European Journal of Neuroscience, Vol. 24, pp. 3578–3588, 2006 doi:10.1111/j.1460-9568.2006.05235.

Neurogenic correlates of an olfactory discrimination task in

the adult olfactory bulb

Nathalie Mandairon,1,2 Joëlle Sacquet,1,2 Samuel Garcia,1,2 Nadine Ravel,2,3 François Jourdan1,2 and Anne Didier1,2
1
Laboratoire de Neurosciences et Systèmes Sensoriels, CNRS UMR 5020, Université Claude Bernard Lyon1, 50 avenue Tony
Garnier, 69366, LYON cedex 07, France
2
Institut Fédératif des Neurosciences de Lyon, IFNL, France
3
Institut des Sciences Cognitives, UMR CNRS 5015, 67 Boulevard Pinel, 69675 Bron Cedex, France

Keywords: adult mouse, BrdU, discrimination, neurogenesis, olfactory memory, Zif268

Abstract
In the main olfactory bulb, stimuli are coded within the spatio-temporal pattern of mitral cells’ activity. Granule cells are interneurons
that shape the mitral cells’ activity, and are continuously generated in the adult main olfactory bulb. However, the role of granule cell
renewal remains elusive. We show here that an associative olfactory discrimination task reduces the survival of newborn neurons.
However, when the olfactory task involves perceptually related odorants, the learning process is slower and does not induce such a
reduction in the number of new neurons. Mapping newborn cells within the granule cell layer of the main olfactory bulb reveals a
clustered distribution that evolves with learning as a function of odorant similarity and partly overlaps with the immediate-early gene
Zif268 expression pattern. These data provide insight into the functional mechanisms underlying olfactory discrimination learning,
and promote the importance of neurogenesis as a cellular basis for the restructuring of odor images in the main olfactory bulb.

Introduction
The olfactory inputs to the main olfactory bulb (MOB) are spatially
segregated based on the expression of molecular odorant receptors by
sensory neurons (Buck, 1996), imparting a chemotopic and dynamic
activation map on the MOB on the sensory input stage (glomeruli)
(Kauer & White, 2001; Korsching, 2002). When a similar glomerular
map is obtained for two odorants, these two odorants are perceptually
similar, attesting to the perceptual correlates of glomerular activation
maps (Rubin & Katz, 1999; Linster et al., 2001; Schaefer et al.,
2001a). Sensory information is then transmitted to a population of
mitral cells (MCs), the output neurons of the MOB. Within the MOB,
MCs activate granule cells (GRs), interneurons that, in turn, inhibit
MCs through dendro-dendritic reciprocal synapses (Shepherd &
Greer, 1990). These dendro-dendritic interactions mediate lateral
inhibition and are thought to contribute to the synchronization of MC
firing. GRs are thought to play a key role in olfactory coding by
shaping the temporal aspects of spatially distributed MC responses to
odors (Lowe, 2003; Schoppa & Urban, 2003; Lledo et al., 2005).
In the MOB of rodents (Lois & Alvarez-Buylla, 1994) and primates
(Kornack & Rakic, 2001), new GRs are formed throughout life from
neural stem cells located in the subventricular zone (SVZ) of the
lateral ventricle (Alvarez-Buylla & Garcia-Verdugo, 2002), which
migrate from the SVZ to the MOB and functionally integrate into the
neuronal network (Carlen et al., 2002; Belluzzi et al., 2003; Carleton
et al., 2003; Magavi et al., 2005). The survival of newborn GRs is
dependent on sensory activity (Petreanu & Alvarez-Buylla, 2002;
Correspondence: Dr N. Mandairon, Department of Neurobiology and Behavior, Cornell
University, Ithaca, NY 14853, USA.
E-mail: nm243@cornell.edu
Received 12 April 2006, revised 5 October 2006, accepted 16 October 2006
Rochefort et al., 2002; Mandairon et al., 2003, 2006; Yamaguchi &
Mori, 2005), but the function of GRs renewal remains largely
speculative (Lledo et al., 2005). A modeling study based on the
activity-dependent survival of newborn GRs (Cecchi et al., 2001)
suggests that they could adjust the tuning of the odor-specific spatio-
temporal pattern of bulbar activity for better discrimination. This view
was reinforced by data linking a low rate of neurogenesis to an
impairment of olfactory discrimination (Gheusi et al., 2000; Enwere
et al., 2004). Better survival of newly formed neurons is also
associated with enhanced short-term olfactory memory (Rochefort
et al., 2002). As is the case in other neurogenic regions (Nottebohm,
2002; Shors et al., 2002; Kempermann et al., 2004), bulbar
neurogenesis may thus subserve learning-related functions, but there
is no direct evidence for such a role.
In the present study, we asked whether olfactory learning modulates
the rate of neurogenesis within the MOB. A discrimination learning
task involving perceptually similar or dissimilar odorants was used to
test the hypothesis that modulation of neurogenesis may be prominent
when the task is made more difficult. Neurogenesis was assessed by
counting and mapping newborn GRs in the MOB of mice submitted to
the learning task. We show that: (i) learning reduces the survival of
newborn neurons only when dissimilar odorants are used; (ii) the
density of the newborn neurons is inversely correlated with the
behavioral performances; and (iii) mapping bromodeoxyuridine
(BrdU)-positive cells within the granule cell layer (GrL) of the
MOB reveals clusters of newborn cells whose positions within the
GrL are at least partly odor-driven and modified by the learning
process. These data provide the first evidence for the influence of
olfactory learning on the rate and distribution within the GrL of bulbar
neurogenesis, pointing to newborn neurons as important cellular
targets of olfactory learning-induced plasticity.

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd

Materials and methods
Animals
Adult male C57Bl ⁄ 6J mice (Charles River, L’Arbresles, France), aged
8 weeks at the beginning of the experiments, were housed under a
12 h light : dark cycle in an environmentally controlled room. All
behavioral training was conducted in the afternoon (14:00–17:00). All
efforts were made to minimize the number of animals used, and the
experimental procedures were in accordance with the European
Community Council Directive of 24 November 1986 (86 ⁄ 609 ⁄ EEC),
and the French Ethical Committee.
Behavioral experiments
Experimental set up
All behavioral experiments were conducted in a specially designed
computer-assisted two-hole-board apparatus (40 · 40 cm), allowing
detection of the beginning of each trial (when the mouse is placed in
the start area facing the holes), and monitoring of the sequence and
duration of nose poking into the holes (3 cm diameter, 4.5 cm deep).
A polypropylene swab embedded in fine plastic mesh was placed at
the bottom of the hole, covered with bedding. For trials involving
odors, the swab was impregnated with 20 lL of pure odorant. The
bedding was replaced after every trial.
Odorant set
The odorant set consisted of the (+) and (–) enantiomers of limonene
[(+)L and (–)L] and propionic acid (PA) (Sigma-Aldrich, Saint Louis,
MO, USA). Purity reported by the manufacturers was > 95% for (–)L,
> 97% for (+)L and 99% for PA.
Experiment 1: olfactory habituation ⁄ dishabituation
Food and water were continuously available during the course of the
experiment. Mice (n ¼ 30) were first familiarized with the experi-
mental set up by a 3-min trial on the hole-board with no odorant.
Habituation ⁄ dishabituation was assessed for two pairs of odorants:
(+)L ⁄ (–)L and (+)L ⁄ PA. Each training session consisted of four
successive trials (3-min duration, 15-min interval). In the first three
trials, one hole was odorized with one odor of the pair. On the fourth
trial, animals were exposed to the other odorant of the pair.
All data analyses on time spent sniffing during odor presentation
trials were performed with Systat software (SSI, Richmond, CA,
USA). Only mice that investigated the odorized hole for at least 8 s
during its first presentation were included in the analysis. Data
obtained for each pair of odors were averaged across animals and
analysed by anova followed by Tukey post hoc tests to determine
whether the investigation time elicited by the test odor was
significantly different from that elicited by the habituated odor during
its fourth presentation. The level of significance was set to 0.05.
Experiment 2: olfactory discrimination
Shaping. Naı̈ve mice (n ¼ 60), different from the mice used in
Experiment 1, were first trained to retrieve a reward (small bit of
sweetened cereal, Kelloggs, Battle Creek, MI, USA) by digging
through the bedding. The mouse was put in the start area and was
allowed to dig for 2 min. During the first few trials the reward was
placed on the top of the bedding of one of the holes. After several
successful retrievals, the reward was buried deeper into the bedding.
Shaping was considered to be complete when a mouse could
successfully retrieve a reward that was deeply buried in the bedding
(8–12 trials).
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
Neurogenic correlates of an olfactory discrimination task 3579
Conditioning. During the discrimination learning experiments, water
was continuously available, but the mice were food-deprived for
8 h before the sessions. We determined whether adult mice could
learn to discriminate between (+)L and (–)L [(+)L ⁄ (–)L, n ¼ 15],
which are known to elicit overlapping glomerular patterns of
activity and, thus, are difficult to discriminate (Linster et al., 2001).
For comparison, we tested the ability of mice to discriminate
between (+)L and PA [(+)L ⁄ PA, n ¼ 15]. Discrimination learning
consisted of six sessions (one per day) of four trials (total number
of trials ¼ 24). Each of the two holes was randomly odorized with
one odorant of the pair, and the reward was systematically
associated with (+)L. The same pairs of odorants were used in
two pseudo-conditioning groups (n ¼ 15 each) in which the
reinforcement was randomly associated with either odorant of the
pair. The number and duration of visits to the holes (nose pokes)
were recorded. The latency (time to find the reward) was measured
as an index of learning.
Data analysis
Mice that did not perform the task (no visit to any hole for one
entire session (four consecutive trials) were excluded from the
experiment. For the remaining animals [conditioning (+)L ⁄ PA n ¼
15; (+)L ⁄ (–)L n ¼ 10; pseudo-conditioning (+)L ⁄ (–)L n ¼ 11;
(+)L ⁄ PA n ¼ 10], anova for repeated measures were applied to
the latency values to assess the learning process (time effect) and
between-groups differences. For each pair of odorants, conditioned
and pseudo-conditioned groups were compared by bilateral Stu-
dent’s t-tests (Systat software). The level of significance was set to
0.05.
Newborn cells mapping in the MOB
BrdU administration
Mice to be used for conditioning or pseudo-conditioning were injected
with BrdU (Sigma; 50 mg ⁄ kg in saline three times daily at 2-h
intervals) 30 days before behavioral tests and 45 days before death.
Histology
Under deep anesthesia (pentobarbital, 0.2 mL ⁄ 30 g), five mice
randomly selected from each of the four behavioral groups were
killed by intracardiac perfusion of 50 mL of cold fixative (parafor-
maldehyde 4% in phosphate-buffered saline, pH 7.4). Brains were
removed, post-fixed, frozen rapidly and then stored at )20 C before
sectioning with a cryostat (Jung).
BrdU immunocytochemistry
The protocol has been previously described (Mandairon et al., 2003).
Briefly, sections were incubated in Target Retrieval Solution (Dako,
Glostrup, Denmark) for 20 min at 98 C. After cooling, they were
treated with pepsin (0.43 U ⁄ mL in 020 min.1-N HCl, Sigma) for
3 min. Sections were transferred to a blocking solution [5% normal
horse serum (Sigma) with 5% bovine serum albumin (BSA) and
0.125% Triton X-100], and were then incubated overnight in a mouse
anti-BrdU antibody (1 ⁄ 100, Chemicon, Temecula, CA, USA) at 4 C
followed by a biotinylated anti-mouse secondary antibody (1 ⁄ 200,
Vector Laboratories, Burlingame, CA, USA) for 2 h. The sections
were then processed through an avidin–biotin–peroxidase complex
(ABC Elite Kit, Vector Laboratories). Following dehydration in
graded ethanols, the sections were defatted in xylene and cover-
slipped in DPX (Fluka, Sigma).
3580 N. Mandairon et al.
BrdU-positive cell mapping
All cell counts were conducted blind with regards to the mouse status.
In each animal, every fifth section was analysed (thickness ¼ 14 lm,
sampling interval ¼ 70 lm). Within each section analysed, every
BrdU-positive cell was counted in the GrL of the left MOB using
mapping software (Mercator, Explora Nova, La Rochelle, France)
coupled to a Zeiss microscope. Maps of BrdU-positive cells were
constructed as follows, inspired by the work of Schaefer et al. (2001b).
The GrL was divided into 36 sectors of 10  with a reference axis
drawn parallel to the most ventral aspect of the subependymal layer of
the MOB. The cell density (number of labeled profiles ⁄ lm2) was
calculated for each sector. Measurements were then merged into arrays
of 10  · 70-lm bins (Fig. 5A and B). A z-score (normalization of the
matrix to a mean ¼ 0 and SD ¼ 1) was calculated for each animal.
The most rostral aspect of the accessory olfactory bulb served as an
anatomical landmark to align the sections across animals (Matlab v.6).
For visualization of the BrdU-positive cell density maps, arrays were
averaged across animals within each group, and a colored image plot
of the data was constructed in Matlab v.6.
Data analysis
The mean BrdU-positive cell density of each array was calculated and
averaged within each experimental group. Between-groups compar-
isons were performed by bilateral Student’s t-tests.
Comparisons of the distribution of BrdU-positive cell density
between the pseudo-conditioned groups exposed to the related and
dissimilar pairs of odorants were achieved by a t-test performed on
z-scored individual arrays and using a sliding area successively
centered on each bin and covering 25 adjacent bins.
To compare the magnitude of the changes induced by conditioning
with the (+)L ⁄ (–)L vs. the (+)L ⁄ PA pair of odorants, differences
between the pseudo-conditioned and conditioned patterns needed to be
normalized with regards to the pseudo-conditioned pattern. The
pseudo-conditioned array was subtracted from the conditioned array
for each pair (Excel software). In the resulting matrices, the number of
bins having a value of BrdU-positive cell density greater than 1 SD
was counted and divided by the number of bins exhibiting a BrdU-
positive cell density greater than 1 SD in the corresponding pseudo-
conditioned array. The 1 SD criteria was chosen on the basis of the
matrices distribution histograms (not shown), as it includes popula-
tions of bins above mean values that represent a measure of the high-
density areas. This ratio indicates the proportion of high-density foci
found in the conditioning groups that were not present in the pseudo-
conditioning groups. This index of change induced by learning was
similarly calculated for the number of bins with values less than
)1 SD in the subtraction-generated matrices, thus indicating the
proportion of high-density foci present in the pseudo-conditioned
group that were not retrieved in the conditioned group. Statistical
comparisons were performed by t-tests for comparison of proportions.
Double-labeling BrdU-NeuN
Sections were treated as described above except that they were
incubated simultaneously with rat anti-BrdU (1 : 100, Harlan Sera lab,
Loughborough, UK) and a mouse anti-NeuN (1 : 500, Chemicon)
overnight at 4 C. Immunoreactivities were revealed with a goat anti-
rat antibody coupled to Texas Red (Vector Laboratories) and a horse
anti-mouse coupled to FITC (DAKO). In three–four animals of each
experimental group, 20–30 BrdU-positive cells were examined for
NeuN double-labeling. Double-labeling was analysed by confocal
scanning microscopy (Zeiss, at the Centre Commun de Quantimétrie,
Université Claude Bernard-Lyon1). All labeled cells were examined
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
along the z-axis to ensure proper detection of double-labeled cells. A
percentage of double-labeled cells was calculated for each group and
compared using anova. The number of newly formed neurons in the
GrL was extrapolated from the percentage of double-labeled cells, the
BrdU-positive cell density and the volume of the GrL.
Zif268 expression mapping
Odorant stimulation
Naı̈ve animals, different from the ones used in the behavioral
procedure, were submitted to (+)L (n ¼ 5) or PA (n ¼ 5) stimulation
1 h before death (Inaki et al., 2002). Briefly, mice were put in a 1-L
hard plastic chamber. Deodorized air was generated by compressed air
passing through activated charcoal and distilled water. The deodorized
air was applied at a rate of 5 L ⁄ min to the chamber from the top inlet.
For odorant stimulation, the deodorized air was passed through
undiluted odorant solution. The odorant-containing air was then
diluted with the deodorized air (1 : 100) and the outlet of the flow-
meter was connected to the chamber. The odorant-containing air was
applied for 5 min followed by 5 min application of deodorized air.
This step was repeated three times. One hour after the last odor
exposure, mice were killed by perfusion as described above.
Zif268 immunocytochemistry
The MOB of odorant-stimulated mice was coronally sectioned
(14 lm). Every fifth section was processed for Zif268 immunostain-
ing. Sections were transferred to 10% normal goat serum (Sigma)
with 2% BSA and 0.1% Triton X-100 for 1 h to block non-specific
binding and were then incubated overnight in a rabbit anti-Zif268
antibody (1 ⁄ 1000, Santa Cruz Biotechnology, Santa-Cruz, CA, USA)
at room temperature for 16 h. Sections were then incubated in a
biotinylated anti-rabbit secondary antibody (1 ⁄ 200, Vector Laborat-
ories) for 2 h. The remaining treatments were similar to those for the
BrdU labeling.
Zif268 expression mapping
Zif268-positive cells were counted automatically in the GrL (Mercator
Pro, Explora Nova, La Rochelle, France). Zif268-positive cell density
maps were constructed using the same procedure as for BrdU-positive
cell density.
The overlap between BrdU-positive cell maps and Zif268 expres-
sion maps was assessed as the number of bins exhibiting values within
15% of the highest values in both the Zif268 and BrdU-positive maps.
This threshold was set as best fitted to the clusters delineated by visual
inspection of the Zif268 expression maps. Because the absolute value
of overlap depends on the threshold, relative overlap should be
considered. T-tests for comparisons of proportions were used to
compare the overlap between groups.
Results
Experiment 1: olfactory habituation ⁄ dishabituation
Two pairs of odorants were selected, one perceptually similar and the
other perceptually dissimilar, based on the ability of mice to
discriminate the two odorants of the pair in a habituation–dishabit-
uation task (Linster et al., 2001). Mice for this experiment were
divided into two groups. In the first group, we tested if they were able
to spontaneously discriminate between the two enantiomers of
limonene (n ¼ 13), and in the second one we tested if they were
able to discriminate between (+)L and PA (n ¼ 17).
European Journal of Neuroscience, 24, 3578–3588

Fig. 1. Behavioral habituation and discrimination. One group of mice was
exposed to the (+)limonene (L) for three trials (Hab1–3) and to (–)L as the test
odor on the fourth trial. Another group was similarly exposed to (+)L for three
trials and to propionic acid (PA) on the test trial. Both groups showed
habituation to (+)L, but only the group exposed to PA on the test trial showed
dishabituation, indicating that mice discriminate (+)L from PA but not from
(–)L. (Tukey, *P < 0.05)
Habituation was evident, because in the two groups there was a
significant effect of trial number (Fig. 1): (+)L ⁄ (–)L, F3,24 ¼ 14.03,
P < 0.0001; (+)L ⁄ PA, F3,52 ¼ 5.76, P ¼ 0.002), and the investiga-
tion response during trial 1 was significantly higher than the response
during trial 3 (P < 0.05 for all odor pairs, Tukey) (Fig. 1). Multiple
comparisons testing showed that the response to test odor was
significantly higher than the response to habituated odor for the
dissimilar odorants group (P ¼ 0.03), but not for the similar odorants
group (P ¼ 0.93, Tukey), indicating that while mice discriminated
between (+)L and PA, they did not discriminate between the
enantiomers of limonene (Fig. 1).
Experiment 2: olfactory discrimination
Two groups of food-restricted mice (n ¼ 30) were submitted to a
discrimination learning task using the pair of perceptually similar
odorants, the (+)L ⁄ (–)L, or the pair of distinct odorants (+)L and PA.
(+)L was positively reinforced in both groups by a food reward. Two
control groups (one for each pair of odorants, n ¼ 30) were subjected
to the same procedure, but the reinforcement was randomly associated
with either odor of the pair (pseudo-conditioning). The evolution of
latency across the learning sessions is presented for conditioned and
pseudo-conditioned animals with the pair of different odorants or the
pair of similar odorants (Fig. 2). An anova analysis using the group
(conditioning vs. pseudo-conditioning) and the pair of odorants
[(+)L ⁄ (–)L vs. (+)L ⁄ PA] as independent factors, and the session of
learning as a dependent factor indicates a difference between
conditioned and pseudo-conditioned animals (group effect,
F1,42 ¼ 16.57, P < 0.0005), which is modulated by the pair of
odorants used (group · pair of odorants interaction, F1,42 ¼ 5.57,
P ¼ 0.023). The latency decreased with learning (day effect
F5,210 ¼ 7.57, P < 0.0005), and this effect depends on the group
(day · group interaction, F5,210 ¼ 4.68, P < 0.0005). Indeed, a
simple comparison between the conditioning and pseudo-conditioning
groups indicates that latency decreased across sessions in the
conditioning (F5,115 ¼ 10.851, P < 0.0005) but not in the pseudo-
conditioning group (F5,95 ¼ 0.99, P ¼ 0.42). This decrease in latency
is also influenced by the pair of odorants used (day · pair of odorants,
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
Neurogenic correlates of an olfactory discrimination task 3581
Fig. 2. The olfactory discrimination learning rate depends on the perceptual
proximity of odorants. Animals pseudo-conditioned with the (+)limonene
(L) ⁄ propionic acid (PA) pair or the (+)L ⁄ –(L) pair of odorants showed no
improvement in performance. In the conditioned groups, (+)L was systemat-
ically reinforced. Latency decreased with days for the (+)L ⁄ PA pair and for the
(+)L ⁄ (–)L pair (see text for statistics). In conditioned animals, performances
diverged from pseudo-conditioning on training day 1 for the (+)L ⁄ PA pair and
on training day 5 for the (+)L ⁄ (–)L pair (bilateral t-tests, *P < 0.05,
**P < 0.001, ***P < 0.0005).
F5,201 ¼ 2.50, P ¼ 0.03). Taken together these data indicate that
latency decreased selectively across learning for the conditioned
animals and that this decrease is different for the two pairs of odorants
used. Comparison of pseudo-conditioned vs. conditioned animals for
each pair of odorants further confirms the latter conclusion. Latency in
the conditioning group differed from pseudo-conditioning on Day 1
for the animals learning the pair of different odorants, and only on Day
5 for animals conditioned with the similar pair (bilateral t-tests). To
verify that the animals were not guided by the odor of the reward, it
was removed from the (+)L odorized hole for the last trial of the last
day of training in the group conditioned with (+)L ⁄ PA. In the absence
of reward, 83.5 ± 4.4% of the exploration was to the hole odorized
with (+)L. Taken together, these data indicate that the learning process
was successful in both conditioning groups, but significantly slower in
the group conditioned with (+)L ⁄ (–)L.
Discrimination learning affects newborn cell survival as a
function of odorant proximity
Under physiological conditions, many GRs in the MOB die between
15 and 45 days after their birth in the SVZ (Petreanu & Alvarez-
Buylla, 2002; Winner et al., 2002; Mandairon et al., 2006). This
phenomenon is enhanced in olfactory-deprived mice (Petreanu &
Alvarez-Buylla, 2002; Yamaguchi & Mori, 2005; Mandairon et al.,
2006), suggesting that the long-term survival of newborn cells is
dependent upon activity. To test the influence of olfactory discrim-
ination learning on the fate of newly formed cells during this
timeframe, the DNA synthesis marker BrdU was injected 30 days
before behavioral training. Five days after the last training session, five
animals from each experimental group were randomly selected, killed
(Fig. 3A), and then subjected to BrdU-positive cells counts in GrL (see
Materials and methods). BrdU-positive cell density was reduced in the
3582 N. Mandairon et al.

Fig. 3. Learning affects neurogenesis as a function of odorant proximity. (A) Experimental paradigm. Bromodeoxyuridine (BrdU) was administered i.p.
(50 mg ⁄ kg, three injections with a 2-h interval), 30 days before the onset of behavioral experiments, and mice were killed (S) 45 days after administration of BrdU.
(B) BrdU-positive cell density in the GrL of the MOB, 45 days after BrdU for the pseudo-conditioned (PC) and conditioned (C) groups exposed to the pair of
odorants (+)limonene (L) ⁄ propionic acid (PA) or (+)L ⁄ (–)L. (C) Number of newborn neurons in the GrL of the MOB. Same experimental groups and legends as in
(B). A reduction in BrdU-positive cell density and in the number of newborn neurons can be observed in the group learning to discriminate (+)L ⁄ PA but not (+)L ⁄
(–)L, *P < 0.05 (bilateral Student’s t-test). (D) Double-staining showing a BrdU-NeuN colabeled cell. (E) Pearson correlation between the latency on Day 3 of
conditioning and BrdU-positive cell density. Filled squares: animals conditioned with the (+)L ⁄ (–)L pair; filled circles: animals conditioned with the (+)L ⁄ PA pair.

group conditioned with (+)L ⁄ PA compared with the corresponding
pseudo-conditioned group (bilateral t-test P ¼ 0.008). In contrast, no
such reduction was observed in the group conditioned with the pair of
closely related odorants (+)L ⁄ (–)L (same test, P ¼ 0.19) (Fig. 3B).
These changes in cell density occurred with no significant change in
GrL volumes (not shown). A confocal analysis of the double-labeling
of BrdU-positive cells with the neuronal marker NeuN (Fig. 3D)
indicated that the proportion of neurons among BrdU-positive cells
showed no statistical difference among the experimental groups
(anova, F3,8 ¼ 2.5, P ¼ 0.13), and averaged 84 ± 7.6% (mean ± -
SEM). Extrapolation of the number of neurons based on the
percentage of double-labeled cells obtained in each group yielded a
reduction in neuronal survival in the (+)L ⁄ PA conditioned group
(bilateral t-test, P ¼ 0.007) (Fig. 3C), which was not observed in the
(+)L ⁄ (–)L conditioned group (bilateral t-test, P ¼ 0.11).
In order to further characterize the relationship between changes in
neurogenesis and learning, we correlated latencies in the discrimin-
ation learning task of the individual animals with their respective
number of BrdU-positive cells. We found a correlation on Day 3 of
learning, large reductions in BrdU-labeled cell survival being
associated with short latencies (Pearson correlation, r ¼ )0.66,
P ¼ 0.037, n ¼ 10) (Fig. 3E). The number of BrdU-positive cells
retrieved in the OB thus correlates with performance measured in the
course of the learning process (Day 3), but not with the initial (Day 1,
r ¼ 0.02) or the final level of performance (Day 6, r ¼ 0.14).
To assess possible differences in the sampling of the odorants due to
different appetence of the mice towards the odorants used, we
measured the time spent exploring each odorant of each pair. In the
pseudo-conditioned groups (Fig. 4A and B), all odors were explored
to the same level (F3,38 ¼ 0.88, P ¼ 0.45). In the conditioned groups
(Fig. 4C and D), the reinforced odor was explored more than the other
odor of the pair (F1,18 ¼ 49.23, P < 0.0005 for the similar pair;
F1,28 ¼ 56.18, P < 0.0005 for the different pair) (Fig. 4C and D).
Comparisons between the two conditioned groups indicated that
exploration times for (+)L were higher in the group learning the
different pair than in the group learning the similar pair

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
 Neurogenic correlates of an olfactory discrimination task 3583

Fig. 4. Mean time spent each day exploring the odorants. The two odorants of each pair are explored to the same level during pseudo-conditioning (A and B). In
conditioned animals, the reinforced odorant [(+)limonene (L)] is more explored than the other odor of the pair [(–)L or propionic acid (PA)] (C and D). Animals
exposed to the (+)L ⁄ PA pair spent more time exploring the odorants than the animals exposed to (+)L ⁄ (–)L. See text for statistics.

(F1,23 ¼ 19.48, P < 0.0005). The same was true for PA vs. (–)L
(F1,23 ¼ 4.86, P ¼ 0.038).
The exploration time for (+)L or for the non-reinforced odors,
averaged across the six sessions, did not correlate with the number of
BrdU-positive cells (r ¼ –0.06, P ¼ 0.85; r ¼ 0.24, P ¼ 0.5,
respectively).
Spatial distribution of BrdU-positive cells and learning-induced
changes
BrdU-positive cells were counted on a series of sections taken from
the GrL (sampling interval ¼ 70 lm), and each section was divided
into 36 sectors of 10  with the reference axis drawn on the ventral
aspect of the subependymal layer of the MOB (Schaefer et al., 2001b)
(Fig. 5A, and Materials and methods). The density of BrdU-positive
cells was calculated for each sector. These data were combined for
each animal in an array where each bin has the value of the labeled cell
density measured in one sector. One column of the data array
represents the 36 sectors of one section and the x-axis represents the
rostro-caudal distance (Fig. 5B). The individual arrays were normal-
ized by calculating a z-score and averaged within groups, before
representations as color-code maps (Matlab). In all groups, BrdU-
positive cells showed uneven densities within the GrL (Fig. 5D, E, G
and H). An example of BrdU-positive cells labeling in two different
regions of the GrL is shown in Fig. 5C. In the (+)L ⁄ (–)L pseudo-
conditioned group, two main areas of the GrL showed a high density
of BrdU-positive cells: one anterior, ventro-lateral; and one posterior,
ventro-medial (Fig. 5D). The BrdU-positive cell map obtained from
the (+)L ⁄ PA pseudo-conditioned group also showed accumulation of
BrdU-positive cells in the anterior ventro-lateral and posterior ventro-
medial GrL and, in addition, small clusters were scattered along the
rostro-caudal axis of the dorso-medial GrL (Fig. 5E). These two maps
thus exhibited both similar features and significant differences as
revealed by t-test comparisons (Fig. 5F).
In the two conditioning groups, maps of clustered BrdU-positive
cells (Fig. 5G and H, respectively) were also different (Fig. 5I), and
these differences were not at the same locations as in pseudo-
conditioning groups, suggesting that the pattern of BrdU-positive cell
clustering has evolved with learning differently for the two pairs.
Differences of the conditioning maps with the corresponding pseudo-
conditioning maps can be visualized on color-coded maps resulting
from the point-to-point subtraction of the pseudo-conditioning maps
from the corresponding conditioning maps [Fig. 5J and K for (+)L ⁄
(–)L and (+)L ⁄ PA, respectively]. High-density foci of BrdU-positive
cells were present after learning, but they were not found in pseudo-
conditioning groups (red areas). For instance, an increased density of
BrdU-positive cells could be observed in the medial and anterior part
of the MOB in the (+)L ⁄ (–)L conditioned map. Conversely, some
high-density foci that were present in the pseudo-conditioned groups
were not present after learning (dark blue areas). A prominent
difference of this type was the almost complete disappearance, in the
group conditioned with (+)L ⁄ (–)L, of the posterior ventro-medial
accumulation of BrdU-positive cells observed in the pseudo-condi-
tioned group.

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
3584 N. Mandairon et al.

To address the functional significance of the learning-specific
patterning of newly formed GRs, we then asked whether these
changes in BrdU-positive cell distribution were quantitatively different
for the two pairs of odorants. We calculated an index of learning-
induced changes in BrdU-positive cell maps in which positive
differences as given by the subtracted map (high-density bins present
in the conditioned map and not in the pseudo-conditioned map) were
normalized with regards to the number of high-density bins observed
in the corresponding pseudo-conditioned map (see Materials and
methods). The same was done for negative differences between
conditioned and pseudo-conditioned maps (high-density foci present
in pseudo-conditioned but not conditioned maps). Positive and

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
 Neurogenic correlates of an olfactory discrimination task 3585

negative differences were considered only if they reached a
value ¼ 1 SD or ¼ )1 SD, respectively. For both odorant pairs, the
bins exhibiting such changes represented about 20% of the total
number of bins forming the BrdU-positive cell pseudo-conditioned
maps (22.6% and 20.1%, respectively, P ¼ 0.13, t-test for comparison
of proportions). The same amount of spatial reorganization thus
occurred with learning for both pairs of odorants. However, within the
 20% of bins exhibiting learning-induced changes, more high-
density bins appeared upon conditioning in the group trained with
(+)L ⁄ (–)L than in the group trained with (+)L ⁄ PA (induction,
Fig. 5L). In contrast, high-density foci present in the pseudo-
conditioned group, although absent after conditioning (suppression,
Fig. 5L), were about 40% more abundant for the (+)L ⁄ PA pair than for
the (+)L ⁄ (–)L pair.
Acquisition of an olfactory task is thus able to shape the distribution
of newborn GRs in a way that is different from exposure to the odor
pairs with random reinforcement and depends on the difficulty of the
discrimination task.
Relationship between the distribution of BrdU-positive cells and
the chemotopic organization of the MOB
The clustered distribution of BrdU-positive cells in the pseudo-
conditioned and conditioned groups raised the possibility of odor-
driven survival of newborn cells.
Initial evidence of such an odor-related pattern is the significant
differences observed between the maps obtained for the two pairs
(Fig. 5F and I). Because the only difference between these groups was
exposure to different pairs of odorants [(+)L ⁄ (–)L vs. (+)L ⁄ PA], the
differential patterning of the BrdU-positive cells strongly suggests that
the survival of newly formed cells is odor-dependent.
To further substantiate this hypothesis, BrdU-positive maps need to
be compared with a 3D map of odorant-induced activity in the GrL.
Such maps were not available from the literature. In order to visualize
the neuronal network involved in (+)L and PA processing in the GrL,
naı̈ve mice were exposed to (+)L or PA 1 h before death, and Zif268
expression was mapped in the GrL by immunocytochemistry, using
the same mapping method as for BrdU-positive cell density. We found
a patterned expression of Zif268 in response to exposure to (+)L or PA
(Fig. 6A and B). T-test comparisons indicated significant differences
between the (+)L- and PA-induced maps (Fig. 6C), strongly arguing in
favor of an odor-specific and activity-dependent component of the
expression of Zif268 in the GrL.
We then analysed the overlap between these activation maps and
maps of BrdU-positive cell obtained in pseudo-conditioned and
conditioned animals (see Materials and methods). The overlap
between the BrdU-positive cell maps obtained in the group pseudo-
conditioned or conditioned with (+)L ⁄ (–)L was greater with the (+)L-
induced than with the PA-induced Zif268 map (Fig. 6D). We also
analysed the overlap between the BrdU-positive cell maps obtained in
the (+)L ⁄ PA groups and the (+)L- or PA-induced Zif268 maps. We
found that the pseudo-conditioned BrdU-positive cell map overlapped
more with the (+)L- than with the PA-induced Zif268 maps. In the
learning condition, the BrdU-positive cell map showed a similar
overlap with the (+)L- and PA-induced Zif268 maps (Fig. 6E).
Discussion
The present study addresses the influence of an olfactory discrimin-
ation task on the survival and spatial distribution of newborn neurons
in the MOB. We also look at the influence of the task difficulty on
learning and on the modulation of neurogenesis.
We first focused on the characterization of the discrimination
performances of adult mice. Despite the fact that the two enantiomers
of limonene were not spontaneously discriminated, adult mice could
learn to discriminate them as shown by their ability to perform the
associative task when (+)L and (–)L had to be discriminated.
The learning curve obtained in the group conditioned with the pair
of related odorants exhibited a 4-day lag compared with the learning
curve obtained when two distinct odorants were used. This time lag
was probably due to the poor ability of the animals to effectively
discriminate the two enantiomers of limonene, which are perceptually
similar.
Neurogenesis is a complex process, which includes proliferation,
differentiation and survival or death of the newborn cells (Alvarez-
Buylla & Garcia-Verdugo, 2002). In the present study, the behavioral
procedure took place 30 days after labeling of dividing cells with
BrdU. This time window was chosen because in a previous study
(Mandairon et al., 2006), we have shown that adult-born neurons
survival is dependent upon sensory inputs from 15 to 45 days after
birth (Mandairon et al., 2006). In addition, this paradigm of BrdU
administration prevents any effect of conditioning or pseudo-condi-
tioning on proliferation or migration of labeled cells. Indeed, although
round divisions of previously marked progenitors occur during the
days following DNA synthesis marker administration, changes in the
number of BrdU-positive cells induced 30 days post-BrdU can only
reflect changes in survival and not in proliferation rate or migration of
labeled progenitors along the rostral migratory stream, which is
achieved in 10–15 days (Lois & Alvarez-Buylla, 1994; Carleton et al.,
2003). Our data, however, do not exclude an effect of learning on

Fig. 5. Clustered distribution of BrdU-positive cells in the MOB and learning-induced changes. (A) Mapping of BrdU-positive cells in the GrL of the MOB.
BrdU-positive cells were counted in the GrL on every fifth section of the MOB (14 lm thickness, sampling interval ¼ 70 lm). The GrL was divided into 36 sectors
of 10  with the reference axis drawn on the ventral aspect of the subependymal layer of the MOB. (B) The volume of the GrL was represented by an array where
each bin has the value of BrdU-positive cells density in one 10 -sector, and each column corresponds to one section. Individual maps were z-scored and
comparisons were performed by t-tests (see Materials and methods). (C) Example of BrdU-positive cells in the GrL. (D and E) Averaged maps of BrdU-positive cells
density in the two pseudo-conditioned (PC) groups, exposed, respectively, to (+)limonene (L) ⁄ (–)L or (+)L ⁄ propionic acid (PA) showing a clustered distribution of
BrdU-positive cells. The scale gives the normalized (z-scored) density of BrdU-positive cells. (F) The comparison by t-tests of the two pseudo-conditioned maps
presented in (C) and (D) reveals significant differences in the location of clusters of BrdU-positive cells. The right scale gives the P-values. Significant differences
due to high-density foci present in the (+)L ⁄ (–)L map and not in the (+)L ⁄ PA map appear as dark blue areas. The opposing situation appears in red. (G and H)
Averaged maps of BrdU-positive cell density after conditioning (C) the (+)L ⁄ (–)L (G) or (+)L ⁄ PA (H). The scale gives the normalized (z-scored) density of BrdU-
positive cells. (I) Comparison by t-tests of the two conditioned maps presented in (G) and (H) showing significant differences in the location of clusters of BrdU-
positive cells. The right scale gives the P-values. (J and K) Locations of conditioning-induced changes given by subtraction of the pseudo-conditioned map from the
conditioned map (C ) PC) for the (+)L ⁄ (–)L pair (J) and the (+)L ⁄ PA pair (K). Changes for both pairs of odorants include new high-density foci (red areas) and
suppression of high-density foci (dark blue areas). The scale gives the values of the differences between the conditioned and the pseudo-conditioned maps.
(L) Conditioning-induced changes in BrdU-positive cell density. The index of conditioning-induced changes shows more new high-density foci (induction) of
BrdU-positive cells in animals conditioned with the pair of related odorants (+)L ⁄ (–)L, while suppression of high-density foci is higher in animals conditioned with
the pair of different odorants (+)L ⁄ PA. **P < 0.01, ***P < 0.0001, t-test for comparison of proportions.

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
3586 N. Mandairon et al.

Fig. 6. Zif268 expression maps in response to odorant stimulation partly overlap with BrdU-positive cell maps. (A) Map of Zif268 expression 1 h after
stimulation by (+)limonene (L). (B) Map of Zif268 expression 1 h after stimulation by PA. (C) Comparison of the (+)L- and propionic acid (PA)-induced Zif268
expression maps by t-test. The right scale gives the P-values. (D) Overlap of high-density foci of the BrdU maps obtained in the pseudo-conditioned (PC) or
conditioned (C) groups exposed to (+)L ⁄ (–)L, with (+)L-induced (gray bars) or PA-induced (black bars) Zif268 expression maps. ***P < 0.005, t-test for
comparison of proportions. (E) Same analysis as in (D) for the groups exposed to the (+)L ⁄ PA pair of odorants.

proliferation or migration of unlabeled progenitors (young cells
formed after BrdU administration).
For the learning task involving the (+)L ⁄ PA pair, newborn cell
density was lower in the conditioned group than the pseudo-
conditioned group (i.e. in animals exposed to the odorants with no
pairing of the reinforcement with one odorant of the pair). In contrast,
animals conditioned with the pair (+)L ⁄ (–)L did not show any
significant reduction in newborn neuron survival compared with the
pseudo-conditioned group. As shown by the behavioral data,
discrimination was more difficult for the pair of similar odorants than
for the pair of distinct odorants, resulting in a slower learning process
for the similar pair. A functional relationship between newborn cell
survival and learning is substantiated by the correlation found between
the behavioral performances in the course of the learning process, and
the rate of survival of newborn cells.
All together, these data suggest an effect of the associative
discrimination learning process on newborn GRs survival and the
influence of the perceptual proximity of olfactory cues used in the
discrimination task on these learning-induced changes in neurogen-
esis. One interpretation could be that more newborn cells are required
to learn the more difficult discrimination task. Thus, a better survival
of newborn cells may be critical for shaping MOB responses during a
complex discrimination task. This is consistent with the report that any
condition reducing the number of newborn neurons in the MOB alters
olfactory discrimination ability (Gheusi et al., 2000; Enwere et al.,
2004).
In the hippocampus, learning has been shown to increase or
decrease the survival of newborn cells depending on the age of the
cells monitored (Dobrossy et al., 2003; Ambrogini et al., 2004; Leuner
et al., 2004). It has been proposed that learning favors the survival of
recently born cells, while older cells undergo increased death. Our data
from the MOB are consistent with this view of the complex influence
of learning on neuronal turnover. Therefore, an alternative interpret-
ation of our data could be that the reduced survival of 30-day-old
neurons observed following learning of the distinct odorants reflects
an increased neuronal turnover that could facilitate learning.
The measure of the time spent sniffing the odors revealed that
within each pair the two odorants were similarly explored by pseudo-
conditioned animals, suggesting they have similar levels of appetence.
In the two groups of conditioned animals, as can be expected, the
learned odor was more explored. More surprising was the finding that
the sniffing times were higher in the group learning the different pairs
of odorants than in the group learning the similar pair. Although the
reason for this difference remains unclear, a putative explanation could
be that the easier task creates a higher level of motivation to do the
task, inducing more time spent exploring the odors. Because olfactory
stimulation influences newborn cell survival, these data raised the
question of the influence of these differences in exploration time on
the rate of neuronal turnover. Such influence is not supported by the
absence of correlation between the time spent exploring the odors and
the rate of neurogenesis. In addition, it is worth noting that only life-
long (Petreanu & Alvarez-Buylla, 2002) or long-term (several weeks)
odor deprivation (Mandairon et al., 2003, 2006) or stimulation
(Rochefort et al., 2002) have been reported to affect the rate of
neurogenesis, while in our study the exploration of odors was in the
order of magnitude of 1–3 s per session. In the context of our study, a
significant effect of the exploration time on neurogenesis cannot thus
be favored. Rather, our data support the hypothesis that short
exposures to odors may affect neurogenesis if they are associated
with a positive reinforcement.
Zif268 is an immediate-early gene (IEG) with neuronal expression,
driven by activity-dependent plasticity (Knapska & Kaczmarek,
2004). The expression of IEGs is widely used to map neural activity
in the brain, particularly in sensory systems, including the glomerular

ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588

layer of the MOB (Inaki et al., 2002), where it maps the odor-induced
pattern of activation. We found that Zif268 expression in naive
animals was also a reliable marker to map odor-related activity in the
GrL. We then reasoned that if BrdU-positive cells clustered in areas of
the GrL were involved in processing the odors to which the animals
were exposed, the BrdU-positive cell map resulting from pseudo-
conditioning or conditioning with the pair (+)L ⁄ (–)L should show
more overlap with the (+)L-induced than with the PA-activation maps,
assuming that (+)L and (–)L elicit very similar bulbar patterns of
activation (Linster et al., 2001). We found this to be exactly the case,
providing support to the hypothesis of a regulation of neurogenesis in
areas of odor activation. For BrdU maps obtained from animals
conditioned with the pair (+)L ⁄ PA, our prediction was that they would
have similar overlap with the (+)L- and the PA-induced Zif268
expression pattern. This was true for conditioned but not pseudo-
conditioned animals. This discrepancy may arise from overlap and ⁄ or
interactions of activity-driven survival induced by the two dissimilar
odorants (Luo & Katz, 2001; Linster & Cleland, 2004), obscuring the
relationships of the BrdU-positive cell map with the Zif268 maps of
each single odor. Finally, the partial overlap between Zif268- and
BrdU-positive cell maps also suggests that by recruiting new
inhibitory interneurons to the OB network, learning may induce
changes in the activation maps.
The functional implications of changes in the number and spatial
distribution of newborn cells might be important for olfactory coding
and learning. It has been proposed that olfactory coding relies on the
temporal coherence of spatially distributed activity in the MOB,
largely under the control of GRs (Laurent et al., 2001; Lledo et al.,
2005), and it is important to note that the spatio-temporal output of the
MOB is affected by learning (Kay & Laurent, 1999; Fletcher &
Wilson, 2003; Ravel et al., 2003; Martin et al., 2004). Any change in
the set of GRs responding to olfactory stimulation may thus be
expected to affect the functional image of the odorant within the
MOB. The learning-induced changes in neurogenesis reported here
provide a cellular substrate for modulating odorant representation
within the MOB, which should be considered in further studies of
bulbar functional plasticity.
Acknowledgements
This study was supported by CNRS. N.M. was supported by a doctoral grant
from the ‘Fondation de France’ (Fondation Roudnitska). We are grateful to Drs
C. Linster and S. Laroche for their critical reading of the manuscript. We thank
W. Lipski for correcting the English language of the paper.
Abbreviations
BrdU, bromodeoxyuridine; BSA, bovine serum albumin; GRs, granule cells;
GRL, granule cell layer; IEG, immediate-early gene; L, limonene; MCs, mitral
cells; MOB, main olfactory bulb; PA, propionic acid; SVZ, subventricular zone.
References
Alvarez-Buylla, A. & Garcia-Verdugo, J. (2002) Neurogenesis in adult
subventricular zone. J. Neurosci., 22, 629–634.
Ambrogini, P., Orsini, L., Mancini, C., Ferri, P., Ciaroni, S. & Cuppini, R.
(2004) Learning may reduce neurogenesis in adult rat dentate gyrus.
Neurosci. Lett., 359, 13–16.
Belluzzi, O., Benedusi, M., Ackman, J. & LoTurco, J. (2003) Electrophysio-
logical differentiation of new neurons in the olfactory bulb. J. Neurosci., 23,
10411–10418.
Buck, L.B. (1996) Information coding in the vertebrate olfactory system. Annu.
Rev. Neurosci., 19, 517–544.
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience, 24, 3578–3588
Neurogenic correlates of an olfactory discrimination task 3587
Carlen, M., Cassidy, R., Brismar, H., Smith, G., Enquist, L. & Frisen, J. (2002)
Functional integration of adult-born neurons. Curr. Biol., 12, 606–608.
Carleton, A., Petreanu, L., Lansford, R., Alvarez-Buylla, A. & Lledo, P. (2003)
Becoming a new neuron in the adult olfactory bulb. Nat. Neurosci., 6, 507–518.
Cecchi, G., Petreanu, L., Alvarez-Buylla, A. & Magnasco, M. (2001)
Unsupervised learning and adaptation in a model of adult neurogenesis.
J. Comput. Neurosci., 11, 175–182.
Dobrossy, M., Drapeau, E., Aurousseau, C., Le Moal, M., Piazza, P. & Abrous,
D. (2003) Differential effects of learning on neurogenesis: learning increases
or decreases the number of newly born cells depending on their birth date.
Mol. Psychiatry, 8, 974–982.
Enwere, E., Shingo, T., Gregg, C., Fujikawa, H., Ohta, S. & Weiss, S. (2004)
Aging results in reduced epidermal growth factor receptor signaling,
diminished olfactory neurogenesis, and deficits in fine olfactory discrimina-
tion. J. Neurosci., 24, 8354–8365.
Fletcher, M. & Wilson, D. (2003) Olfactory bulb mitral-tufted cell plasticity:
odorant-specific tuning reflects previous odorant exposure. J. Neurosci., 23,
6946–6955.
Gheusi, G., Cremer, H., McLean, H., Chazal, G., Vincent, J. & Lledo, P. (2000)
Importance of newly generated neurons in the adult olfactory bulb for odor
discrimination. Proc. Natl Acad. Sci. USA, 97, 1823–1828.
Inaki, K., Takahashi, Y., Nagayama, S. & Mori, K. (2002) Molecular-feature
domains with posterodorsal-anteroventral polarity in the symmetrical sensory
maps of the mouse olfactory bulb: mapping of odourant-induced Zif268
expression. Eur. J. Neurosci., 15, 1563–1574.
Kauer, J.S. & White, J. (2001) Imaging and coding in the olfactory system.
Annu. Rev. Neurosci., 24, 963–979.
Kay, L. & Laurent, G. (1999) Odor- and context-dependent modulation of
mitral cell activity in behaving rats. Nat. Neurosci., 2, 1003–1009.
Kempermann, G., Wiskott, L. & Gage, F. (2004) Functional significance of
adult neurogenesis. Curr. Opin. Neurobiol., 14, 186–191.
Knapska, E. & Kaczmarek, L. (2004) A gene for neuronal plasticity in the
mammalian brain: Zif268 ⁄ Egr-1 ⁄ NGFI-A ⁄ Krox-24 ⁄ TIS8 ⁄ ZENK? Prog.
Neurobiol., 74, 183–211.
Kornack, D. & Rakic, P. (2001) The generation, migration, and differentiation
of olfactory neurons in the adult primate brain. Proc. Natl Acad. Sci. USA,
98, 4752–4757.
Korsching, S. (2002) Olfactory maps and odor images Curr. Opin. Neurobiol.,
12, 387–392.
Laurent, G., Stopfer, M., Friedrich, R.W., Rabinovich, M.I., Volkovskii, A.
& Abarbanel, H.D. (2001) Odor encoding as an active, dynamical
process: experiments, computation, and theory. Annu. Rev. Neurosci., 24,
263–297.
Leuner, B., Mendolia-Loffredo, S., Kozorovitskiy, Y., Samburg, D., Gould, E.
& Shors, T. (2004) Learning enhances the survival of new neurons beyond
the time when the hippocampus is required for memory. J. Neurosci., 24,
7477–7481.
Linster, C. & Cleland, T.A. (2004) Configurational and elemental odor mixture
perception can arise from local inhibition. J. Comput. Neurosci., 16, 39–47.
Linster, C., Johnson, B., Yue, E., Morse, A., Xu, Z., Hingco, E., Choi, Y.,
Choi, M., Messiha, A. & Leon, M. (2001) Perceptual correlates of neural
representations evoked by odorant enantiomers. J. Neurosci., 21, 9837–9843.
Lledo, P., Gheusi, G. & Vincent, J. (2005) Information processing in the
mammalian olfactory system. Physiol. Rev., 85, 281–317.
Lois, C. & Alvarez-Buylla, A. (1994) Long-distance neuronal migration in the
adult mammalian brain. Science, 264, 1145–1148.
Lowe, G. (2003) Electrical signaling in the olfactory bulb. Curr. Opin.
Neurobiol., 13, 476–481.
Luo, M. & Katz, L. (2001) Response correlation maps of neurons in the
mammalian olfactory bulb. Neuron, 32, 1165–1179.
Magavi, S., Mitchell, B., Szentirmai, O., Carter, B. & Macklis, J. (2005) Adult-
born and preexisting olfactory granule neurons undergo distinct experience-
dependent modifications of their olfactory responses in vivo. J. Neurosci.,
25, 10729–10739.
Mandairon, N., Jourdan, F. & Didier, A. (2003) Deprivation of sensory inputs
to the olfactory bulb up-regulates cell death and proliferation in the
subventricular zone of adult mice. Neuroscience, 119, 507–516.
Mandairon, N., Sacquet, J., Jourdan, F. & Didier, A. (2006) Long-term fate and
distribution of newborn cells in the adult mouse olfactory bulb: influences of
olfactory deprivation. Neuroscience, 1, 443–451.
Martin, C., Gervais, R., Hugues, E., Messaoudi, B. & Ravel, N. (2004)
Learning modulation of odor-induced oscillatory responses in the rat
olfactory bulb: a correlate of odor recognition? J. Neurosci., 24, 389–397.
Nottebohm, F. (2002) Neuronal replacement in adult brain. Brain Res. Bull., 57,
737–749.
3588 N. Mandairon et al.
Petreanu, L. & Alvarez-Buylla, A. (2002) Maturation and death of adult-born
olfactory bulb granule neurons: role of olfaction. J. Neurosci., 22, 6106–
6113.
Ravel, N., Chabaud, P., Martin, C., Gaveau, V., Hugues, E., Tallon-Baudry,
C., Bertrand, O. & Gervais, R. (2003) Olfactory learning modifies the
expression of odour-induced oscillatory responses in the gamma (60–90
Hz) and beta (15–40 Hz) bands in the rat olfactory bulb. Eur. J. Neurosci.,
17, 350–358.
Rochefort, C., Gheusi, G., Vincent, J. & Lledo, P. (2002) Enriched
odor exposure increases the number of newborn neurons in the
adult olfactory bulb and improves odor memory. J. Neurosci., 22,
2679–2689.
Rubin, B.D. & Katz, L.C. (1999) Optical imaging of odorant representations in
the mammalian olfactory bulb. Neuron, 23, 499–511.
Schaefer, M., Young, D.A & Restrepo, D. (2001a) Olfactory fingerprints for
major histocompatibility complex-determined body odors. J. Neurosci., 21,
2481–2487.
ª The Authors (2006). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd
Schaefer, M.L., Young, D.A. & Restrepo, D. (2001b) Olfactory fingerprints for
major histocompatibility complex-determined body odors. J. Neurosci., 21,
2481–2487.
Schoppa, N. & Urban, N. (2003) Dendritic processing within olfactory bulb
circuits. Trends Neurosci., 26, 501–506.
Shepherd, G. & Greer, C. (1990) Olfactory bulb. In Shepherd, G.M. (Ed.), The
Synaptique Organization of the Brain, Oxford University Press, New York,
pp. 133–169.
Shors, T., Townsend, D., Zhao, M., Kozorovitskiy, Y. & Gould, E. (2002)
Neurogenesis may relate to some but not all types of hippocampal-dependent
learning. Hippocampus, 12, 578–584.
Winner, B., Cooper-Kuhn, C., Aigner, R., Winkler, J. & Kuhn, H. (2002) Long-
term survival and cell death of newly generated neurons in the adult rat
olfactory bulb. Eur. J. Neurosci., 16, 1681–1689.
Yamaguchi, M. & Mori, K. (2005) Critical period for sensory experience-
dependent survival of newly generated granule cells in the adult mouse
olfactory bulb. Proc. Natl Acad. Sci. USA, 102, 9697–9702.
European Journal of Neuroscience, 24, 3578–3588