[Skip to Content]
[Skip to Content Landing]
Download PDF
Figure 1.
Anatomical location of the olfactory area, the 3-dimensional finite element mesh of the nasal and pharyngeal cavities, and the time variance in the airflow velocities using the flow simulation. A, Coronal computed tomographic (CT) images of the human nasal cavity. The medial wall of the nasal cavity is formed by the septum and the lateral wall and by mucosal swellings called the inferior, middle, and superior turbinates. These lateral components impart complicated contours to the nasal cavity. The olfactory area is located in the upper part of the nasal epithelium (arrow). The left and right portions of the figure depict the anterior and middle coronal slices, respectively. B, Lateral view of the nasal cavity. The olfactory epithelium has been thought to be located below the olfactory bulb (area i). However, based on histological and electro-olfactogram studies, Leopold et al reported in 2000 that the human olfactory epithelium was probably more anterior in location than previously assumed (area ii). C, A 3-dimensional finite element mesh of the nasal and pharyngeal cavities. The model partially lacks the uppermost portion because it is closely based on the original CT images. The final nasal cavity mesh contains 1 207 673 elements. D, Time variance in the airflow velocities, defined by the sine wave function. We assumed that during respiration, the tidal volume was 500 mL and breathing frequency, 12/min (inspiratory phase, 2.5 seconds; resting phase, 0.1 second; expiratory phase, 2.5 seconds; and resting phase, 0.1 second). The tidal volume during sniffing was assumed to be 200 mL (sniffing phase, 0.5 second, and resting phase, 0.1 second).

Anatomical location of the olfactory area, the 3-dimensional finite element mesh of the nasal and pharyngeal cavities, and the time variance in the airflow velocities using the flow simulation. A, Coronal computed tomographic (CT) images of the human nasal cavity. The medial wall of the nasal cavity is formed by the septum and the lateral wall and by mucosal swellings called the inferior, middle, and superior turbinates. These lateral components impart complicated contours to the nasal cavity. The olfactory area is located in the upper part of the nasal epithelium (arrow). The left and right portions of the figure depict the anterior and middle coronal slices, respectively. B, Lateral view of the nasal cavity. The olfactory epithelium has been thought to be located below the olfactory bulb (area i). However, based on histological and electro-olfactogram studies, Leopold et al14 reported in 2000 that the human olfactory epithelium was probably more anterior in location than previously assumed (area ii). C, A 3-dimensional finite element mesh of the nasal and pharyngeal cavities. The model partially lacks the uppermost portion because it is closely based on the original CT images. The final nasal cavity mesh contains 1 207 673 elements. D, Time variance in the airflow velocities, defined by the sine wave function. We assumed that during respiration, the tidal volume was 500 mL and breathing frequency, 12/min (inspiratory phase, 2.5 seconds; resting phase, 0.1 second; expiratory phase, 2.5 seconds; and resting phase, 0.1 second). The tidal volume during sniffing was assumed to be 200 mL (sniffing phase, 0.5 second, and resting phase, 0.1 second).

Figure 2.
Representative time-lapse images of the streamlines in the left nasal flow. Time-lapse images were obtained at every 1/16th cycle during nasal respiration and after 0.1, 0.2, 0.4, and 0.5 second of sniffing. The area of the left nasal cavity is shown in light gray. A, Streamline profiles during left nasal inspiration (cycles 1/16-7/16). B, Streamline profiles during left nasal expiration (cycles 9/16-15/16). C, Streamline profiles during left nasal sniffing (0.1-0.5 second). D, Olfactory areas that were assumed in previous (i) and recent (ii) periods.

Representative time-lapse images of the streamlines in the left nasal flow. Time-lapse images were obtained at every 1/16th cycle during nasal respiration and after 0.1, 0.2, 0.4, and 0.5 second of sniffing. The area of the left nasal cavity is shown in light gray. A, Streamline profiles during left nasal inspiration (cycles 1/16-7/16). B, Streamline profiles during left nasal expiration (cycles 9/16-15/16). C, Streamline profiles during left nasal sniffing (0.1-0.5 second). D, Olfactory areas that were assumed in previous (i) and recent (ii) periods.

Figure 3.
Representative time-lapse images of the streamlines in the right nasal flow. Time-lapse images were obtained at every 1/16th cycle during nasal respiration and after 0.1, 0.2, 0.4, and 0.5 second of sniffing. The area of the right nasal cavity is shown in dark gray. A, Streamline profiles of right nasal inspiration (cycles 1/16-7/16). B, Streamline profiles of right nasal expiration (cycles 9/16-15/16). C, Streamline profiles during right nasal sniffing (0.1-0.5 second). D, Olfactory areas that were assumed in previous (i) and recent (ii) periods. This is a mirror image of Figure 2D.

Representative time-lapse images of the streamlines in the right nasal flow. Time-lapse images were obtained at every 1/16th cycle during nasal respiration and after 0.1, 0.2, 0.4, and 0.5 second of sniffing. The area of the right nasal cavity is shown in dark gray. A, Streamline profiles of right nasal inspiration (cycles 1/16-7/16). B, Streamline profiles of right nasal expiration (cycles 9/16-15/16). C, Streamline profiles during right nasal sniffing (0.1-0.5 second). D, Olfactory areas that were assumed in previous (i) and recent (ii) periods. This is a mirror image of Figure 2D.

Figure 4.
Comparison of the streamlines of the nasal flow during the midinspiratory and midexpiratory phases and at the end of the sniffing phase in both nasal cavities in the anterolateral view. The right (A) and left (B) nasal flows during the midexpiratory phase. The right (C) and left (D) nasal flows during the midinspiratory phase. The right (E) and left (F) nasal flows at the end of the sniffing phase.

Comparison of the streamlines of the nasal flow during the midinspiratory and midexpiratory phases and at the end of the sniffing phase in both nasal cavities in the anterolateral view. The right (A) and left (B) nasal flows during the midexpiratory phase. The right (C) and left (D) nasal flows during the midinspiratory phase. The right (E) and left (F) nasal flows at the end of the sniffing phase.

Figure 5.
Streamline profiles during the inspiratory phase and sniffing showing different airflow velocities in the nostril. The top and bottom panels represent the left and right sides of the nasal cavity, respectively.

Streamline profiles during the inspiratory phase and sniffing showing different airflow velocities in the nostril. The top and bottom panels represent the left and right sides of the nasal cavity, respectively.

1.
Proetz  AW Air currents in the upper respiratory tract and their clinical importance. Ann Otol Rhinol Laryngol 1951;60 (2) 439- 467
PubMed
2.
Swift  DLProctor  DF Access of air to the respiratory tract. Brain  JDProctor  DFReid  LMRespiratory Defense Mechanism. New York, NY Marcel Dekker1977;63- 91
3.
Girardin  MBilgen  EArbour  P Experimental studies of velocity fields in a human nasal fossa by laser anemometry. Ann Otol Rhinol Laryngol 1983;92 (3, pt 1) 231- 236
PubMed
4.
Hornung  DELeopold  DAYoungentob  SL  et al.  Airflow patterns in a human nasal model. Arch Otolaryngol Head Neck Surg 1987;113 (2) 169- 172
PubMedArticle
5.
Hahn  IScherer  PWMozell  MM Velocity profiles measured for airflow through a large scale model of the human nasal cavity. J Appl Physiol 1993;75 (5) 2273- 2287
PubMed
6.
Keyhani  KScherer  PWMozell  MM Numerical simulation of airflow in the human nasal cavity. J Biomech Eng 1995;117 (4) 429- 441
PubMedArticle
7.
Keyhani  KScherer  PWMozell  MM A numerical model of nasal odorant transport for the analysis of human olfaction. J Theor Biol 1997;186 (3) 279- 301
PubMedArticle
8.
Subramaniam  RPRichardson  RBMorgan  KT  et al.  Computational fluid dynamics simulations of inspiratory airflow in the human nose and nasopharynx. Inhal Toxicol 1998;10 (2) 91- 120Article
9.
Kimbell  JSSubramaniam  PRGross  EASchlosser  PMMorgan  KT Dosimetry modeling of inhaled formaldehyde. Toxicol Sci 2001;64 (1) 100- 110
PubMedArticle
10.
Zhao  KScherer  PWHajiloo  SADalton  P Effect of anatomy on human nasal air flow and odorant transport patterns. Chem Senses 2004;29 (5) 365- 379
PubMedArticle
11.
Tarabichi  MFanous  N Finite element analysis of airflow in the nasal valve. Arch Otolaryngol Head Neck Surg 1993;119 (6) 638- 642
PubMedArticle
12.
Müller-Wittig  WMlynsji  GWeinhold  IBockholt  UVoss  G Nasal airflow diagnosis: comparison of experimental studies and computer simulations. Stud Health Technol Inform 2002;85311- 317
PubMed
13.
Weinhold  IMlynski  G Numerical simulation of airflow in the human nose. Eur Arch Otorhinolaryngol 2004;261 (8) 452- 455
PubMedArticle
14.
Leopold  DAHummel  TSchwob  JEHong  SCKnecht  MKobal  G Anterior distribution of human olfactory epithelium. Laryngoscope 2000;110 (3, pt 1) 417- 421
PubMedArticle
15.
Ishimaru  T Optical recording of the intrinsic signal from the human olfactory cleft. Ann Otol Rhinol Laryngol 2007;116 (5) 335- 341
PubMed
16.
Ishikawa  SNakayama  TWatanabe  MMatsuzawa  T Visualization of flow resistance in physiological nasal respiration. Arch Otolaryngol Head Neck Surg 2006;132 (11) 1203- 1209
PubMedArticle
17.
Zhao  KDalton  PYang  GCScherer  PW Numerical modeling of turbulent and laminar airflow and odorant transport during sniffing in the human and rat nose. Chem Senses 2006;31 (2) 107- 118
PubMedArticle
18.
Croce  CFodil  RDurand  M  et al.  In vitro experiments and numerical simulations of airflow in realistic nasal airway geometry. Ann Biomed Eng 2006;34 (6) 997- 1007
PubMedArticle
19.
Koelega  HS Olfaction and sensory asymmetry. Chem Senses 1979;4 (1) 89- 95Article
20.
Youngentob  SLKurtz  DBLeopold  DAMozell  MMHornung  DE Olfactory sensitivity: is there laterality? Chem Senses 1982;7 (1) 11- 21Article
21.
Damm  MVent  JSchmidt  M  et al.  Intranasal volume and olfactory function. Chem Senses 2002;27 (9) 831- 839
PubMedArticle
22.
Royet  JPPlailly  J Lateralization of olfactory processes. Chem Senses 2004;29 (8) 731- 745
PubMedArticle
23.
Olofsson  JKBroman  DAGilbert  PEDean   PNordin  SMurphy  C Laterality of the olfactory event-related potential response. Chem Senses 2006;31 (7) 699- 704
PubMedArticle
24.
Sobel  NPrabhakaran  VDesmond  JE  et al.  Sniffing and smelling: separate subsystems in the human olfactory cortex. Nature 1998;392 (6673) 282- 286
PubMedArticle
25.
Hamada  TYamaguchi  M Evoked and oscillatory neuromagnetic responses to sniffing odor in human subjects. Behav Brain Res 2001;123 (2) 219- 223
PubMedArticle
26.
Mainland  JSobel  N The sniff is part of the olfactory percept. Chem Senses 2006;31 (2) 181- 196
PubMedArticle
27.
Johnson  BNMainland  JDSobel  N Rapid olfactory processing implicates subcortical control of an olfactomotor system. J Neurophysiol 2003;90 (2) 1084- 1094
PubMedArticle
28.
Kurtz  DBZhao  KHornung  DEScherer  P Experimental and numerical determination of odorant solubility in nasal and olfactory mucosa. Chem Senses 2004;29 (9) 763- 773
PubMedArticle
Original Article
February 16, 2009

Flow Mechanisms in the Human Olfactory GrooveNumerical Simulation of Nasal Physiological Respiration During Inspiration, Expiration, and Sniffing

Author Affiliations

Author Affiliations: Department of Otolaryngology, Kanazawa Municipal Hospital, Kanazawa, Japan (Dr Ishikawa); Intelligent Fluid System Division, Institute of Fluid Science, Tohoku University, Sendai, Japan (Dr Nakayama); and Center for Information Science, Japan Advanced Institute of Science and Technology, Nomi, Japan (Drs Watanabe and Matsuzawa).

Arch Otolaryngol Head Neck Surg. 2009;135(2):156-162. doi:10.1001/archoto.2008.530
Abstract

Objectives  To visualize the velocity and the streamline of physiological unsteady nasal flow and sniffing using the computational fluid dynamics method and to compare the inspiratory phase, expiratory phase, and sniffing flow patterns of the olfactory area.

Design  An anatomically correct 3-dimensional nasal and pharyngeal cavity was constructed from computed tomographic images of a healthy adult nose and pharynx. The unsteady state Navier-Stokes and continuity equations were solved numerically on inspiratory and expiratory nasal flow and sniffing.

Setting  Numerical simulation application.

Main Outcome Measures  The detailed velocity distribution and streamline distribution of nasal airflow were visualized using the computational fluid dynamics method (an imaging technology for regional flow factors [velocity and streamline]).

Results  The inspiratory flow passes through a wider olfactory area than the expiratory flow, and the sniffing flow passes through the widest olfactory area without increasing the velocity of the airflow. In addition, a recirculating flow strongly promotes olfactory function.

Conclusion  The computational fluid dynamics model allows for the investigation of the flow mechanisms in the human olfactory groove.

Human olfaction is a process that involves the transfer of odorant molecules to the olfactory epithelium located in the upper part of the human nasal cavity (olfactory groove). Although molecules can reach the olfactory epithelium by diffusion, olfaction essentially requires some type of nasal airflow, usually as a part of inhalation. While significant advances have been made in understanding functional odor reception, the manner in which odorant molecules reach the human olfactory epithelium is not completely understood. Sniffing is an almost universally performed maneuver when a person is presented with an olfactory stimulus. It is possible that its purpose is to momentarily increase the number of olfactory molecules reaching the olfactory groove by causing a transient change in the airflow pattern in the nose. However, the effect of sniffing on the in vivo airflow pattern remains unknown. The basic questions that have not been adequately addressed are regarding our ability to smell an odor during the inspiratory phase but not during the expiratory phase and the need to sniff an odor for better odor perception.

Airflow profiles in human nasal cavities have been investigated by a number of researchers using in vitro models. The earliest nasal physical models were usually cast from the noses of human cadavers. Quantitative measurement in these casts was performed by visualizing smoke airflow1 and by using miniature Pitot tubes,2 laser Doppler velocimetry,3 and radioactive tracers.4 To increase the spatial resolution and accuracy of measurement, enlarged models of the nasal cavity have also been constructed based on magnetic resonance images and computed tomographic (CT) images. Hahn et al5 constructed an anatomically accurate, enlarged-scale (×20) physical model of the right human nasal cavity from coronal CT images. Although these studies provided valuable descriptive and quantitative information on airflow patterns in the nasal passages, measuring the precise airflow profiles in the olfactory groove has proven to be difficult.

With the development of a technique based on computational fluid dynamics (CFD), the characteristics of human nasal flow, including those of human olfactory flow, could be predicted using mathematical models of the nose created from CT images.613 These theoretical results obtained under steady-flow conditions are in agreement with human experimental data.

It is generally agreed that the human olfactory epithelium is located high in the nasal cavity, predominantly on the dorsal aspects of the nasal vault, septum, and superior turbinate (Figure 1A). However, based on histological and electro-olfactogram studies, Leopold et al14 reported that the human olfactory epithelium is probably more anterior in location than was previously assumed (Figure 1B). Moreover, Ishimaru15 made the same observation based on an optical recording of the intrinsic signal from the human olfactory groove.

The principal goals of the CFD analyses are to shed light on how the human nose functions, to test the current model with regard to olfactory function, and to search for signatures that may clarify the nature of the human nose. These goals can be achieved by accurate simulation of human physiological respiration. We describe herein a nasal respiratory simulation of physiological (unsteady) flow conditions and of sniffing, which was performed by using an anatomically correct 3-dimensional (3-D) nasal and nasopharyngeal cavity model constructed from human CT images. Based on new evidence from a recent study14 of the distribution of the human olfactory epithelium, we followed the location and velocity of the uppermost streamline (UMSL) near the human olfactory area during the postprocessing for CFD.

METHODS

Anatomical data have been presented in our previous study.16 To construct an anatomically correct 3-D model of the nasal and pharyngeal cavities, 141 sections from a CT image of a single subject were used after obtaining informed consent. The images were obtained at intervals of 1 mm from the anterior tip of the nose to the posterior end of the pharynx. The 3-D geometry of the nasal and pharyngeal cavities was reconstructed based on the DICOM (digital imaging and communications in medicine) data of the CT scan sections by using the imaging software Intage 3.1 (KGT Co, Tokyo, Japan). After necessary smoothing and artifact correction, we created a 3-D mesh that was denser near the walls and coarser in regions where the velocity gradients are small; a mesh-generating software (Gambit; Fluent Inc, Lebanon, New Hampshire) was used for this purpose. The final mesh representing the nasal cavity contained 1 207 673 elements (Figure 1C).

In this study, we used the same model as that used previously, with modeling assumptions and parameters adjusted to fit the observed properties of physiological human nasal respiration and sniffing.16 The air was assumed to be a Newtonian, homogeneous, and incompressible fluid. Navier-Stokes equations and the continuity equation were used in the study. Navier-Stokes equations are a set of equations that describe the motion of fluid substances. These equations establish that the changes in the momentum of the particles of a fluid are merely a product of the changes in the pressure and dissipative viscous forces acting inside the fluid. The continuity equation is an equation for the conservation of mass. The Fluent 6.1.22 simulation package (Fluent Inc) computation code was used. The simulated airflow was maintained in an unsteady state during both the inspiratory and expiratory phases and during sniffing. We assumed the tidal volume to be 500 mL and the breathing frequency to be 12/min (inspiratory phase, 2.5 seconds; resting phase, 0.1 second; expiratory phase, 2.5 seconds; and resting phase, 0.1 second) during nasal respiration. During sniffing, we assumed the tidal volume to be 200 mL (sniffing phase, 0.5 second, and resting phase, 0.1 second). The time variance in the airflow velocity in the nostril was defined by the sine wave function (respiration, v = 2 sine (2πt/5) m/s, and sniffing, v = 4 sine (2πt/2) m/s, where t indicates time and v, velocity) (Figure 1D). The following boundary conditions were set. The nasal cavity wall was set at a no-slip condition. The airflow velocity in the nares was considered to be representative of all the velocities. The pressure value at the mesopharynx was assumed to be 0. The airflow velocity distribution in the nares was set at a uniform value. Initially, 1 cycle was simulated before the final (second) run was completed to obtain an accurate description of the solution. The convergence of the solution was measured within 0.001 relative to continuity and velocity.

RESULTS

We show the results of the simulation as a streamline profile with airflow velocity at every 1/16th cycle of nasal respiration and after 0.1, 0.2, 0.4, and 0.5 second of sniffing. During the left inspiratory phase (Figure 2A), as the velocity increased, the UMSL moved up toward the olfactory area (Figure 2A, cycles 1/16-3/16) and reached the highest level at a point in the 4/16th cycle of respiration (videos 1 and 2 are available athttp://archoto.com). When the airflow velocity in the nostril was 3.2 m/s, that in the UMSL was 0.3 m/s. The UMSL did not move down as the velocity decreased (Figure 2A, cycles 5/16-7/16). During the left expiratory phase (Figure 2B), as the velocity increased, the UMSL moved down from the olfactory area (Figure 2B, cycles 9/16-11/16) and reached the lowest level at a point in the 12/16th cycle of respiration (videos 1 and 2 are available at http://archoto.com). When the airflow velocity in the nostril was 2.5 m/s, that in the UMSL was 0.4 m/s. The UMSL did not move up as the velocity decreased (Figure 2B, cycles 13/16-15/16). The location of the UMSL during the midinspiratory phase (Figure 2A, cycle 4/16) was higher than that during the midexpiratory phase (Figure 2A, cycle 12/16). During the left sniffing phase (Figure 2C), as the airflow velocity increased, the UMSL moved up to the olfactory area (video 3 is available athttp://archoto.com). When the airflow velocity in the nostril was 8.0 m/s, that in the UMSL was 0.3 m/s. The location of the UMSL during sniffing was higher than that during inspiration (Figure 3, A and C).

During the right inspiratory phase (Figure 3A), as the velocity increased, a recirculating flow (RCF) was generated in the olfactory area (Figure 3A, cycles 1/16-3/16) and reached the highest level at a point in the 4/16th cycle of respiration (videos 4 and 5 are available at http://archoto.com). When the airflow velocity in the nostril was 3.2 m/s, that in the UMSL was 0.2 m/s. The shape and location of the RCF did not change during the inspiratory phase (Figure 3A, cycles 5/16-7/16). During the right expiratory phase (Figure 3B), as the velocity increased, the UMSL moved down from the olfactory area (Figure 3B, cycles 9/16-11/16) and reached the lowest level at a point in the 12/16th cycle of respiration (videos 4 and 5 are available athttp://archoto.com). When the airflow velocity in the nostril was 2.5 m/s, that in the UMSL was 0.3 m/s. The UMSL did not move up as the velocity decreased (Figure 3B, cycles 13/16-15/16). The location of the UMSL during the midinspiratory phase (Figure 3A, cycle 4/16) was higher than that during the midexpiratory phase (Figure 3A, cycle 12/16). During the right sniffing phase (Figure 3C), as the velocity increased, RCF was generated in the olfactory area, and the UMSL moved up to the olfactory area (video 6 is available athttp://archoto.com). When the airflow velocity in the nostril was 8.5 m/s, that in the UMSL was 0.3 m/s. The location of the UMSL during sniffing was higher than that during inspiration (Figure 3, A and C), and the streamlines of the RCF observed during sniffing were denser than those observed during inspiration (Figure 3, A and C). Figure 4 shows the comparison of the streamlines of the nasal flow during the midinspiratory and midexpiratory phases and at the end of the sniffing phase in both nasal cavities in the anterolateral view.

Figure 5 shows the streamline profiles during the first half of the inspiratory phase and during sniffing and the airflow velocity in the nostril. This figure indicates the relationship between the height of the UMSL and the airflow velocity in the nostril; when the latter increases, the UMSL moves up in continuity.

COMMENT

Experiments performed on narrow-slit flows such as those in the human olfactory groove by using noselike models have drawbacks because of the limited accuracy of measuring equipment. To circumvent these difficulties, the CFD method has been used for analyzing human olfactory flows.7,17,18

The primary objectives of this study were (1) to visualize the velocity and streamlines during physiological unsteady nasal flow and sniffing by using the CFD method and (2) to compare the flow patterns in the olfactory area during the inspiratory, expiratory, and sniffing phases. To obtain accurate results, various computational conditions were used in this model. We assumed the tidal volume to be 500 mL, and the time variance in the airflow velocity was defined by the sine wave function for simulating physiological nasal respiration. During sniffing, we assumed the tidal volume to be 200 mL, and the time variance in the airflow velocity in the nostril was defined by the quarter of the sine wave function. Anatomical reconstruction of the CT images was performed using sections that were 1 mm apart, and a denser mesh containing more than 1.2 million elements was created for the nasal and pharyngeal model in our study; this was considered to be adequately accurate.

Although there are no precise experimental data on olfactory flows, the velocity and streamline profiles in the olfactory region during inspiration in this study are largely consistent with those obtained by Zhao et al10 and Croce et al18 in their numerical simulations. Croce et al18 reported a velocity of 0.4 m/s in the olfactory groove during the midinspiratory phase when that in the nostril was 3.1 m/s. We calculated the velocity to be 0.3 m/s (left) and 0.2 m/s (right) in the olfactory groove during the midinspiratory phase when that in the nostril was 3.2 m/s.

The most interesting finding, which was obtained only under unsteady-flow simulations, was the displacement of the UMSL during respiration. As the velocity increased, the UMSL moved closer to the olfactory area, and it moved away from the olfactory area as the velocity decreased. This may explain why we smell an odor during the inspiratory phase but not during the expiratory phase. The difference in the flow patterns during the inspiratory and expiratory phases was attributed to the structural difference between the anterior and posterior nasal components. As previously shown (Figure 1A), the anterior nasal cavity is narrow, and the turbinate head directs the flow in different directions. In contrast, the posterior end is wide and circular. Hence, the expiratory airflow is directed along the turbinate.

Another interesting finding was the existence of the RCF in the olfactory groove. In our study, an RCF was generated unilaterally (right side). Zhao et al10 reported a similar unilateral RCF, whereas Croce et al18 reported bilateral RCFs in the olfactory groove on steady-state CFD analysis of the nose.18 Croce et al18 described that airflows can be divided into 2 parts. The first part of the airflow, which has relatively high velocity, passes through the lower half of the nasal cavities and therefore appears to be primarily concerned with respiration. The second part of the airflow, which has a much lower velocity, passes through the upper half of the nasal cavities and is probably concerned with olfaction. We agree with their conclusion. In our study, the location of the UMSL in the RCF-positive flow (right) was higher than that in the RCF-negative flow (left) during the inspiratory phase and sniffing. We considered the existence of the RCF to be advantageous for olfactory flow. In addition, we thought that these advantages are the distribution and circulation of air over a wide olfactory area, thereby amplifying olfactory stimuli. As shown in Figure 3A, the fast, narrow flow along the middle meatus (the space between the middle and inferior turbinates) generates a detached RCF in the anterior part of the nasal cavity. Several factors have been proposed regarding the lateralization of olfactory sensitivity, including anatomical difference of nostril,19 handedness and nasal side sensitivity,20 intranasal volume,21 and laterality of the processing in central nervous system.22,23 We thought that the existence of RCF might affect the laterality of olfactory sensitivity.

Our study visualized the streamlines during sniffing. Figure 4A confirms that sniffing is merely an extension of inspiration. Although there are many functional magnetic resonance imaging studies indicating the role of the central nervous system in sniffing,2427 our results clarify that human sniffing is a mechanism involving airflows. On both sides, the position of the UMSL during sniffing was higher than that during inspiration. Although the velocity in the nostril during the sniffing phase was higher than that during the inspiratory phase, the velocity of the UMSL during sniffing was almost the same as that during the inspiratory phase; only the height of the UMSL differed. Zhao et al17 reported in a numerical study that the transport of odorants onto the olfactory mucosa during sniffing was remarkably similar to that during resting breathing. The purpose of sniffing is thought to be to increase the number of olfactory molecules that reach the olfactory groove by causing a transient change in the airflow pattern in the nose. However, based on our findings, we hypothesize that sniffing behavior enables the odorant flow to cover a wider olfactory area than that during usual inspiration.

There are certain limitations to our study. First, experimental verification of our results was not carried out. However, flow experiments in complex and narrow structures such as the olfactory groove are difficult; therefore, it is essential to devise new experimental methods. Second, our study is a single-model study performed under a single flow condition; therefore, the results cannot be generalized. More simulations using other models with different shapes and calculation conditions are required. Third, our study did not consider the diffusion effects of odorants. However, we believe that diffusion effects will not affect our conclusion because the velocities of most odorants are very low (<0.001 m/s) compared with the velocity of the UMSL in our simulation (0.3 m/s).28

By introducing entirely unsteady calculations into the nasal CFD study, we can obtain many important findings regarding flow mechanisms in the human olfactory groove under several respiratory conditions. The CFD study is an important tool for understanding the functions of the human nose.

Back to top
Article Information

Corresponding Author: Shigeru Ishikawa, MD, PhD, Department of Otolaryngology Kanazawa Municipal Hospital, 3-7-3 Heiwa-machi Kanazawa, Ishikawa 921-8105, Japan (ishikawa@po.incl.ne.jp).

Submitted for Publication: April 14, 2007; final revision received September 15, 2007; accepted October 30, 2007.

Author Contributions: Dr Ishikawa had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ishikawa and Matsuzawa. Acquisition of data: Watanabe. Analysis and interpretation of data: Nakayama. Drafting of the manuscript: Ishikawa. Critical revision of the manuscript for important intellectual content: Nakayama, Watanabe, and Matsuzawa. Administrative, technical, and material support: Ishikawa, Nakayama, and Watanabe. Study supervision: Matsuzawa.

Financial Disclosure: None reported.

Additional Contributions: The staff at the Department of Radiology, Kanazawa Municipal Hospital, helped in obtaining computed tomographic scans. Tadashi Ishimaru, MD, PhD, provided helpful discussions.

References
1.
Proetz  AW Air currents in the upper respiratory tract and their clinical importance. Ann Otol Rhinol Laryngol 1951;60 (2) 439- 467
PubMed
2.
Swift  DLProctor  DF Access of air to the respiratory tract. Brain  JDProctor  DFReid  LMRespiratory Defense Mechanism. New York, NY Marcel Dekker1977;63- 91
3.
Girardin  MBilgen  EArbour  P Experimental studies of velocity fields in a human nasal fossa by laser anemometry. Ann Otol Rhinol Laryngol 1983;92 (3, pt 1) 231- 236
PubMed
4.
Hornung  DELeopold  DAYoungentob  SL  et al.  Airflow patterns in a human nasal model. Arch Otolaryngol Head Neck Surg 1987;113 (2) 169- 172
PubMedArticle
5.
Hahn  IScherer  PWMozell  MM Velocity profiles measured for airflow through a large scale model of the human nasal cavity. J Appl Physiol 1993;75 (5) 2273- 2287
PubMed
6.
Keyhani  KScherer  PWMozell  MM Numerical simulation of airflow in the human nasal cavity. J Biomech Eng 1995;117 (4) 429- 441
PubMedArticle
7.
Keyhani  KScherer  PWMozell  MM A numerical model of nasal odorant transport for the analysis of human olfaction. J Theor Biol 1997;186 (3) 279- 301
PubMedArticle
8.
Subramaniam  RPRichardson  RBMorgan  KT  et al.  Computational fluid dynamics simulations of inspiratory airflow in the human nose and nasopharynx. Inhal Toxicol 1998;10 (2) 91- 120Article
9.
Kimbell  JSSubramaniam  PRGross  EASchlosser  PMMorgan  KT Dosimetry modeling of inhaled formaldehyde. Toxicol Sci 2001;64 (1) 100- 110
PubMedArticle
10.
Zhao  KScherer  PWHajiloo  SADalton  P Effect of anatomy on human nasal air flow and odorant transport patterns. Chem Senses 2004;29 (5) 365- 379
PubMedArticle
11.
Tarabichi  MFanous  N Finite element analysis of airflow in the nasal valve. Arch Otolaryngol Head Neck Surg 1993;119 (6) 638- 642
PubMedArticle
12.
Müller-Wittig  WMlynsji  GWeinhold  IBockholt  UVoss  G Nasal airflow diagnosis: comparison of experimental studies and computer simulations. Stud Health Technol Inform 2002;85311- 317
PubMed
13.
Weinhold  IMlynski  G Numerical simulation of airflow in the human nose. Eur Arch Otorhinolaryngol 2004;261 (8) 452- 455
PubMedArticle
14.
Leopold  DAHummel  TSchwob  JEHong  SCKnecht  MKobal  G Anterior distribution of human olfactory epithelium. Laryngoscope 2000;110 (3, pt 1) 417- 421
PubMedArticle
15.
Ishimaru  T Optical recording of the intrinsic signal from the human olfactory cleft. Ann Otol Rhinol Laryngol 2007;116 (5) 335- 341
PubMed
16.
Ishikawa  SNakayama  TWatanabe  MMatsuzawa  T Visualization of flow resistance in physiological nasal respiration. Arch Otolaryngol Head Neck Surg 2006;132 (11) 1203- 1209
PubMedArticle
17.
Zhao  KDalton  PYang  GCScherer  PW Numerical modeling of turbulent and laminar airflow and odorant transport during sniffing in the human and rat nose. Chem Senses 2006;31 (2) 107- 118
PubMedArticle
18.
Croce  CFodil  RDurand  M  et al.  In vitro experiments and numerical simulations of airflow in realistic nasal airway geometry. Ann Biomed Eng 2006;34 (6) 997- 1007
PubMedArticle
19.
Koelega  HS Olfaction and sensory asymmetry. Chem Senses 1979;4 (1) 89- 95Article
20.
Youngentob  SLKurtz  DBLeopold  DAMozell  MMHornung  DE Olfactory sensitivity: is there laterality? Chem Senses 1982;7 (1) 11- 21Article
21.
Damm  MVent  JSchmidt  M  et al.  Intranasal volume and olfactory function. Chem Senses 2002;27 (9) 831- 839
PubMedArticle
22.
Royet  JPPlailly  J Lateralization of olfactory processes. Chem Senses 2004;29 (8) 731- 745
PubMedArticle
23.
Olofsson  JKBroman  DAGilbert  PEDean   PNordin  SMurphy  C Laterality of the olfactory event-related potential response. Chem Senses 2006;31 (7) 699- 704
PubMedArticle
24.
Sobel  NPrabhakaran  VDesmond  JE  et al.  Sniffing and smelling: separate subsystems in the human olfactory cortex. Nature 1998;392 (6673) 282- 286
PubMedArticle
25.
Hamada  TYamaguchi  M Evoked and oscillatory neuromagnetic responses to sniffing odor in human subjects. Behav Brain Res 2001;123 (2) 219- 223
PubMedArticle
26.
Mainland  JSobel  N The sniff is part of the olfactory percept. Chem Senses 2006;31 (2) 181- 196
PubMedArticle
27.
Johnson  BNMainland  JDSobel  N Rapid olfactory processing implicates subcortical control of an olfactomotor system. J Neurophysiol 2003;90 (2) 1084- 1094
PubMedArticle
28.
Kurtz  DBZhao  KHornung  DEScherer  P Experimental and numerical determination of odorant solubility in nasal and olfactory mucosa. Chem Senses 2004;29 (9) 763- 773
PubMedArticle
×