: 1986. pp. 117–125.
36. Melara R.D, Marks L.E, Lesko K.E. Optional processes in similarity judgments. Perception & Psychophysics. 1992;51(2):123–133. [PubMed: 1549431]
37. Mobini S, Chambers L.C, Yeomans M.R. Interactive effects of flavour-flavour and flavour-consequence learning in development of liking for sweet-paired flavours in humans. Appetite. 2007;48:20–28.
38. Myers K.P, Sclafani A. Conditioned enhancement of flavor evaluation reinforced by intragastric glucose: I. Intake acceptance and preference analysis. Physiology & Behavior. 2001;74:481–493. [PubMed: 11790408]
39. Nguyen D.H, Valentin D, Ly M.H, Chrea C, Sauvageot F. When does smell enhance taste? Effect of culture and odorant/tastant relationship. 2002 Paper presented at the European Chemoreception Research Organisation conference, Erlangen, Germany.
40. Nordgren L.F, Dijksterhuis A.P. The devil is in the deliberation: Thinking too much reduces preference consistency. Journal of Consumer Research. 2008;36:39–46.
41. Pangborn R.M. Individual variation in affective responses to taste stimuli. Psychonomic Science. 1970;21(2):125–126.
42. Pfieffer J.C, Hollowood T.A, Hort J, Taylor A.J. Temporal synchrony and integration of sub-threshold taste and smell signals. Chemical Senses. 2005;30:539–545. [PubMed: 16049150]
43. Pierce J, Halpern B. Orthonasal and retronasal odorant identification based upon vapor phase input from common substances. Chemical Senses. 1996;21(5):529–543. [PubMed: 8902282]
44. Prescott J. Comparisons of taste perceptions and preferences of Japanese and Australian consumers: Overview and implications for cross-cultural sensory research. Food Quality and Preference. 1998;9(6):393–402.
45. Prescott J. Flavour as a psychological construct: Implications for perceiving and measuring the sensory qualities of foods. Food Quality and Preference. 1999;10:349–356.
46. Prescott J. Effects of added glutamate on liking for novel food flavors. Appetite. 2004a;42(2):143–150. [PubMed: 15010178]
47. Prescott J. Psychological processes in flavour perception. In: Taylor A.J, Roberts D, editors. Flavour perception. London: Blackwell Publishing: 2004b. pp. 256–277.
48. Prescott J, Johnstone V, Francis J. Odor/taste interactions: Effects of different attentional strategies during exposure. Chemical Senses. 2004;29:331–340. [PubMed: 15150146]
49. Prescott J, Murphy S. Inhibition of evaluative and perceptual odour-taste learning by attention to the stimulus elements. Quarterly Journal of Experimental Psychology. 2009;62:2133–2140. [PubMed: 19565428]
50. Prescott J, Kim K.-O, Lee S.M. Analytic approaches to evaluation modify hedonic responses. Food Quality and Preference. 2011;22:391–393.
51. Prescott J, Wilkie J. Pain tolerance selectively increased by a sweet-smelling odor. Psychological Science. 2007;18(4):308–311. [PubMed: 17470253]
52. Rozin P. “Taste-smell confusions” and the duality of the olfactory sense. Perception & Psychophysics. 1982;31(4):397–401. [PubMed: 7110896]
53. Sakai N, Kobayakawa T, Gotow N, Saito S, Imada S. Enhancement of sweetness ratings of aspartame by a vanilla odor presented either by orthonasal or retronasal routes. Perceptual and Motor Skills. 2001;92:1002–1008. [PubMed: 11565908]
54. Schifferstein H.N. J, Verlegh P.W.J. The role of congruency and pleasantness in odor-induced taste enhancement. Acta Psychologica. 1996;94:87–105. [PubMed: 8885712]
55. Small D.M, Gerber J.C, Mak Y.E, Hummel T. Differential neural responses evoked by orthonasal versus retronasal odorant perception in humans. Neuron. 2005;47:593–605. [PubMed: 16102541]
56. Spence C, Squire S. Multisensory integration: Maintaining the perception of synchrony. Current Biology. 2003;13::R519–R521. [PubMed: 12842029]
57. Spence C, Parise C. Prior-entry: A review. Consciousness and Cognition. 2010;19:364–379. [PubMed: 20056554]
58. Stein B.E, Huneycutt W.S, Meredith M.A. Neurons and behavior: The same rules of multisensory integration apply. Brain Research. 1988;448:355–358. [PubMed: 3378157]
59. Steiner J.E, Glaser D, Hawilo M.E, Berridge K.C. Comparative expression of hedonic impact: Affective reactions to taste by human infants and other primates. Neuroscience & Biobehavioral Reviews. 2001;25:53–74. [PubMed: 11166078]
60. Stevenson R.J, Boakes R.A. Sweet and sour smells: The acquisition of taste-like qualities by odors. In: Calvert G, Spence C.B, Stein B, editors. Handbook of multisensory processes. Cambridge, MA: MIT Press; 2004. pp. 69–83.
61. Stevenson R.J, Boakes R.A, Prescott J. Changes in odor sweetness resulting from implicit learning of a simultaneous odor-sweetness association: An example of learned synesthesia. Learning and Motivation. 1998;29:113–132.
62. Stevenson R.J, Case T.I. Preexposure to the stimulus elements, but not training to detect them, retards human odour-taste learning. Behavioural Processes. 2003;61:13–25. [PubMed: 12543479]
63. Stevenson R.J, Prescott J, Boakes R.A. The acquisition of taste properties by odors. Learning & Motivation. 1995;26:1–23.
64. Stevenson R.J, Prescott J, Boakes R.A. Confusing tastes and smells: How odors can influence the perception of sweet and sour tastes. Chemical Senses. 1999;24:627–635. [PubMed: 10587495]
65. Sumby W.H, Polack I. Visual contribution to speech intelligibility in noise. Journal of the Acoustical Society of America. 1954;26:212–215.
66. Sundquist N, Stevenson R.J, Bishop I.R.J. Can odours acquire fat-like properties? Appetite. 2006;47:91–99. [PubMed: 16723173]
67. Todrank J, Bartoshuk L.M. A taste illusion: Taste sensation localised by touch. Physiology & Behavior. 1991;50:1027–1031. [PubMed: 1805264]
68. Treisman A. Feature binding, attention and object perception. Philosophical Transactions of the Royal Society, London B. 1998;353:1295–1306. [PMC free article: PMC1692340] [PubMed: 9770223]
69. Treisman A. How the deployment of attention determines what we see. Visual Cognition. 2006;14:411–443. [PMC free article: PMC1832144] [PubMed: 17387378]
70. van der Klaauw N.J, Frank R.A. Scaling component intensities of complex stimuli: The influence of response alternatives. Environment International. 1996;22(1):21–31.
71. Voirol E, Daget N. Comparative study of nasal and retronasal olfactory perception. Lebensmittel-Wissenschaft und-Technologie. 1986;19:316–319.
72. von Bekesy G. Olfactory analogue to directional hearing. Journal of Applied Physiology. 1964;19:369–373. [PubMed: 14173530]
73. White T.L, Prescott J. Chemosensory cross-modal Stroop effects: Congruent odors facilitate taste identification. Chemical Senses. 2007;32:337–341. [PubMed: 17307760]
74. Whitehead M.C, Beeman C.S, Kinsella B.A. Distribution of taste and general sensory nerve endings in fungiform papillae of the hamster. American Journal of Anatomy. 1985;173:185–201. [PubMed: 20726120]
75. Yeomans M.R, Gould N, Mobini S, Prescott J. Acquired flavor acceptance and intake facilitated by monosodium glutamate in humans. Physiology & Behavior. 2008a;93:958–66. [PubMed: 18279899]
76. Yeomans M.R, Leitch M, Gould N.J, Mobini S. Differential hedonic, sensory and behavioral changes associated with flavor-nutrient and flavor-flavor learning. Physiology & Behavior. 2008b;93:798–806. [PubMed: 18201736]
77. Yeomans M.R, Mobini S. Hunger alters the expression of acquired hedonic but not sensory qualities of food-paired odors in humans. Journal of Experimental Psychology: Animal Behavior Processes. 2006;32(4):460–466. [PubMed: 17044748]
78. Yeomans M.R, Mobini S, Elliman T.D, Walker H.C, Stevenson R.J. Hedonic and sensory characteristics of odors conditioned by pairing with tastants in humans. Journal of Experimental Psychology: Animal Behavior Processes. 2006;32(3):215–228. [PubMed: 16834490]
79. Yeomans M.R, Prescott J, Gould N.G. Individual differences in responses to tastes determine hedonic and perceptual changes to odours following odour/taste conditioning. Quarterly Journal of Experimental Psychology. 2009;62(8):1648–1664. [PubMed: 19180363]
80. Zellner D.A, Rozin P, Aron M, Kulish C. Conditioned enhancement of human's liking for flavor by pairing with sweetness. Learning and Motivation. 1983;14:338–350.