points.It should be possible,if we knew where to look,to find neural
changes that correspond to the duration of the sensitive period.Song-
birds have a specialized network of neurons in their brains dedicated to
song learning and production (Nottebohm,Stokes,and Leonard 1976).
The cell bodies of these neurons are organized into a series of clusters
referred to as song control nuclei.The axons of these neurons project to
adjacent song control nuclei to form synaptic connections.Lesion studies,
electrophysiological recordings,and histological studies reveal that these
nuclei form two circuits,an anterior forebrain pathway involved in song
learning,and a posterior motor pathway involved in song production
(reviewed by Brenowitz and Kroodsma 1996).
Nuclei of the anterior forebrain pathway undergo several changes
during the period of song learning.Studies of zebra finches and canaries
reveal that new neurons are added to the nucleus referred to as the
higher vocal center (HVC),suggesting that these new neurons may
encode new memories acquired by listening to the songs of others
(Sohrabji,Nordeen,and Nordeen 1993;Kirn et al.1994).This observa-
tion suggests that learning ceases when neurogenesis is complete (Not-
tebohm 1981). In other areas of the anterior forebrain pathway,
connections between nuclei appear to decrease during song learning by
reducing the number of synapses (Herrmann and Arnold 1991),decreas-
ing the number of receptors for chemicals used to communicate across
the synapse (NMDA receptors;Aamodt et al.1992),or decreasing the
number of dendritic spines where synapses are formed (Wallhausser-
Franke,Nisdorf-Bergweiler,and DeVoogd 1995).These observations
indicate a different mechanism from the one mentioned above;namely,
learning involves simplifying connectivity between neurons (Changeux
and Danchin 1976).This model suggests that unused synapses are elim-
inated during song learning,thus paring down the initial network of con-
nections.Further learning would not be possible once synapse selection
was complete.A similar model was proposed for imprinting in chickens
(Wallhausser and Scheich 1987).
Before we can begin to understand the neural basis of song learning
and regulation of the sensitive period,we must distinguish changes in the
brain that are a consequence of maturation from those that are func-
tionally related to learning,regardless of age.The critical question is
whether the neural changes described above would be observed if birds
were prevented from learning to sing.This experiment was done in two
ways.One method compared the neural development of deafened and
hearing birds of the same age (Aamodt,Nordeen,and Nordeen 1995;
Burek,Nordeen,and Nordeen 1991).The other compared the brains of
tutored birds that had begun practicing song with those of song-deprived
birds of the same age that had not begun to sing (Wallhausser-Franke,68 Carol WhalingMUS5 9/14/99 11:59 AM Page 68