Statistical localization of human olfactory cortex

NeuroImage 66 (2013) 333–342

Contents lists available at SciVerse ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/ynimg

Statistical localization of human olfactory cortex Janina Seubert a, Jessica Freiherr b, Jelena Djordjevic c, Johan N. Lundström a, d, e,⁎ a

Monell Chemical Senses Center, Philadelphia, PA, USA Clinic for Diagnostic and Interventional Neuroradiology, RWTH Aachen University, Aachen, Germany c Neurology and Neurosurgery, McGill University, Montreal, QC, Canada d Department of Psychology, University of Pennsylvania, PA, USA e Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden b

a r t i c l e

i n f o

Article history: Accepted 19 October 2012 Available online 24 October 2012 Keywords: Olfaction Brain imaging Activation likelihood estimation Piriform cortex

a b s t r a c t Functional neuroimaging methods have been used extensively during the last decades to explore the neural substrates of olfactory processing. While a general consensus on the functional anatomy of olfactory cortex is beginning to emerge, the mechanisms behind the functions of individual processing nodes still remain debated. Further, it remains unclear to which extent divergent findings result from differences in methodological approaches. Using Activation Likelihood Estimation (ALE), the aim of the present study was to statistically combine all published data on functional neuroimaging of olfaction to provide a probability map reflecting the state of the field to date. Additionally, we grouped studies according to various methodological approaches to investigate whether these systematically affected the reported findings. A total of 45 studies (69 contrasts, 594 foci) met our inclusion criteria. Significant ALE peaks for odor against baseline were observed in areas commonly labeled as primary and secondary olfactory cortex, such as the piriform and orbitofrontal cortex, amygdala, anterior insula, and ventral putamen. In addition, differences were observed in the extent to which different methods were able to induce activation in these different nodes of the olfactory network. © 2012 Elsevier Inc. All rights reserved.

Introduction Over the last 25 years, our understanding of basic sensory processing and neurobiological substrates of the human sense of smell has increased notably. This progress has in particular been facilitated by methodological advances in functional neuroimaging techniques such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). While the substantial rise in the number of functional neuroimaging articles published on olfactory processing has contributed significantly to localization inferences of primary and secondary olfactory cortex, it has also brought considerable challenges to the scientific community. In particular, the diversity of methods and experimental paradigms, statistical analyses, and approaches to data interpretation often render between-study comparisons and the integration of findings a complex and evasive task. Considerable disagreement persists concerning the best approaches for studying olfactory mediated brain activations, and potential influences of these methodological differences on experimental results have not yet conclusively been investigated. In the following, we will provide a brief descriptive overview of the anatomy and connectome of the peripheral and central ⁎ Corresponding author at: Monell Chemical Senses Center, 3500 Market Street, Philadelphia, PA 19104, USA. Fax: +1 267 519 4690. E-mail address: [email protected] (J.N. Lundström). 1053-8119/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neuroimage.2012.10.030

primate olfactory system. Within a meta-analytical context of published functional neuroimaging data, using the statistical activation likelihood estimation (ALE) method, we will then identify areas of the cortical olfactory network which are consistently activated across human neuroimaging studies, and quantify functional differences between frequently employed approaches. The early portion of the olfactory sensory pathway has been well mapped out using neural tracing methods and anatomical studies in non-human animals. In primates, sensory processing of odors starts at the olfactory mucosa situated on the roof of the nasal cavity. Here, the odor molecules bind to the primary sensing cells, the olfactory receptor neurons. Their axons form the olfactory nerve, projecting to the tufted and mitral cells of the olfactory bulb (Firestein, 2001). From there, the largest portion of neuronal input is received by the piriform cortex. However, several other structures also receive direct projections from the olfactory bulb, including the anterior olfactory nucleus, the olfactory tubercle, a small anteromedial part of the enthorhinal cortex, the periamygdaloid cortex as well as several areas within the amygdala, including the anterior cortical nucleus and the nucleus of the lateral olfactory tract (Carmichael et al., 1994; Price, 1985). Together, these structures receiving direct input from the olfactory bulb have traditionally been labeled as olfactory cortex (Price, 2003). Neuroanatomical and electrophysiological studies in primates consistently demonstrate that the areas traditionally labeled as primary olfactory sensory areas project to a secondary series of structures,

334

J. Seubert et al. / NeuroImage 66 (2013) 333–342

including the caudal orbitofrontal cortex (OFC), the agranular insula, the hippocampus, but also the dorsomedial nucleus of the thalamus, medial and lateral hypothalamus, and ventral striatum and pallidum (Carmichael et al., 1994; Price, 2003). Among these, the region that receives the majority of corticocortical projections from the piriform cortex is the caudal OFC (Carmichael et al., 1994; Rolls et al., 1996), which as such has traditionally been considered to constitute secondary olfactory cortex. In addition to this direct link, this region also receives indirect projections from several areas of the primary olfactory cortex through a relay in the dorsomedial nucleus of the thalamus (Buck, 2000; Powell et al., 1965). Extensive projections subsequently connect the caudal OFC to other anatomical subsections of the orbitofrontal cortex (Price, 2003). In contrast to this detailed structural understanding of the olfactory neural pathways, the functional contributions of the main processing nodes within the cortical olfactory network have only recently been systematically explored by means of functional neuroimaging techniques, and are to date far less coherently defined. Several reviews have attempted to offer a synthesis of the functional neuroimaging findings reported for the olfactory sense. These reviews have been either narrative (Gottfried, 2006; Royet and Plailly, 2004; Savic, 2002; Sela and Sobel, 2010; Yeshurun and Sobel, 2010; Zald and Pardo, 2000; Zelano and Sobel, 2005) or have made attempts to present the findings visually, plotting the activations reported in individual studies on the same anatomical template (Djordjevic and Jones-Gotman, 2006; Gottfried and Zald, 2005b; Zatorre and Jones-Gotman, 2000). While literature reviews are well suited to find common activations between studies based on a given variable of interest, much of the three-dimensional spatial information that voxel-based data consists of is lost. Function–location meta-analyses like ALE, on the other hand, merge data from many datasets to see whether consistent patterns arise that may not be evident on the basis of individual reports. By means of formal statistical integration of the available data, they are thus able to not only visualize common activation between studies, but also to provide a formal estimate of activation likelihood. The goals of the present meta-analysis were twofold. First, we identified olfactory neuroimaging studies that were sufficiently similar in methodology to allow for their combination into a quantitative estimate of activation likelihood. This not only allowed us to increase our understanding of the regions commonly involved in the processing of olfactory information, but also, to provide the chemosensory imaging community with a probability map of the olfactory network that can be used as an independent inclusive mask in statistical analyses of future neuroimaging data. Second, we divided the included studies according to variations in experimental parameters to estimate their potential impact on the reported neural representation of odor processing. In particular, we assessed the effects of cued versus non-cued odor presentation, passive smelling versus active tasks, and among the passive smelling tasks, the difference between studies asking subjects to practice velopharyngeal closure (a technique minimizing subject-induced airflow through the nasal passages) and studies not instructing subjects to practice this technique. Finally, we investigated the effects associated with the restriction of the subject sample to male or female subjects only. Method Identification of papers Suitable papers were identified by means of a two-step procedure. First, we searched the Medline and PsycINFO databases to identify human olfactory functional imaging journal articles either published or in press at the end of September 2012. Keywords used were positron emission tomography and functional magnetic resonance imaging (including common acronyms and synonyms such as PET, fMRI, regional cerebral blood flow, BOLD, etc.) which were cross-referenced

with the search terms odor*, odour*, olfact*, or smell* using the wildcard option (asterisk in this case) to capture all possible endings of the terms. As a second step, the reference lists of the original research articles resulting from this search were explored using tools accessible in Web of Science to find additional articles that were not identified by the Medline and PsycINFO searches. Inclusion criteria The contrasts reported in the identified articles had to fulfill ten criteria to be included in the meta-analysis. 1) The stimulus had to be odorous only, i.e. no additional interacting stimuli such as tastants were allowed to be present. We did not, however, exclude contrasts of odorants that had the potential to activate both the olfactory and the trigeminal system, unless this was explicitly stated by the authors of the article. 2) The contrast had to be of an odorous stimulus contrasted against an ‘odorless’ baseline. Direct comparisons between two conditions both including olfactory processing were excluded. 3) We included contrasts regardless of the task performed by the subject during or after scanning. The inclusion of contrasts independent of task allows maximum benefits from the use of statistical probability methods. Activations not mediated by olfactory processing will be identified as outliers by the ALE analyses due to the inconsistency in their activations across studies. 4) The odorant stimulus had to be administered orthonasally. 5) Whole-brain data needed to be reported in a direct contrast, i.e. contrasts reporting only results of region of interest analyses (ROI), volume of interest analyses (VOI), or significant small volume corrections (SVC) were excluded. Also, studies that did not acquire signals from the whole brain, or reported only correlations of BOLD signal change with other measures, such as behavioral data, were excluded. 6) We only included contrasts of healthy young subjects, i.e. contrasts based on special populations, such as aged individuals, were excluded. 7) No more than five contrasts from any given study were included to avoid overrepresentation of one individual experiment (on average, 1.5 contrasts were included per study). 8) The article had to report all peaks and contain sufficient explanation of both experimental and control task to allow for a proper evaluation of the method. In case of missing information, studies were included if authors provided the missing methodological information via email. 9) Only results reported in a standardized stereotaxic space, i.e. MNI or Talairach space, were included. 10) We only included contrasts originating from group-based comparisons and not from single subject analyses. BOLD signal had to be acquired from, and averaged across, at least five subjects for the contrast to be included. Deactivations were omitted due to the infrequency with which they are reported, and contrasts using between-group comparisons were omitted because they do not allow for a separation of group- and odor-dependent effects. Procedure and statistical calculations All analyses were performed using the Java-based version of the ALE software (GingerALE 2.2; http://www.brainmap.org/ale), an automated analysis software that has been described in detail elsewhere (Eickhoff et al., 2009; Eickhoff et al., 2012; Laird et al., 2005; Turkeltaub et al., 2002; Turkeltaub et al., 2012). In brief, one of the major benefits of this method compared to many others is that ALE analyzes the given coordinates to search for concordance, modeling each of the reported foci as the center of a 3D Gaussian probability distribution by permutation testing. These distributions are then used to create a whole-brain statistical map that estimates the likelihood of activation for each individual voxel as determined by the entire set of studies included (Laird et al., 2005). As a first step prior to statistical analyses, the anatomical template used for group statistics in each included article was noted. Using the GingerALE transformation tool, the reported coordinates were then transformed from their original template space into MNI space to ensure that all data

J. Seubert et al. / NeuroImage 66 (2013) 333–342

was described within a common stereotaxic space. A whole-brain ALE map was created by modeling a Gaussian probability distribution centered at each reported activation peak coordinate. A voxel-wise calculation of the probability that each activation was located within that particular voxel was then performed using a 3D Gaussian filter function with an empirically determined FWHM (full-width, halfmaximum) value (Eickhoff et al., 2009). Histogram permutation testing was subsequently used to test against the null hypothesis that activation foci are distributed uniformly across the brain (Eickhoff et al., 2012; Turkeltaub et al., 2012). Resulting statistical maps were thresholded at p b .001 (uncorrected) at the voxel level. In order to reduce the risk of false rejections of the null hypothesis due to the presence of multiple comparisons, we restricted our analyses additionally at the topological inference level, reporting only clusters meeting a permutation-based cluster-level threshold of p b .05. Using clusters of spatially contiguous voxels as units of analysis has been shown to provide better sensitivity than pure voxel-based corrections for multiple comparisons, improving signal to noise ratio and providing information about location and shape of continuous clusters, which form the fundamental level of interest in functional neuroimaging (Heller et al., 2006). We additionally investigated whether various experimental factors had an impact on activation of olfactory-related areas (Lundstrom et al., 2011). To this end, the included papers were separated into subgroups on the basis of three experimental variables. For each of these variables, pairwise comparisons between the subgroups were conducted. These variables included the cuing of odor delivery (odor delivered with cued sniffing vs. non-cued delivery), breathing technique used (natural breathing vs. velopharyngeal closure), and the cognitive component of the experimental task. For this comparison, we categorized articles into three subgroups, namely passive smelling, odor detection, and higher-order odor processing (HOOP). A task was classified as a passive smelling task if subjects were not instructed to perform any behavioral task to indicate that they had detected the odor, and they were not instructed to memorize the individual odor trials, or give a response at a later stage. The detection task category included all studies where the subject was either instructed to respond to each odor, or was explicitly instructed to try to detect whether an odor was present or not in each trial. Finally, tasks were classified as HOOP if subjects were instructed to directly discriminate between odors, identify an odor, or perform any other form of higher-order cognitive processing when presented with an odor. An additional comparison was directed at identification of the effects of the study participants' sex. For this contrast, we contrasted studies which only included male participants against studies which only included female participants. We conducted these contrasts on the basis of all included studies given that the contrast algorithm implemented in GingerALE 2.2 has been shown to reliably handle contrasts with an unequal sample size of included experiments or uneven number of foci (Eickhoff et al., 2012; Rottschy et al., 2012). The main datasets for each individual category against baseline were loaded and subtracted from one another, and then statistically compared for convergence as described in detail elsewhere (Laird et al., 2005). Given that the underlying statistical maps themselves were already thresholded at the voxel- and cluster-level, an additional conservative threshold at the subtraction level would unreasonably inflate the false negative rate. Thus, a voxel-level threshold of p b .01 was used for all subtraction analyses. Results Literature Review The systematic review of the available literature identified a total of 133 articles that acquired cortical activity elicited by odorous stimuli in humans using either PET or fMRI, spanning in time from 1992 to

335

our cut-off date (September 2012). Of these 133 articles, 45 articles contained at least one contrast satisfying all our criteria. These 45 articles yielded a total of 69 contrasts with a sum of 594 foci (Table 1). Twenty-one articles used fMRI and 24 articles used PET imaging. Of the included papers, 21 used a correction for multiple comparisons at the peak- or cluster-level. The remaining 24 studies used an uncorrected p-value threshold. The least conservative reported threshold for any included peak value was .01 uncorrected (two studies). Significant ALE values for odor processing The initial ALE meta-analysis including all ‘Odor vs. Baseline’ contrasts revealed the highest probability for activations, as demonstrated by the ALE values, in the piriform cortex. As evident from Table 2 and Fig. 1, the peaks in the piriform cortex were bilateral and of larger magnitude than any other reported peak values. The highest probability of activation in the piriform cortex was observed in the ventral portion of the posterior piriform cortex; however, in more dorsal views, the same cluster extended towards anterior portions of the piriform cortex as well as the putamen. Large probabilities for activation were also noted bilaterally in central portions of the OFC as well as the insular and peri-insular cortex. Effects of cued odor delivery As a next step, we contrasted activations originating from studies instructing participants to sniff at a certain time point (cued odor delivery) with activations originating from studies that either employed sniff-triggered odor delivery or odor delivery non-synchronized to participants’ sniffing (non-cued delivery). Assessing which brain areas were more likely to be activated under cued relative to noncued conditions, we found significantly higher likelihood for activations within a cluster that overlapped with the cluster with the highest ALE value within the overall odor against baseline analysis, peaking in the bilateral anterior piriform cortex (Table 3, Fig. 2A). In the reversed contrast, i.e. the non-cued relative to the cued condition, we found a higher likelihood for activation of the right OFC and superior frontal gyrus (Fig. 2B). Effects of attempted odor detection To assess task-specific odor activations, we contrasted activations from studies that instructed participants to respond upon odor detection with studies instructing participants to just passively smell the odor without performing a specified task. As predicted, the odor detection tasks rendered higher activation likelihood values in areas associated with conscious odor processing, such as the anterior and central portions of the OFC and the anterior insular cortex, extending into the medial putamen and nucleus accumbens (Figs. 2B, C). The reverse contrast, i.e. passive smelling against detection, did not produce any results meeting our predefined statistical threshold. When comparing simple detection tasks to higher-order olfactory processing, we found that the HOOP studies rendered lower activation likelihood values within the amygdala and hippocampus than the lower-level detection tasks (Fig. 2A), while the reverse contrast yielded no significant result. Effects of breathing techniques We also assessed potential differences in functional activations between studies employing the method of velopharyngeal closure and studies where subjects were instructed to breathe naturally. Velopharyngeal closure is a breathing method during which the soft palate is elevated in order to eliminate subject-induced airflow into the nasal cavity (Kobal, 1981). This method was initially introduced with the goal to reduce breathing-induced motor artifacts in electroencephalography

336

Table 1 Publications included in the meta-analyses. Question marks indicate that no information was given in the publication and that we were unsuccessful in contacting the authors for information. Stereotactic space (software used)

Method used

Filter for analyses (mm)

Number of subjects

Stimulated nostril

Sex

Odor stimuli used

Number of foci

Albrecht et al., HBM, 2009 Bengtsson et al., Neuroreport, 2001

MNI (SPM2) MNI (SPM99)

fMRI PET

8 10

19 23

Monorhinal Birhinal

Nicotine Vanilla, mixed

36 9

Bensafi et al., Neuroscience, 2008 Berglund et al., PNAS, 2006

MNI (SPM2) MNI (SPM99)

fMRI PET

7 10

8 24

Monorhinal Birhinal

H2S Mixed

Boyle et al., Chem Senses, 2007 Boyle et al., Cerebral Cortex, 2009 Ciumas et al., Neuroimage, 2008 Dade et al., Neuroimage, 2001 Dade et al., Brain, 2002 Djordjevic et al., Neuroimage, 2005 Gottfried and Dolan, Neuron, 2003 Gottfried et al., J Neurosc, 2002 Herz et al., Neuropsychologia, 2004 Hillert et al., HBM, 2007 Kareken et al., Neuropsychology, 2003 Kareken et al., Neuroimage, 2004 Kjelvik et al., J Neurophysiol, 2012 Lombion et al., HBM, 2009 Lundstrom et al., Cerebral Cortex, 2008 Osterbauer et al., J Neurophysiol, 2005 Plailly et al., HBM, 2007 Poellinger et al., Neuroimage, 2001 Qureshy et al., J Neurophysiol, 2000 Reske et al., Behav Neurosc, 2010 Royet et al., J Neurosc, 2000 Royet et al., Neuroimage, 2001 Savic and Berglund, HBM, 2004 Savic and Gulyas, Neuroreport, 2000 Savic et al., Neuron, 2000 Savic et al., HBM, 2002 Savic et al., PNAS, 2005

MNI (SPM5) MNI (SPM2) MNI (SPM2) MNI (DOT) MNI (DOT) MNI (DOT) MNI (SPM 99) MNI (SPM99) TAL (Afni) MNI (SPM99) MNI (SPM99) MNI (SPM99) MNI (FSL 4.0) TAL (BrainVoyager) MNI (DOT) MNI (FSL) MNI (SPM99) MNI (SPM96) MNI (SPM96) MNI (SPM2) MNI (SPM96) MNI (SPM96) MNI (SPM99) MNI (SPM96) MNI (SPM96) MNI (SPM99) MNI (SPM99)

fMRI PET PET PET PET PET fMRI fMRI fMRI PET PET PET fMRI fMRI PET fMRI PET fMRI PET fMRI PET PET PET PET PET PET PET

7 12 10 14 14 14 8 8 6 10 8 8 5 4 12 5 12 ? 16 10 20 20 10 10 10 10 10

15 12 21 12 12 12 17 17 5 12 11 15 17 15 15 9 16 9 8 15 12 12 14 18 18 12 24

Monorhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Birhinal Monorhinal Birhinal Birhinal Birhinal Monorhinal Monorhinal Birhinal Birhinal

Seubert et al., Neuroimage, 2010 Small et al., Neuroreport, 1997 Sobel et al., J Neurophysiol, 2000 Stankewitz et al., Cephalalgia, 2010 Tabert et al., Neuroimage, 2007 \Treyer et al., Rhinology, 2006 Veldhuizen et al., Front Hum Neurosc, 2010 Wang et al., J Gerontology, 2005 Weismann et al., Neuroimaging Clin N Am, 2001 Wicker et al., Neuron, 2003 Wiesmann et al., Neuroimage, 2006 Yousem et al., Am J Neuroradiol, 1999 Zatorre et al., Nature, 1992 Zatorre et al., Neuroreport, 2000

MNI (SPM5) MNI (DOT) MNI (mrVista) MNI (SPM5) TAL (FSL 5.4) MNI (SPM99) MNI(SPM5) MNI (SPM2) MNI (SPM96b)

fMRI PET fMRI fMRI fMRI PET fMRI fMRI fMRI

8 14 ? 10 8 15 6 8 6

44 10 8 20 10 9 19 11 14

Monorhinal Birhinal Birhinal Monorhinal Birhinal Birhinal Birhinal Birhinal Birhinal

Mixed (9 W) Women only (12 W), Men only (11 M) Women only Women only (12 W), Men only (12 M) Men only Mixed (6 W) Mixed (11 W) Mixed (6 W) Mixed (6 W) Mixed (6 W) Mixed (13 W) Mixed (10 W) Women only Women only Mixed (7 W) Mixed (8 W) Women only Women only Women only Mixed (5 W) Mixed (8 W) Mixed (4 W) Men only Women only Men only Men only Men only Women only Women only Women only Women only (12 W), Men only (12 M) Mixed (23 W) Mixed (5 W) Mixed (4 W) Mixed (9 W) Mixed (5 W) Men only Mixed (15 W) Mixed (6 W) Men only

MNI MNI MNI MNI MNI

fMRI fMRI fMRI PET PET

10 8 10 20 14

6 22 5 11 12

Monorhinal Monorhinal Birhinal Birhinal Birhinal

(SPM99) (SPM2) (SPM96) (DOT) (DOT)

Mixed (3 W) Mixed (9 W) Mixed (3 W) Mixed (6 W) Mixed (6 W) Σ 286M/348W

Number of contrasts

Cued odor delivery

Task

1 4

No — VC Yes

Passive Passive

4 4

1 2

No — VC Yes

Passive Passive

PEA Citral, Pyridine, Mixed Mixed Mixed Mixed Mixed Mixed PEA, Vanilla, 4MP Mixed Mixed Mixed Mixed Mixed PEA, IAA Mixed Mixed Mixed PEA Mixed Rotten yeast, vanilla Mixed Mixed Mixture Mixture Mixed Vanilla Mixed

3 10 5 18 31 16 11 9 4 12 13 13 28 8 3 6 10 45 5 26 3 17 9 5 5 2 4

1 2 2 1 3 1 1 3 1 2 1 2 1 1 1 1 3 2 1 2 1 5 2 1 1 1 2

No — VC Yes Yes No — Nat No — Nat Yes Yes Yes No — Nat Yes Yes Yes/VC No— Nat No — Nat Yes No — Nat Yes No — Nat No — Nat No — Nat No — Nat No — Nat Yes Yes Yes Yes Yes

Passive Detection Passive HOOP HOOP Detection Detection Undecided Detection Passive Detection Passive Detection Passive Detection Detection Detection/HOOP Detection Passive Passive HOOP HOOP Passive Passive Passive Passive Passive

Vanilla, H2S Mixed Mixed PEA Various PEA Mixed Lavender, spearmint Mixed

3 6 32 16 11 6 49 16 10

1 1 1 1 1 1 1 1 1

No — Nat No — Nat Yes Yes No — Nat No — Nat Yes No — Nat No — VC

Passive Detection Detection Detection Detection Undecided Passive Passive Passive

Eugenol PEA Eugenol, PEA Mixture Mixed

26 15 13 6 11 Σ 594

? No No No No

Undecided Passive Passive Passive HOOP

2 1 1 1 2 69

— VC — Nat — Nat — Nat

Abreviations: MNI = MNI/ICBM coordinates, TAL = Talairach coordinates, SPM/DOT/Afni/FSL = Different imaging analyses software, H2S = Hydrogen Sulfide, PEA = Phenylethyl Alcohol, 4MP = 4-methyl-pentanoic acid, IAA = Iso-Amyl-Acetate, EB= Ethyl Butanol, VC = Velopharyngeal closure, Nat = Natural breathing, HOOP = Higher order olfactory processing.

J. Seubert et al. / NeuroImage 66 (2013) 333–342

Publication

J. Seubert et al. / NeuroImage 66 (2013) 333–342

337

Table 2 Location of significant ALE maxima for reported peak activations in Odor vs Baseline contrasts. The ‘Percent contributing’ column indicates the percentage of the included papers that contributed to this specific result. Anatomical label

ALE values

Left hemisphere

Percent contributing

x

y

z

0.1

−22

0

−14

69

Ventral insula/planum polare

0.02 0.03 0.02 0.02 0.02

−36 −30 −38 −36 −40

0 18 12 14 2

8 6 −2 −8 −14

Prefrontal cortex Orbitofrontal cortex

0.03

−24

30

−10

Fronto-temporal junction Piriform cortex Insular cortex Anterior insula

Middle frontal gyrus

0.02

−4

18

50

Right hemisphere

Percent contributing

x

y

z

0.09

22

2

−12

69

62 62 62 62 69

0.02 0.02 0.02

36 28 40

24 16 12

−2 8 −2

58 58 58

62

0.04

28

34

−12

58

0.02

20

48

−10

58

0.02 0.02 0.02 0.02 0.02 0.02 0.02

10 6 6 4 42 32 44

36 36 28 14 34 46 40

−2 20 38 56 28 8 0

58 9 18 18 11 7 7

Lateral orbital Gyrus/frontomarginal Gyrus Anterior cingulate Gyrus Superior frontal gyrus

ALE values

18

Fig. 1. Localization of significant ALE values superimposed on a standard anatomical template in MNI space in neurological convention (left side of the brain on left side of the picture). Numbers above brain slices indicate stereotactic coordinates in axial orientation and blue lines indicates location of displayed slices.

338

J. Seubert et al. / NeuroImage 66 (2013) 333–342

Table 3 Location of significant ALE maxima for reported peak activations in difference contrasts comparing different methodological approaches. Contrasts where no significant differences were observed are not shown in the table. Subpeaks within anatomically homogeneous areas are not reported. The column labeled as PC indicates the percentage of the included papers that contributed to this specific result. Contrast/anatomical label

ALE values

Left hemisphere x

Cued > Uncued Anterior Piriform Cortex

y

ALE values

z

Right hemisphere x

z

23

1

−5

55

3.72 3.72

21 8

43 24

−19 40

4 21

3.72 3.54

24 29

30 15

−15 9

29 36

2.54

30

32

−6

16

2.62

49

2

43

34

Females > Males Putamen Lateral amygdala

3.72 2.82

23 26

−5 −6

2 −16

46 46

Males > Females Middle frontal gyrus

3.24

31

49

3

9

1

−9

y

PC

3.72

2.79

−16

PC

5

Uncued > Cued Orbitofrontal cortex Superior frontal gyrus Detection > Passive Orbitofrontal cortex Anterior insula Nucleus accumbens

2.73 2.82 2.71

−20 −28 −12

26 12 4

−10 0 −6

7 14 14

Detection > HOOP Lateral amygdala Nucleus accumbens Head of the hippocampus

3.09 3.72 3.09

−22 −13 −20

−8 4 −10

−23 −4 −16

43 43 43

Natural Breathing > Velopharyngeal closure Orbitofrontal cortex Velopharyngeal closure > Natural Breathing Medial frontal gyrus Frontal operculum 3.09

−40

16

based event-related potential (ERP) studies. It has since been suggested, however, that piriform activation may be elicited by sniffing even in absence of an odorant stimulus, potentially forming part of a feedback mechanism which regulates sniffing and odor sampling (Sobel et al., 1998). This would imply that the active prevention of sniffing may reduce observed activation within the olfactory networks. To assess the influence of this method, we only included experiments during which no sniff cue was presented, since an instruction to sniff is incompatible with the practice of velopharyngeal closure. The remaining studies were then separated by breathing method. We found that studies practicing natural breathing presented reliably higher activations likelihoods in the right OFC (Fig. 2B). Studies that did practice velopharyngeal closure, on the other hand, elicited stronger activation in a cluster in the frontal operculum extending into the inferior insula. Effects of participant sex A final contrast was conducted to test for differences between studies which only included male participants and studies which included only female participants. Earlier studies were frequently restricted to male subjects only in an effort to reduce anatomical and functional variability. On the other hand, olfactory research has sometimes been restricted to female subjects because of reports that women outperform men on tests of olfactory functioning. We found that studies including only female participants were more likely to demonstrate activation in a cluster centered on the amygdala and extending into the putamen and hippocampus. Studies with male subjects only, on the other hand, were more likely to report activation of the middle frontal gyrus. The statistical parametric map (SPM) of the main analysis, as well as of the comparisons between the different stimulation methods, can be downloaded from our website (http://flavor.monell.org/~jlundstrom/ index_ALE.html).

−16

17

Discussion The primary aim of this study was to delineate the neural correlates of olfactory processing by taking advantage of the statistical power of a meta-analytical approach. As expected, the present ALE analysis on published papers contrasting the presentation of odorous stimuli against the inhalation of odorless air revealed the highest likelihood for activation in the cortical structures that are commonly referred to as primary and secondary olfactory cortex, in particular the piriform cortex and the OFC. Additionally, midbrain and cortical areas that have functionally and anatomically been shown to have strong connections with olfactory processing networks were reliably activated. The task-related differences in activation patterns support current models of olfactory percept formation, which suggest that value-added computations are emerging at each level of the processing hierarchy.

Representation of olfactory stimulation within the piriform cortex The piriform cortex, situated at the junction of the frontal and temporal lobes, showed the highest consistency across studies for a strong and spatially continuous activation in the overall contrast of odor against baseline. The highest likelihood for activation in this meta-analysis was observed in the posterior portion of the piriform cortex, with the activation cluster extending, at a more dorsal view, into the anterior portion. The cluster further extended into the amygdala and ventral striatum as well as the medial putamen, all regions known to be closely interconnected (Carmichael et al., 1994) and subserving affective learning and memory functions. One of the previously published analyses spatially mapping odor-evoked activation in the piriform cortex (Djordjevic and Jones-Gotman, 2006) reports a localized mean in close spatial proximity to the peak demonstrated in

J. Seubert et al. / NeuroImage 66 (2013) 333–342

339

Fig. 2. Significant ALE difference contrasts superimposed on an anatomical template in MNI space in neurological convention (left side of the brain on left side of the picture) of the difference between activations obtained in studies using different experimental parameters. Stereotactic coordinates for each section are reported in the bottom right corner. Lighter colors represent higher ALE values up to a maximum of 5. A. Significant differences in piriform cortex activation. B. Significant differences in OFC activation C. Significant differences in anterior insula activation. Abbreviations: HOOP — higher order olfactory processing, NB — natural breathing, VPC — velopharyngeal closure.

our analysis, in the temporal portion of piriform cortex bilaterally (y = 3, z = − 16). Numerous studies to date have convincingly demonstrated the functional relevance of the anatomical sub-division of the piriform cortex between an anterior (frontal) and a posterior (temporal) portion. While projections to the anterior piriform cortex still appear to loosely reflect chemosensory composition (Davison and Ehlers, 2011), posterior piriform cortex coding appears to be more biased towards subjective perceptual similarities and differences between odorants (Chapuis and Wilson, 2012; Howard et al., 2009; Zelano et al., 2011). Recently, it has further been suggested that ensemble activation in the anterior piriform cortex is most strongly shaped by the expected odor quality (Veldhuizen and Small, 2011; Zelano et al., 2009; Zelano et al., 2011), while posterior piriform activation reflects actual stimulus identity. Within this conceptual framework, it is of interest to note that we identified anterior piriform cortex to be more likely activated when subjects received a cue to expect odor onset than when this was not the case. This finding stresses the importance of attentional processes for anterior piriform activation (Zelano et al., 2005). Previous research has suggested that nasal airflow may provide a feedback mechanism driving directed sniffing and targeted odor sampling (Sobel et al., 1998; Sobel et al., 2000). The existence of such a mechanism has recently been supported by the finding that the intrinsic state of excitatory networks within piriform cortex affects how olfactory bulb input is processed, selectively enhancing or suppressing it (Franks et al., 2011). In line with these findings, it was suggested that this intimate relationship between sniff and odor perception might make olfactory activation difficult to observe in studies

which contrast odor against a sniff-only baseline (for a detailed review, see Bensafi, 2012). On a meta-analytical level, however, we can clearly demonstrate that activation of anterior piriform cortex increases when the environment is actively monitored as a result of expectancy. In line with the notion of attentional modulation of the piriform cortex, results of previous studies indicate that piriform cortex activity may be highest during detection tasks, and decrease when more complex judgments are performed during odor perception (Kareken et al., 2003; Qureshy et al., 2000; Savic et al., 2000). The significant difference between detection and higher-order cognitive tasks that we report on a meta-analytical level supports these findings. However, our localized peak is located within the amygdala, posterior to the previously reported accounts of individual studies, and extends into the hippocampus; both are areas frequently implicated in the cognitive evaluation of olfactory stimulus material (Igarashi et al., 2012). A possible interpretation of these results is that feedback loops from direct projections to the amygdala may be engaged in such automatized evaluative behaviors during detection tasks. The execution of a concurrent cognitively demanding task might then deploy attention away from cognitive odor assessment, thus potentially reducing the amygdala response. This view is further strengthened by our finding that women, who have been shown to outperform men in odor identification and odor memory tasks (Doty et al., 1985; Larsson et al., 2004), showed stronger activity in the same region. Future studies are, however, needed to determine the accuracy of this hypothesis and it should be noted that many behavioral and neuroimaging studies have failed to demonstrate sex differences in odor processing; rather, effects might be mediated by sex-dependent differences in

340

J. Seubert et al. / NeuroImage 66 (2013) 333–342

trigeminal processing (Lundstrom and Hummel, 2006; Lundstrom et al., 2005). Representation of olfactory stimulation within orbitofrontal cortex The notion that crucial steps in the formation of the odor percept occur within the OFC is increasingly supported by the literature. In this area, higher-order cognitive processes, including experiencedependent modulation, affective coding (Royet et al., 2001; Zald and Pardo, 1997), and influences from other sensory modalities, i.e. multimodal convergence (Gottfried and Dolan, 2003; Rolls and Baylis, 1994), help shape the odor signal coming from the piriform cortex into the final conscious percept that we experience when we smell an odor. Anatomically, its role as a central node in the formation of the final olfactory percept is supported by a dense network of both afferent and efferent connections which link this area to other sensory systems (Carmichael and Price, 1995). The OFC is thus thought to integrate the olfactory percept with input from the many cortical and subcortical regions responsible for basic sensory processing, as well as with cognitive information extending beyond sensory input alone. Previously published meta-analyses based on voxel-coordinate mapping localized the parts of the OFC that respond to olfactory stimulation bilaterally in an area near the transverse orbital sulci (Gottfried, 2007; Gottfried and Zald, 2005). Recently, it has further been demonstrated that lesions to this area can result in a loss of the ability to consciously perceive odors in spite of intact early sensory pathways (Li et al., 2010). Peak activations for odor against baseline in our meta-analysis closely converge with these previous reports. Differential activation of the OFC was seen when contrasting results from non-cued against those from cued studies, and in tasks which required active detection rather than mere passive exposure. These findings support the view of OFC being crucially involved in monitoring and spontaneous adaptation to changes in the olfactory environment, a task that gains importance when the environment needs to be monitored for sudden changes. Activation within the OFC has also been repeatedly demonstrated in tasks requiring flexible odor representations for the adjustment of behavior, such as perceptual decision-making and confidence judgment (Bowman et al., 2009; Kepecs et al., 2008), perceptual learning, valence judgment, and aversive conditioning (Gottfried, 2007; Rolls et al., 2003), as well as integration of information from other sensory modalities (Gottfried and Dolan, 2003; Price, 2008). Most recently, specific odor-evoked spatial and temporal response patterns in the OFC have been convincingly demonstrated to underlie the formation of these adaptable odor representations (Gottfried and Zelano, 2011; Zelano et al., 2011). In this context it is further interesting that, in line with the findings reported by Kareken et al. (2004), natural sniffing was associated with higher activation than velopharyngeal closure in the olfactory OFC, thus providing further support to the notion that sniffing may constitute a crucial component to the deployment of olfactory attention or a low-level readiness response. The exact location of the secondary olfactory cortex within the OFC suggested by our ALE analysis does not fully overlap with the area expected from the animal model. Studies in rodents and primates have consistently marked the caudal OFC to play this role in olfaction, while converging data coming from the human neuroimaging work seem to suggest a slightly anterior region within the OFC, namely the area surrounding the orbital transverse sulcus (Gottfried and Zald, 2005). Whether this represents a true difference between species, or a slight displacement resulting from some systematic bias elicited by the use of functional neuroimaging techniques, remains to be explored. Representation of olfactory stimulus material in higher order association cortex High activation likelihood values were also observed within areas that are not traditionally thought to belong to the olfactory system,

but are increasingly understood to form part of a wider network involved in chemosensory processing. In particular, we found high ALE values in the anterior insula, the middle frontal, superior frontal and cingulate gyri. Anterior insular cortex, although primarily considered an important processing node of the gustatory cortical network (Faurion et al., 2005; Small et al., 1999; Veldhuizen et al., 2011), has consistently been reported to respond to odorous stimuli ever since the first olfactory neuroimaging study of olfaction (Zatorre et al., 1992). It is now increasingly thought of as a chemosensory convergence zone involved in flavor perception (de Araujo et al., 2003; Lundstrom et al., 2011; Rolls et al., 2003; Seo and Hummel, 2011). As an important recipient of amygdala projections (Carlson et al., 2011), the anterior insula is also likely involved in the elicitation of avoidance behavior and has frequently been reported to differentially activate to negative valence and dislikes, both in olfaction and trigeminal perception alone (Albrecht et al., 2010; Royet et al., 2003), as well as during multisensory integration (Seubert et al., 2010; Wicker et al., 2003). In line with this idea of the anterior insula as a cognitiveevaluative area, we found that activation likelihood values were particularly high for studies requiring a task to be performed while smelling the olfactory stimulus. The middle frontal gyrus is an area which has been implicated in top-down modulation of the response to conceptually overlapping stimuli (Schacter et al., 2007) and could thus be involved in cognitive labeling of odors. It is of interest to note that this area tended to be more strongly activated in male subjects, who tend to have more difficulty retrieving labels to match odors to verbal descriptors (Oberg et al., 2002). Due to the small sample size of male-only studies included here, future studies will, however, be needed to determine whether these sex-dependent activation differences are indeed stable observations. In particular, investigations of the patterns of connectivity modulation between these upstream cognitive areas and cortical as well as subcortical chemosensory target regions will provide unique insights into their functional relevance to the formation of the olfactory percept. Limitations While providing a powerful method to quantify statistical overlap between studies, the ALE method, like all meta-analyses, suffers from some inherent drawbacks. A main concern is that studies may show subtle methodological differences which cannot be appropriately assessed by meta-analysis, such as an inherent tendency to report a larger or smaller number of foci or the potential for masking of perceived non-relevant voxels during the statistical pre-processing steps within the individual studies. Nonetheless, these, along with other potential errors and differences between studies in the chosen data acquisition and analysis parameters, can be assumed to be largely non-systematic. Moreover, since the sample size and number of reported foci per study are factored into the analysis by the ALE algorithm (Eickhoff et al., 2009), no individual study has the capacity to disproportionately bias the observed results. A systematic overestimation of results could, however, occur due to the fact that absence of activations cannot be statistically assessed within our present analyses models and that hard to interpret deactivations are rarely reported. Moreover, one should also note the prevailing tendency in all fields of science that results which conform with the dominating opinion within the specific research field are published while conflicting results may never even be prepared for publication. Another inherent concern is found in the constant development of methods for neuroimaging. Newer studies will likely be more methodologically rigorous than older studies, as more lenient thresholds become less likely to pass peer-review. Every meta-analyst of neuroimaging data therefore has to carefully weigh the cost and benefit of employing rigorous selection criteria that adhere to the current standard, but which would likely result in an overrepresentation of recent work, or to accept peer-review at the time of manuscript publication as the main

J. Seubert et al. / NeuroImage 66 (2013) 333–342

criterion for scientific rigor. The latter comes at the cost of including more liberally reported studies. In the present analysis, we decided to follow the latter approach. Future meta-analyses which continue to incorporate the increasing body of literature will thus be required to determine whether observed differences persist over time. It should further be noted that, due to their location in direct proximity to air-filled cavities, primary and secondary olfactory areas are especially prone to susceptibility artifacts. The strong and consistent reported activations suggest that there is no underrepresentation of these areas in the published literature; however, one should keep in mind that studies which fail to report activation in these wellestablished areas could be less likely to pass peer review and thus be underrepresented in this meta-analysis. Finally, our meta-analysis does not allow for the discrimination of more subtle differences between specific cognitive tasks — tasks which likely have differential effects on observed activation patterns. While we emphasized group size over number of potential sub-comparisons to ensure a relative robustness of the reported data, the ongoing accumulation of olfactory neuroimaging studies will over time likely allow for more detailed task comparison in future meta-analyses. This will in particular increase our power to investigate higher-order cognitive processing of odors. Conclusion In the meta-analysis reported here, we statistically quantified the anatomical overlap between functional neuroimaging studies of olfaction. Our results show a robust activation of core regions of the olfactory network, starting with regions receiving direct projections from the olfactory bulb, such as the piriform cortex and amygdala, and extending into higher-order association areas such as the orbitofrontal cortex and the anterior insula. By means of meta-analytical statistics, we were able to provide robust support for current models of cortical olfactory processing, highlighting areas most likely to be influenced by experimental parameters such as sniff, the presence or absence of cues, and the cognitive load of the task completed by the subject. The task-dependent commonalities and differences in neural processing demonstrate that the olfactory system is a complex and dynamically regulated system. Acknowledgments This material is based upon work supported by the U. S. Army Research Office under grant number W911NF-11-1-0087 and the Swedish Research Council (2009-2337), both awarded to JNL, and partly by the Natural Sciences and Engineering Research Council of Canada (NSERC), grant 355938-08 awarded to JD. JS is supported by a postdoctoral fellowship awarded by the German Research Foundation (DFG SE 2147/1-1). References Albrecht, J., Kopietz, R., Linn, J., Sakar, V., Anzinger, A., Schreder, T., et al., 2009. Activation of olfactory and trigeminal cortical areas following stimulation of the nasal mucosa with low concentrations of S(-)-nicotine vapor-an fMRI study on chemosensory perception. Hum. Brain Mapp. 30 (3), 699–710. Albrecht, J., Kopietz, R., Frasnelli, J., Wiesmann, M., Hummel, T., Lundstrom, J.N., 2010. The neuronal correlates of intranasal trigeminal function—an ALE meta-analysis of human functional brain imaging data. Brain Res. Rev. 62, 183–196. Bengtsson, S., Berglund, H., Gulyas, B., Cohen, E., Savic, I., 2001. Brain activation during odor perception in males and females. Neuroreport 12 (9), 2027–2033. Bensafi, M., 2012. The role of the piriform cortex in human olfactory perception: insights from functional neuroimaging studies. Chemosens. Percept. 5, 4–10. Bensafi, M., Iannilli, E., Gerber, J., Hummel, T., 2008. Neural coding of stimulus concentration in the human olfactory and intranasal trigeminal systems. Neuroscience 154 (2), 832–838. Berglund, H., Lindstrom, P., Savic, I., 2006. Brain response to putative pheromones in lesbian women. Proc. Natl. Acad. Sci. U. S. A. 103 (21), 8269–8274 (PMCID: 1570103).

341

Bowman, N.E., Howard, J.D., Kording, K.P., Gottfried, J.A., 2009. Perceptual decisionmaking in the human olfactory brain. Chem. Senses 34, A56–A57. Boyle, J.A., Djordjevic, J., Olsson, M.J., Lundstrom, J.N., Jones-Gotman, M., 2009. The human brain distinguishes between single odorants and binary mixtures. Cereb. Cortex. 19 (1), 66–71. Boyle, J.A., Heinke, M., Gerber, J., Frasnelli, J., Hummel, T., 2007. Cerebral activation to intranasal chemosensory trigeminal stimulation. Chem. Senses. 32 (4), 343–353. Buck, L.B., 2000. The molecular architecture of odor and pheromone sensing in mammals. Cell 100, 611–618. Carlson, J.M., Greenberg, T., Rubin, D., Mujica-Parodi, L.R., 2011. Feeling anxious: anticipatory amygdalo-insular response predicts the feeling of anxious anticipation. Soc. Cogn. Affect. Neurosci. 6, 74–81. Carmichael, S.T., Clugnet, M.C., Price, J.L., 1994. Central olfactory connections in the macaque monkey. J. Comp. Neurol. 346, 403–434. Carmichael, S.T., Price, J.L., 1995. Sensory and premotor connections of the orbital and medial prefrontal cortex of macaque monkeys. J. Comp. Neurol. 363, 642–664. Chapuis, J., Wilson, D.A., 2012. Bidirectional plasticity of cortical pattern recognition and behavioral sensory acuity. Nat. Neurosci. 15, 155–161. Ciumas, C., Lindstrom, P., Aoun, B., Savic, I., 2008. Imaging of odor perception delineates functional disintegration of the limbic circuits in mesial temporal lobe epilepsy. Neuroimage 39 (2), 578–592. Dade, L.A., Zatorre, R.J., Evans, A.C., Jones-Gotman, M., 2001. Working memory in another dimension: functional imaging of human olfactory working memory. Neuroimage 14 (3), 650–660. Dade, L.A., Zatorre, R.J., Jones-Gotman, M., 2002. Olfactory learning: convergent findings from lesion and brain imaging studies in humans. Brain 125 (Pt 1), 86–101. Davison, I.G., Ehlers, M.D., 2011. Neural circuit mechanisms for pattern detection and feature combination in olfactory cortex. Neuron 70, 82–94. de Araujo, I.E., Rolls, E.T., Kringelbach, M.L., McGlone, F., Phillips, N., 2003. Taste-olfactory convergence, and the representation of the pleasantness of flavour, in the human brain. Eur. J. Neurosci. 18, 2059–2068. Djordjevic, J., Jones-Gotman, M., 2006. Olfaction and the temporal lobes. In: Brewer, W.J., Castle, D., Pantelis, C. (Eds.), Olfaction and The Brain. University Press, Cambridge, pp. 28–49. Djordjevic, J., Zatorre, R.J., Petrides, M., Boyle, J.A., Jones-Gotman, M., 2005. Functional neuroimaging of odor imagery. Neuroimage 24 (3), 791–801. Doty, R.L., Applebaum, S., Zusho, H., Settle, R.G., 1985. Sex differences in odor identification ability: a cross-cultural analysis. Neuropsychologia 23, 667–672. Eickhoff, S.B., Bzdok, D., Laird, A.R., Kurth, F., Fox, P.T., 2012. Activation likelihood estimation meta-analysis revisited. Neuroimage 59, 2349–2361. Eickhoff, S.B., Laird, A.R., Grefkes, C., Wang, L.E., Zilles, K., Fox, P.T., 2009. Coordinate-based activation likelihood estimation meta-analysis of neuroimaging data: a randomeffects approach based on empirical estimates of spatial uncertainty. Hum. Brain Mapp. 30, 2907–2926. Faurion, A., Kobayakawa, T., Cerf-Ducastel, B., 2005. Cerebral imaging in taste. Chem. Senses 30 (Suppl. 1), i230–i231. Firestein, S., 2001. How the olfactory system makes sense of scent. Nature 413, 211–218. Franks, K.M., Russo, M.J., Sosulski, D.L., Mulligan, A.A., Siegelbaum, S.A., Axel, R., 2011. Recurrent circuitry dynamically shapes the activation of piriform cortex. Neuron 72, 49–56. Gottfried, J.A., 2006. Smell: central nervous processing. Adv Otorhinolaryngol. 63, 44–69. Gottfried, J.A., 2007. What can an orbitofrontal cortex-endowed animal do with smells? Ann. N. Y. Acad. Sci. 1121, 102–120. Gottfried, J.A., Deichmann, R., Winston, J.S., Dolan, R.J., 2002. Functional heterogeneity in human olfactory cortex: an event-related functional magnetic resonance imaging study. J. Neurosci. 22 (24), 10819–10828. Gottfried, J.A., Dolan, R.J., 2003. The nose smells what the eye sees: crossmodal visual facilitation of human olfactory perception. Neuron 39, 375–386. Gottfried, J.A., Zald, D.H., 2005. On the scent of human olfactory orbitofrontal cortex: meta-analysis and comparison to non-human primates. Brain Res. Brain Res. Rev. 50, 287–304. Gottfried, J.A., Zelano, C., 2011. The value of identity: olfactory notes on orbitofrontal cortex function. Ann. N. Y. Acad. Sci. 1239, 138–148. Heller, R., Stanley, D., Yekutieli, D., Rubin, N., Benjamini, Y., 2006. Cluster-based analysis of FMRI data. Neuroimage 33, 599–608. Herz, R.S., Eliassen, J., Beland, S., Souza, T., 2004. Neuroimaging evidence for the emotional potency of odor-evoked memory. Neuropsychologia 42 (3), 371–378. Hillert, L., Musabasic, V., Berglund, H., Ciumas, C., Savic, I., 2007. Odor processing in multiple chemical sensitivity. Hum. Brain Mapp. 28 (3), 172–182. Howard, J.D., Plailly, J., Grueschow, M., Haynes, J.D., Gottfried, J.A., 2009. Odor quality coding and categorization in human posterior piriform cortex. Nat. Neurosci. 12, 932–938. Igarashi, K.M., Ieki, N., An, M., Yamaguchi, Y., Nagayama, S., Kobayakawa, K., Kobayakawa, R., Tanifuji, M., Sakano, H., Chen, W.R., Mori, K., 2012. Parallel mitral and tufted cell pathways route distinct odor information to different targets in the olfactory cortex. J. Neurosci. Off. J. Soc. Neurosci. 32, 7970–7985. Kareken, D.A., Mosnik, D.M., Doty, R.L., Dzemidzic, M., Hutchins, G.D., 2003. Functional anatomy of human odor sensation, discrimination, and identification in health and aging. Neuropsychology 17, 482–495. Kareken, D.A., Sabri, M., Radnovich, A.J., Claus, E., Foresman, B., Hector, D., Hutchins, G.D., 2004. Olfactory system activation from sniffing: effects in piriform and orbitofrontal cortex. Neuroimage 22, 456–465. Kepecs, A., Uchida, N., Zariwala, H.A., Mainen, Z.F., 2008. Neural correlates, computation and behavioural impact of decision confidence. Nature 455, 227–231. Kjelvik, G., Evensmoen, H.R., Brezova, V., Haberg, A.K., 2012. The human brain representation of odor identification. J. Neurophysiol. 108 (2), 645–657.

342

J. Seubert et al. / NeuroImage 66 (2013) 333–342

Kobal, G., 1981. Electrophysiologische Untersuchungen des menschlichen Geruchssinns. Thieme Verlag, Stuttgart. Laird, A.R., Fox, P.M., Price, C.J., Glahn, D.C., Uecker, A.M., Lancaster, J.L., Turkeltaub, P.E., Kochunov, P., Fox, P.T., 2005. ALE meta-analysis: controlling the false discovery rate and performing statistical contrasts. Hum. Brain Mapp. 25, 155–164. Larsson, M., Nilsson, L.G., Olofsson, J.K., Nordin, S., 2004. Demographic and cognitive predictors of cued odor identification: evidence from a population-based study. Chem. Senses 29, 547–554. Li, W., Lopez, L., Osher, J., Howard, J.D., Parrish, T.B., Gottfried, J.A., 2010. Right orbitofrontal cortex mediates conscious olfactory perception. Psychol. Sci. 21, 1454–1463. Lombion, S., Comte, A., Tatu, L., Brand, G., Moulin, T., Millot, J.L., 2009. Patterns of cerebral activation during olfactory and trigeminal stimulations. Hum. Brain Mapp. 30 (3), 821–828. Lundstrom, J.N., Boesveldt, S., Albrecht, J., 2011. Central processing of the chemical senses: an overview. ACS Chem. Neurosci. 2, 5–16. Lundstrom, J.N., Boyle, J.A., Zatorre, R.J., Jones-Gotman, M., 2008. Functional neuronal processing of body odors differs from that of similar common odors. Cereb. Cortex. 18 (6), 1466–1474. Lundstrom, J.N., Frasnelli, J., Larsson, M., Hummel, T., 2005. Sex differentiated responses to intranasal trigeminal stimuli. Int. J. Psychophysiol. 57, 181–186. Lundstrom, J.N., Hummel, T., 2006. Sex-specific hemispheric differences in cortical activation to a bimodal odor. Behav. Brain Res. 166, 197–203. Oberg, C., Larsson, M., Bäckman, L., 2002. Differential sex effects in olfactory functioning: the role of verbal processing. J. Int. Neuropsychol. Soc. 8, 691–698. Osterbauer, R.A., Matthews, P.M., Jenkinson, M., Beckmann, C.F., Hansen, P.C., Calvert, G.A., 2005. Color of scents: chromatic stimuli modulate odor responses in the human brain. J. Neurophysiol. 93 (6), 3434–3441. Plailly, J., Radnovich, A.J., Sabri, M., Royet, J.P., Kareken, D.A., 2007. Involvement of the left anterior insula and frontopolar gyrus in odor discrimination. Hum. Brain Mapp. 28 (5), 363–372. Poellinger, A., Thomas, R., Lio, P., Lee, A., Makris, N., Rosen, B.R., et al., 2001. Activation and habituation in olfaction–an fMRI study. Neuroimage 13 (4), 547–560. Powell, T.P., Cowan, W.M., Raisman, G., 1965. The central olfactory connexions. J. Anat. 99, 791–813. Price, J.L., 1985. Beyond the primary olfactory cortex: olfactory-related areas in the neocortex, thalamus and hypothalamus. Chem. Senses 10, 239–258. Price, J.L., 2003. The olfactory system. In: Paxinos, G. (Ed.), The human nervous system. Academic, San Diego, pp. 1198–1212. Price, J.L., 2008. Multisensory convergence in the orbital and ventrolateral prefrontal cortex. Chemosens. Percept. 1, 103–109. Qureshy, A., Kawashima, R., Imran, M.B., Sugiura, M., Goto, R., Okada, K., Inoue, K., Itoh, M., Schormann, T., Zilles, K., Fukuda, H., 2000. Functional mapping of human brain in olfactory processing: a PET study. J. Neurophysiol. 84, 1656–1666. Reske, M., Kellermann, T., Shah, N.J., Schneider, F., Habel, U., 2010. Impact of valence and age on olfactory induced brain activation in healthy women. Behav. Neurosci. 124 (3), 414–422. Rolls, E.T., Baylis, L.L., 1994. Gustatory, olfactory, and visual convergence within the primate orbitofrontal cortex. J. Neurosci. 14, 5437–5452. Rolls, E.T., Critchley, H.D., Treves, A., 1996. Representation of olfactory information in the primate orbitofrontal cortex. J. Neurophysiol. 75, 1982–1996. Rolls, E.T., Kringelbach, M.L., de Araujo, I.E., 2003. Different representations of pleasant and unpleasant odours in the human brain. Eur. J. Neurosci. 18, 695–703. Rottschy, C., Langner, R., Dogan, I., Reetz, K., Laird, A.R., Schulz, J.B., Fox, P.T., Eickhoff, S.B., 2012. Modelling neural correlates of working memory: a coordinate-based meta-analysis. Neuroimage 60, 830–846. Royet, J.P., Hudry, J., Zald, D.H., Godinot, D., Gregoire, M.C., Lavenne, F., Costes, N., Holley, A., 2001. Functional neuroanatomy of different olfactory judgments. Neuroimage 13, 506–519. Royet, J.P., Plailly, J., 2004. Lateralization of olfactory processes. Chem. Senses 29, 731–745. Royet, J.P., Plailly, J., Delon-Martin, C., Kareken, D.A., Segebarth, C., 2003. fMRI of emotional responses to odors: influence of hedonic valence and judgment, handedness, and gender. Neuroimage 20, 713–728. Royet, J.P., Zald, D., Versace, R., Costes, N., Lavenne, F., Koenig, O., et al., 2000. Emotional responses to pleasant and unpleasant olfactory, visual, and auditory stimuli: a positron emission tomography study. J. Neurosci. 20 (20), 7752–7759. Savic, I., 2002. Imaging of brain activation by odorants in humans. Curr. Opin. Neurobiol. 12, 455–461. Savic, I., Berglund, H., 2004. Passive perception of odors and semantic circuits. Hum. Brain Mapp. 21 (4), 271–278. Savic, I., Berglund, H., Lindstrom, P., 2005. Brain response to putative pheromones in homosexual men. Proc. Natl. Acad. Sci. U. S. A. 102 (20), 7356–7361 (PMCID: 1129091). Savic, I., Gulyas, B., 2000. PET shows that odors are processed both ipsilaterally and contralaterally to the stimulated nostril. Neuroreport 11 (13), 2861–2866. Savic, I., Gulyas, B., Berglund, H., 2002. Odorant differentiated pattern of cerebral activation: comparison of acetone and vanillin. Hum. Brain Mapp. 17 (1), 17–27. Savic, I., Gulyas, B., Larsson, M., Roland, P., 2000. Olfactory functions are mediated by parallel and hierarchical processing. Neuron 26, 735–745. Schacter, D.L., Wig, G.S., Stevens, W.D., 2007. Reductions in cortical activity during priming. Curr. Opin. Neurobiol. 17, 171–176.

Sela, L., Sobel, N., 2010. Human olfaction: a constant state of change-blindness. Exp. Brain Res. 205, 13–29. Seo, H.S., Hummel, T., 2011. Auditory–olfactory integration: congruent or pleasant sounds amplify odor pleasantness. Chem. Senses 36, 301–309. Seubert, J., Kellermann, T., Loughead, J., Boers, F., Brensinger, C., Schneider, F., Habel, U., 2010. Processing of disgusted faces is facilitated by odor primes: a functional MRI study. Neuroimage 53, 746–756. Small, D.M., Jones-Gotman, M., Zatorre, R.J., Petrides, M., Evans, A.C., 1997. Flavor processing: more than the sum of its parts. Neuroreport 8 (18), 3913–3917. Small, D.M., Zald, D.H., Jones-Gotman, M., Zatorre, R.J., Pardo, J.V., Frey, S., Petrides, M., 1999. Human cortical gustatory areas: a review of functional neuroimaging data. Neuroreport 10, 7–14. Sobel, N., Prabhakaran, V., Desmond, J.E., Glover, G.H., Goode, R.L., Sullivan, E.V., Gabrieli, J.D., 1998. Sniffing and smelling: separate subsystems in the human olfactory cortex. Nature 392, 282–286. Sobel, N., Prabhakaran, V., Zhao, Z., Desmond, J.E., Glover, G.H., Sullivan, E.V., Gabrieli, J.D., 2000. Time course of odorant-induced activation in the human primary olfactory cortex. J. Neurophysiol. 83, 537–551. Stankewitz, A., Voit, H.L., Bingel, U., Peschke, C., May, A., 2010. A new trigemino-nociceptive stimulation model for event-related fMRI. Cephalalgia 30 (4), 475–485. Tabert, M.H., Steffener, J., Albers, M.W., Kern, D.W., Michael, M., Tang, H., et al., 2007. Validation and optimization of statistical approaches for modeling odorant-induced fMRI signal changes in olfactory-related brain areas. Neuroimage 34 (4), 1375–1390. Treyer, V., Koch, H., Briner, H.R., Jones, N.S., Buck, A., Simmen, D.B., 2006. Male subjects who could not perceive the pheromone 5a-androst-16-en-3-one, produced similar orbitofrontal changes on PET compared with perceptible phenylethyl alcohol (rose). Rhinology 44 (4), 278–282. Turkeltaub, P.E., Eden, G.F., Jones, K.M., Zeffiro, T.A., 2002. Meta-analysis of the functional neuroanatomy of single-word reading: method and validation. Neuroimage 16, 765–780. Turkeltaub, P.E., Eickhoff, S.B., Laird, A.R., Fox, M., Wiener, M., Fox, P., 2012. Minimizing within-experiment and within-group effects in Activation Likelihood Estimation meta-analyses. Hum. Brain Mapp. 33, 1–13. Veldhuizen, M.G., Albrecht, J., Zelano, C., Boesveldt, S., Breslin, P., Lundstrom, J.N., 2011. Identification of human gustatory cortex by activation likelihood estimation. Hum. Brain Mapp. 32, 2256–2266. Veldhuizen, M.G., Nachtigal, D., Teulings, L., Gitelman, D.R., Small, D.M., 2010. The insular taste cortex contributes to odor quality coding. Front. Hum. Neurosci. 4 (PMCID: 2917218). Veldhuizen, M.G., Small, D.M., 2011. Modality-specific neural effects of selective attention to taste and odor. Chem. Senses 36, 747–760. Wang, R.Y., Yu, S.M., Yang, Y.R., 2005. Treadmill training effects in different age groups following middle cerebral artery occlusion in rats. Gerontology 51 (3), 161–165. Wiesmann, M., Kopietz, R., Albrecht, J., Linn, J., Reime, U., Kara, E., et al., 2006. Eye closure in darkness animates olfactory and gustatory cortical areas. Neuroimage 32 (1), 293–300. Weismann, M., Yousry, I., Heuberger, E., Nolte, A., Ilmberger, J., Kobal, G., et al., 2001. Functional magnetic resonance imaging of human olfaction. Neuroimaging Clin. N. Am. 11 (2), 237–250, viii. Wicker, B., Keysers, C., Plailly, J., Royet, J.P., Gallese, V., Rizzolatti, G., 2003. Both of us disgusted in My insula: the common neural basis of seeing and feeling disgust. Neuron 40, 655–664. Yeshurun, Y., Sobel, N., 2010. An odor is not worth a thousand words: from multidimensional odors to unidimensional odor objects. Annu. Rev. Psychol. 61 (219–241), C211–C215. Yousem, D.M., Maldjian, J.A., Hummel, T., Alsop, D.C., Geckle, R.J., Kraut, M.A., et al., 1999. The effect of age on odor-stimulated functional MR imaging. AJNR Am. J. Neuroradiol. 20 (4), 600–608. Zald, D.H., Pardo, J.V., 1997. Emotion, olfaction, and the human amygdala: amygdala activation during aversive olfactory stimulation. Proc. Natl. Acad. Sci. U. S. A. 94, 4119–4124. Zald, D.H., Pardo, J.V., 2000. Functional neuroimaging of the olfactory system in humans. Int. J. Psychophysiol. 36, 165–181. Zatorre, R.J., Jones-Gotman, M., 2000. Functional imaging of the chemical senses. In: Toga, A.W., Mazziotta, J.C. (Eds.), Brain Mapping: The Applications. Academic Press, San Diego, pp. 403–424. Zatorre, R.J., Jones-Gotman, M., Evans, A.C., Meyer, E., 1992. Functional localization and lateralization of human olfactory cortex. Nature 360, 339–340. Zatorre, R.J., Jones-Gotman, M., Rouby, C., 2000. Neural mechanisms involved in odor pleasantness and intensity judgments. Neuroreport 11 (12), 2711–2716. Zelano, C., Bensafi, M., Porter, J., Mainland, J., Johnson, B., Bremner, E., Telles, C., Khan, R., Sobel, N., 2005. Attentional modulation in human primary olfactory cortex. Nat. Neurosci. 8, 114–120. Zelano, C., Mohanty, A., Gottfried, J.A., 2011. Olfactory predictive codes and stimulus templates in piriform cortex. Neuron 72, 178–187. Zelano, C., Montag, J., Khan, R., Sobel, N., 2009. A specialized odor memory buffer in primary olfactory cortex. PLoS One 4, e4965. Zelano, C., Sobel, N., 2005. Humans as an animal model for systems-level organization of olfaction. Neuron 48, 431–454.