| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Laboratoire Neurosciences Sensorielles, Comportement, et Cognition, Université Claude Bernard Lyon 1 and Centre National de la Recherche Scientifique Unité Mixte de Recherche 5020, Lyon, France; 2Department of Neurobiology, The Weizmann Institute of Science, Rehovot, Israel; and 3Helen Wills Neuroscience Institute, University of California–Berkeley, Berkeley, California
Submitted 28 March 2007; accepted in final form 29 September 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Kosslyn et al. 1995), auditory (Halpern and Zatorre 1999; Zatorre and Halpern 1993), and motor (Jeannerod 1995; Jeannerod and Frak 1999) systems, its existence in olfaction remains controversial (Bensafi and Rouby 2007; Elmes 1998; Herz 2000; Stevenson and Case 2005). The existence of mental imagery for smells is supported by studies showing similarity in the relative contributions of real and imagined odors to the perception of an odor mixture (Cain and Algom 1997) and similarity in perceptual grouping of perceived and imagined smells (Carrasco and Ridout 1993). Additional support comes from the observed improvement of recognition memory (Lyman and McDaniel 1990), taste detection (Djordjevic et al. 2004a), and odor detection (Djordjevic et al. 2004b) when tested concurrently with odor imagery. Furthermore, psychophysical evidence suggests that during odor imagery, humans spontaneously activate the olfactomotor system in an odorant-specific manner: when imagining a pleasant smell, humans spontaneously take a large sniff and when imagining an unpleasant smell they take a smaller sniff. When sniffing behavior is modified by either preventing or encouraging sniffing, olfactory imagery is modulated accordingly (Bensafi et al. 2003a, 2005).
A prevalent theme in mental imagery research is the idea that the neural mechanisms underlying imagery generation are shared with those that process corresponding real sensations (Farah 1989). Psychophysical (Laeng and Teodorescu 2002), neuropsychological (Farah 1988), and imaging (Kosslyn et al. 1999) methods have all demonstrated similarities between perception and mental imagery, to the extent that common neural substrates are implied in both processes (Kosslyn et al. 1995, 1999, 2001). Similarly in olfaction, Djordjevic et al. (2005) used positron emission tomography (PET) to show that mental imagery of odors is associated with increased activation in a number of regions associated with the processing of odors including orbitofrontal cortex, anterior insula, and piriform cortex.
A potential barrier to the interpretation of imaging of imagery results (in all modalities) has to do with task demands and attention. Trying to imagine an odor may drive the olfactomotor system (Bensafi et al. 2003a, 2005), which alone is sufficient to drive activity in olfactory cortex (Sobel et al. 1998, 2000). Furthermore, merely shifting attention to olfaction may further modulate patterns of neural activity throughout the olfactory system (de Araujo et al. 2005; Geisler and Murphy 2000; Sabri et al. 2005; Zelano et al. 2005).
To address this issue, we took advantage of the hedonic-specific patterns of neural activity evident in olfactory cortex (Anderson et al. 2003; Gottfried et al. 2002a; Heining et al. 2003; Rolls et al. 2003; Royet et al. 2003; Wicker et al. 2003). We hypothesized that attention to olfaction and general task demands can be held equal when imagining either pleasant or unpleasant odorants. Thus if olfactory imagery is generated through activation of sensory circuits, then differences in neural activity between imagining pleasant versus unpleasant odors should mimic the differences between smelling the same pleasant versus unpleasant odorants.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sixteen subjects (8 women and 8 men, mean age 26.8 ± 4.38 yr) participated in the experiment after giving informed consent to procedures approved by the UC Berkeley Committee for the Protection of Human Subjects. Participants were screened for abnormal olfaction, history of neurological disease or injury, history of nasal insult (broken nose or surgery), or magnetic resonance (MR) counterindications. Due to technical and motion-related problems, 2 of the 16 subjects were excluded from further analysis.
Odorants and olfactometry
Odorants were delivered by a computer-controlled air-dilution olfactometer described in detail elsewhere (Johnson and Sobel 2007; Johnson et al. 2003). This olfactometer switches between odorant presence and absence in <2 ms, with no nonolfactory cues as to the alteration. The endpoint of the olfactometer is a nasal mask where the odorant is vacuumed away at the same rate at which it is supplied. The mask is coupled to a pneumatotachograph that provides a highly accurate constant real-time measurement of airflow in the nose. The pneumatotachograph signal was processed with a spirometer (AD Instruments, Grand Junction, CO), amplified (PowerLab 4SP, AD Instruments) and digitally recorded at 100 Hz using Chart version 4.1 software (AD Instruments). This measurement was used to validate that subjects were following task instructions (i.e., sniffing at the tone and for its duration, and otherwise orally respirating), and as a potential measure of interest in later analysis (i.e., comparing sniffing parameters across conditions). Sniffs were preprocessed by removing baseline offsets and aligned in time by setting the point where the sniff entered the inspiratory phase as time 0. Sniff inspired volume, sniff duration, and maximum flow rate were calculated for all sniffs. Volume was calculated by the trapezoidal Riemann sum method. Both the volume integration and sniff duration ended at the first data point where the sniff returned to zero flow. To allow cross-subject comparison, each subject's airflow values were divided by the maximum value within that subject, resulting in normalized values reaching a maximum of one. The olfactometer, digitized auditory instruction generator, recording of respiratory data, and the MRI scanner itself, were all linked through one TTL pulse that ensured accurate time-locking of all experimental components. The two odorants that were used were ammonium sulfide (4 x 10–4 vol/vol; Sigma–Aldrich, St. Louis, MO), which smells like rotten eggs, and strawberry oil (10–2 vol/vol; Lhasa Karnak, Berkeley, CA), which smells like strawberries.
Experimental design
We used an event-related design consisting of five conditions distributed across five functional scans. The five conditions were a no-odorant event, consisting only of clean air ("Clean" condition) and a 2 x 2 design in which Task ("sensory" vs. "imagery") was crossed with Hedonic Value ("Pleasant" vs. "Unpleasant"). The pleasant odor was "Strawberry" (strawberry oil) and the unpleasant odor was "Rotten Eggs" (ammonium sulfide). Thus the five conditions were Clean (no odorant), Sensory-Pleasant ("strawberry"), Sensory-Unpleasant ("rotten eggs"), Imagery-Pleasant ("strawberry"), and Imagery-Unpleasant ("rotten eggs"). To avoid the risk of any residual odors in the masks during imagery scans that would contaminate our results, we divided the five scans into two types: "sensory" scans (three scans) and "imagery" scans (two scans) that were run in blocks. Subjects wore new, unused masks during each of the blocks. Three of the trial types (the Clean and the two Sensory conditions) were pseudorandomized across the "sensory" scans block and the two remaining trial types (the Imagery conditions) were pseudorandomized across the "imagery" scans block. The temporal order of the "sensory" and the "imagery" blocks was counterbalanced across subjects. Conditions were randomized with the constraint that each scan should contain equal numbers of trials of each condition.
During each trial subjects received task instructions through earphones generated by digitally recorded voice (Fig. 1). During the "sensory" scans, each trial began with an auditory primer "Please prepare to sniff," followed by a countdown ("3, 2, 1") and a tone. Subjects were asked to sniff for the duration of the tone (1.67 s) during which odorants were diffused into the mask subjects wore. The odorants were diffused for 3 s. During the "imagery" scans, each trial began with an auditory primer "Please prepare to sniff," followed by a countdown ("3", "2"), the name of the stimulus to be imagined ("Rotten Eggs" or "Strawberry") and a tone. Here also, subjects were asked to sniff for the duration of the tone (1.67 s). There were 22 trials per condition type and the stimulus onset asynchrony (SOA) was 30 s.
|
Imaging parameters
All the raw structural and functional imaging data, along with corresponding files describing the temporal structure of the data for each subject, will be made publicly available at http://www.weizmann.ac.il/neurobiology/worg/materials.html.
The experiment was conducted on a 4-T Varian Inova magnet. A custom-built full-head receive coil was used for signal reception. A T2*-sensitive echo planer sequence was used with parameters of TR = 500 ms, TE = 28 ms, and flip angle = 20°. The spatial resolution was set by a 64 x 64 voxel matrix covering a 19.2 x 19.2-cm field of view, resulting in a functional in-plane resolution of 3 mm and through-plane resolution of 3.5 mm. Two interleaves were collected for each frame, with total acquisition time of 1,000 ms per frame. The interleaves were interpolated during image reconstruction, resulting in an effective temporal resolution of 500 ms per frame. Fifteen frames were collected before task onset at the beginning of each scan to achieve dynamic equilibrium. Eight 3.5-mm-thick slices were acquired at an oblique plane traversing from the frontal pole to the temporal pole (typically 30° clockwise to the anterior commissure–posterior commissure plane) with a 0.5-mm gap between slices. This slice orientation was chosen to cover the entire primary olfactory cortex while minimizing partial voluming artifacts (Sobel et al. 1997). To prevent head motion, a custom-formed bite bar was fit to the individual dental impression of each subject. This bite bar was also fit with a pyrolitic graphite implant aimed at reducing ventral temporal susceptibility artifacts (Wilson et al. 2002). Full brain T1-weighted flow compensated spin-warp anatomy images (TR = 500 ms, minimum TE, isotropic 0.875-mm voxels) were acquired as a substrate on which to overlay functional data.
Image analysis
Data were analyzed using mrVista software (http://white.stanford.edu/software/). This analysis package has been extensively developed and used to probe sensory processing with fMRI (Boynton et al. 1996; Engel et al. 1994), and has been used by us for olfaction (Porter et al. 2005; Zelano et al. 2005). First, in-plane anatomical images were aligned to the high-resolution anatomical volume of each subject's brain so that all MR images (across multiple scanning sessions) from a given subject were coregistered to an accuracy of 1 mm (Nestares and Heeger 2000). The fMRI images from each scan were tested for head movements. We then defined regions of interest (ROIs) anatomically for each subject.
Defining ROIs
We combined a structural and functional restriction to define ROIs. We first outlined the expected subdivisions in piriform cortex (PIR; frontal portion, "PirF"; temporal portion, "PirT"; and olfactory tubercule, "Tu") based on an atlas that is particularly detailed in this respect (Mai et al. 1997). This delineation was performed before any further analysis and in the absence of functional results (see Fig. 4A). To eliminate voxels within the anatomical boundary that are functionally unresponsive (e.g., white matter, cerebrospinal fluid, etc.) we further functionally restricted this region to only those voxels that responded hemodynamically to the control "Clean" condition (see next section for details). A similar principle was applied to ROIs created in the insula (anterior and posterior portions) and orbitofrontal cortex (anterior, lateral, medial, and posterior portions). In this case, demarcation of the expected subdivisions of insula and orbitofrontal cortex was based on the Duvernoy atlas (Duvernoy 1991).
|
|
|
The average time series for each trial within each ROI for each subject was calculated using three steps:
First, for each subject we calculated an activation mask to filter out voxels for which we had no signal. We produced images of the average response across all time points at each voxel. Because voxels in gray and white matter have a significantly different mean response than that of voxels in bone or air, we were able to filter voxels based on their mean response to include only voxels for which we had signal. This eliminated voxels that were in regions of high susceptibility, particularly near the ventral frontal and temporal surface of cortex.
Second, for each subject we produced a noise mask similar to the first that calculated the SD of the response at each voxel. This mask also discriminated between regions of high susceptibility and brain tissue, and further excluded voxels with high noise, such as voxels on large blood vessels.
Third, we restricted each subjects' anatomical ROI to those voxels that responded to the control "Clean" trials. This was calculated by correlating the response at each voxel after an event with a hemodynamic response function (HRF) derived from the average responses of subjects in a separate study (Anderson et al. 2003). This HRF is derived from a large number of subjects and has been used as a standardized HRF in previous studies (Anderson et al. 2003; Porter et al. 2005; Zelano et al. 2005). Because it is data derived, it is likely to be a better representation of the actual HRF observed in our data than the use of an idealized HRF based on a gamma or Poisson density function as is sometimes done (Friston et al. 1994). By calculating the correlation of each voxel to an expected hemodynamic response function, and the statistical significance of this correlation, we were able to produce a statistical parametric map of the responsiveness of each voxel to the functional condition. To restrict for "Clean" responsive voxels, we excluded all voxels whose correlation to the HRF had a statistical significance value higher than P = 0.01. Supplementary Fig. 81 and Table 1 show example ROIs and the effects of this restriction procedure.
|
10 s before the time of odorant onset, for the whole time series (the baseline response before sniff) so that all time series had comparable baselines. Finally, we calculated the average of all trials of a certain condition to derive an average time series for that ROI. fMRI response was defined as the area under the hemodynamic response curves in the window 2 to 8 s after odorant presentation. We then calculated the average across all trials of a given condition (in the same way described for the time courses) to get an average area under the curve for each ROI and each condition of interest. Thus, for each event and ROI we had a single number, the area under the time series curve, as a measure of the magnitude of fMRI response. To account for small variations in sniffing from trial to trial, this value was adjusted for the sniff parameters of sniff volume, maximum flow rate, and sniff duration, which were modeled out of the fMRI response before the ANOVA by linear regression on a subject-by-subject basis. For each subject in each ROI, we regressed fMRI response against these three sniff parameters and took the residual values as our adjusted fMRI response. Thus, the adjusted fMRI response data are that part of the fMRI response that cannot be explained by the sniff data. These adjusted numbers are used in subsequent analyses. These adjusted integrals are shown as the bar graphs in Figs. 4, 5, and 7. For each ROI, we performed a three-way ANOVA with subjects as the blocking variable and Task ("sensory" or "imagery") and Hedonic Value ("pleasant" or "unpleasant") as the grouping variables. We looked for main effects as well as second-order interactions between these factors.
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
As expected, "Strawberry" was rated as more pleasant than "Rotten Eggs" during both odor imagery [F(1,13) = 55.513, P < 0.0001) (Strawberry: mean = 6.10, SE = 0.28; Rotten Eggs: mean = 3.03, SE = 0.32) and real sensation [F(1,13) = 101.409, P < 0.0001] (Strawberry: mean = 6.30, SE = 0.17; Rotten Eggs: mean = 2.60, SE = 0.27). Stimuli were not different in odor intensity [F(1,13) = 0.002, P > 0.05] (Strawberry: mean = 5.21, SE = 0.29; Rotten Eggs: mean = 5.19, SE = 0.27) and imagined odor vividness [F(1,13) = 1.644, P > 0.05] (Strawberry: mean = 4.21, SE = 0.40; Rotten Eggs: m = 4.60, SE = 0.33) (Fig. 2). It is possible that grouping the imagery scans and the perception scans may have had an impact on the experience of vividness depending on the order in which the two blocks were presented. To address this concern, we compared the vividness ratings of both groups of subjects (those who had the imagery scans first vs. those who had the sensory scans first). The statistical analysis did not reveal any significant difference between groups [F(1,12) = 3.083, P > 0.05], suggesting that the order of scans did not affect the experience of vividness in perception versus imagery across these orderings.
Effects of airflow
Figure 3A illustrates the average sniff trace during the five experimental conditions. A significant main effect of conditions was observed for sniff volume [F(4,1535) = 7.873, P < 0.0001], reflecting a 14.16% increase in sniff volume for the pleasant odor relative to the unpleasant odor following real odor sensation [t(614) = 3.287, P < 0.0011] and a 10.94% increase in sniff volume for the pleasant odor relative to the unpleasant odor following odor imagery [t(614) = 2.392, P < 0.0171] (Fig. 3B). Significant main effects of these conditions were also observed for maximum flow rate [F(4,1535) = 22.782, P < 0.0001] and sniff duration [F(4,1535) = 17.250, P < 0.0001], reflecting a 9.95% increase in maximum flow rate for the pleasant odor relative to the unpleasant odor following odor imagery [t(614) = 3.962, P < 0.0171] (Fig. 3C) and a 7.30% increase in sniff duration for the pleasant odor relative to the unpleasant odor following real odor sensation [t(614) = 4.126, P < 0.0011] (Fig. 3D). Airflow parameters (sniff volume, maximum flow rate, and sniff duration) were collected for every trial and were used to adjust the fMRI data on a subject-by-subject basis.
Activity induced by the sensory and the imagery conditions in subdivisions of piriform cortex
As described in METHODS, we used previously defined landmarks to divide PIR into the olfactory tubercule, frontal PIR and temporal PIR (Fig. 4A). Distinct response profiles were observed across these subregions (Fig. 4, B and C). In all analyses, the fMRI response measurements were adjusted with covariates for the sniff data, so that the differences in sniff volume, sniff duration, and maximum flow rate were factored out.
Consistent with previous results, there was significant functional heterogeneity across the three subregions of piriform cortex (Gottfried et al. 2002a; Porter et al. 2005; Zelano et al. 2005). Left and right regions were analyzed separately for all ROIs. The only region that showed a significantly different response for real and imagined stimuli was PirF. In both right [F(1,13) = 27.22, P < 0.05] and left [F(1,13) = 32.75, P < 0.05] PirF, the effect of Hedonic Value was accompanied by an effect of Task with greater activity in the sensory than that in the imagery trials. A significant effect of Hedonic Value was observed in left PirF [F(1,13) = 15.51, P < 0.05] with greater activity, in both sensation and mental imagery induced in the unpleasant trials (Fig. 4D). Such effects were not observed in left (Fig. 5A) and right (Fig. 5B) PirT and left (Fig. 5C) and right (Fig. 5D) olfactory tubercule (P > 0.05 in all cases). In the right PirF, a significant Hedonic Value x Task interaction reflected that the effect of Hedonic Value was mostly carried by the sensory condition [F(1,13) = 5.27, P < 0.05] (Fig. 4E).
In sum, the unpleasant perceived odor induced greater activity than the pleasant perceived odor in both left and right frontal PIR, whereas the unpleasant imagined odor induced greater activity than the pleasant imagined odor in the left frontal PIR.
Activity induced by the sensory and the imagery conditions in additional regions
Next, we examined the responses of a number of odor responsive regions outside piriform cortex. We created ROIs within subregions of orbitofrontal cortex (OFC) and insula (INS) (Fig. 6, A and B) based on anatomical landmarks, and measured responses separately within each ROI. Left and right regions were analyzed separately. In all analyses, the fMRI response measurements were adjusted with covariates for the sniff data, so that the differences in sniff volume and maximum flow rate were factored out.
Insula ROIs
Two insula regions were defined in each hemisphere: a posterior region (pINS) and an anterior region (aINS). Parts of these regions in a single slice are shown in Fig. 6B. Average responses in these ROIs showed effects of both Task and Hedonic Value (Fig. 7, A–D).
Effects of Task were observed in both right aINS [F(1,13) = 8.84, P < 0.05] and right pINS [F(1,13) = 11.57, P < 0.05], with greater activity during Sensory than during Imagery conditions, but not in either region in the left hemisphere. Conversely, effects of Hedonic Value were observed in both left aINS [F(1,13) = 16.09, P < 0.05] and left pINS [F(1,13) = 11.25, P < 0.05] but not in the right hemisphere regions. In both regions, responses were greater for unpleasant than for pleasant stimuli for both real and imagined odors.
A general characterization of the insula response suggests a left versus right hemisphere dissociation in which the left regions reflected odor content of an odor percept, regardless of its source (real vs. imagined) and the right hemisphere regions were responsive more to real odors.
Orbitofrontal ROIs
Four orbitofrontal regions were defined in each hemisphere: an anterior region (aOFC), a lateral region (lOFC), a medial region (mOFC), and a posterior region (pOFC). Parts of these regions in a single slice are shown in Fig. 6A. Average responses in these ROIs showed effects of both Task and Hedonic Value (Fig. 6, E–L).
Effects of task were observed in all OFC regions in both hemispheres: left aOFC [F(1,13) = 18.42, P < 0.05], right aOFC [F(1,13) = 21.36, P < 0.05], left lOFC [F(1,13) = 7.43, P < 0.05], right lOFC [F(1,13) = 15.32, P < 0.05], left mOFC [F(1,13) = 26.81, P < 0.05], right mOFC [F(1,13) = 62.88, P < 0.05], left pOFC [F(1,13) = 7.97, P < 0.05], and right pOFC [F(1,13) = 15.5, P < 0.05]. In all instances the response to real odors was greater than that for imagined odors.
Effects of Hedonic Value were observed in left [t(13) = 2.19, P < 0.05] and right [t(13) = 2.52, P < 0.05] pOFC and in left mOFC [t(13) = 2.45, P < 0.05] for the real but not the imagined odors.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
, suggests that some form of odor imagery exists and that it relies in part on neural mechanisms common with real olfactory perception.
This observed greater neural activity in primary sensory areas, during both perception and imagery of the unpleasant odor, may reflect the primacy of hedonic judgments—and particularly judgments of negative hedonic value—in olfactory processing (Rouby et al. 2002). Hedonic categorization at an early processing stage (i.e., in frontal piriform cortex) may reflect the survival value in deciding very quickly whether the environmental stimulus is noxious or dangerous. This view is consistent with the fact that humans react more rapidly to aversive and dangerous odors than to appetitive odors (Bensafi et al. 2003b; Jacob et al. 2003) and that unpleasant stimuli have a greater impact on everyday life and are usually more arousing than pleasant stimuli (Lang et al. 1993).
The heterogeneity of response within piriform cortex during odor perception and odor imagery seen here was consistent with previous results of heterogeneity during odor imagery (Djordjevic et al. 2005) and during hedonic processing of odorants (Gottfried et al. 2002a; Zelano et al. 2007). Together, these findings support the notion of generally higher-order top-down processing in frontal piriform, and more bottom-up stimulus-driven and attention-independent activity in temporal piriform (Gottfried et al. 2002a; Zelano and Sobel 2005; Zelano et al. 2005).
Beyond piriform cortex we observed activity in a number of regions, notably OFC and insula. Real sensation of odorants induced more activity than odor imagery in all regions of the OFC in both hemispheres. In posterior OFC, unpleasant odorants induced more activity than pleasant odorants. This is consistent with one previous study (Gottfried et al. 2002a), but may be inconsistent with several others that did not find hedonic-value–dependent activity in posterior OFC (Anderson et al. 2003; Gottfried et al. 2002b; Rolls et al. 2003). However, considering that OFC activity is modulated by a considerable variety of factors, such as sensory-specific satiety (O'Doherty et al. 2000), motivation (Small et al. 2001), and lexical knowledge (de Araujo et al. 2005), the differences between our results and those of previous studies may be the result of methodological variables or differences in study design. For example, whereas subjects in the studies reported by Anderson et al. (2003) and Rolls et al. (2003) performed respectively a detection task and hedonic and intensity judgments during scanning, in our study subjects did not perform any task after odorant onset.
Whereas OFC activity was driven mainly by perception and not imagery, activity in the insula was nearly identical in its extent and power under both conditions. Furthermore, the left insular activity in this study was strongly valence dependent, being greater during unpleasant real and imagined odors. This hedonic-specific pattern is consistent with previous findings of increased left insular activity during identification of unpleasant odors (Royet et al. 2003), during unpleasant emotions evoked by odors, and during perception of the same unpleasant emotion in others (Wicker et al. 2003). It is also consistent with electrical stimulation results that report that stimulation of the anterior insula through implanted depth electrodes produced unpleasant sensations in the throat spreading up to the mouth, lips, and nose (Krolak-Salmon et al. 2003). Although when taken together these results highlight an increased insular response related to unpleasant stimuli, further work is necessary to determine whether this preferential insular response to unpleasant stimuli reflects valence alone or whether it is confounded by related entities such as valence-independent arousal, which contribute to the general insular role in flavor processing (Small et al. 2001).
Another general feature of our data was that imagery-induced brain activity was strongly lateralized to the left in piriform cortex. This is consistent with previous results in olfaction (Djordjevic et al. 2005) and resembles similar lateralization observed during visual imagery (Sack et al. 2005). The relatively greater response in these left lateralized regions to unpleasant than pleasant odors seen in our study is in line with emotion theories that suggest a more generalized left hemisphere dominance for unpleasant stimuli (Royet et al. 2003) and support the notion of an overall leftward lateralization in imagery generation (D'Esposito et al. 1997). It is tempting to link the lateralization in imagery-induced activity to lateralization in language mechanisms. The role of language in olfactory cortex activity was recently studied by Gonzalez et al. (2006), who observed that reading words with strong olfactory associations activates olfactory regions. Our study, however, cannot answer the question of the role of linguistic labels on piriform cortex activity in that all stimuli were easily namable.
Although our results present evidence of olfactory imagery in primary olfactory cortex, it is possible that additional brain regions outside of our image acquisition were also involved. We could have opted to scan the entire brain, at the cost of significantly increased susceptibility artifacts in the ventral portion of the brain that contains primary olfactory cortex. Rather, we elected to optimize our scanning parameters to obtain good signal in the classic olfactory areas, at a cost of scanning only a limited portion of brain. Thus ours is not a comprehensive study of brain mechanisms involved in odor imagery, but is a targeted investigation of olfactory cortex. What role additional regions such as parietal and more dorsal regions may play during olfactory imagery is a topic for further investigation.
A second caveat about our results is the possible effect of other factors that may have been fortuitously correlated with hedonic value in our study. Any pair of odors will differ along a large number of dimensions, such as hedonic valence, familiarity, edibility, linguistic associations, and so forth, and whereas we designed our study with the intention of comparing hedonic value, it is possible that it was some other difference between the pair we used that created the pattern of differences we observed. Whereas such a concern limits interpretation of our results regarding the larger issue of odor hedonics, it does not undermine our main result, which is that patterns of behavioral and neural activity during imagery of odors paralleled that during perception.
Although subjects were asked to maintain constant sniffs across conditions, sniffs during imagery and sensation of the pleasant odor were slightly but significantly larger than those for the unpleasant odor. That this behavioral difference persisted despite instructions to the contrary points to the robustness of this behavioral mechanism (Bensafi et al. 2003a, 2005). Although the differences in sniffing were small, we nevertheless entered these differences as regressors into our brain imaging analysis, to prevent any possible related confounds. An additional finding here was that although increased sniff volume for the pleasant odor was found for both odor sensation and imagery conditions, increased sniff duration for the pleasant odor was found only during sensory trials. Moreover, increased sniff flow rate for the pleasant odor was found only during imagery trials. One explanation of this may be that during the sensory trials, participants did not know until after they encountered the odor whether it was pleasant or unpleasant. Thus they increased sniff volume (for the pleasant odor) by increasing sniff duration for pleasant rather than unpleasant odors. In contrast, during the imagery trials participants knew beforehand the hedonic value of the odor they had to imagine, and thus increased their sniff volume (for the pleasant odor) at the beginning of the trial, by increasing their inspired flow rate.
Another issue worth considering in interpreting our results is a potential bias introduced in our procedure for selecting voxels of interest to analyze. In our procedure, we analyzed voxels responsive to the "clean" no-odor condition for their response to the pleasant and unpleasant stimuli. This strategy could be a potential problem if there are voxels that respond to odors that do not respond during sniffing. This question has been the subject of several studies (Sobel et al. 1998; Zelano et al. 2005) that suggest that although active sniffing (sniffing in which a potential odor is sought and processed) leads to neural activity in olfactory cortex (including primary olfactory cortex)—even when no odor is actually present—passive inspiration, where the subject has no interest or anticipation of a potential odor, leads to significantly less activity in these regions. Because the sniffing in the sensory conditions of our study was active in this sense (subjects had no cue as to which odor was to be presented on any given trial), we would expect activity in olfactory cortex even during the "clean" trial when no odor was presented. Indeed, we found significant activity in response to sniffing alone (no odorant) in all olfactory brain areas we examined; had we not, our procedure would have led to the selection of a random and extremely small set of voxels, which would have shown a random pattern of response in the other conditions. The fact that this was not the case, that in fact the pattern of response (in terms of its hedonic-value response) was consistent with that in previous reports (Royet et al. 2003; Wicker et al. 2003) and was paralleled during imagery suggests that the voxel selection procedure based on clean trials was meaningful.
A final question concerns how our findings relate to competing theories of mental imagery. The two main theories are the "perceptual anticipation theory," in which mental images may arise when one anticipates an object or scene so strongly that a depictive representation of the stimulus is created in early sensory cortex (Kosslyn and Thompson 2003) and the "propositional theory," in which mental images are not images, but descriptions (Pylyshyn 2003). Under the propositional theory, one might argue that propositional representations are not expected to activate the earliest stage of olfactory processing because they are essentially linguistic. Our finding of a parallel hedonic-value–dependent pattern of neural activity at the earliest stage of processing in both olfactory sensation and olfactory mental imagery is consistent with the perceptual anticipation theory. Nevertheless, because linguistic odor labels may modulate patterns of activity in human primary olfactory cortex (Gonzalez et al. 2006), we cannot claim that the current findings weigh definitively against the propositional theory. Ultimately, absent detailed mechanistic models of sensory processing and imagery, these competing theories are irresolvable. That said, our findings lend support to the view that olfactory imagery exists and that generating imagery of sensory events involves activation of neural substrates common to real perception. This has been well documented in vision (Kosslyn and Thompson 2003), and its current iteration in olfaction suggests that this is a general principle in brain function.
|
FOOTNOTES |
|---|
1 The online version of this article contains supplemental data. 
Address for reprint requests and other correspondence: M. Bensafi, Université Claude Bernard Lyon 1, Laboratoire Neurosciences Sensorielles, Comportement, et Cognition, UMR 5020 CNRS, 50 avenue Tony Garnier, 69366 Lyon cedex 07, France (E-mail: bensafi{at}olfac.univ-lyon1.fr)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bensafi M, Porter J, Pouliot S, Mainland J, Johnson B, Zelano C, Young N, Bremner E, Aframian D, Khan R, Sobel N. Olfactomotor activity during imagery mimics that during perception. Nat Neurosci 6: 1142–1144, 2003a.[CrossRef][Web of Science][Medline]
Bensafi M, Pouliot S, Sobel N. Odorant-specific patterns of sniffing during imagery distinguish "bad" and "good" olfactory imagers. Chem Senses 30: 521–529, 2005.
Bensafi M, Rouby C. Individual differences in odor imaging ability reflect differences in olfactory and emotional perception. Chem Senses 32: 237–244, 2007.
Bensafi M, Rouby C, Farget V, Bertrand B, Vigouroux M, Holley A. Perceptual, affective, and cognitive judgments of odors: pleasantness and handedness effects. Brain Cogn 51: 270–275, 2003b.[CrossRef][Web of Science][Medline]
Boynton GM, Engel SA, Glover GH, Heeger DJ. Linear systems analysis of functional magnetic resonance imaging in human V1. J Neurosci 16: 4207–4221, 1996.
Cain WS, Algom D. Perceptual and mental mixtures in odor and in taste: are there similarities and differences between experiments or between modalities? Reply to Schifferstein (1997). J Exp Psychol Hum Percept Perform 23: 1588–1593, 1997.[CrossRef][Web of Science][Medline]
Carrasco M, Ridout JB. Olfactory perception and olfactory imagery: a multidimensional analysis. J Exp Psychol Hum Percept Perform 19: 287–301, 1993.[CrossRef][Web of Science][Medline]
de Araujo IE, Rolls ET, Velazco MI, Margot C, Cayeux I. Cognitive modulation of olfactory processing. Neuron 46: 671–679, 2005.[CrossRef][Web of Science][Medline]
D'Esposito M, Detre JA, Aguirre GK, Stallcup M, Alsop DC, Tippet LJ, Farah MJ. A functional MRI study of mental image generation. Neuropsychologia 35: 725–730, 1997.[CrossRef][Web of Science][Medline]
Djordjevic J, Zatorre RJ, Jones-Gotman M. Effects of perceived and imagined odors on taste detection. Chem Senses 29: 199–208, 2004a.
Djordjevic J, Zatorre RJ, Petrides M, Boyle JA, Jones-Gotman M. Functional neuroimaging of odor imagery. Neuroimage 24: 791–801, 2005.[CrossRef][Web of Science][Medline]
Djordjevic J, Zatorre RJ, Petrides M, Jones-Gotman M. The mind's nose: effects of odor and visual imagery on odor detection. Psychol Sci 15: 143–148, 2004b.[CrossRef][Web of Science][Medline]
Duvernoy H. The Human Brain. New York: Springer-Verlag, 1991.
Elmes DG. Is there an inner nose? Chem Senses 23: 443–445, 1998.[Abstract]
Engel SA, Rumelhart DE, Wandell BA, Lee AT, Glover GH, Chichilnisky EJ, Shadlen MN. fMRI of human visual cortex (Correspondence). Nature 369: 525, 1994.[CrossRef][Medline]
Farah MJ. Is visual imagery really visual? Overlooked evidence from neuropsychology. Psychol Rev 95: 307–317, 1988.[CrossRef][Web of Science][Medline]
Farah MJ. The neural basis of mental imagery. Trends Neurosci 12: 395–399, 1989.[CrossRef][Web of Science][Medline]
Fletcher PC, Frith CD, Baker SC, Shallice T, Frackowiak RS, Dolan RJ. The mind's eye—precuneus activation in memory-related imagery. Neuroimage 2: 195–200, 1995.[CrossRef][Web of Science][Medline]
Friston KJ, Jezzard P, Turner R. Analysis of functional MRI time-series. Hum Brain Mapp 1: 153–171, 1994.[CrossRef]
Geisler MW, Murphy C. Event-related brain potentials to attended and ignored olfactory and trigeminal stimuli. Int J Psychophysiol 37: 309–315, 2000.[CrossRef][Web of Science][Medline]
Gonzalez J, Barros-Loscertales A, Pulvermuller F, Meseguer V, Sanjuan A, Belloch V, Avila C. Reading cinnamon activates olfactory brain regions. Neuroimage 32: 906–912, 2006.[CrossRef][Web of Science][Medline]
Gottfried JA, Deichmann R, Winston JS, Dolan RJ. Functional heterogeneity in human olfactory cortex : an event-related functional magnetic resonance imaging study. J Neurosci 22: 10819–10828, 2002a.
Gottfried JA, O'Doherty J, Dolan RJ. Appetitive and aversive olfactory learning in humans studied using event-related functional magnetic resonance imaging. J Neurosci 22: 10829–10837, 2002b.
Halpern AR, Zatorre RJ. When that tune runs through your head: a PET investigation of auditory imagery for familiar melodies. Cereb Cortex 9: 697–704, 1999.
Heining M, Young AW, Ioannou G, Andrew CM, Brammer MJ, Gray JA, Phillips ML. Disgusting smells activate human anterior insula and ventral striatum. Ann NY Acad Sci 1000: 380–384, 2003.[CrossRef][Web of Science][Medline]
Herz RS. Verbal coding in olfactory versus nonolfactory cognition. Mem Cogn 28: 957–964, 2000.[Web of Science][Medline]
Jacob TJ, Fraser C, Wang L, Walker V, O'Connor S. Psychophysical evaluation of responses to pleasant and mal-odour stimulation in human subjects; adaptation, dose response and gender differences. Int J Psychophysiol 48: 67–80, 2003.[CrossRef][Web of Science][Medline]
Jeannerod M. Mental imagery in the motor context. Neuropsychologia 33: 1419–1432, 1995.[CrossRef][Web of Science][Medline]
Jeannerod M, Frak V. Mental imaging of motor activity in humans. Curr Opin Neurobiol 9: 735–739, 1999.[CrossRef][Web of Science][Medline]
Johnson BN, Mainland JD, Sobel N. Rapid olfactory processing implicates subcortical control of an olfactomotor system. J Neurophysiol 90: 1084–1094, 2003.
Johnson BN, Sobel N. Methods for building an olfactometer with known concentration outcomes. J Neurosci Methods 160: 231–245, 2007.[CrossRef][Web of Science][Medline]
Kosslyn SM, Behrmann M, Jeannerod M. The cognitive neuroscience of mental imagery. Neuropsychologia 33: 1335–1344, 1995.[CrossRef][Web of Science][Medline]
Kosslyn SM, Ganis G, Thompson WL. Neural foundations of imagery. Nat Rev Neurosci 2: 635–642, 2001.[CrossRef][Web of Science][Medline]
Kosslyn SM, Pascual-Leone A, Felician O, Camposano S, Keenan JP, Thompson WL, Ganis G, Sukel KE, Alpert NM. The role of area 17 in visual imagery: convergent evidence from PET and rTMS. Science 284: 167–170, 1999.
Kosslyn SM, Thompson WL. When is early visual cortex activated during visual mental imagery? Psychol Bull 129: 723–746, 2003.[CrossRef][Web of Science][Medline]
Krolak-Salmon P, Henaff MA, Isnard J, Tallon-Baudry C, Guenot M, Vighetto A, Bertrand O, Mauguiere F. An attention modulated response to disgust in human ventral anterior insula. Ann Neurol 53: 446–453, 2003.[CrossRef][Web of Science][Medline]
Laeng B, Teodorescu DS. Eye scanpaths during visual imagery reenact those of perception of the same visual scene. Cogn Sci 26: 207–231, 2002.[CrossRef]
Lang PJ, Greenwald MK, Bradley MM, Hamm AO. Looking at pictures: affective, facial, visceral, and behavioral reactions. Psychophysiology 30: 261–273, 1993.[Web of Science][Medline]
Lyman BJ, McDaniel MA. Memory for odors and odor names: modalities of elaboration and imagery. J Exp Psychol Learn Mem Cogn 16: 656–664, 1990.[CrossRef][Web of Science]
Mai J, Assheurer J, Paxinos G. Atlas of the Human Brain. San Diego, CA: Academic Press, 1997.
Nestares O, Heeger DJ. Robust multiresolution alignment of MRI brain volumes. Magn Reson Med 43: 705–715, 2000.[CrossRef][Web of Science][Medline]
O'Doherty J, Rolls ET, Bowtell R, McGlone F, Kobal G, Renner B, Ahne G. Sensory-specific satiety-related olfactory activation of the human orbitofrontal cortex. Neuroreport 11: 399–403, 2000.[Web of Science][Medline]
Porter J, Anand T, Johnson B, Khan RM, Sobel N. Brain mechanisms for extracting spatial information from smell. Neuron 47: 581–592, 2005.[CrossRef][Web of Science][Medline]
Pylyshyn Z. Return of the mental image: are there really pictures in the brain? Trends Cogn Sci 7: 113–118, 2003.[CrossRef][Web of Science][Medline]
Rolls ET, Kringelbach ML, de Araujo IE. Different representations of pleasant and unpleasant odours in the human brain. Eur J Neurosci 18: 695–703, 2003.[CrossRef][Web of Science][Medline]
Rouby C, Schaal B, Dubois D, Gervais R, Holley A. Olfaction, Taste and Cognition. New York: Cambridge Univ. Press, 2002.
Royet JP, Plailly J, Delon-Martin C, Kareken DA, Segebarth C. fMRI of emotional responses to odors: influence of hedonic valence and judgment, handedness, and gender. Neuroimage 20: 713–728, 2003.[CrossRef][Web of Science][Medline]
Sabri M, Radnovich AJ, Li TQ, Kareken DA. Neural correlates of olfactory change detection. Neuroimage 25: 969–974, 2005.[CrossRef][Web of Science][Medline]
Sack AT, Camprodon JA, Pascual-Leone A, Goebel R. The dynamics of interhemispheric compensatory processes in mental imagery. Science 308: 702–704, 2005.
Small DM, Zatorre RJ, Dagher A, Evans AC, Jones-Gotman M. Changes in brain activity related to eating chocolate: from pleasure to aversion. Brain 124: 1720–1733, 2001.
Sobel N, Prabhakaran V, Desmond JE, Glover GH, Goode RL, Sullivan EV, Gabrieli JD. Sniffing and smelling: separate subsystems in the human olfactory cortex. Nature 392: 282–286, 1998.[CrossRef][Medline]
Sobel N, Prabhakaran V, Desmond JE, Glover GH, Sullivan EV, Gabrieli JD. A method for functional magnetic resonance imaging of olfaction. J Neurosci Methods 78: 115–123, 1997.[CrossRef][Web of Science][Medline]
Sobel N, Prabhakaran V, Zhao Z, Desmond JE, Glover GH, Sullivan EV, Gabrieli JD. Time course of odorant-induced activation in the human primary olfactory cortex. J Neurophysiol 83: 537–551, 2000.
Stevenson RJ, Case TI. Olfactory imagery: a review. Psychon Bull Rev 12: 244–264, 2005.[Web of Science][Medline]
Wicker B, Keysers C, Plailly J, Royet JP, Gallese V, Rizzolatti G. Both of us disgusted in My insula: the common neural basis of seeing and feeling disgust. Neuron 40: 655–664, 2003.[CrossRef][Web of Science][Medline]
Wilson JL, Jenkinson M, Jezzard P. Optimization of static field homogeneity in human brain using diamagnetic passive shims. Magn Reson Med 48: 906–914, 2002.[CrossRef][Web of Science][Medline]
Zatorre RJ, Halpern AR. Effect of unilateral temporal-lobe excision on perception and imagery of songs. Neuropsychologia 31: 221–232, 1993.[CrossRef][Web of Science][Medline]
Zelano C, Bensafi M, Porter J, Mainland J, Johnson B, Bremner E, Telles C, Khan R, Sobel N. Attentional modulation in human primary olfactory cortex. Nat Neurosci 8: 114–120, 2005.[CrossRef][Web of Science][Medline]
Zelano C, Montag J, Johnson B, Khan R, Sobel N. Dissociated representations of irritation and valence in human primary olfactory cortex. J Neurophysiol 97: 1969–1976, 2007.
Zelano C, Sobel N. Humans as an animal model for systems-level organization of olfaction. Neuron 48: 431–454, 2005.[CrossRef][Web of Science][Medline]
This article has been cited by other articles:
|
|
 |
|
  A.M. Kleemann, R. Kopietz, J. Albrecht, V. Schopf, O. Pollatos, T. Schreder, J. May, J. Linn, H. Bruckmann, and M. Wiesmann Investigation of Breathing Parameters during Odor Perception and Olfactory Imagery Chem Senses, January 1, 2009; 34(1): 1 - 9. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Rinck, C. Rouby, and M. Bensafi Which format for odor images? Chem Senses, January 1, 2009; 34(1): 11 - 13. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
