Adult neurogenesis is associated with the maintenance of a stereotyped, learned motor behavior

Abstract

Adult neurogenesis is thought to provide neural plasticity used in forming and storing new memories. Here we show a novel relationship between numbers of new neurons and the stability of a previously learned motor pattern. In the adult zebra finch, new projection neurons are added to the nucleus HVC and become part of the motor pathway for producing learned song. However, new song learning occurs only in juveniles and the behavioral impact of adding new neurons to HVC throughout life is unclear. We report that song changes after deafening are inversely correlated with the number of new neurons added to HVC, suggesting that adult neurogenesis in this context may contribute to behavioral stability. More broadly, we propose that new neuron function may depend on the site of integration and can vary as widely as promoting, or restricting, behavioral plasticity.

Figure 1.

A, Spectral derivatives of the songs of three birds, generated with Sound Analysis Pro. The song of hearing control Bird 1 does not change in song structure 4 weeks after the initial recording (left). The song of Bird 2 shows a low degree of song degradation 4 weeks after deafening (middle). The song of Bird 3 demonstrates a high degree of post-deafening song degradation at 4 weeks (right). Song notes are numbered below the figures to show same notes across recordings. Accuracy scores are a measure of the degree to which post-4 week songs are acoustically the same as the pre-deaf songs. B–E, New neuron number was positively correlated with song stability measured as similarity (B), and accuracy (C) between pre-deaf and post-deaf songs. Post-deaf declines in similarity (D) and accuracy (E) were inversely correlated with new neuron numbers.

Figure 2.

A–C, Post-deafening day 27 changes in the song features Weiner entropy (A), entropy variance (B), and pitch (C) were negatively correlated with new neuron numbers. The zero point in all graphs represents no change between pre- and post-deaf song feature.

Figure 3.

A, There was an inverse trend between new neurons and pre-deaf singing. We found no correlations between new neurons and singing rates after deafening at any time point sampled, or throughout the entire post-deaf period. B, C, Shown are new neurons corresponding to singing rates on post-deaf days 14 (B) and 27 (C).

Figure 4.

There was no difference in new neuron number between age-matched deafened birds and hearing controls.