Experimental Brain Research

, Volume 159, Issue 3, pp 405–408

Odor-induced changes in taste perception


    • Montreal Neurological InstituteMcGill University
    • Montreal Neurological Institute
  • R. J. Zatorre
    • Montreal Neurological InstituteMcGill University
  • M. Jones-Gotman
    • Montreal Neurological InstituteMcGill University
Research Note

DOI: 10.1007/s00221-004-2103-y

Cite this article as:
Djordjevic, J., Zatorre, R.J. & Jones-Gotman, M. Exp Brain Res (2004) 159: 405. doi:10.1007/s00221-004-2103-y


We investigated odor-induced changes in taste perception (OICTP), by examining the influence of strawberry and soy sauce odors on perceived sweetness (Experiment 1) and saltiness (Experiment 2). We explored whether taste-smell interactions occur at the central level, by delivering odorants (strawberry, soy sauce, odorless water) and tastants (sucrose, sodium chloride) separately, and whether effects of imagined odors are comparable to those of physically presented odors. We found specific taste-smell interactions: sweetness enhancement induced by strawberry odor and saltiness enhancement induced by soy sauce odor. These interactions were elicited with separate delivery of olfactory and gustatory stimuli. Secondly, we found a similar but rather limited effect with the imagined odors: imagined strawberry enhanced perceived sweetness of water solutions, and imagined soy sauce enhanced perceived saltiness of weak sodium chloride solutions. We concluded that OICTP is a centrally mediated phenomenon, and that imagined odors can to some extent induce changes in perceived taste intensity comparable to those elicited by perceived odors.


Taste-smell interactionOdor imagerySweetnessSaltinessOdorantTastant


Effects that odorants exert on taste perception have been shown to be odorant- and tastant-specific (Frank and Byram 1988; Frank et al. 1993; Schifferstein and Verglegh 1996; Stevenson et al. 1999). The phenomenon that a sipped solution tastes sweeter when presented simultaneously with a “sweet smelling” or “sweet congruent” odorant came to be known as odor-induced enhancement of taste perception. Subsequent studies showed that specific taste-smell interactions occur with taste qualities other than sweetness, such as bitterness (Frank et al. 1993), sourness (Shaffer and Frank 1990; Stevenson et al. 1999), and saltiness (Shaffer and Frank 1990).

In previous studies of specific odor-induced changes in taste perception (OICTP), odorants and tastants were presented together in mixtures that were sipped. Such stimulus delivery did not permit the question of origin of this effect (i.e. peripheral or central) to be resolved. Sakai et al. (2001) showed that sweetness enhancement occurs regardless of whether odorants and tastants are delivered mixed or separately, however without any control odorants and tastants. Dalton et al. (2000) contributed to this question by demonstrating that taste-induced changes in odor perception occur in the brain and not in the mouth: during separate presentation of congruent odorants and tastants, olfactory sensitivity was better than the baseline sensitivity or the one obtained with noncongruent tastants. In the present study, we sought to provide such evidence in the other direction, for odor-induced changes in taste perception.

The first question addressed was whether OICTP are a peripheral or a central phenomenon, by separately delivering odorants (nasally) and tastants (orally). The demonstration of this phenomenon with separate delivery of gustatory and olfactory stimuli would imply its central origin, whereas its absence with such a delivery would suggest its peripheral nature. Previously, we demonstrated OICTP as measured by detection of weak tastants (Djordjevic et al. 2004a), and here we examine whether this phenomenon can be demonstrated with another measure of taste perception, i.e. perceived intensity of specific taste qualities. The second question addressed was whether imagined odors could induce similar specific changes in taste perception as those induced by real odors. Finding that imagined odors have a comparable effect to real odors on another perceptual process (perceived taste intensity) would be consistent with the existence of odor imagery as a sensory-specific form of mental imagery, which is currently an open debate (Elmes 1998).

Experiment 1


Forty healthy volunteers were randomly assigned to one of two conditions (20 to each): perception (13 women, mean age = 20.1 years, range 17–23) or imagery (14 women, mean age = 19.6 years, range 17–26). All participants gave informed consent and the study was approved by the McGill University Research Ethics Board, and was in accordance with standards of the 1964 Declaration of Helsinki. Olfactory stimuli were strawberry (10% diluted in double distilled water), a commercially available soy sauce, and water as an odorless stimulus. Gustatory stimuli were aqueous solutions of sucrose (sugar) in three concentrations: zero (0 M), weak (0.1 M), and strong (0.56 M). Participants sniffed or imagined the odorant (depending on the condition), sipped the tastant, and then rated the sweetness of each odorant-tastant combination on a 21-point scale. The interval between trials was 30 s. Each stimulus combination was presented twice and the average of the two ratings was used for analysis.


A (3 × 3) repeated measures ANOVA was conducted for the two dependent variables (perceived sweetness induced by presented odors, and perceived sweetness induced by imagined odors), with the two factors being odor (strawberry, soy sauce, no odor) and sugar concentration (zero, weak, strong). Results for presented odors are shown in Fig. 1A. Given that the interaction between odor and sugar concentration was significant (F =3.58, p =.02), we conducted tests of simple effects. The effect of odor was significant within all three concentrations of sucrose (p <.000). Post hoc pairwise comparisons revealed a different pattern of differences: within the zero and the weak concentration, solutions were perceived as sweeter with strawberry than with no odor or soy sauce (p <.003 or less). Within the strong concentration, solutions were perceived as sweeter with strawberry than with soy sauce (p <.01), and with no odor than with soy sauce (p <.01).
Fig. 1

Perceived sweetness as a function of presented (A) or imagined (B) odors (no odor, strawberry, soy sauce) and sugar concentrations (zero, weak, strong)

We conducted the same analyses for perceived sweetness induced by imagined odors (Fig. 1B). The predicted effect of imagined odors was not significant (p =.24), nor was its interaction with sugar concentration (p =.62). We conducted planned comparisons as we had made a priori predictions of specific effects. Only one of the differences predicted in advance was significant, within the zero concentration: water solutions (zero sugar concentration) were rated as sweeter with imagined strawberry odor than with no odor (p =.04).

Test-retest reliability (correlation between the first and second rating by each participant) was satisfactory, r =.81 for perception and .83 for imagery, both p <.000, indicating reasonable measurement of perceived sweetness.

In order to examine individual differences in odor imagery, we calculated for each participant an index of sweetness enhancement induced by imagined strawberry odor (combining enhancement with strawberry versus no odor and strawberry versus soy sauce, and collapsing these over three sugar concentrations). We also established that good odor imagery ability should be associated with a minimum average rating difference of 1, on a 21-point scale. We found that only four participants (20%) showed significant sweetness enhancement elicited by imagined strawberry odor. Their mean enhancement score was 2.73, whereas the mean for the whole group was 0.50. Notably, mean sweetness enhancement elicited by presented strawberry odor was 2.81.

Experiment 2


The same 40 volunteers participated in Experiment 2, and the same olfactory stimuli were used. The gustatory stimulus was sodium chloride (salt), given in three concentrations: zero (0 M), weak (0.056 M), and strong (0.32 M). The procedure was similar in all respects to the procedure in Experiment 1, but we measured odor-induced effects on perceived saltiness.


Experiment 2 has the same factorial design (3 × 3) as Experiment 1. Results for the effects of presented odors on perceived saltiness are shown in Fig. 2A. The interaction between odor and salt concentration was significant, F(4,76)=3.27, p =.03. The effect was significant within zero (F(2,38)=15.63, p <.01) and weak (F(2,38)=9.14, p <.01), but not within the strong concentration of sodium chloride (p =.29). Post hoc comparisons revealed that within the zero concentration, presented solutions (water) were perceived as saltier with soy sauce than with strawberry and no odor (p <.01). Within the weak salt concentration, solutions were perceived as saltier with soy sauce than with no odor (p <.01).
Fig. 2

Perceived saltiness as a function of presented (A) or imagined (B) odors (no odor, strawberry, soy sauce) and salt concentrations (zero, weak, strong)

We conducted similar analyses for imagined odors (Fig. 2B). The interaction between imagined odor and salt concentration was significant, F(4,76)=3.77, p <.01. The only significant simple effect was within the weak salt concentration, F(2,38)=8.03, p <.01. Pairwise comparisons revealed that weak sodium chloride solutions were perceived as saltier when combined with imagined soy sauce than with no odor, p <.000.

Test-retest reliability was again reasonable: r =.82 for perception and .83 for imagery, both p <.000.

We calculated an index of saltiness enhancement elicited by imagined soy sauce for each participant, and we used the same cut-off of 1-point difference to define good odor imagery. We found that eight participants (40%) showed significant saltiness enhancement associated with imagined soy sauce odor (all four subjects showing sweetness enhancement by imagined strawberry were in this group). Their mean enhancement score was 2.17, whereas the mean for the whole group was 0.77. Mean saltiness enhancement elicited by the presented soy sauce odor was 2.12.


In the present study, we demonstrated OICTP using separate delivery of odorants and tastants: strawberry odor enhanced perceived sweetness but not saltiness, and soy sauce odor enhanced perceived saltiness but not sweetness. We found that imagined odors had a more limited effect on taste perception than perceived odors: imagined strawberry odor somewhat enhanced perceived sweetness of water, and imagined soy sauce odor enhanced perceived saltiness of weak sodium chloride solutions.

Results from Experiment 1 are consistent with previous findings on odor-induced enhancement of perceived sweetness, and several studies had shown this effect with a strawberry odorant (Frank and Byram 1988; Frank et al. 1993; Schifferstein and Verglegh 1996; Stevenson et al. 1999). Suppression of saltiness has been reported by sweet-congruent odors (Shaffer and Frank 1990), but to our knowledge this is the first study that reported an odorant-specific enhancement of perceived saltiness induced by a salty-congruent odor (Experiment 2).

We found that OICTP were weaker in the strongest taste conditions, which parallels results of Schifferstein and Verlegh (1996). This is also consistent with previous studies that used weaker rather than stronger stimuli to demonstrate taste-smell interactions (Dalton et al. 2000; Djordjevic et al. 2004a). We believe that this may be related to the ceiling effect with strongest tastants, which are either easily detected or rated as strong, so simultaneous presentation of odorants has no further effects on their perception.

Furthermore, we explicitly addressed the question of whether these changes are a central or a peripheral phenomenon. By presenting odorants nasally and tastants orally, we minimized the possibility of the interaction taking place in the mouth, via some peripheral mechanism, as there was no physical contact between odorants and tastants, nor between odorants and taste receptors. Consequently, these results strongly support the notion that OICTP take place centrally, consistent with some previous findings (Dalton et al. 2000; Sakai et al. 2001). Neuroimaging studies have started to elucidate neural substrates of olfactory-gustatory integration, and first findings suggest multiple brain regions, including orbitofrontal cortex and anterior insula (de Araujo et al. 2003, Small et al. 1997), as sites for such integration. Together, psychophysical and neuroimaging findings provide converging evidence that integration of olfactory and gustatory sensory information can take place at a central level of processing.

We also attempted to elicit comparable changes in taste perception by imagined odors in both experiments. We found that imagined strawberry odor enhanced perceived sweetness of water and that imagined soy sauce odor enhanced perceived saltiness of weak sodium chloride. These results are consistent with the possibility that some odor imagery occurred, although the evidence is rather limited. Although we recognize that the present results do not provide strong support for the existence of odor imagery, we are reluctant to dismiss this idea, based on the partial evidence reported here, and on results reported previously by others (Elmes 1998) and ourselves (Djordjevic et al. 2004a, 2004b).

It is possible that the paradigm used here was not optimal for capturing the effects of odor imagery. One factor contributing to the modest effects with imagined odors could be the large individual differences that exist in odor imagery ability. In fact, we found that only 20–40% of our subjects were capable of imagining odors, consistent with our previous findings (Djordjevic et al. 2004a, 2004b). Whereas mean sweetness and saltiness enhancement elicited by imagined odors were low for the whole group, a subset of participants showed enhancement effects comparable to those elicited by presented odors. Another factor could be that subjects assigned to the odor imagery condition did not undergo pre-training in odor imagery. Some investigators have suggested that imagery training can facilitate odor (Sugiyama et al. 2003) and flavor/food imagery (Richardson and Patterson 1986). Furthermore, several studies that convincingly demonstrated odor imagery used some form of imagery training, usually via repeated exposure of odorants associated with their names or with other codes (Algom et al. 1993; Djordjevic et al. 2004a, 2004b). Finally, the measure we used—intensity ratings—might have contributed to our equivocal results on imagery. Intensity rating is a subjective measure of perception shown to be liable to influences of knowledge and expectations (Schifferstein 1997). This attempt to capture odor imagery—an elusive mental process—with a subjective measure, resulted in finding a very modest effect. Our previous findings (Djordjevic et al. 2004a) provide support to the claim that the effect of odor imagery on taste perception is more likely to be elicited with an objective measure of taste perception, such as detection of weak tastants.


We wish to thank Elisabeth Perreau-Linck, Zia Lakdawalla, and Julie Boyle for help in testing participants. The strawberry flavorant was donated by Bell Flavors & Fragrances Canada. This study was supported in part by grant MOP 57846, awarded to MJG and RJZ by the Canadian Institutes of Health Research.

Copyright information

© Springer-Verlag 2004