CHEMISTRY & BIODIVERSITY Vol.

5 (2008) 853

Retronasal Perception of Odors

by Thomas Hummel
Smell & Taste Clinic, Department of Otorhinolaryngology, University of Dresden Medical School,
Fetscherstrasse 74, D-01307 Dresden
(phone: 49-351-458-4189; fax: 49-351-458-4326; e-mail: thummel@mail.zih.tu-dresden.de)

Odors often produce different sensations when presented in front of the nose or intraorally, when
eaten. It is a long-standing question whether these differences in sensations are due, for example, to the
additional mechanical sensations elicited by the food in the mouth or additional odor release during
mastication. To study this phenomenon in detail, a stimulation technique has been developed that allows
controlled ortho- or retronasal presentation of odorous stimuli. Results from psychophysical, electro-
physiological, and imaging studies suggest that there are clear differences in the perception of ortho- and
retronasal stimuli. This 6duality of the sense of smell7 is also observed in a clinical context where some
patients exhibit good retronasal olfactory function with little or no orthonasal function left, and vice
versa. The differences between ortho- and retronasal perception of odors are thought to be, at least partly,
due to absorption of odors to the olfactory epithelium, which appears to differ in relation to the direction
of the airflow across the olfactory epithelium.

Introduction. Aroma stimuli reach the olfactory epithelium usually by two

pathways: via the nose, during sniffing, and via the mouth, during eating or drinking
(Fig. 1). These two pathways are typically referred to as orthonasal and retronasal
olfaction [1] [2]. As retronasal sensations are frequently referred to as 6taste7, smell
taste confusions have been described frequently [3] [4]. In fact, in case people lose their
sense of smell, they would often describe their smell loss as a 6loss of taste function7 [5].
In addition, the smell taste confusions are so profound that there are even languages
(e.g., Swiss German or Franconian German) that do not have a word for 6smelling7 but
only refer to it as 6tasting7. As an interesting aspect regarding the specific biological role
of retronasal odor perception, recent work [6] suggests that perceived satiation is
modulated by altering the extent of retronasal aroma release.

Material and Methods. To investigate differences between orthonasal and retronasal perception of
odors, presentation of odorous stimuli is a key issue. Halpern pointed out that ortho- and retronasal
stimuli should reach the olfactory mucosa by the two distinct pathways, without producing additional
gustatory or mechanical stimuli [7]. Thus, studies based on the application of liquid or solid stimuli to the
oral cavity cannot allow direct comparisons to stimuli that are presented in front of the nose, simply
because the oral administration of odors may produce gustatory, thermal, and mechanical sensations
which may interact with olfactory mediated sensations [8] [9]. To avoid this situation, some researchers
placed odors in containers presented to the oral cavity (e.g., [10]), while others asked subjects to sniff the
headspace of an odorous liquid or inhale the same headspace through the mouth, followed by nasal
exhalation [11] (cf. [12] [13]). However, a major limitation of these methods is the unknown odor
concentration in the oral cavity and the mechanical stimulation of intraoral surfaces [8]. In fact, these

C 2008 Verlag Helvetica Chimica Acta AG, ZFrich

854 CHEMISTRY & BIODIVERSITY Vol. 5 (2008)

Fig. 1. Schematic drawing of the placement of the tubes for orthonasal (light yellow) and retronasal
stimulation (light green) with the flow of air indicated by open circles. The drawing is superimposed on a
(low-resolution) MR scan of a healthy subject in order to show the anatomy of the nasal cavity in relation
to the placement of the tubes.

complex interactions, for example, between oral mechano-sensations and taste have received more and
more attention during the recent years [14 18].
To allow for a more defined retronasal stimulation, a new device was developed [19] which allows the
release of odors directly into the epipharynx above the soft palate (Fig. 1). This avoids concomitant oral
gustatory, thermal, and mechanical stimulation, thus permitting the study of ortho- or retronasal olfaction
in isolation. Two plastic tubes of 3.3 mm outer diameter, each, are cut from sterile suction catheters made
from soft polyvinyl chloride. These tubes are placed in the nasal cavity under endoscopic control such
that the opening of one of the tubes is just beyond the nasal valve (ca. 1.5 cm from the naris) and the
opening of the other tube is in the epipharynx (ca. 7.5 cm from the naris). For stimulus presentation, the
tubes are connected to outlets of a computer-controlled air-dilution olfactometer (OM6b; Burghart, D-
Wedel; Fig. 2). This stimulator allows application of rectangular-shaped pulses of chemical stimuli.
Mechanical stimulation is avoided by embedding the stimuli into a constant flow of odorless, humidified
air of controlled temperature (80% relative humidity, total flow 8 l/min, 368) [20]. Odors presented
through either tube reach the olfactory epithelium at approximately the same concentration and with the
same time course [21].

Results. In the following, the results from studies will be described that have been
obtained using the techniques described above for the specific presentation of ortho-
and retronasal olfactory stimuli. Thresholds to orthonasal stimuli are lower compared
to retronasal thresholds.
A first experiment aimed to evaluate the perception of odors presented
orthonasally and retronasally [19]. Results from this study showed that orthonasal
 CHEMISTRY & BIODIVERSITY Vol. 5 (2008) 855

Fig. 2. Photograph of a six-channel olfactometer (Burghart Instruments, D-Wedel). The olfactometer is

based on the principles of air-dilution olfactometry: odorous stimuli are embedded in a constant flow of
odorless air, warmed up to body temperature and humidified ( > 80% relative humidity), which is
important especially for retronasal presentation due to the high sensitivity of the epipharynx to changes
in temperature, humidity, or airflow.

thresholds for both lavender and chocolate were lower than retronasal thresholds. This
observation confirmed work by Voirol and Daget [11]. Their experiments indicated
higher thresholds and, accordingly, decreased suprathreshold responsiveness for
retronasal presentation of air-phase vanillin and citral compared to orthonasal
stimulation. On a suprathreshold level, it has also been reported [22] that the ability
to identify odors is less efficient when stimuli are presented retronasally [13]. Using
olfactory threshold measurements, greater impairment has been reported for
retronasal than for orthonasal perception in elderly people [12].

Retronasally Presented Olfactory Stimuli Can Be Localized. Other ways to

investigate whether ortho- and retronasal stimuli are perceived differently relate to the
856 CHEMISTRY & BIODIVERSITY Vol. 5 (2008)

work of von Skramlik [23], indicating that olfactory stimuli cannot be localized if the
stimulus is applied unilaterally (but see also [24]). Based on this hypothesis, it was
thought that subjects should not be able to localize the site of stimulation when odors
are presented ortho- or retronasally. However, while the trigeminal stimulus CO2 could
be localized, as expected, subjects were also able to localize the pure olfactory
stimulant H2S, and most of them also localized 2-phenylethanol, a rose-like odorant
[21]. Importantly, these results were not related to differences in stimulus intensity.

Cortical Responses Reflect Differences between Orthonasal and Retronasal

Stimuli. To investigate potential distinctions between the early sensory processing
of food- and non-food-related odors, chocolate and lavender were applied to an event-
related potential paradigm which is superior to other experimental approaches in terms
of the temporal resolution of the response [25]. In this setup, potentials are extracted
from the EEG. Results indicated that, when an odorant unrelated to food was
presented in an unusual site, i.e., retronasally, the response was larger compared to the
presentation of the same odorant at an orthonasal site [26]. This was the other way
round for a food-related odor. These changes clearly indicate differences of
information processing depending on the route of odor presentation. In addition, they
also seem to reflect strong contextual confounds in the perception of ortho- or
retronasal stimuli.

Orthonasal and Retronasal Stimuli Produce Different Patterns of Brain Activation.

Imaging techniques allow to investigate differences between ortho- and retronasal
olfaction at the cerebral level [27] [28] (Fig. 3). Response patterns to orthonasal and
retronasal stimulation were investigated in a study based on Functional Magnetic
Resonance Imaging (FRMI). 6Non-food7 odorants, different regarding their phys-
icochemical properties (butanol: hydrophilic; farnesol: lipophilic), and their associa-
tion with food (chocolate) or non-food (lavender) were delivered both ortho- and
retronasally. Activation due to retronasal stimulation was found at the base of the
central sulcus, corresponding to the primary representation of the oral cavity [29],
possibly reflecting that retronasal odors are referred to the mouth.
Retronasal vs. orthonasal presentation of chocolate produced preferential activa-
tion in the medial orbitofrontal cortex and in the perigenual cingulate, superior
temporal gyrus, and posterior cingulate cortex. Comparison of orthonasally vs.
retronasally presented chocolate-produced activations in the thalamus, right caudo-
lateral orbitofrontal cortex and right hippocampus, frontal operculum bilaterally,
temporal operculum/ventral bilaterally, right temporal-parietal operculum at the level
of the supramarginal gyrus, left anterodorsal insula, and right anterior insula. These
results clearly indicate that the neural processing of an odor is influenced by route of
administration. Because these observations were made for the food-related odor
'chocolate(, it was hypothesized that the observed difference in activation may relate to
differential reward circuits for food but not non-food odors [30]. Results from this
study were also interpreted such that orthonasal olfaction was more correlated to the
anticipatory phase in food reward, whereas retronasal olfaction related to the
consummatory phase, receipt of a reward, which finally resulted in the different
patterns of activation.
 CHEMISTRY & BIODIVERSITY Vol. 5 (2008) 857

Fig. 3. Results from functional magnetic resonance imaging (FMRI) a typical response is shown to
orthonasal activation with an unpleasant odor. Activation is produced in the piriform cortex as marked by
open circles. Results from a study in a group of healthy subjects.

Interactions Orthonasal and Retronasal Odorous Stimuli and Texture of Oral

Stimuli. During oral processing of foods and swallowing, the senses involved interact
in a nonlinear way meaning that texture and taste influence the perception of odors
(e.g., [31 33]). Results obtained by Visschers et al. [34] suggest that the intensity of
aroma decreased with increasing consistency of the consumed food. In addition, it has
been shown that retronasal odorous stimuli increase the intensities of thickness and
creaminess of oral stimuli [35]. Importantly, this interaction was only visible when the
presentation of the odor coincided with swallowing. Overall, the results from these
experiments showed that the investigation of multi-modal interactions appears to
depend on 1) the questions being asked, 2) the odors being used (for example, whether
they are contextually congruent with the orally administered texture or not), and 3)
also on the time during oral processing, including swallowing, when measurements are
being obtained.
Some patients experience odors orthonasally, but not retronasally, and vice versa.
During the last years, a series of clinical studies reported quantitative differences
between ortho- and retronasal smelling. Retronasal olfactory loss has been described in
858 CHEMISTRY & BIODIVERSITY Vol. 5 (2008)

the absence of orthonasal deficits [36] [37]. In addition, better retronasal than
orthonasal olfactory function, meaning impaired smell with preserved flavor percep-
tion, has been reported in patients with nasal polyposis [38]. This is thought to be
related, at least in part, to the presence of mechanical obstruction in the anterior
portion of the olfactory cleft [39]. In addition, orthonasal olfactory loss with little or no
changes in retronasal odor perception has been reported even in the absence of nasal
polyposis (e.g., [40]). Specifically, the differential loss of ortho- or retronasal olfactory
function has been documented using psychophysical [41] and electrophysiological tools
[42 44].
These results suggest on a clinical level that orthonasal and retronasal olfaction is
processed differently. Consequently, clinical testing of retronasal olfaction seems to be
of interest, which is possible using relatively simple test kits validated for such purposes
[41] (Fig. 4).

Fig. 4. Retronasal olfactory testing. Photograph of a simple, but validated test for the psychophysical
investigation of retronasal olfactory perception [41]. Twenty stimuli, for example, spices from a grocery
etc., are presented intraorally in compressible vials (0.05 g per application). Subjects identify the odors
using lists of four descriptors for each odor. Normative data are available to qualify the test score.

Discussion. There are numerous possible reasons for differences between ortho-
and retronasal olfaction. Differences in airflow patterns are thought to contribute to
perceptual differences between ortho- and retronasal presentation of odors [45]. The
direction of odor movement across the olfactory epithelium could lead to differences in
the processing of odorous information. This hypothesis is also supported by work
emphasizing the significance of nasal air-flow on the perception of odors [46 48].
 CHEMISTRY & BIODIVERSITY Vol. 5 (2008) 859

These differences between air-flows induced by inhalation and exhalation are

additionally highlighted by mathematical models of intranasal airflow [49].
Furthermore, the interaction with other sensory systems was also hypothesized to
account for perceptual differences between ortho- and retronasal olfactory stimuli. A
suppression of retronasal aroma by taste has been reported [13]. Small et al. reported
activation in different brain areas for interaction between taste and orthonasally or
retronasally presented stimuli [50] [51]. They observed deactivation in the orbito-
frontal cortex, insula, and anterior cingulate cortex for taste-orthonasal stimulus
association, and supra-additive responses at the same regions for the combination of
taste and retronasal stimuli. However, Sakai et al. [52] described odor-induced taste
enhancement for both orthonasal and retronasal olfaction, and Cerf-Ducastel and
Murphy did not see major differences for brain activation for orthonasal or retronasal
delivery of odorous stimuli [53].
With regard to the trigeminal activation induced by most odorants [54], there is also
the possibility that the interaction between the trigeminal and the olfactory system
relates to the differential perception of ortho- and retronasal stimuli [55 57]. To be
specific, there is evidence indicating that the sensitivity for chemosensory stimuli is
higher in the anterior compared to the posterior portion of the nose, while this is the
other way around for mechanical stimuli [58].
It has also been suggested [59] that colors specifically enhance odor intensity, when
odors were presented orthonasally, and suppress the intensity of retronasal olfactory
stimuli. In addition, contextual variables appear to shape the perception of ortho- or
retronasal stimuli. Specifically, experiments based on olfactory event-related potentials
suggest differences between food- and non-food-related odors such that stimuli
produce larger amplitudes when presented in an unusual context, e.g., lavender odor
being presented retronasally [26].
Finally, ideas from Mozell may help to explain differences between ortho- and
retronasal stimuli [45]. The proposed 6chromatographic model7 of olfaction suggests
that the direction of odorant flow across the olfactory epithelium may result in a
specific pattern of activation. A physiological basis for this effect may include the zonal
organization of olfactory receptor neurons within the olfactory epithelium [60]. Thus, it
seems possible that orthonasal presentation of an odor produces a pattern of mucosal
activation that is very different from the pattern induced through retronasal
presentation of the same odor [61] [62].

Conclusions. A number of highly controlled experiments suggest differences in

processing of orthonasal and retronasal olfactory information. A basis for this
phenomenon may be found in the model proposed by Mozell [45] [63] [64] hypothesiz-
ing that the sorption of odors to the olfactory epithelium in relation to the direction of
airflow changes the pattern of mucosal activation, and consequently, the perception of
the same odor in relation to the route of presentation.
This research was supported in part by a grant from the Deutsche Forschungsgemeinschaft (DFG
HU441/2).
860 CHEMISTRY & BIODIVERSITY Vol. 5 (2008)

REFERENCES
[1] P. Rozin, Perception Psychophysics 1982, 31, 397.
[2] G. M. Shepherd, Ann. N.Y. Acad. Sci. 2007, 1121, 87.
[3] C. Murphy, W. S. Cain, L. M. Bartoshuk, Sens. Processes 1977, 1, 204.
[4] C. Murphy, W. S. Cain, Physiol. Behav. 1980, 24, 601.
[5] D. A. Deems, R. L. Doty, R. G. Settle, V. Moore-Gillon, P. Shaman, A. F. Mester, C. P. Kimmelman,
V. J. Brightman, J. B. J. Snow, Arch. Otolaryngol. Head Neck Surg. 1991, 117, 519.
[6] R. M. Ruijschop, A. E. Boelrijk, J. A. de Ru, C. de Graaf, M. S. Westerterp-Plantenga, Br. J. Nutr.
2007, 22, 1.
[7] B. P. Halpern, ChemoSense 2004, 6, 1.
[8] D. G. Land, in 6Perspectives on the effects of interactions on flavor perception: an overview7, R. J.
McGorrin, J. Leland, eds., ACS Symp. Ser. 633, Washington, 1994, pp. 2.
[9] A. Welge-LFssen, J. Drago, M. Wolfensberger, T. Hummel, Brain Res. 2005, 1038, 69.
[10] B. C. Sun, B. P. Halpern, Chem. Senses 2005, 30, 1.
[11] E. Voirol, N. Daget, Food Sci. Technol. 1986, 19, 316.
[12] V. B. Duffy, W. S. Cain, A. M. Ferris, Chem. Senses 1999, 24, 671.
[13] K. J. Burdach, J. H. Kroeze, E. P. Koster, Percept. Psychophys. 1984, 36, 205.
[14] A. BFttner, P. Schieberle, in 6Changes in the concentrations of key fruit odorants induced by
mastication7, Eds. D. D. Roberts, A. J. Taylor, ACS Symp. Ser. 763, Washington, 2000, pp. 87.
[15] A. BFttner, J. Agric. Food Chem. 2002, 50, 7105.
[16] M. Hodgson, R. S. Linforth, A. J. Taylor, J. Agric. Food Chem. 2003, 51, 5052.
[17] A. M. Haahr, A. Bardow, C. E. Thomsen, S. B. Jensen, B. Nauntofte, M. Bakke, J. Adler-Nissen,
W. L. Bredie, Physiol. Behav. 2004, 82, 531.
[18] M. D. Hodgson, J. P. Langridge, R. S. Linforth, A. J. Taylor, J. Agric. Food Chem. 2005, 53, 1700.
[19] S. Heilmann, T. Hummel, Behav. Neurosci. 2004, 118, 412.
[20] G. Kobal, 6Elektrophysiologische Untersuchungen des menschlichen Geruchssinns7, Thieme Verlag,
Stuttgart, 1981.
[21] D. M. Small, J. C. Gerber, Y. E. Mak, T. Hummel, Neuron 2005, 47, 593.
[22] J. Pierce, B. P. Halpern, Chem. Senses 1996, 21, 529.
[23] E. von Skramlik, Z. Sinnesphysiol. 1924, 56, 69.
[24] J. Porter, T. Anand, B. Johnson, R. M. Khan, N. Sobel, Neuron 2005, 47, 581.
[25] T. Hummel, G. Kobal, in 6Olfactory Event-Related Potentials7, Eds. S. A. Simon, M. A. L. Nicolelis,
CRC Press, Boca Raton, Florida, USA, 2001, pp. 429.
[26] T. Hummel, S. Heilmann, Dairy J. 2008, in press.
[27] D. M. Small, R. J. Zatorre, A. Dagher, A. C. Evans, M. Jones-Gotman, Brain 2001, 124, 1720.
[28] D. M. Small, J. Prescott, Exp. Brain Res. 2005, 166, 345.
[29] H. Yamashita, Y. Kumamoto, T. Nakashima, T. Yamamoto, A. Inokuchi, S. Komiyama, Eur. Arch.
Oto-Rhino-Laryngol. 1999, 256, S42.
[30] K. C. Berridge, Neurosci. Biobehav. Rev. 1996, 20, 1.
[31] R. J. Stevenson, J. Prescott, R. A. Boakes, Chem. Senses 1999, 24, 627.
[32] T. A. Hollowood, R. S. T. Linforth, A. J. Taylor, Chem. Senses 2002, 27, 583.
[33] K. G. C. Weel, A. E. M. Boelrijk, A. C. Alting, P. J. J. M. van Mil, J. J. Burger, H. Gruppen, A. G. J.
Voragen, G. Smit, J. Agric. Food Chem. 2002, 50, 5149.
[34] R. W. Visschers, M. A. Jacobs, J. Frasnelli, T. Hummel, M. Burgering, A. E. Boelrijk, J. Agric. Food
Chem. 2006, 54, 5509.
[35] J. H. F. Bult, R. A. de Wijk, T. Hummel, Neurosci. Letters 2007, 411, 6.
[36] B. J. Cowart, B. P. Halpern, E. K. Varga, Chem. Senses 1999, 24, 608.
[37] B. J. Cowart, B. P. Halpern, D. Rosen, C. T. Klock, E. D. Pribitkin, Chem. Senses 2003, 28, A65.
[38] B. N. Landis, R. Giger, A. Ricchetti, I. Leuchter, M. Hugentobler, T. Hummel, J. S. Lacroix,
Laryngoscope 2003, 113, 1993.
[39] O. Pfaar, B. N. Landis, J. Frasnelli, K. B. Huttenbrink, T. Hummel, Chem. Senses 2006, 31, 27.
[40] W. Ogle, Med. Chir. Trans. 1870, 53, 263.
 CHEMISTRY & BIODIVERSITY Vol. 5 (2008) 861

[41] S. Heilmann, G. Strehle, K. Rosenheim, M. Damm, T. Hummel, Arch. Otolaryngol. Head Neck
Surg. 2002, 128, 414.
[42] B. N. Landis, J. Frasnelli, J. Reden, J. S. Lacroix, T. Hummel, Arch. Otolaryngol. Head Neck Surg.
2005, 131, 977.
[43] P. Rombaux, A. Mouraux, B. Bertrand, G. Nicolas, T. Duprez, T. Hummel, Laryngoscope 2006, 116,
901.
[44] P. Rombaux, H. Weitz, A. Mouraux, G. Nicolas, B. Bertrand, T. Duprez, T. Hummel, Arch.
Otolaryngol. Head Neck Surg. 2006, 132, 1346.
[45] M. M. Mozell, Nature 1964, 203, 1181.
[46] D. A. Leopold, Laryngoscope 1988, 98, 1232.
[47] N. Sobel, R. M. Khan, A. Saltman, E. V. Sullivan, J. D. Gabrieli, Nature 1999, 402, 35.
[48] M. Damm, J. Vent, M. Schmidt, P. Theissen, H. E. Eckel, J. Lotsch, T. Hummel, Chem. Senses 2002,
27, 831.
[49] K. Zhao, P. W. Scherer, S. A. Hajiloo, P. Dalton, Chem. Senses 2004, 29, 365.
[50] D. M. Small, M. Jones-Gotman, R. J. Zatorre, M. Petrides, A. C. Evans, Neuroreport 1997, 8, 3913.
[51] D. M. Small, J. Voss, Y. E. Mak, K. B. Simmons, T. R. Parrish, D. R. Gitelman, J. Neurophysiol. 2004,
92, 1892.
[52] N. Sakai, T. Kobayakawa, N. Gotow, S. Saito, S. Imada, Percept. Mot. Skills 2001, 92, 1002.
[53] B. Cerf-Ducastel, C. Murphy, Chem. Senses 2001, 26, 625.
[54] C. A. Elsberg, I. Levy, E. D. Brewer, Bull. Neurol. Inst. N.Y. 1935, 4, 270.
[55] W. S. Cain, C. Murphy, Nature 1980, 284, 255.
[56] T. Hummel, A. Livermore, Int. Arch. Occ. Env. Health 2002, 75, 305.
[57] J. Frasnelli, T. Hummel, Int. J. Psychophysiol. 2007, 65, 177.
[58] J. Frasnelli, S. Heilmann, T. Hummel, Neurosci. Lett. 2004, 362, 65.
[59] B. J. Koza, A. Cilmi, M. Dolese, D. A. Zellner, Chem. Senses 2005, 30, 643.
[60] K. J. Ressler, S. L. Sullivan, L. B. Buck, Cell 1993, 73, 597.
[61] V. Vedin, B. Slotnick, A. Berghard, Eur. J. Neurosci. 2004, 20, 1858.
[62] J. W. Scott, H. P. Acevedo, L. Sherrill, M. Phan, J. Neurophysiol. 2007, 97, 1941.
[63] M. M. Mozell, M. Jagodowicz, Science 1973, 181, 1247.
[64] P. F. Kent, M. M. Mozell, S. J. Murphy, D. E. Hornung, J. Neurosci. 1996, 16, 345.

Received February 8, 2008