Involvement of newborn neurons in olfactory associative learning? The operant or non-operant component of the task makes all the difference

Abstract

New neurons are continuously generated in the adult mammalian olfactory bulb. The role of these newborn neurons in olfactory learning has been debated. Blocking the addition of neurons has been reported either to result in memory alteration or to have no effect at all (Imayoshi et al., 2008; Breton-Provencher et al., 2009; Lazarini et al., 2009; Sultan et al., 2010). These discrepancies may have arisen from differences in the behavioral paradigms used: operant procedures indicated that neurogenesis blockade had substantial effects on long-term memory (Lazarini et al., 2009; Sultan et al., 2010) whereas other methods had little effect (Imayoshi et al., 2008; Breton-Provencher et al., 2009). Surprisingly, while operant learning is known to modulate the survival of new neurons, the effect of non-operant learning on newborn cells is unknown. Here we use mice to show that compared with operant learning, non-operant learning does not affect cell survival, perhaps explaining the current controversy. In addition, we provide evidence that distinct neural substrates at least partly underlie these two forms of learning. We conclude that the involvement of newborn neurons in learning is subtly dependent on the nature of the behavioral task.

Figure 1.

Experimental design. A, Operant conditioning. Ai, Two groups of animals (conditioned and pseudo-conditioned; n = 10 per group) were trained for 5 d. BrdU was administered 13 d before training to label newborn cells arriving in the OB at the beginning of training. All mice were killed on day 5, 1 h after the last trial. Aii, Experimental apparatus used for operant conditioning. B, Non-operant conditioning. Bi, Two groups of animals (conditioned and pseudo-conditioned; n = 10 per group) were trained for 4 d and tested for learning on day 5, 1 h before they were killed. Again, BrdU was administered 13 d before training. Bii, Experimental apparatus used for non-operant conditioning.

Figure 2.

Only operant conditioning modulates newborn cell survival. A, Operant conditioning. Ai, Behavioral performances. Aii, BrdU-positive cell density in the granule cell layer is increased in the conditioned compared with the pseudo-conditioned group. Aiii, Counts of BrdU+/Zif268+ cells indicate that newborn cells are functionally more solicited in the conditioned than the pseudo-conditioned group. Aiv, The overall expression of Zif268 in the granule cell layer is similar between the conditioned and pseudo-conditioned groups. *p < 0.05, ***p < 0.005 for difference between conditioned and pseudo-conditioned. B, Non-operant conditioning. Bi, Behavioral performances. Bii, BrdU-positive cell density in the granule cell layer is not modulated by conditioning. Biii, The percentage of BrdU+/Zif268+ cells is similar between the conditioned and pseudo-conditioned groups. Biv, The overall expression of Zif268 in the granule cell layer is similar between the conditioned and pseudo-conditioned groups (***p < 0.005). C, Conditioned group; PC, pseudo-conditioned group. All cell densities are cell/μm². Zif268-positive cell density in orbital cortex (Ci), infralimbic cortex (Cii), and prelimbic cortex (Ciii). *p < 0.05.