In adulthood, songbird species vary considerably in the extent to which the
y rely on auditory feedback to maintain a stable song structure. The contin
ued recruitment of new neurons into vocal motor circuitry may contribute to
this lack of resiliency in song behavior insofar as new neurons that are n
ot privy to auditory instruction could eventually corrupt established neura
l function. In a first step to explore this possibility, we used a comparat
ive approach to determine if species differences in the rate of vocal chang
e after deafening in adulthood correlate positively with the extent of HVc
neuron addition. We confirmed previous reports that deafening in adulthood
changes syllable phonology much more rapidly in bengalese finches than in z
ebra finches. Using [H-3]thymidine autoradiography to identify neurons gene
rated in adulthood, we found that the proportion of new neurons in the HVc
one month after labeling was nearly twice as great in bengalese than in zeb
ra finches. Moreover, among the subset of HVc vocal motor neurons that proj
ect to the robust nucleus of the archistriatum, the incidence of [H-3]thymi
dine-labeled neurons was nearly three times as great in bengalese than in z
ebra finches. This correlation between the proportion of newly added neuron
s and the rate of song deterioration supports the hypothesis that HVc neuro
n addition may disrupt stable adult song production if new neurons cannot b
e "trained" via auditory feedback. (C) 2000 John Wiiey & Sons, Inc. J Neuro
biol 43: 79-88, 2000.