Involvement of right piriform cortex in olfactory familiarity judgments Jane Plailly, a, * Moustafa Bensafi, a Mathilde Pachot-Clouard, b Chantal Delon-Martin, b David A. Kareken, c Catherine Rouby, a Christoph Segebarth, b and Jean-P. Royet a,d a Neurosciences et Syste `mes Sensoriels, Universite ´ Claude Bernard Lyon 1, UMR CNRS 5020, IFR 19, Institut Fe ´de ´ratif des Neurosciences de Lyon, 50 Avenue Tony Garnier, 69366 Lyon Cedex 07, France b Unite ´ mixte INSERM/Universite ´ Joseph Fourier U594, LRC-CEA, Ho ˆpital Michallon, 38043 Grenoble, France c Neuropsychology Section, Department of Neurology, Indiana University School of Medicine, Indianapolis, IN 46202, USA d CERMEP, 69003 Lyon, France Received 16 April 2004; revised 12 October 2004; accepted 26 October 2004 Available online 9 December 2004 Previous studies have shown activation of right orbitofrontal cortex during judgments of odor familiarity. In the present study, we sought to extend our knowledge about the neural circuits involved in such a task by exploring the involvement of the right prefrontal areas and limbic/ primary olfactory structures. Fourteen right-handed male subjects were tested using fMRI with a single functional run of two olfactory conditions (odor detection and familiarity judgments). Each condition included three epochs. During the familiarity condition, subjects rated whether odors were familiar or unfamiliar. During the detection condition, participants decided if odors were present. When contrasting the familiarity with the detection conditions, activated areas were found mainly in the right piriform cortex (PC) and hippocampus, the left inferior frontal gyrus and amygdala, and bilaterally in the mid-fusiform gyrus. Further analyses demonstrated that the right PC was more strongly activated than the left PC. This result supports the notion that the right PC is preferentially involved in judgments of odor familiarity. D 2004 Elsevier Inc. All rights reserved. Keywords: Olfaction; Familiarity judgment; Recognition memory; Piriform cortex; fMRI Introduction Hemispheric asymmetry is well-established for high-level brain functions such as language and spatial attention (e.g., Broca, 1863; Weintraub and Mesulam, 1987). Hemispheric predominance also exists in sensory functions such as hand somatosensory represen- tation (Soros et al., 1999) and temporal and spectral auditory resolution (Zatorre et al., 2002). Studies in olfaction lead to similar conclusions. Early cerebral imaging studies showed functional lateralization of olfactory processes in the right hemisphere, especially in orbitofrontal cortex (OFC) (Zatorre et al., 1992); most subsequent studies confirm this result (Dade et al., 1998; Sobel et al., 1998; Yousem et al., 1997). However, Zald and colleagues reported stronger activation in the left OFC and amygdala for very aversive odors, pointing to these areas as being important in the emotional processing of olfactory information (Zald and Pardo, 1997; Zald et al., 1998). Using positron emission tomography (PET), Royet et al. (1999, 2001) found that judgments of odor familiarity preferentially activated right OFC, whereas hedonic judgments principally activated left OFC. Beyond OFC, recent cerebral imaging data have extended these observations to piriform cortex (PC). In a functional magnetic resonance imaging (fMRI) study, we showed that left piriform– amygdala region activation was associated with subjects’ ratings of the odorants’ emotional intensities (Royet et al., 2003), a result consistent with Gottfried et al.’s (2002) and Anderson et al.’s (2003) findings. Convergent results from Dade et al.’s (2002) lesion and PET studies further show that the piriform region mediates olfactory long-term recognition memory, giving support to the notion that this area may be more than primary sensory cortex (e.g., Schoenbaum and Eichenbaum, 1995). Specifically, Dade et al. (2002) found that the extent of piriform activity corresponded with different cognitive demands, in which PC activity followed a continuum between mnemonic encoding (no significant activity), to short-term recognition (weak bilateral activity), to long-term recognition (strong bilateral activity). These authors further suggested that piriform activity could be related to odor familiarity, which would require that subjects compare odors with previously stored olfactory representations, and thus represent a type of long-term olfactory reference memory. In the present study, we explored that question by specifically asking whether PC is involved in the processing of odor familiarity, and if so, whether familiarity-evoked activations are lateralized, as previously suggested (Royet et al., 1999, 2001). We studied 1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2004.10.028 * Corresponding author. Fax: +33 4 37 28 76 01. E-mail address: [email protected] (J. Plailly). Available online on ScienceDirect (www.sciencedirect.com). www.elsevier.com/locate/ynimg NeuroImage 24 (2005) 1032 – 1041
10
Embed
Involvement of right piriform cortex in olfactory familiarity judgments
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
www.elsevier.com/locate/ynimg
NeuroImage 24 (2005) 1032–1041
Involvement of right piriform cortex in olfactory
familiarity judgments
Jane Plailly,a,* Moustafa Bensafi,a Mathilde Pachot-Clouard,b Chantal Delon-Martin,b
David A. Kareken,c Catherine Rouby,a Christoph Segebarth,b and Jean-P. Royeta,d
aNeurosciences et Systemes Sensoriels, Universite Claude Bernard Lyon 1, UMR CNRS 5020, IFR 19, Institut Federatif des Neurosciences de Lyon,
50 Avenue Tony Garnier, 69366 Lyon Cedex 07, FrancebUnite mixte INSERM/Universite Joseph Fourier U594, LRC-CEA, Hopital Michallon, 38043 Grenoble, FrancecNeuropsychology Section, Department of Neurology, Indiana University School of Medicine, Indianapolis, IN 46202, USAdCERMEP, 69003 Lyon, France
Received 16 April 2004; revised 12 October 2004; accepted 26 October 2004
Available online 9 December 2004
Previous studies have shown activation of right orbitofrontal cortex
during judgments of odor familiarity. In the present study, we sought to
extend our knowledge about the neural circuits involved in such a task
by exploring the involvement of the right prefrontal areas and limbic/
primary olfactory structures. Fourteen right-handed male subjects
were tested using fMRI with a single functional run of two olfactory
conditions (odor detection and familiarity judgments). Each condition
included three epochs. During the familiarity condition, subjects rated
whether odors were familiar or unfamiliar. During the detection
condition, participants decided if odors were present. When contrasting
the familiarity with the detection conditions, activated areas were found
mainly in the right piriform cortex (PC) and hippocampus, the left
inferior frontal gyrus and amygdala, and bilaterally in the mid-fusiform
gyrus. Further analyses demonstrated that the right PC was more
strongly activated than the left PC. This result supports the notion that
the right PC is preferentially involved in judgments of odor familiarity.
Behavioral data recorded for the three sets of odors (a, b, c) during the detection and familiarity judgment tasks
Parameter Task Response Set a Set b Set c
Mean number of stimulations Detection 13.33 F 1.87 14.11 F 3.02 13.89 F 1.96
Familiarity 12.89 F 2.20 13.33 F 2.24 13.44 F 2.65
Response accuracy Detection 0.962 F 0.050 0.857 F 0.104 0.930 F 0.078
Reaction time Detection Yes 1.597 F 0.462 1.698 F 0.624 1.625 F 0.487
No 2.125 F 0.602 2.092 F 0.758 2.033 F 0.620
Familiarity Yes 2.141 F 0.668 1.985 F 0.540 2.153 F 0.635
No 2.220 F 0.652 2.346 F 0.730 2.359 F 0.819
Proportion of familiar odors Familiarity 0.545 F 0.214 0.520 F 0.188 0.509 F 0.180
For reaction time, data are given according to whether the odors were detected or not (Yes or No), or whether they were recognized as being familiar or not (Yes
or No).
J. Plailly et al. / NeuroImage 24 (2005) 1032–1041 1035
limbic olfactory regions with much less detail. Activated areas were
indicated using the MNI coordinate system.
Specific effects for the familiarity judgment task were calculated
by comparing the signals during the F and D conditions using the
general linear model (Friston et al., 1995b). Intrasubject analyses
were first performed, followed by a random effects analysis which
extent statistical inferences into the healthy population. This two-
stage analysis accounted first for intrasubject variance (scan-to-
scan), and second for intersubject variance. In the first step, scan-to-
scan variance was separately modeled for each subject by creating a
summary contrast image from weighted parameter estimates that
represented each scan condition. In the second step, these contrast
images were then analyzed using a basic model one-sample t tests
to assess the F–D contrast against a null hypothesis.
A cluster analysis was further performed to compare activation
between the right and left PC. A region of interest (ROI)
corresponding to the right PC was defined by selecting an 8-
mm-diameter sphere centered on coordinates (30, 2, �16) of the
activation cluster obtained in the F–D contrast image from the
group analysis. An identical ROI was centered on the contralateral
coordinate in the left hemisphere (�30, 2, �16). Using the
MarsBar SPM toolbox (Brett et al., 2002), we then obtained a
mean activity level within both ROIs for each one of 12 subjects. A
statistical analysis was then performed to compare the activity
levels of left and right PC.
Fig. 2. Reaction times represented as a function of the olfactory task
(Detection vs. Familiarity) and of the type of response (Yes vs. No). Data
were normalized with respect to the number of stimulations per epoch. The
vertical bars show the standard errors of the means. *, significant difference
(P b 0.007).
Results
Behavioral data
Response accuracy was determined for the detection task only,
since the familiarity judgment depends on personal experience.
During the scanning day, two subjects scored very low response
accuracy (65% and 68%, respectively), due to a very small number
of odors detected (29% and 41%, respectively). The aberration of
both these values was rated with the Grubbs’ test (Dagnelie, 1975),
which indicated that they were indeed outliers (t = 5.414 and t =
5.052, respectively, for a theoretical value of t0.99815 = 3.663). One
of these subjects also did not provide behavioral responses during
the three epochs of the familiarity task. These two subjects were
therefore excluded from further analysis. For the 12 remaining
subjects, the mean numbers and response accuracies of odors
delivered per epoch for the three odor sets (Da, Db, Dc) of the
detection task are given in Table 2. A one-way ANOVA with
repeated measurements performed on response accuracy showed a
significant main effect of set factor [F(2,22) = 5.985, P = 0.0084],
indicating that odors of the Da and Dc sets were more easily
detected than those of the Db set.
For the F condition, the number of odorants judged as being
familiar or unfamiliar by the subjects was determined for the three
odor sets (Fa, Fb, and Fc). For each subject, data were normalized
with respect to the number of stimulations per epoch. The mean
number of stimulations and the mean ratios of familiar odorants
delivered per epoch are given in Table 2. A one-way ANOVAwith
repeated measurements performed on the ratios of familiar odor-
ants showed no significant effect of set factor [F(2,22) = 0.455, P =
0.640], indicating that the same proportion of familiar odorants was
found in the three odor sets.
Subjects’ reaction times for odor detection and familiarity
judgments were also calculated (Table 2). The number of
stimulations delivered per epoch depended on the subject’s
breathing rhythm. As this factor could affect reaction times by
creating a shorter interstimulus response period in those who
breathe more rapidly, the data were normalized with respect to the
number of stimulations, and analyzed as a function of task, odor
sets, and response (Yes or No) factors. A three-way ANOVA with
repeated measurements showed a significant effect of the judgment
task [F(1,11) = 33.380, P b 0.0001], of response [F(1,11) =
65.965, P b 0.0001], but no significant effect of odor set factor
[F(2,22) = 1.020, P = 0.3772]. Significant task � response
[F(2,22) = 10.153, P = 0.0087] (see Fig. 2) and task � set �
J. Plailly et al. / NeuroImage 24 (2005) 1032–1041 1039
et al., 2001), the right hippocampal activation during the familiarity
judgment task was not anticipated in the present study. Inconsistent
activation of the hippocampal formation is not specific to olfaction,
and has been characteristic of other studies of memory (Andreasen
et al., 1995; Shallice et al., 1994; Tulving et al., 1996). Our data are
nevertheless consistent with a recent finding showing hippocampal
activation during the retrieval of olfactory episodic memories
(Gottfried et al., 2004).
Lesion studies recently examined whether the brain structures
that comprise the medial temporal lobe memory system (i.e., the
hippocampal and parahippocampal regions) differ in how they
support recollective and familiarity components (Manns et al.,
2003; Yonelinas et al., 2002). Our data do not permit distinguish-
ing these aspects of memory, as the subjects may well have had
consciously evoked memories from the odorants. They are,
however, consistent with the idea that both regions probably
contribute to olfactory recognition memory.
Activation of the left inferior frontal gyrus, in the pars
orbitalis, during familiarity judgments further supports the
hypothesis that this region is involved in the selection and
integration of semantic information in a modality-independent
manner (Homae et al., 2002; Kareken et al., 2003). In a recent
study, Savic and Berglund (2004) found that left frontal and right
parahippocampal region activation positively correlated with
familiarity ratings, showing the engagement of semantic circuits
during passive smelling of familiar odorants. Along the same line,
the mid-fusiform gyrus activation in the current study might lead
to similar interpretations, since it has further been associated with
visual, tactile and auditory recognition and categorization of
objects (Adams and Janata, 2002; Joseph and Gathers, 2003;
Stoeckel et al., 2003). Its involvement in olfactory object
recognition therefore reinforces the idea of the polymodal nature
of this area (Adams and Janata, 2002) and its role in semantic
processing (Price, 2000; Wagner et al., 1998).
Conclusion
Complementing previous PET studies that demonstrate right
OFC involvement in odor familiarity judgments, the present fMRI
study shows that right PC is also activated during this task, an
activation that may be related to olfactory recognition memory. In
previous fMRI and PET studies, we demonstrated that a neural
network in the left hemisphere, involving the OFC and primary
olfactory areas, mediated olfactory hedonic perception (Royet et al.,
2000, 2003). It thus appears that odor processing activates a large
neural network involving both hemispheres. Nevertheless, this
network possesses hemispheric predominance depending on the
type of olfactory task performed (see Royet and Plailly, 2004, for
review). The present data provide further evidence that the right
hippocampal region, left inferior frontal gyrus and mid-fusiform
gyrus take part in recognition memory, likely cross-modally to
assist in gathering relevant associations to enable identification of
olfactory percepts.
Acknowledgments
We thank the technical team (M. Vigouroux, B. Bertrand, and
V. Farget) for designing and building the stimulation and recording
materials and J.P. Lomberget and M.B. Sanglerat for medical
examinations of subjects participating in the study. We are grateful
to the companies Givaudan, International Flavors and Fragrances,
Lenoir, Davenne, and Perlarom for supplying the odorants used in
this study. This work was supported by research grants from the
dRegion Rhone-AlpesT and the dGIS Sciences de la Cognition,T thedCentre National de la Recherche Scientifique,T and the dUniversiteClaude-Bernard de Lyon.T
References
Adams, R.B., Janata, P., 2002. Comparison of neural circuits underlying
auditory and visual object categorization. NeuroImage 16, 361–377.
Anderson, A.K., Christoff, K., Stappen, I., Panitz, D., Ghahremani, D.G.,