Olfactory perceptual learning requires adult neurogenesis

Abstract

Perceptual learning is required for olfactory function to adapt appropriately to changing odor environments. We here show that newborn neurons in the olfactory bulb are not only involved in, but necessary for, olfactory perceptual learning. First, the discrimination of perceptually similar odorants improves in mice after repeated exposure to the odorants. Second, this improved discrimination is accompanied by an elevated survival rate of newborn inhibitory neurons, preferentially involved in processing of the learned odor, within the olfactory bulb. Finally, blocking neurogenesis before and during the odorant exposure period prevents this learned improvement in discrimination. Olfactory perceptual learning is thus mediated by the reinforcement of functional inhibition in the olfactory bulb by adult neurogenesis.

Fig. 1.

Olfactory enrichment improves discrimination. (A) Design of the experiment. Spontaneous discrimination between +/−limonene, pentanol/butanol, and decanal/dodecanone was tested before and after an odor enrichment period. Experimental groups were enriched by introducing +/−limonene or decanal/dodecanone into the home cage for 1-h periods over 10 days. A control group was not enriched (no odor). (B) Behavioral discrimination before (Bi) and after (Bii-Biv) enrichment. The two odorants of each pair are confused before the enrichment period (Bi), and in the control nonenriched group (Bii). After enrichment with +/−limonene, +/−limonene and pentanol/butanol are discriminated (Biii), and after enrichment with decanal/dodecanone, only this pair of odorants is discriminated (Biv). [*, P < 0.05, ***, P < 0.0001, in response magnitude between trials 4 (Hab4) and 5 (Otest)]. The data are expressed as mean values ± SEM.

Fig. 2.

Olfactory enrichment improves the survival of newborn cells involved in odor processing. (A) Experimental paradigm. BrdU was administered 8 days before the enrichment period and mice were killed 25 days after administration of BrdU. Animals were exposed to + limonene or decanal 1 h before sacrifice. (B) Olfactory enrichment induces an increase of BrdU-positive cell density in the granule (GrL) but not glomerular (GL) cell layers of the OB (*, P < 0.05). (C) Quantification of BrdU/NeuN double-labeling in the GrL showed no effect of the enrichment on the neuronal fate of newborn cells. (D) By using BrdU/Calbindin double-labeling in the GL, no effect of the enrichment on the phenotype of newborn cells in the GL was found. (E) Enrichment with +/−limonene increases the percentage of newborn neurons responding to +limonene compared to newborn neurons responding to decanal or to newborn neurons that respond to +limonene in control non enriched animals. (F) Similarly, enrichment with decanal/dodecanone increases the percentage of newborn neurons responding to decanal compared to newborn neurons responding to +limonene or to newborn neurons that respond to decanal in control non enriched animals. (*, P < 0.05). The data are expressed as mean values ± SEM.

Fig. 3.

Olfactory enrichment increases inhibition in the olfactory bulb. (A) Example of GAD65/67 immunolabeling in the OB. (B) Enrichment with +/−limonene or decanal/dodecanone increases the expression of GAD65/67 in the granule cell layer of the OB compared to control nonenriched animals. (**, P < 0.001 for difference from control group). (C) Paired-pulse stimulation. The graph shows an example of population EPSPs in the granule cell layer in response to two LOT stimulations separated by 20 ms. The response to the second pulse is significantly decreased. (D) Paired-pulse inhibition quantified by the ratio between the rising slopes of responses to the second and first LOT pulse in control and +/−limonene enriched mice. The data are expressed as mean values ± SEM.

Fig. 4.

Neurogenesis is necessary for perceptual learning. (A) Experimental design. Saline or AraC was locally infused 3 days before the administration of BrdU and lasting for 21 days. (B) +/−limonene enriched mice that received AraC have a significant reduction of BrdU-positive cell density in the granule cell layer of the OB. (C) The strong reduction of bulbar neurogenesis in the AraC group blocks the enrichment-induced improvement of discrimination that occurs in the saline group (*, P < 0.05; **, P < 0.001, ***, P < 0.0001). (D) The expression of GAD65/67 in the granule cell layer of the OB is decreased in the AraC group compared to the saline group. (*, P < 0.05; **, P < 0.001; ***, P < 0.0001). The data are expressed as mean values ± SEM.