University of Southern California Dissertations and Theses

Neural mechanisms of sensorimotor learning in cortico-basal ganglia pathways

(USC Thesis Other)

Neural mechanisms of sensorimotor learning in cortico-basal ganglia pathways

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
NEURAL MECHANISMS OF SENSORIMOTOR LEARNING IN
CORTICO-BASAL GANGLIA PATHWAYS

BY
Jennifer McGrady Achiro

A dissertation submitted in partial fulfillment
of the requirements for the degree of
DOCTOR OF PHILOSOPHY IN NEUROSCIENCE
at the
UNIVERSITY OF SOUTHERN CALIFORNIA
2013

Dedication

To Ricky
ii
Acknowledgements

I would like to thank my advisor, Sarah Bottjer, for her dedicated and caring mentorship
throughout my time as a PhD student.
Thank you to my past and present dissertation committee members Judith Hirsch, Bartlett Mel,
Dion Dickman, Louis Goldstein and Tansu Celikel for their support and valuable input.
I thank my collaborator on the work presented in chapter 3, John Shen, for his help with coding
and analysis.
Thank you to past and present lab members Vanessa Miller-Sims, Priscilla Logerot, Rachel
Yuan, Amy Garrison and John Shen for stimulating scientific discussions as well as support as
friends.
Thank you to the caring support staff Gloria Wan, Beatriz Gil, Vanessa Clark, and Mallory
Redel.
Thanks to the Neuroscience Graduate Program and the Hearing and Communication
Neuroscience Training Program for devoting time and resources into my training.
Special thanks to my husband, Ricky Achiro, who deserves a PhD-spouse award for his
unending encouragement and love.
Thank you to my wonderful family who nurtured my ambitions and always inspires me.
Thanks to my great friends who I love and cherish.
This work was supported by NINDS grant NS 037547, NIDCD Training Fellowship DC 009975,
and NINDS Training Fellowship NS 073323.
iii
Table of Contents

Dedication ...................................................................................................................................... ii
Acknowledgements ...................................................................................................................... iii
Chapter 1: Introduction ................................................................................................................1
Introduction to birdsong ...............................................................................................................1
Structure and function of developing song-learning circuits .......................................................7
Insights from studies in adult songbirds .....................................................................................18
Insights from studies in mammalian cortico-basal ganglia circuits ...........................................24
Presented contributions ..............................................................................................................27
Chapter 2: Neural representation of a target auditory memory in a cortico-basal ganglia
pathway ........................................................................................................................................28
Abstract ......................................................................................................................................28
Introduction ................................................................................................................................29
Materials & Methods ..................................................................................................................33
Results ........................................................................................................................................43
Discussion ..................................................................................................................................72
Chapter 3: Chronic recordings in LMAN CORE and SHELL in singing juvenile birds .........80
Abstract ......................................................................................................................................80
Introduction ................................................................................................................................81
Materials & Methods ..................................................................................................................85
Results ........................................................................................................................................91
Discussion .................................................................................................................................113
Chapter 4: Conclusions .............................................................................................................118
References ...................................................................................................................................120
iv
Chapter 1: Introduction
INTRODUCTION TO BIRDSONG
For over 70 years the study of vocal learning in songbirds, pioneered by William Thorpe and
Peter Marler (Marler, 1956, Thorpe, 1958), has been transformative in the research of such areas
as adult neurogenesis, the role of hormones in development and behavior, and genetic factors
underlying complex behavior. Moreover, songbirds provide an ideal model for studying
mechanisms of sensorimotor integration and imitative motor learning. In a process analogous to
the acquisition of human speech, songbirds learn a specific vocal pattern by listening to adult
male “tutor” song during a sensitive period early in life (Doupe and Kuhl, 1999, Brainard and
Doupe, 2013). Like humans, songbirds learn to imitate those communication sounds by
vocalizing and using sensory feedback to compare incipient babbling to the memory of tutor
sounds (Konishi, 1965, Marler, 1970). This dissertation addresses one major unresolved
question with regard to mechanisms of vocal learning: what neural circuits carry out
comparisons of vocal feedback to the tutor sounds, and how is this comparison achieved?
Songbirds (order Passeriformes and suborder Oscines) are one of three groups of birds,
including hummingbirds and parrots, which have evolved the ability to learn the sounds used for
vocal communication during development. Indeed, vocal learning is a rare trait; evidence
suggests that humans, bats, cetaceans and elephants are the only other species to possess this
complex social behavior. Details of the vocal learning process differ between songbird species.
However in many species, including zebra finches (Taeniopygia guttata), after fledging young
male birds memorize the song of the adult male who feeds and socially interacts with them,
usually their biological father (Immelmann, 1969, Böhner, 1983, Eales, 1985, Clayton, 1987,
Böhner, 1990, Catchpole and Slater, 1995, Mann and Slater, 1995, Roper and Zann, 2006).
1
After this period of sensory learning (from ~20-35 days post hatch; dph; Fig. 1), young birds no
longer require the experience of hearing the tutor in order to learn to imitate the song (Konishi,
1965) (and above references). Young zebra finches begin to produce variable babbling sounds
(subsong) at approximately 35 dph, and after two months of practice as the birds reach adulthood
(90 dph), they are able to produce a stereotyped song motif that is a close imitation of the tutor
song. During the period of sensorimotor integration, auditory and likely other sensory feedback
is required for learning to occur (Konishi, 1965, Konishi, 2004, Wild, 2004, Ashmore et al.,
2008, Bottjer and To, 2012). Researchers have uncovered much about how adult song is
generated, however the neural mechanisms and substrates underlying the sensorimotor song
learning period remain obscure.

2

Figure 1. Timeline of vocal learning for humans and zebra finches. The sensory learning period
for humans likely begins at birth and continues until mature speech is achieved. The sensitive
period for sensory learning in zebra finches begins at approximately 20 dph (days post hatch),
and by 35 dph, young birds have had sufficient experience with the tutor to make a good
imitation of his song later in development. Zebra finches are “close-ended” learners, meaning
that they learn only one song early in life and continue to sing that song in a highly stereotyped
fashion throughout adulthood. Beginning at around six months, humans begin producing
babbling sounds and therefore enter the sensorimotor integration period. Young zebra finches
begin babbling, that is singing subsong, at approximately 35 dph. As sensorimotor integration
progresses, humans are able to produce words at around twelve months and zebra finches
produce relatively stable and recognizable syllables, called plastic song, at approximately 50
dph. By adulthood, zebra finches produce a stable stereotyped copy of the tutor song. Below the
timeline are photographs of zebra finches: newly hatched, approximately 25 dph, and adult,
from left to right. Figure after (Kuhl, 2004).
3
As for most animals, unlearned vocalizations of songbirds (such as certain vocal calls)
are produced by projections from midbrain regions to hindbrain centers (Simpson and Vicario,
1990, Fukushima and Aoki, 2002, Jurgens, 2002). However, song is controlled by a distinct set
of forebrain nuclei (in birds, neural components are generally organized into nuclei as opposed to
the laminar structure of mammalian cortex, however evidence points to remarkably similar
functional organization of cortical and subcortical circuits between the two classes (Karten,
1997, Reiner et al., 2005, Wang et al., 2010, Atoji and Karim, 2012)). In adults, a motor
pathway consisting of a projection from the cortical area HVC (high vocal center) to the
premotor cortical area, RA (robust nucleus of the arcopallium), is necessary and sufficient for the
production of song in adults (Nottebohm et al., 1982, Simpson and Vicario, 1990, Yu and
Margoliash, 1996, Hahnloser et al., 2002, Fujimoto et al., 2011) (Fig. 2). A separate population
of projection neurons in HVC innervates Area X in the basal ganglia (Nordeen and Nordeen,
1988). The cortico-basal ganglia circuit including Area X, the dorsolateral thalamus (DLM) and
the cortical nucleus LMAN (lateral magnocellular nucleus of the anterior nidopallium) is
required for song learning in juveniles and plasticity in adults (Bottjer et al., 1984, Sohrabji and
Nordeen, 1990, Scharff and Nottebohm, 1991, Warren et al., 2011). Increasing evidence
suggests that higher level auditory cortex is also crucial for song learning, not only as the conduit
for auditory feedback, but also as an important substrate for the tutor song memory (Terpstra et
al., 2004, Phan et al., 2006, London and Clayton, 2008, Gobes et al., 2010, Hahnloser and
Kotowicz, 2010) (Miller-Sims & Bottjer, in press).

4

Figure 2. Schematic of the song motor areas and cortico-basal ganglia circuit. One population
of neurons in the cortical region HVC sends an excitatory projection to striatal neurons in the
basal ganglia (Area X; contains both striatal and pallidal neurons). Another population of HVC
neurons sends an excitatory projection to the vocal motor cortex, RA, which then projects to
downstream motor circuits. Pallidal neurons in Area X send an inhibitory projection to the
thalamus, DLM. Projection neurons in DLM send an excitatory projection to the cortical
nucleus, LMAN, which outputs through an excitatory projection to RA and sends a collateral
axon back onto striatal neurons in Area X. The HVC-RA circuit is sufficient to drive song
production in adults. The LMAN-Area X-DLM circuit drives song production in very young
birds and is necessary for song learning in juveniles and song plasticity in adults. Abbreviations:
HVC, high vocal center; RA, robust nucleus of the arcopallium; LMAN, lateral magnocellular
nucleus of the anterior nidopallium; DLM, dorsolateral medial thalamus.
5
Work from our lab discovered that the LMAN-Area X-DLM cortico-basal ganglia
pathway is organized in separate parallel loops, reminiscent of mammalian basal ganglia loop
architecture (Johnson and Bottjer, 1992, Johnson et al., 1995, Iyengar et al., 1999, Miller-Sims
and Bottjer, 2012)(Fig. 3A). LMAN is composed of two sub-regions, a core and a surrounding
shell, which give rise to independent topographic pathways that traverse the forebrain in parallel
(Johnson and Bottjer, 1992, Iyengar et al., 1999, Pinaud and Mello, 2007, Person et al., 2008).
In adults, LMAN
CORE
projects to RA whereas LMAN
SHELL
projects to a region adjacent to RA
within motor cortex, AI
d
(dorsal intermediate arcopallium). Rather than innervating hindbrain
motor neurons as RA does, AI
d
forms recurrent feedback and feedforward loops including
projections through limbic and reward circuitry (Johnson et al., 1995, Bottjer et al., 2000). In
addition, coordination between the two hemispheres is critical for the production of song (Vu et
al., 1998, Schmidt et al., 2004, Long and Fee, 2008) but direct inter-hemispheric connections
between the basal ganglia loops are present only for the SHELL pathway (Fig. 3B) (Johnson et al.,
1995, Bottjer et al., 2000). Below, I review the current literature concerning the functional and
structural features of the LMAN circuits and describe evidence for the hypothesis that the
LMAN
SHELL
pathway is critical for the process of sensorimotor integration, in which the young
bird must compare self-generated vocal-related feedback to the memory of the tutor song.
Finally, I relate findings in song sensorimotor integration to the mammalian cortico-basal ganglia
literature.

6

Figure 3. Schematic of the LMAN
SHELL
circuits. A, Ipsilateral connections of the CORE (gray)
and SHELL (red) pathways. Both pathways make parallel connections from LMAN to basal
ganglia (Area X), to thalamus and back to LMAN. In adults, LMAN
CORE
outputs to RA, and
LMAN
SHELL
outputs to an adjacent region in motor cortex (AI
d
). In juveniles but not adults,
LMAN
CORE
projection neurons send a collateral axon into AI
d
. AI
d
projects to many regions,
including to the thalamus and VTA. B, Inter-hemispheric connections of the SHELL pathway.
AI
d
projects to both ipsi- and contra-lateral VTA. A portion of the amygdala projects to
contralateral LMAN
SHELL
, and receives a projection from ipsilateral LMAN
SHELL
. Dotted line
represents the midline. Abbreviations: LMAN, lateral magnocellular nucleus of the anterior
nidopallium; RA, robust nucleus of the arcopallium; AI
d
, dorsal intermediate arcopallium; DLM,
dorsolateral medial thalamus, DL, dorsolateral, VM, ventromedial; VTA, ventral tegmental area;
amyg, amygdala.

7
STRUCTURE AND FUNCTION OF DEVELOPING SONG-LEARNING CIRCUITS
Before the period of sensory learning (memorization of tutor song) occurs, the basic connections
of the cortico-basal ganglia pathways are in place (Johnson and Bottjer, 1992, Mooney and Rao,
1994, Foster and Bottjer, 1998). However, dramatic changes in the organization of vocal control
pathways occur during development. In many ways, the song control system presents an
extraordinary example of postnatal circuit development. During the sensory learning and early
sensorimotor integration period there is an increase in the size of HVC and Area X due to the
addition of new neurons (Bottjer et al., 1985, Bottjer et al., 1986, Nordeen and Nordeen, 1988).
Additionally, the size of LMAN
SHELL
grows dramatically during the sensory learning period and
regresses by the late sensorimotor integration period (Johnson and Bottjer, 1992, Johnson et al.,
1995); such growth suggests these neural systems are critical during this period of development.
Axonal refinement during the sensory learning period is equally dramatic. During this
period, the projection from LMAN
CORE
to RA undergoes major refinement of its topographic
organization through a process which is dependent on normal auditory experience (Iyengar et al.,
1999, Iyengar and Bottjer, 2002b, Miller-Sims and Bottjer, 2012). At this point, very few axons
from HVC have grown into RA, and indeed most will not enter RA until the period of
sensorimotor integration begins (Konishi and Akutagawa, 1985, Mooney and Rao, 1994, Foster
and Bottjer, 1998). A new discovery has revealed that RA-projecting LMAN
CORE
neurons send a
collateral axon into AI
d
during the sensory learning period but not in adulthood (Miller-Sims and
Bottjer, 2012), indicating that copies of signals from CORE to RA are transiently routed into the
SHELL pathway during learning. In addition, individual thalamic (DLM) axon arbors in LMAN
are pruned during the sensory learning period in LMAN
CORE
, and more dramatically in
8
LMAN
SHELL
(Iyengar and Bottjer, 2002a)(Fig. 4), indicating that this refinement may represent a
correlate of learning.

9

Figure 4. Refinement of thalamic axons in LMAN
SHELL
, from (Iyengar and Bottjer, 2002a,
Bottjer, 2004). Left panel shows single axon reconstructions of DLM
VM
neurons projecting to
LMAN
SHELL
in a 20 dph, 35 dph and an adult bird. The right panel shows illustrations of the
borders of LMAN
SHELL
and terminating DLM axons, demonstrating the growth and regression of
LMAN
SHELL
during development as well as the refinement of the spatial extent and branching of
its afferent connection. Scale bar = 200 µm.

10
Neural recordings in Area X of the CORE cortico-basal ganglia pathway during the
sensory learning period (20-35 dph) were made during playback of various auditory stimuli
(Doupe, 1997). During this stage, Area X neurons displayed no selectivity for the tutor song
over other adult songs, and these neurons also responded to white noise and tones. Such
unselective responses reflect immature auditory tuning; Area X neurons in adults show robust
selectivity for the bird’s own song (Doupe and Solis, 1997, Kojima and Doupe, 2007, Kojima
and Doupe, 2008). Although little is known of the specific functions of the LMAN circuits
during the sensory learning period, blockade of NMDA (N-methyl-D-aspartate) receptors in
LMAN (LMAN written without a subscript indicates manipulation likely included both CORE and
SHELL) during tutoring sessions prevented birds from producing an accurate copy of the tutor
song later in development (Basham et al., 1996). However, for that study the birds were isolated
and artificially tutored at ages older than natural sensory learning ages (32-52 dph), and so many
were likely also singing during this period, raising the possibility that it was the disruption of
LMAN during sensorimotor integration that resulted in poor learning.
During the sensorimotor integration period (~35-90 dph), young birds first produce
subsong, which consists of highly variable sounds akin to babbling in humans, and then around
50 dph begin singing “plastic song”, characterized by more spectrally structured and identifiable
syllables (Johnson et al., 2002, Derégnaucourt et al., 2004, Aronov et al., 2011) (Fig. 5).
Evidence suggests that subsong is driven by the LMAN
CORE
to RA projection; inactivation or
cooling of LMAN, but not HVC, disrupts subsong production (Aronov et al., 2008, Aronov et
al., 2011) and lesions of LMAN at this stage prevent song learning (Bottjer et al., 1984).
Recordings from DLM
DL
in birds producing subsong revealed that some of the thalamic neurons
that project to LMAN
CORE
exhibit increases in firing rate just prior to syllable onsets and rate
11
suppression prior to syllable offsets (Goldberg and Fee, 2012). Similarly, recordings from
LMAN
CORE
in birds producing subsong revealed that many neurons showed an increase in firing
before syllable onset (Aronov et al., 2008). This supports the idea that the CORE cortico-basal
ganglia pathway functions as a premotor driver of early vocalizations.
12

Figure 5. Song spectrograms of zebra finch song throughout development, from (Bolhuis and
Gahr, 2006). Top row shows spectrogram of the tutor song. The fifth row up to the second row
show spectrograms of the developing song of the son of the tutor, from subsong at 40 dph, to
plastic song at 60 and 80 dph, and then stereotyped adult song at 100 dph.

13
In contrast to the motor role of the CORE pathway during subsong, lesions of AI
d
of the
SHELL pathway in juveniles did not produce an immediate disruption of song or any long-term
effect on syllable phonology. However, AI
d
-lesioned birds were eventually unable to copy tutor
song syllables or produce stable song sequences (Bottjer and Altenau, 2009). The lack of
immediate effect followed by an eventual impairment of imitative learning suggests that the
SHELL pathway may be involved in evaluating or instructing motor learning based on comparing
self-generated feedback to the tutor memory. In addition, the presence of LMAN
CORE
neurons’
collateral axons to AI
d
(Miller-Sims and Bottjer, 2012) (Fig. 6) indicates that the SHELL pathway
receives an efference copy of the subsong motor signal, which could be used in vocal evaluation.

14

Figure 6. Example of three individual RA-projecting LMAN
CORE
neurons from 35 dph birds
which send a collateral axon into AI
d
, from (Miller-Sims and Bottjer, 2012). Below outlines of
RA and AI
d
is an illustration of the tracer injection location within LMAN
CORE
for each traced
axon.
15
Later in sensorimotor integration, when the young birds are singing plastic song,
evidence suggests that the motor driver of song gradually shifts from LMAN
CORE
to HVC.
Lesions of LMAN at these later ages causes some disruption in song imitation, but less severe
than lesions in birds producing subsong (Bottjer et al., 1984). However, lesions or inactivation
of HVC disrupt plastic song but do not abolish song production, such that vocalizations seem to
acquire the characteristics of subsong (Aronov et al., 2008, Aronov et al., 2011). This result
suggests that although LMAN
CORE
may not be the primary driver of plastic song, it may continue
to contribute variability characteristic of subsong to motor output. One idea is that LMAN
CORE

functions during this phase to inject variability into song for reinforcement learning (Doya and
Sejnowski, 2000, Kao and Brainard, 2006, Fiete et al., 2007, Fee and Goldberg, 2011). In
reinforcement learning, variability is required to explore possible actions that could lead to
reward, i.e. variable song patterns represent an exploration of motor space so that at some point a
vocalization will be made that is a good match to the tutor song memory and can be reinforced.
Indeed, songbirds have the ability to locally regulate variability in their songs during learning,
and do so according to how well a particular song element is learned (Ravbar, 2012).
Supporting the idea that LMAN
CORE
injects variability into the song motor program,
inactivation of LMAN during the plastic song phase resulted in a substantial reduction in
variability of song and fewer bursts in firing in RA (Ölveczky et al., 2005, Ölveczky et al.,
2011)( Fig. 7). In addition, gaps in firing of HVC neurons correlated with increased variability
in plastic song (Day et al., 2008). LMAN
CORE
neurons burst more during plastic song than during
a baseline period, however spike trains were highly variable with respect to time alignment in the
song (Ölveczky et al., 2005). In contrast, during this age, Area X striatal neurons’ activity
correlates well to specific (10 ms) time points in song (Goldberg and Fee, 2010), as does the
16
activity of DLM
DL
neurons, which no longer show robust syllable onset and offset responses in
plastic song birds as they did in subsong birds (Goldberg and Fee, 2012), suggesting that timing
information, presumably generated in HVC (Hahnloser et al., 2002), is preserved though the
basal ganglia and thalamus, but is not evident at the level of LMAN
CORE
. This may indicate that
LMAN
CORE
receives input that provides information regarding the time in the song, and can use
that input to locally direct variability to specific song elements by bursting in a variable manner
at different, but regulated points in the song. In reinforcement learning, such a role would be
deemed an experimenter/actor (Barto et al., 1983, Doya and Sejnowski, 2000, Fiete et al., 2007).

17

Figure 7. Spectrograms of plastic song before, during and after LMAN inactivation, from
(Ölveczky et al., 2005). Left and right panels show four iterations (rows) each of the production
of the syllables A-D. Because this bird was at the plastic song phase, the song is somewhat
variable across renditions. The middle pane shows four iterations of the song during LMAN
inactivation. The song has become significantly more stereotyped from rendition to rendition.
18
Other insights into the function of the LMAN circuits have come from tests of auditory
tuning throughout development. In anesthetized birds, neurons in LMAN
CORE
develop auditory
selectivity to each bird’s own song during the plastic song stage (Doupe, 1997, Solis and Doupe,
1997). Interestingly, many neurons in LMAN
CORE
respond to both the bird’s own song and the
tutor song, even if the songs are quite different acoustically (Solis and Doupe, 1997, 1999, 2000,
Yazaki-Sugiyama and Mooney, 2004), indicating that single neurons may receive input from
sources tuned to either the bird’s own song or the tutor song. Behavioral evidence discussed
above indicates that a neural substrate representing a “critic” evaluates motor output by
comparing it to the tutor song representation. Neurons tuned to the bird’s own song may provide
a representation of the current motor program (Doupe et al., 2004). Neurons tuned to the tutor
song could function as a critic if their responsiveness is proportional to how well the auditory
feedback of the bird’s own song matches the tutor song, and a strong response could be used to
generate a reinforcement signal. However, until now (see Chapter 2), no substantial population
of neurons tuned to the tutor song had been discovered.
INSIGHTS FROM STUDIES IN ADULT SONGBIRDS
Although there are potentially many differences in the structure and function of the LMAN
circuits in adult birds compared to juveniles, studies of song plasticity and behavioral state in
adults can provide valuable information concerning mechanisms of learning and behavior. Adult
birds continue to use sensory feedback from singing to maintain their stereotyped song:
deafening results in a slow degradation of song (Nordeen and Nordeen, 1992, Lombardino and
Nottebohm, 2000). However, if LMAN is lesioned at the time that deafening occurs, no
degradation of song behavior occurs (Brainard and Doupe, 2000, Nordeen and Nordeen, 2010),
indicating that LMAN circuits continue to be required for plasticity in adults.
19
During singing, neural activity in LMAN
CORE
of adult birds is more tightly temporally
aligned to song than in juveniles, however neuron firing remains somewhat variable across
iterations of song (Kao et al., 2008) (Fig. 8). In addition, LMAN lesions in adulthood do not
disrupt song, but rather result in a reduction in the already low variability of the stereotyped song
(Kao and Brainard, 2006, Hampton et al., 2009). These results indicate that neurons in
LMAN
CORE
continue to contribute to motor variability in adulthood. Interestingly, however,
microstimulation in adult LMAN
CORE
during singing induced specific directional changes in
fundamental frequency that depended on location in the nucleus (Kao et al., 2005), supporting
the notion that LMAN
CORE
has the ability to alter specific song features. Such a function could be
used in juvenile birds to regulate specific song features during learning.

20

Figure 8. LMAN
CORE
neural activity during singing in a juvenile and adult bird, from (Ölveczky
et al., 2005, Kao et al., 2008). A, Top shows a spectrogram of three syllables of a 67 dph bird’s
plastic song. Below is a raster plot showing times in which a single LMAN
CORE
neuron fired
during singing. Scale bar = 100 ms. B, Top shows the spectrogram of an adult bird’s
stereotyped song, and below is the corresponding raster plot of a single LMAN
CORE
neuron’s
activity during singing. Scale bar = 100 ms.
21
Adult birds produce song in two contexts: in a performance state, singing to a female in
courtship (directed singing), and in a practice state, singing not directed to a female (undirected
singing) (Sossinka and Böhner, 1980, Zann, 1996). Directed singing is characterized by an
increased tempo and a reduction in variability in the song. The increased tempo of directed song
appears to be produced through increased heart rate and temperature (Cooper and Goller, 2006,
Aronov and Fee, 2012), however the increased variability of undirected song requires a
functional LMAN
CORE
circuit, in that the song of LMAN-lesioned birds appears “directed” even
in the absence of a female (Kao and Brainard, 2006, Kao et al., 2008). These results support the
idea the adult song motor program is produced through the HVC-RA circuit, and variability is
introduced through the LMAN
CORE
circuit. In juveniles, if the LMAN circuits have more control
over song, it follows that the songs would be more variable. Indeed, older juveniles that are
exposed to responsive adult females are able to produce a stereotyped directed song which is
more developed in structure than is ever normally seen at those ages (Kojima and Doupe, 2011),
indicating that the juveniles were able to access a more mature song program, presumably
through a more HVC-dominated motor signal. It follows that sensory feedback appears to be
evaluated more when the bird is singing undirected song (and therefore utilizing LMAN circuits)
than when singing directed song (Sakata and Brainard, 2009).
Since adult birds continue to use sensory feedback to maintain their songs, with the
appropriate training, adult birds can be stimulated to alter their songs. Such methods represent
an important tool to investigate mechanisms underlying plasticity, and can, with the benefits of
experimental control, inform research aimed at understanding natural song learning in juveniles.
Initiated by work from Michael Brainard’s lab, it was discovered that birds could adaptively
change their song if the production of a particular syllable feature triggered either delayed or
22
altered auditory feedback, or an aversive stimulus such as a loud noise burst. Interestingly, birds
could alter specific features of their song, such as pitch to avoid the triggered stimulus (Sakata
and Brainard, 2006, Tumer and Brainard, 2007, Sober and Brainard, 2009), indicating that adults
were actively evaluating feedback during singing and adjusting the song motor program
accordingly. When the altered feedback/aversive stimulus was removed, birds’ songs returned to
baseline, suggesting that adults iteratively compare their song to an internal template. This
plasticity is dependent on a functioning LMAN circuit: when LMAN was inactivated, or
LMAN
CORE
’s input to RA blocked, there was no evidence of adaptive changes in the song due to
training (Warren et al., 2011, Charlesworth et al., 2012). However, over the course of days,
changes in the motor program appear to become independent of LMAN, such that only recent
learning is affected by LMAN inactivation (Warren et al., 2011) (Fig. 9). Furthermore, after the
training stimulus is removed, LMAN is required for the song’s recovery to baseline. This pattern
of results evokes the idea that LMAN (presumably both CORE and SHELL) is required for any
plasticity in the song motor program, that plasticity requires sensory feedback and an internal
template, and that changes in the motor program are eventually stored outside of the LMAN
circuits, likely in the HVC-RA pathway.

23

Figure 9. Training induced changes in song syllable fundamental frequency and loss of recent
learning with LMAN inactivation in an adult, from (Warren et al., 2011). An aversive stimulus
(white noise) was triggered when a particular song syllable had a fundamental frequency above
or below a set threshold. Over the course of days, the bird was able to alter the fundamental
frequency of that syllable to avoid the aversive stimulus (black circles). When LMAN was
inactivated with muscimol however, recent changes in the fundamental frequency were lost (red
circles).

24
INSIGHTS FROM STUDIES IN MAMMALIAN CORTICO-BASAL GANGLIA
CIRCUITS
In mammals, multiple parallel basal ganglia loops (such as those formed by the core and shell of
LMAN) have been shown to represent different aspects of movement control and sequence
motor learning (Alexander et al., 1986, Parent and Hazrati, 1995, Joel and Weiner, 1997, Brown
et al., 1999, Nakahara et al., 2001), and damage to these pathways, as occurs in Parkinson’s
disease, causes gross disruption of motor sequencing and language fluency (Cooper et al., 1991).
Basal ganglia loops are critical for speech production in humans (Lieberman, 2001, Wildgruber
et al., 2001) and speech disorders such as stuttering have been attributed to abnormalities in the
basal ganglia (Giraud et al., 2008). Furthermore, autism, a disorder characterized by deficient
communication and speech acquisition, is linked to the FoxP2 gene which is highly expressed in
the birdsong basal ganglia pathway during the sensitive period for vocal learning (Haesler et al.,
2004, Haesler et al., 2007, Thompson et al., 2013). Therefore, connecting findings in mammalian
cortico-basal ganglia research with those in birdsong is a vital part of developing hypotheses
relating to sensorimotor integration.
Motor learning typically begins as a goal-oriented process requiring memory and
attention, and is dependent on the mammalian associative cortico-basal ganglia pathway
(dorsomedial striatal loop), whereas well-learned motor behaviors require the sensorimotor
pathway (dorsolateral striatal loop) (Samejima and Doya, 2007, Graybiel, 2008, Yin et al., 2008,
Yin et al., 2009, Ashby et al., 2010, Redgrave et al., 2010, Thorn et al., 2010, Gremel and Costa,
2013)(Fig. 10). Neurons in both loops are active during learning, however the activity is
differentially modulated in the two pathways. Neurons in the associative loop showed increased
modulation early in learning during goal-directed tasks (Yin et al., 2009, Gremel and Costa,
2013), and specifically at points in the task in which the animal is choosing an action and when
25
waiting for the reward outcome (Histed et al., 2009, Thorn et al., 2010, Kim et al., 2013),
suggesting that this loop is involved in evaluating motor performance. In contrast, neurons in the
sensorimotor loop show increasing activity throughout training (Yin et al., 2009, Thorn et al.,
2010), specifically during action boundaries and do not show the sustained reward-related
activity associative loop neurons do (Thorn et al., 2010, Kim et al., 2013), indicating that this
circuit regulates motor aspects of learning. Therefore learning entails the integrative product of
multiple circuits whose function reflect specific aspects and the progression of learning.

26

Figure 10. Schematic of the major cortico-basal ganglia loops in mammals, from (Graybiel,
2008). Top section shows medial view of cortex, middle section shows lateral view of cortex,
and bottom section shows striatum. Arrows indicate the progression of areas’ contribution from
early to late motor learning. Abbreviations: SMA, supplementary motor area; ACC, anterior
cingulate cortex; MI, primary motor cortex; SI, primary somatosensory cortex; OFC,
orbitofrontal cortex; CN, caudate nucleus; P, putamen; VS, ventral striatum.

27
PRESENTED CONTRIBUTIONS
Together, results from previous work on avian and mammalian cortico-basal ganglia pathways
have led me to develop the following framework for the sensorimotor learning period in juvenile
songbirds. During learning, the CORE pathway functions in a manner similar to the mammalian
sensorimotor loop, in that its function is related to motor aspects of learning. From evidence
recounted above, it is clear that LMAN
CORE
contributes to changes in the motor program of song.
In contrast, the SHELL pathway functions similarly to the mammalian associative loop,
monitoring motor behavior and its relation to the goal outcome. Results showing that lesions of
the SHELL pathway prevent copying of tutor syllables, in addition to the anatomical evidence that
the SHELL pathway receives efference copy of the subsong motor signal and that LMAN
SHELL

grows during the learning period, support this idea. However, until the data presented in this
dissertation, no recordings in the SHELL pathway had been reported. As a first test of
LMAN
SHELL
’s involvement with evaluating performance during sensorimotor integration, I
looked for evidence that a representation of the tutor song was available to this pathway in
juvenile birds (Chapter 2). I then compared the neural activity of LMAN
CORE
and LMAN
SHELL

neurons during singing at the subsong and plastic song stages of learning (Chapter 3).

28
Chapter 2: Neural representation of a target auditory memory in a cortico-
basal ganglia pathway
ABSTRACT
Vocal learning in songbirds, like speech acquisition in humans, entails a period of sensorimotor
integration during which vocalizations are evaluated via auditory feedback and progressively
refined to achieve an imitation of memorized vocal sounds. This process requires the brain to
compare feedback of current vocal behavior to a memory of target vocal sounds. We report the
discovery of two distinct populations of neurons in a cortico-basal ganglia circuit of juvenile
songbirds (zebra finches, Taeniopygia guttata) during vocal learning: one in which neurons are
selectively tuned to memorized sounds and another in which neurons are selectively tuned to
self-produced vocalizations. These results suggest that neurons tuned to learned vocal sounds
encode a memory of those target sounds whereas neurons tuned to self-produced vocalizations
encode a representation of current vocal sounds. The presence of neurons tuned to memorized
sounds is limited to early stages of sensorimotor integration: following learning, the incidence of
neurons encoding memorized vocal sounds was greatly diminished. In contrast to this circuit,
neurons known to drive vocal behavior through a parallel cortico-basal ganglia pathway show
little selective tuning until late in learning. One interpretation of these data is that
representations of current and target vocal sounds in the SHELL circuit are used to compare
ongoing patterns of vocal feedback to memorized sounds, whereas the parallel CORE circuit has a
motor-related role in learning. Such a functional subdivision is similar to mammalian cortico-
basal ganglia pathways in which associative-limbic circuits mediate goal-directed responses
while sensorimotor circuits support motor aspects of learning.
29
INTRODUCTION
Both humans and songbirds display the rare trait of vocal learning; juveniles hear and
memorize vocal sounds of adult “tutors” during a sensitive period of development and then
translate the neural memory of those sounds into a motor program (Doupe and Kuhl, 1999). The
existence of a stable memory of tutor sounds in songbirds was first demonstrated behaviorally by
seminal work in the 1950s (Marler, 1956, Thorpe, 1958, Marler, 1970). Acquiring a neural
representation of tutor song is the first essential step in the process of vocal learning, and once
formed this memory guides sensorimotor integration as each bird learns to imitate its tutor’s song
by vocalizing and using auditory feedback to compare its incipient vocalizations to that tutor
memory (Konishi, 1965).
Parallel cortico-basal ganglia loops in mammals process distinct but related aspects of
sequential motor and reward-based learning (Hikosaka et al., 1999, O'Doherty et al., 2004,
Tanaka et al., 2004, Yin et al., 2009, Thorn et al., 2010). In songbirds, the cortical region LMAN
(lateral magnocellular nucleus of the anterior nidopallium) is required for vocal learning (Bottjer
et al., 1984, Scharff and Nottebohm, 1991, Aronov et al., 2008). LMAN is composed of two
subregions, a CORE and a surrounding SHELL, whose pathways make parallel connections to form
recurrent loops through the basal ganglia and through other cortical regions (Bottjer, 2004) (Fig.
1A). Neurons in LMAN
CORE
project to vocal motor cortex and drive motor variability that is
characteristic of babbling in juvenile birds to facilitate trial-and-error learning (Ölveczky et al.,
2005, Warren et al., 2011, Ravbar, 2012). However, strong evidence for a neural representation
unique to tutor song in this circuit has been lacking. Neurons in the SHELL region of LMAN
project to an adjacent area of motor cortex necessary for accurate temporal sequencing and
copying of tutor syllables but not for normal vocal production (Johnson et al., 1995, Iyengar et
30
al., 1999, Bottjer et al., 2000, Bottjer and Altenau, 2010). Furthermore, the SHELL pathway
receives a strong transient projection from LMAN
CORE
only in juvenile birds (Miller-Sims and
Bottjer, 2012) (Fig. 1B), indicating that CORE neurons convey efference copy to LMAN
SHELL

circuitry as birds are engaged in vocal learning. This pattern of results suggests that neurons in
LMAN
SHELL
may contribute to evaluating whether current vocal behavior matches the tutor song
memory.
31

Figure 1. Parallel cortico-basal ganglia pathways in songbirds. A, Parallel circuits are formed
by the axonal connections of CORE (gray) and SHELL (red) regions of LMAN (lateral
magnocellular nucleus of the anterior nidopallium). These parallel projections form recurrent
loops through both the basal ganglia (Area X, a nucleus containing both striatal and pallidal
neurons), and through other cortical regions: RA (robust nucleus of arcopallium) and AI
d
(dorsal
intermediate arcopallium) which are located in the analog of mammalian motor cortex. AI
d
was
referred to as Ad (dorsal arcopallium) in previous papers from our lab, but we have changed the
terminology here to conform to the nomenclature suggested by Reiner et al. (2004). CORE and
SHELL regions of LMAN receive input from separate subgroups of neurons in the thalamic
nucleus DLM (dorsolateral thalamus, VM = ventromedial, DL = dorsolateral). B, Anterograde
label in coronal sections of RA and AI
d
following iontophoretic injections of biotinylated dextran
amine into LMAN
CORE
. Individual CORE neurons send axons exclusively into RA in adult birds,
but send numerous collateral branches into AI
d
in juvenile (35 days post hatch) birds (Miller-
32
Sims and Bottjer, 2012) (dorsal is up, medial is left; scale bar = 0.5 mm). Thus, LMAN
CORE

neurons make a robust transient projection into the SHELL pathway during sensorimotor
integration that is completely gone by adulthood.

33
Neurons in LMAN
CORE
develop auditory selectivity to each bird’s own song during vocal
learning, but individual CORE neurons in young juveniles respond to both the bird’s own song
and the tutor song even when they have been manipulated to be acoustically dissimilar (Solis and
Doupe, 1997, 1999, 2000, Yazaki-Sugiyama and Mooney, 2004). We predicted that in contrast
to LMAN
CORE
,

LMAN
SHELL
in young birds should receive information regarding learned tutor
sounds and hence contain neurons that respond selectively to the tutor song. In accord with this
prediction, we found that LMAN
SHELL
contains distinct populations of neurons which encode
either the tutor song or self-produced vocal babbling in birds in early stages of sensorimotor
integration. In adult birds that have completed the process of vocal learning, the neural
representation of each bird’s own song is strengthened as the representation of the tutor song is
diminished, indicating that these neuronal populations are specialized for the period of vocal
learning.
MATERIALS & METHODS
Subjects
All procedures were performed in accordance with protocols approved by the University of
Southern California Animal Care and Use Committee. 28 urethane-anesthetized male zebra
finches (Taeniopygia guttata) of three different ages were used; 43-47 days post hatch (dph;
mean = 45.5; n = 6), 54-68 dph (mean = 60.9; n = 10), and adult birds (> 90 dph; n = 12). All
birds were bred in group aviaries and remained with their natural parents until at least 35 dph to
ensure that they received normal tutor song exposure and social experience (Immelmann, 1969,
Böhner, 1983, Eales, 1985, Clayton, 1987, Böhner, 1990, Catchpole and Slater, 1995, Mann and
Slater, 1995, Roper and Zann, 2006). Because birds were raised by their parents, they were
34
exposed to their father’s song (the tutor song) during both the early period of social imprinting as
well as during the tutor memorization period; although the presence of the tutor during the
imprinting period does not affect song learning (Roper and Zann, 2006), exposure to the tutor
song during imprinting might influence the development of neural responsiveness.
Previous work in our lab has shown, consistent with the studies cited above, that natural rearing
until 35 dph constitutes sufficient tutor exposure to develop an accurate copy of the tutor song
(Foster and Bottjer, 2001). Most birds raised under these naturalistic conditions (i.e. by their
biological parents within a colony) memorize the song of their father (see references above).
Because juveniles had access to other adult males they could have memorized some aspects of
other males’ songs (Williams, 1990). However, this tendency would result in weaker tuning to
the tutor song of the natural father, and would therefore work against an overall pattern of tutor
tuning. The tutor for each bird was therefore its natural father; because the songs of individual
zebra finches vary considerably, each experimental bird learned a specific and unique tutor song.
Song stimuli
Birds were placed in individual sound attenuation chambers and songs were recorded using
Sound Analysis Pro (Tchernichovski et al., 2000) (44 kHz sampling rate). All song stimuli were
chosen based on how frequently they were produced by each bird, as in previous studies (Solis
and Doupe, 1999, Phan et al., 2006, Kojima and Doupe, 2007); at least 10 song bouts (periods in
which multiple song motifs are sung) for adult song and at least 20 song bouts for juvenile song
were inspected both aurally and visually in order to pick the song motif that represented the most
frequently produced song for each bird. Each song stimulus was high-pass filtered at 400 Hz and
edited to include one or two song motifs (Solis and Doupe, 1997, Rosen and Mooney, 2000,
35
Solis and Doupe, 2000). Because juvenile song is variable, at the onset of each recording session
we tested several (four-five) motifs of each juvenile bird’s own song (BOS) that were produced
within 24 hours prior to the experiment. Each version of BOS elicited similar responses in
LMAN

CORE and SHELL of all birds tested (Solis and Doupe, 1997, 1999, 2000); we chose the
version of BOS which produced the largest neural response and used it thereafter. This song was
also played back as a mirror-image reverse (REV). Conspecific songs (CON) were recorded
from adult breeders from non-adjacent aviaries. Age-matched conspecific songs (AMC) were
recorded from unfamiliar juveniles matched to age ± 1 day of the experimental bird. Although
previous studies have found no intensity preferences in song control nuclei (Margoliash, 1992,
Doupe, 1997) we equalized song stimuli to 68.6 ± 0.3 dB (mean ± s.e.m.) measured with a sound
level meter (Extech Instruments). Twenty iterations of each stimulus were presented; song
stimuli were presented in blocks of 3-5 different stimuli (depending on how many stimuli were
presented for a given bird); the order of stimuli within a block was random (without
replacement) and the inter-stimulus interval was 10 ± 2 sec.
Electrophysiology
One-two days prior to an experiment, birds were anesthetized with 1.5% isoflurane (inhalation)
and a stainless steel post was attached to the skull with dental cement. On the day of the
experiment, birds were anesthetized with an intramuscular injection of 20% urethane in dH
2
O
(100-110 µl, Sigma) and the steel post was fixed to a stereotaxic apparatus so that the ears were
unobstructed during recording. Extracellular recordings were made using single electrodes (0.4-
0.8 MΩ, Carbostar) or linear electrode 16-pad arrays (1.5-5 MΩ, NeuroNexus). Spike data was
amplified (Neuralynx), bandpassed between 300-5000 Hz, and digitized at 20 or 32 kHz using
36
Spike 2 software (Cambridge Electronic Design). Small electrolytic lesions (4-7 µA for 20
seconds) were made above and below LMAN
CORE
for single electrodes and in the bottom-most
and top-most recording pads of linear arrays for electrode tract reconstruction. Coronal sections
(50-µm thick) were Nissl-stained to reveal LMAN
CORE
borders, and for 7 birds an alternate series
was stained for calbindin using standard immunohistochemical procedures to confirm
LMAN
SHELL
borders (Fig. 2A) (Pinaud et al., 2007).
37

Figure 2. Histological and electrophysiological methods used. A, Photomicrograph of LMAN
CORE and SHELL (50 µm thick coronal sections; medial is right). The borders of the core region
can be clearly distinguished in Nissl-stained tissue (left panel). CORE is outlined with dashed
line and asterisks indicate fiduciary lesions. Different sub-groups of thalamic projection neurons
(from DLM; Fig. 1A) terminate in either LMAN
CORE
or LMAN
SHELL
, but all DLM axons express
calbindin in a highly selective fashion (Pinaud et al., 2007). Calbindin staining therefore
encompasses both core and shell regions and the terminal field demarcates the outer border of
LMAN
SHELL
(right panel, dark staining; dashed line is copy of CORE outline to illustrate location
of CORE). Scale bar = 300 µm. B-C, Example multiunit recording with clustered spike
waveforms from LMAN
SHELL
of 43 dph bird. B, Inset at top left shows raw voltage signal of
multiunit recording made with a Carbostar electrode; black box indicates expanded portion
shown to right. LMAN neurons have relatively low firing rates, and therefore ambivalent
waveform shapes due to overlapping spikes were rare. The graph shows a plot of energy versus
38
rising slope of raw waveforms (no corrections applied) from the recording site shown in inset.
This plot shows that raw spike shapes form distinct clusters even when based on only two
features. C, Waveforms were automatically clustered based on six waveform features using
KlustaKwik (Ken Harris, Rutgers University; see Methods). The resulting clusters for this
recording are shown in different colors (gray cluster did not meet signal to noise criterion);
overlaid waveforms of each unit are shown near the plotted cluster. After clustering, each unit
was tested to determine if the firing rate during the presentation of any stimulus was significantly
greater or less than the firing rate during baseline; each unit is labeled with its corresponding
response pattern.

39
Analysis
A recording site was considered for analysis if it was confirmed histologically to be in either
LMAN
CORE
or LMAN
SHELL
(excluding 50 µm on either side of the CORE/SHELL border). The
evoked responses of LMAN neurons tended not to persist throughout the whole stimulus, as has
been observed previously (Doupe, 1997, Solis and Doupe, 1997, 1999, Kojima and Doupe,
2007). Therefore response strengths calculated for song stimuli longer than ~1 sec
underestimated the actual response by averaging across both the early phasic response and the
period of decreased response (c.f. Doupe, 1997, Solis and Doupe, 1999). In order to correct for
stimulus duration bias (e.g. longer songs would yield speciously low response strengths), all
analyses were performed using neural data collected during the first second of song playback (if
the song was > 1 sec; some adult songs were slightly < 1 sec in duration). Our results for
LMAN
CORE
replicate those reported previously (Doupe, 1997, Solis and Doupe, 1997, Rosen and
Mooney, 2000, Solis and Doupe, 2000, Kojima and Doupe, 2007), showing that calculation of
response strength based on the first second did not change the nature of the results. In addition,
because neurons showed either an enhanced or a suppressed response to song playback (see
Results), it was important to determine whether rate increases and rate decreases were evaluated
equally when restricting the analysis to the first second of the response. To address this point
quantitatively, we examined the time course of responses of 20 randomly selected neurons which
showed decreased firing rates during playback. Decreases in firing rate in this sample of neurons
lasted on average 0.13 – 1.20 sec after stimulus onset. This pattern was equivalent to that of
neurons that show increases in firing rate; that is, like rate increases, rate decreases occurred
primarily during the first second of a song stimulus. We also assessed the responses of 20
randomly selected neurons which showed no significant change in firing rate during playback to
40
test if any of these neurons showed a significant increase or decrease in firing rate if the response
window was extended to the entire stimulus length. None of these neurons showed any
significant responses when tested this way. Therefore, restriction of analyses to the first second
does not differentially reject neurons which show decreased firing rates, and extending the
analysis window did not increase the incidence of song-responsive neurons.
Isolated units were clustered manually using MClust-3.5 (A. David Redish, University of
Minnesota; Fig. 2B-C) in MATLAB (Mathworks). To confirm the quality of clustering, all 45
dph recordings were re-clustered automatically using KlustaKwik (Ken Harris, Rutgers
University) based on the maximum and first derivative of energy of the waveforms, the first two
principal components, and the rising and falling slope of the waveforms. Within KlustaKwik,
the minimum and maximum number of clusters required was 4 and 30, respectively, and the
KlustKwik clustering algorithm was used. Those clusters were then manually inspected and
corrected where needed. The two different sorting methods yielded highly similar patterns of
data; the results presented here are from the 45 dph data which were automatically clustered.
Spike sorting methods inevitably include some error (Harris et al., 2000, Pedreira et al., 2012),
however standard precautions were taken to ensure that any error was minimal: single units were
considered for analysis if signal to noise ratio was > 3 (mean ± s.e.m. = 5.9 ± 0.09) and if less
than 1% of spikes had an interspike interval < 2 ms (n’s: CORE 45 dph = 93, 60 dph = 214, adult
= 308; SHELL 45 dph = 114, 60 dph = 172, adult = 265).
We determined whether each single unit was responsive by testing for a significant change in
firing rate from baseline (1 sec prior to stimulus) during presentation of any song stimulus
41
(matched t-test, p < 0.05). To compare differences in firing rates across neurons, standardized
responses (SR) were calculated:
𝑠 𝑡 𝑎𝑛 𝑑 𝑎𝑟 𝑑 𝑖 𝑧 𝑒 𝑑 𝑟𝑒 𝑠 𝑝 𝑜𝑛𝑠 𝑒 𝑠 𝑡 𝑟 𝑒 𝑛 𝑔𝑡 ℎ =
𝑆 − 𝐵 � 𝑉 𝑎𝑟 ( 𝑆 ) + 𝑉 𝑎𝑟 ( 𝐵 ) − 2 ∗ 𝐶 𝑜𝑣 𝑎 𝑟 ( 𝑆 , 𝐵 )

where S = firing rate during stimulus and B = firing rate during baseline so that a positive value
indicates an increased rate to a stimulus and a negative value indicates a decreased rate to a
stimulus. We did not adjust the standardized response strengths for the number of trials (i.e.
multiply standardized response by √# trials). Therefore, when accounting for the number of
trials, all neurons that were considered responsive to a stimulus showed an adjusted standardized
response of at least 2.24. In describing the responses of single units, use of the word “only”
signifies that an individual neuron gave a statistically significant response to only one stimulus
out of 3-5 tested (see Results). For example, “TUT-only” refers to neurons that gave a
significant response only to tutor song and not to any other stimulus. In contrast, “TUT-
responsive” neurons included those that showed a significant response to TUT, regardless of
whether they also responded significantly to other stimuli. Similarly, “BOS only” refers to the
subset of neurons that responded significantly only to BOS and not to any other stimulus, and
“BOS-responsive” refers to the larger group of neurons that responded significantly to BOS
regardless of responses to other song stimuli. When describing the responses of single units,
“excited” and “suppressed” are used as operational terms that signify firing rate increases and
decreases compared to baseline rates, and are not descriptions of the underlying synaptic
mechanisms responsible for changes in firing rate.
42
To measure selectivity, a difference score was calculated for “Song A” as follows: Song A
∆SR
=
SR
Song A
– SR
Song B
, where SR is the standardized response strength. This is an alternative to the
psychometric discriminability index d’ used in previous studies. We chose to standardize
responses before subtracting, as opposed to subtracting response strengths and dividing by the
standard deviations as in d’. As other groups have noted (Solis and Doupe, 1997, Coleman et al.,
2004), this use of d’ scores has limitations, including the fact that the score is sensitive to
response strength and is less useful for revealing the degree to which the response of one
stimulus is greater or less than another. Selectivity scores based on proportions were not
appropriate for the data as there were frequently negative response strengths which would
confound the interpretation of the score. Use of standardized response strength differences
corrects for differences in response strength, allows the interpretation of scores calculated from a
negative response, and retains one of the benefits of d’, insensitivity to sample size. To address
whether d’ scores would produce substantially different patterns of selectivity, we calculated d’
values for 45 dph TUT- and BOS-responsive neurons (Tables 2, 3, below). Comparison of d’
scores with ∆ standardized response (∆SR) scores revealed no substantial differences. As shown
for ∆SR scores, d’ values revealed that neurons in SHELL were more selective for BOS or TUT
than CORE neurons. In order to include all song-responsive neurons in selectivity measures,
regardless of whether cells showed increased or decreased firing rates, we reversed the ΔSR
scores and d’ values of song-suppressed neurons.
Statistics
All continuous data were tested for normality using Kolmogorov-Smirnov and Shapiro-
Wilk tests. t-tests were used to compare means for normally distributed data, and Mann-Whitney
43
tests were used to compare medians for non-normally distributed data. Differences in
proportions were tested using chi-square tests, and differences in distributions were tested with
Kolmogorov-Smirnov Z tests. A Z test for proportions was used to test if the relative
proportions of excited and suppressed 45 dph SHELL neurons were different from chance.
Bonferroni corrections were applied wherever multiple comparisons were made.
RESULTS
Distinct neuronal populations in LMAN
CORE
and LMAN
SHELL
of juvenile birds
To test whether LMAN
SHELL
contains a neural representation of the tutor song during
vocal learning, we recorded from CORE and SHELL regions of LMAN in juvenile zebra finches
(Taeniopygia guttata) that have completed memorization of a tutor song and begun to practice
their song vocalizations (45 dph). We tested each neuron for auditory responsiveness, expressed
as a significant change in firing rate from baseline during playback of each song stimulus (p <
0.05; t-test). A high proportion of neurons in both CORE and SHELL gave an auditory response to
playback of at least one song stimulus; the proportion of CORE neurons that showed a significant
auditory response was higher than that for SHELL neurons (0.89 versus 0.68, p < 0.001, chi-
square; n: CORE = 83/93, SHELL = 77/114; CORE/SHELL auditory neurons per birds for n = 6 birds:
18/17; 0/16; 20/36; 2/0; 32/4; 11/4).
As an initial population analysis, we categorized each auditory neuron according to
whether it exhibited a significant change in firing rate only to the bird’s own song (BOS only),
only to the tutor song (TUT only), or to both stimuli (BOS & TUT) (Fig. 3A; all of these neurons
were tested with TUT, BOS, and a mirror-image reverse of BOS in which temporal structure is
reversed while leaving spectral content intact). For example, “TUT only” neurons showed a
44
statistically significant response only to TUT and not to BOS or reverse BOS (see Methods). A
large proportion of neurons in SHELL showed a significant response only to the tutor song (0.32,
21/65 neurons). In contrast, only a small proportion of CORE neurons responded to TUT only
(0.07, 5/68 neurons; p = 0.003, chi-square). However, approximately equal proportions of CORE
and SHELL neurons responded only to BOS. In addition, a large proportion of CORE neurons
responded to both BOS and TUT whereas few SHELL neurons responded significantly to both
stimuli (0.47 versus 0.15; 32/68 versus 10/65 neurons; p < 0.0001, chi-square). These results
show that LMAN
SHELL
contains two distinct populations of auditory neurons during the period of
sensorimotor integration: those which respond only to the tutor song and a separate population
which responds only to the current version of the bird’s own song; relatively few SHELL neurons
respond to both BOS and TUT. In contrast, the majority of neurons in LMAN
CORE
respond either
only to the bird’s own song or to both BOS and TUT.
45

Figure 3. A population of neurons in LMAN
SHELL
of 45 dph birds responds only to tutor song.
A, Among all neurons that were tested with playback of bird’s own song (BOS), reversed bird’s
own song (REV), and tutor song (TUT): absolute proportions of neurons that responded to BOS
only, TUT only, or both BOS and TUT in LMAN
CORE
(gray; n = 68) and LMAN
SHELL
(red; n =
65). For this analysis, “BOS only” included neurons that showed a significant response to BOS
and not to REV or TUT (see Methods), “TUT only” included neurons that responded only to
TUT and not to BOS or REV, and “BOS & TUT” includes neurons that responded to both BOS
and TUT (regardless if there was a response to REV). SHELL neurons in all three response
categories include cells that showed either rate increases or decreases (see Results); analysis
conducted on only excited neurons yielded highly similar results (data not shown). *** indicates
p ≤ 0.001. B, Among all SHELL neurons included in (A), proportions of neurons that were either
46
excited (filled bars) or suppressed (open bars) by BOS only or TUT only. C, Among all neurons
included in (A), mean proportion of neurons averaged across each bird that responded to BOS
only, TUT only, or both stimuli in LMAN
CORE
(gray) and LMAN
SHELL
(red). Data from individual
birds are plotted as dots (n = 4 birds; not included are data from one bird did not receive the TUT
stimulus and from one bird where we recorded one BOS/TUT responsive neuron). The mean
proportion of TUT only neurons per bird was higher in SHELL than in CORE (p = 0.019, Mann-
Whitney), and the mean proportion of BOS & TUT neurons per bird was higher in CORE than in
SHELL (p = 0.021, Mann-Whitney). * indicates p < 0.05. These data show that the differences
between CORE and SHELL computed across individual birds were also significant and match those
computed across neurons.

47
To more fully characterize the populations of CORE and SHELL neurons, we presented a
subset of neurons with an expanded stimulus set of five different songs including an adult
conspecific song (CON) and an age-matched conspecific song (AMC) in addition to BOS,
reverse BOS (REV) and TUT (Table 1). Despite the inclusion of more song types, these data
confirmed that a large proportion of neurons gave a significant response only to the tutor song in
LMAN
SHELL
compared to a smaller proportion in LMAN
CORE
(0.28 versus 0.04, 11/40 versus 2/52
neurons; p < 0.01, chi-square). A relatively large proportion of CORE neurons responded
significantly to combinations of TUT plus other songs compared to SHELL neurons (sum of rows
2-4; 0.43 versus 0.15; p = 0.005, chi-square). Expressed as a relative percentage, 91% of
individual TUT-responsive neurons in CORE also responded to other songs compared to only
35% in SHELL (p < 0.001, chi-square). The proportion of neurons which responded only to BOS
did not differ between CORE and SHELL (0.15 versus 0.13, 8/52 versus 5/40 neurons), confirming
that each subregion contains a population of neurons that responds only to BOS (cf. Fig. 3A).
Fewer SHELL than CORE neurons were BOS-responsive overall (sum of rows 3-6; 0.73 versus
0.43, p = 0.003, chi-square), such that even though more neurons in CORE than SHELL responded
to BOS plus other stimuli (sum of rows 3, 4 and 6; 0.58 versus 0.30; p = 0.008, chi-square) the
relative percentage of BOS-responsive neurons that also responded to other songs did not differ
between CORE and SHELL (79% versus 70%). Interestingly however, many neurons in CORE were
very broadly tuned (i.e., responded significantly to BOS, TUT, plus at least one other song
stimulus, row 4) compared to SHELL (0.25 versus 0.05, 13/52 versus 2/40 neurons).

48
Table 1
row neuron category CORE SHELL
1 TUT only 0.04 0.28**
2 TUT + REV, CON or AMC 0.10 0.05
3 BOS + TUT only 0.08 0.05
4 BOS + TUT + other songs 0.25 0.05*
5 BOS only 0.15 0.13
6 BOS + REV, CON or AMC 0.25 0.20
sum 2,3,4 / sum 1-4
relative percentage of TUT-
responsive neurons that
respond to other songs 91 35***
sum 3,4,6 / sum 3-6
relative percentage of BOS-
responsive neurons that
respond to other songs 79 70

Table 1. Proportion of auditory neurons in 45 dph CORE and SHELL which were tested with all
five stimuli (bird’s own song (BOS), reverse BOS (REV), tutor song (TUT), adult conspecific
(CON) and age-matched conspecific (AMC); CORE n = 52; SHELL n = 40). For this analysis,
“TUT only” indicates neurons that showed a significant response to TUT and not to BOS, REV,
CON or AMC; “BOS only” indicates neurons that responded significantly to BOS and not to
REV, TUT, CON or AMC. Rows 1-6 sum to 87% for CORE and 76% in SHELL (remaining
neurons did not respond to BOS or TUT). *, **, *** indicate p < 0.05, p < 0.01, and p < 0.001,
respectively for significant differences between LMAN
CORE
and LMAN
SHELL
proportions (chi-
square tests).

49
Figure 4 shows examples of individual TUT-responsive neurons in 45 dph juvenile
LMAN SHELL and CORE during playback of various song stimuli. Rasters for each neuron
indicate that responses in both subregions tended to be somewhat sparse and variable from trial
to trial, as previously reported for LMAN
CORE
(Doupe, 1997). A representative TUT-responsive
LMAN
SHELL
neuron showed a strong response only to TUT, as well as a modest but
nonsignificant response to BOS and an even lower response to the bird’s own song played in
reverse (Fig. 4A). In contrast, a typical TUT-responsive neuron from LMAN
CORE
gave a
significant increased rate to BOS, REV, and TUT (Fig. 4B). This broad responsiveness concurs
with previous studies which found that most TUT-responsive neurons in CORE also respond to
BOS (Solis and Doupe, 1997, 1999, 2000, Yazaki-Sugiyama and Mooney, 2004), as was also
true for TUT-responsive neurons found in other song-control nuclei (Nick and Konishi, 2005a).
Another difference between SHELL and CORE subregions was that song-evoked activity in SHELL
neurons could manifest as either increases or decreases in firing rate, whereas CORE neurons
showed increased rate responses only. Figure 4C shows a SHELL neuron that showed a
suppressed response to TUT but not to BOS or CON (see below).
50

Figure 4. Neurons responsive to the tutor song in 45 dph LMAN
SHELL
and LMAN
CORE
. A,
Recording from LMAN
SHELL
. Top: song spectrograms of a 45 dph bird’s tutor song, its own
song, and reverse of the bird’s own song. Below are raster plots and instantaneous firing rates
for a TUT-excited single unit; overlaid waveforms are shown in inset (see Methods and Fig. 2B-
C for recording and spike sorting details.) scale bar = 0.5 msec. * indicates response
significantly different from baseline activity. B, Recording from LMAN
CORE
of 43 dph bird,
arranged as in (A) showing responses to the tutor song, the bird’s own song and reverse bird’s
own song. C, Recording from LMAN
SHELL
of 45 dph bird showing a TUT-suppressed single unit,
51
arranged as in (A) and (B) showing responses to the tutor song, the bird’s own song and an adult
conspecific song.

52
Figure 5 shows examples of individual BOS-responsive neurons in SHELL and CORE of 45
dph juveniles. The CORE neuron fired strongly in response to BOS, REV, and TUT, whereas the
SHELL neuron showed significant increased firing only to BOS. Thus, as for TUT-responsive
neurons, BOS-responsive neurons in CORE tended to respond to multiple other stimuli, whereas
those in SHELL were likely to respond strongly only to BOS. In this example, the response of the
SHELL neuron distinguished the bird’s own babbling song from that of another young bird (age-
matched conspecific) as well as from the tutor song.

Figure 5. Neurons responsive to the bird’s own song in 45 dph LMAN
SHELL
and LMAN
CORE
. A,
Recording from LMAN
SHELL
. Top: song spectrograms of a 45 dph bird’s own song, its tutor
song, and an age-matched conspecific song. Below are raster plots and instantaneous firing rates
for a BOS-excited single unit from this recording site; overlaid waveforms are shown in inset.
scale bar = 0.5 msec. * indicates a response that is significantly different from baseline activity.
53
B, Recording from LMAN
CORE
of 43 dph bird, arranged as in (A) showing responses to the bird’s
own song, reverse bird’s own song, and the tutor song.

54
Neurons in LMAN
SHELL
of 45 dph birds were either excited or suppressed by playback
As indicated above, many auditory neurons in SHELL, but not CORE, of 45 dph birds
showed suppressed responses to playback of song (Fig. 4C). Among all BOS-responsive SHELL
neurons (i.e., those that showed a significant response to BOS alone or in combination with other
song stimuli), 81% (29/36) were excited while 19% (7/36) showed a suppressed response to
BOS. Among TUT-responsive SHELL neurons, 76% (28/37) showed excitation and 24% (9/37)
showed suppression to TUT. Interestingly, BOS- and TUT-excited SHELL neurons never showed
suppression to any other stimulus. This mutual exclusion was true for suppressed neurons as
well (except for one 1 out of 73 cases in which the neuron showed increased firing to REV and
suppression to TUT). Thus, SHELL neurons encoded a representation of tutor song and bird’s
own song via either increases or decreases in response strength.
Among CORE neurons that were tested with TUT, BOS, and REV, none showed
suppression to any song (n = 68). In contrast, among SHELL neurons that responded only to
BOS, 84% (16/19) showed an enhanced response while 16% (3/19) showed a suppressed
response (p = 0.003, Z test for proportion; Fig. 3B). Among SHELL neurons that responded only
to TUT, 67% (14/21) showed increased firing and 33% (7/21) showed suppression (p = 0.12, Z
test for proportion). This pattern suggests that neurons that responded only to TUT were more
likely to show suppression than were neurons that responded only to BOS in LMAN
SHELL
of 45
dph birds. However, despite the tendency towards a greater incidence of song suppression
within neurons that responded only to TUT, the relative proportion of excited:suppressed
neurons was not significantly different between neurons that responded only to TUT and those
that responded only to BOS (p = 0.20, chi-square).

55
TUT-responsive neurons were more selectively tuned in juvenile SHELL than CORE
To test the relative magnitude of responses across all TUT-responsive neurons, we
compared response strengths to all song stimuli in 45 dph LMAN
SHELL
and LMAN
CORE
.
Standardized responses for TUT-excited SHELL neurons were largest to TUT compared with all
other songs (Fig. 6A, filled red bars; p < 0.05 in all cases, independent t-test, unequal variance,
Bonferroni corrected). In contrast, TUT-excited CORE neurons responded equally well to TUT
and BOS (filled gray bars; p = 1.0, Mann-Whitney, Bonferroni corrected), but like SHELL
neurons, responded less to REV, CON and AMC (p < 0.05 in all cases, Mann-Whitney,
Bonferroni corrected). Furthermore, SHELL neurons responded much less to BOS, REV, CON
and AMC than did those in CORE (p < 0.05 in all cases, independent t-test, unequal variance for
CON comparison). Neurons in SHELL which exhibited suppression to TUT showed similar
response patterns to TUT-excited neurons: response suppression was greatest to TUT compared
with other songs (open red bars; p < 0.01 in all cases, Mann-Whitney, Bonferroni corrected). As
indicated above, TUT-suppressed neurons were not detected in CORE. In summary, TUT-
responsive neurons in CORE and SHELL showed comparable response strength to TUT, but
neurons in SHELL gave a much weaker response to non-tutor songs than did those in CORE.
56

Figure 6. Response strength and selectivity across all TUT-responsive neurons in LMAN
SHELL

and LMAN
CORE
of 45 dph birds. TUT-responsive neurons include all those that showed a
significant response to tutor song (either alone or in combination with other song stimuli,
includes both TUT-excited and TUT-suppressed neurons). A, Standardized response strength for
TUT-responsive neurons in CORE and SHELL to tutor song (TUT), bird’s own song (BOS),
reversed bird’s own song (REV), adult conspecific song (CON), and age-matched conspecific
song (AMC). Filled bars show mean response of neurons that were excited by TUT (CORE n =
41 for TUT, BOS and REV, 24 for CON and AMC; SHELL n = 28 for TUT, BOS and REV, 12
for CON and AMC) and open bars show mean response of neurons that were suppressed by TUT
(CORE n = 0; SHELL n = 9 for TUT, BOS and REV, 7 for CON and AMC). Error bars indicate
s.e.m.; *, **, *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively. B, Histograms
57
showing distributions of selectivity (∆SR) scores for tutor song versus other songs (excited and
suppressed responses combined) in CORE and SHELL (CORE n = 41 for vs. BOS and REV, 24 for
vs. CON and AMC; SHELL n = 37 for vs. BOS and REV, 19 for vs. CON and AMC). The sign of
selectivity scores for suppressed neurons is reversed. Error bars are centered on mean values and
indicate s.e.m. C-D, Cumulative distributions of TUT-BOS ∆SR scores (CORE n = 41; SHELL n =
37) and of CON-AMC ∆SR scores (CORE n = 24; SHELL n = 19); p values shown are from
Kolmogorov-Smirnov Z tests for difference between the distributions. E, TUT selectivity
averaged across birds to compare with means averaged across neurons (B). Bars show mean
selectivity for TUT over other songs across birds in CORE (gray) and SHELL (red). Data for
individual birds are plotted as dots. Mean selectivity scores per bird were higher in SHELL than
in CORE for comparisons to both BOS and REV (p = 0.02 and 0.04, respectively, Mann-Whitney;
n = 5 birds for vs. BOS and vs. REV, one bird had recordings only in CORE and one bird had
recordings only in SHELL; not included are data from one bird which did not receive the TUT
stimulus; n= 2 birds for vs. CON and vs. AMC). These data show that mean values do not differ,
regardless of whether values are computed based on means of individual birds or across all
neurons.

58
To quantify response selectivity between pairs of stimuli, we computed a difference score
for each neuron: Δ standardized response (∆SR) = standardized response
Song A
– standardized
response
Song B
(a score of zero indicates no song preference). Across all TUT-responsive neurons
in SHELL of 45 dph birds, mean selectivity (∆SR) scores revealed a significant preference for
TUT versus all other songs (Fig. 6B; ∆SR significantly different from zero, p < 0.001 in all
cases, one-sample t-test; Fig. 6E shows data for individual birds; see Table 2 for d’ scores).
Neurons in CORE also preferred TUT over REV, CON and AMC (all cases significantly different
from zero, p < 0.05, one-sample t-test). However, there was no significant selectivity for TUT
versus BOS (p = 0.30, one-sample t-test), indicating no song preference between TUT and BOS
in 45 dph CORE neurons. Furthermore, SHELL neurons exhibited much greater selectivity for
TUT over all other songs than did CORE neurons (p always < 0.005, independent t-test). These
results for TUT-responsive neurons demonstrate that SHELL neurons showed greater selectivity
for tutor song compared to those in CORE for all comparisons to other song stimuli.

59
Table 2
TUT-responsive 45 dph neurons
d’ comparison CORE SHELL
TUT-BOS
0.00 ± 0.06

0.26 ± 0.07**
TUT-REV
0.26 ± 0.07

0.48 ± 0.07*
TUT-CON
0.23 ± 0.09

0.63 ± 0.12**
TUT-AMC
0.16 ± 0.08

0.54 ± 0.12**

Table 2. d-prime values (mean ± s.e.m.) for 45 dph CORE and SHELL for TUT-responsive
neurons. n’s: CORE = 41 for vs. BOS and REV, 24 for vs. CON and AMC; SHELL n = 37 for vs.
BOS and REV, 19 for vs. CON and AMC. * and ** indicate p < 0.05 and p < 0.01, respectively
for significant differences between mean d’ values of CORE and SHELL. We used ∆SR scores in
preference to d’ scores for methodological reasons (see Methods). However, both measures
showed good agreement in terms of assessing differences in selectivity between SHELL and CORE
neurons: TUT-selectivity of SHELL neurons was higher than that in CORE neurons.

60
Cumulative distributions of selectivity scores for TUT versus BOS were shifted to the
right for SHELL relative to CORE neurons (Fig. 6C), demonstrating that more neurons preferred
TUT over BOS in SHELL (p = 0.002, Kolmogorov-Smirnov Z). As a control for whether the
selective response to TUT simply reflected a preference for an adult song versus an immature
juvenile song, we also compared cumulative distributions for adult conspecific versus age-
matched conspecific songs (Fig. 6D). SHELL neurons evinced no preference for CON over AMC
(p = 0.56, one-sample t-test), and the distributions of CORE versus SHELL selectivity scores did
not differ (p = 0.62, Kolmogorov-Smirnov Z). Thus, TUT-responsive SHELL neurons show a
specific preference for the tutor song they have memorized, as opposed to a generic preference
for general characteristics of adult song over immature song.
BOS-responsive neurons were more selectively tuned in juvenile SHELL than CORE
We also tested BOS-responsive neurons in SHELL versus CORE of 45 dph birds by
comparing relative response strength across song stimuli. BOS-excited neurons in both CORE
and SHELL showed a stronger response to BOS than to all other songs (Fig. 7A; p < 0.01 for all
cases, Mann-Whitney, Bonferroni corrected). However, SHELL neurons responded less strongly
to REV, TUT and AMC than did CORE neurons (p = 0.005, 0.002, and 0.047, respectively,
Mann-Whitney for REV comparison, independent t-test for TUT and AMC comparisons). The
response to CON did not differ between CORE and SHELL (p = 0.38, independent t-test, unequal
variance). BOS-suppressed neurons were found only in SHELL, and showed greater suppression
to BOS compared to REV (p = 0.002, Mann-Whitney, Bonferroni corrected). Thus, neurons in
CORE and SHELL responded comparably to the bird’s own babbling sounds, but with the
exception of CON, neurons in SHELL gave a much weaker response to non-BOS songs than did
those in CORE.
61

Figure 7. Response strength and selectivity across all BOS-responsive neurons in LMAN
SHELL

and LMAN
CORE
of 45 dph birds. A, Standardized response strength for BOS-responsive neurons
in CORE and SHELL to bird’s own song (BOS), reversed bird’s own song (REV), tutor song
(TUT), adult conspecific song (CON), and age-matched conspecific song (AMC). Filled bars
show mean response of neurons which were excited by BOS (CORE n = 65 for BOS and REV, 55
for TUT, 38 for CON and AMC; SHELL n = 29 for BOS, REV and TUT, 14 for CON and AMC)
and open bars show mean response of neurons which were suppressed by BOS (CORE n = 0;
SHELL n = 7 for BOS and REV, 3 for TUT, 2 for CON and AMC). Error bars indicate s.e.m.; *,
**, *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively. B, Histograms showing
distributions of selectivity (∆SR) scores for BOS over other songs (excited and suppressed
responses combined) in CORE and SHELL (CORE n = 65 for vs. REV, 55 for vs. TUT, 38 for vs.
62
CON and AMC; SHELL n = 36 for vs. REV, 32 for vs. TUT, 16 for vs. CON and AMC). The
sign of selectivity scores for suppressed neurons is reversed. Error bars are centered on mean
values and indicate s.e.m. C-D, Cumulative distributions of BOS-TUT ∆SR scores (CORE n =
55, SHELL n = 32) and AMC-CON ∆SR scores (CORE n = 38, SHELL n = 16); p values shown are
from Kolmogorov-Smirnov Z tests for difference between the distributions. E, BOS selectivity
averaged across birds to compare with means averaged across neurons (B). Bars show mean
selectivity for BOS over other songs across birds in CORE (gray) and SHELL (red). Data for
individual birds are plotted as dots. Mean selectivity scores per bird were higher in SHELL than
in CORE for comparisons to both REV and TUT (p = 0.047 and 0.020, respectively, Mann-
Whitney; n = 6 birds for vs. REV, one bird had recordings only in CORE and one bird had
recordings only in SHELL; n = 5 birds for vs. TUT, one bird had recordings only in CORE and one
bird had recordings only in SHELL, not included are data from one bird which did not receive the
TUT stimulus; n= 2 birds for vs. CON and vs. AMC). These data show that mean values do not
differ, regardless of whether values are computed based on means of individual birds or across
all neurons.

63
As in the case of TUT-responsive neurons, higher selectivity scores among BOS-
responsive SHELL neurons reflected their weaker response to non-preferred songs (Fig. 7B).
BOS-responsive neurons in both CORE and SHELL were selective for BOS over all other songs
(∆SR different from zero, p < 0.001 in all cases, one-sample t-test; Fig. 7E shows data for
individual birds; see Table 3 for d’ scores). However, SHELL neurons showed higher selectivity
than CORE neurons for BOS compared to all other songs (p always < 0.03, Mann-Whitney for
REV and AMC comparisons, independent t-test for TUT and CON comparisons). Likewise,
more neurons in SHELL were selective for BOS over TUT than in CORE as illustrated by the right-
shifted cumulative distribution (p = 0.009, Kolmogorov-Smirnov Z; Fig. 7C). If BOS-
responsive SHELL neurons simply preferred a juvenile song over adult songs, then one would
expect high selectivity scores for AMC-CON comparisons. However, the cumulative
distributions of AMC-CON ΔSR scores revealed no selectivity (Fig. 7D; scores not different
from zero; CORE p = 0.15, SHELL p = 0.86, one sample t-test), and there was no difference
between the distributions (p = 0.277, Kolmogorov-Smirnov Z). These results show that although
response strength to BOS is similar in CORE versus SHELL, responses of SHELL neurons are more
selective due to diminished responses to other song stimuli. Therefore, LMAN
SHELL
of 45 dph
birds contains a population of neurons that encodes a selective representation of vocal babbling
sounds and a distinct population of neurons that is selective to the adult tutor song of each bird.

64
Table 3
BOS-responsive 45 dph neurons
d’ comparison CORE SHELL
BOS-REV
0.41 ± 0.05

0.58 ± 0.06*
BOS-TUT
0.27 ± 0.04

0.49 ± 0.07**
BOS-CON
0.39 ± 0.06

0.62 ± 0.10*
BOS-AMC
0.32 ± 0.05

0.50 ± 0.08

Table 3. d-prime values (mean ± s.e.m.) for 45 dph CORE and SHELL for BOS-responsive
neurons. n’s: CORE = 65 for vs. REV, 55 for vs. TUT, 38 for vs. CON and AMC; SHELL n = 36
for vs. REV, 32 for vs. TUT, 16 for vs. CON and AMC. * and ** indicate p < 0.05 and p < 0.01,
respectively for significant differences between mean d’ values of CORE and SHELL. We used
∆SR scores in preference to d’ scores for methodological reasons (see Methods). However, both
measures showed good agreement in terms of assessing differences in selectivity between SHELL
and CORE neurons: BOS-selectivity of SHELL neurons was higher than that in CORE neurons.

65
Selective tuning to tutor song is restricted to a phase of early sensorimotor integration
The presence of a substantial population of neurons that respond only to the tutor song in
LMAN
SHELL
of 45 dph birds shows that the SHELL pathway has access to a representation of target
vocal sounds during the period of active imitative learning. To test if a representation of the
tutor song persists in SHELL neurons following the conclusion of vocal learning, we recorded
from CORE and SHELL in older juveniles at a later stage of sensorimotor integration when birds
are singing “plastic” song (60 dph), and in adults that are producing a stereotyped song following
the learning period (> 90 dph). These neurons were categorized as in Figure 3 according to
whether BOS, TUT, or both songs evoked a significant response. This developmental analysis
revealed that the high proportion of SHELL neurons showing a significant response only to TUT
at 45 dph was greatly reduced by 60 dph (p = 0.01, chi-square, Bonferroni corrected), and this
proportion remained low in adult birds (Fig. 8A, left panel). The substantial reduction of TUT-
only SHELL neurons by 60 days eliminated the significantly greater incidence of tutor-tuned
neurons in SHELL versus CORE seen at 45 dph. The proportion of CORE neurons that responded
only to TUT did not change between 45 and 60 days (p = 0.52, chi-square, Bonferroni corrected).
66

Figure 8. Developmental changes in responses and selective tuning of LMAN
CORE
and
LMAN
SHELL
neurons. A, Proportion of neurons which responded only to TUT and not to REV
(left panel) in CORE (gray) and SHELL (red) at 45 dph (CORE n = 68; SHELL n = 65), 60 dph (CORE
n = 57; SHELL n = 31) and in adult birds (CORE n = 47; SHELL = 88). Proportion of neurons
which responded only to BOS and not REV or TUT (middle panel) and proportion of neurons
which responded to both BOS and TUT (right panel). *, **, *** indicate p < 0.05, p < 0.01, and
67
p < 0.001, respectively. 45 dph data are the same as those presented in Figure 3. B, Mean BOS-
REV ∆SR scores for BOS-responsive neurons in LMAN
CORE
and LMAN
SHELL
in 45 dph (CORE n =
65; SHELL n = 36), 60 dph (CORE n = 53; SHELL n = 26) and adult birds (CORE n = 82; SHELL n =
100). Error bars indicate s.e.m. C, Mean BOS-TUT ∆SR scores for BOS-responsive neurons in
LMAN
CORE
and LMAN
SHELL
in 45 dph (CORE n = 55; SHELL n = 32), 60 dph (CORE n = 40; SHELL
n = 23) and adult birds (CORE n = 26; SHELL n = 64). D, Mean response strength of all BOS-
responsive neurons to BOS across development. Filled circles show mean standardized
responses of BOS-excited neurons in CORE (gray) and SHELL (red) to BOS in 45 dph (CORE n =
65; SHELL n = 29), 60 dph (CORE n = 49; SHELL n = 21) and adult birds (CORE n = 72; SHELL n =
88). Open circles show mean standardized responses of BOS-suppressed neurons throughout
development in CORE (n = 0 at 45 dph, 4 at 60 dph, 10 in adults) and SHELL (n = 7 at 45 dph, 5 at
60 dph, 12 in adults). There were no significant age differences within either subregion for both
BOS-excited and BOS-suppressed neurons (p > 0.05 in all cases). Error bars indicate s.e.m.

68
Developmental changes in BOS responses showed a different pattern (Fig. 8A, middle
panel). The proportion of neurons that showed a significant response only to BOS increased in
both CORE and SHELL between 45 and 60 days (CORE p = 0.03, SHELL p = 0.07, chi-square,
Bonferroni corrected), and was not different between the two subregions at any age. Therefore,
the neural representation of each bird’s own song became stronger in both subregions of LMAN
as vocal development progressed from vocal babbling to plastic song. The proportion of CORE
neurons that responded to both BOS and TUT decreased substantially between 45 and 60 days
(Fig. 8A, right panel; p = 0.002, chi-square, Bonferroni corrected) and remained low in
adulthood. Although the proportion of SHELL neurons that responded to both BOS and TUT did
not change throughout development, by adulthood a greater proportion of neurons in SHELL
responded to BOS and TUT than in CORE (p = 0.008, chi-square, Bonferroni corrected). In
summary, these data show that with learning a larger proportion of neurons in both subregions
respond significantly only to BOS as song matures, and the proportion of neurons in LMAN
SHELL

that respond only to the tutor song diminishes.
We found that CORE neurons became more selectively tuned to BOS during development, in
accord with previous studies’ reports of selectivity (although this represents the first test of BOS-
selectivity in birds younger than 55 dph) (Doupe, 1997, Solis and Doupe, 1997). The selectivity
of CORE neurons for BOS increased between 45 and 60 dph (p < 0.001 for both REV and TUT
comparisons, independent t-test, unequal variance for TUT comparison), such that there was no
difference in ΔSR scores of SHELL versus CORE neurons by 60 dph (Fig. 8B-C; independent t-
test, unequal variance for TUT comparison). The selectivity of SHELL neurons for BOS over
both REV and TUT did not change between 45 and 60 dph (p > 0.05 in both cases, independent
t-test, Bonferroni corrected), although BOS-REV selectivity did increase between 45 dph and
69
adulthood (p = 0.03, independent t-test, Bonferroni corrected). Response strength to BOS did
not change in either subregion during development (Fig.8D; Mann-Whitney, Bonferroni
corrected), indicating that the developmental increases in BOS-selectivity were due to decreased
responses to other stimuli. These results indicate that SHELL neurons show relatively more
mature levels of selectivity than CORE neurons early in vocal learning, and that the increase in
selectivity throughout development in both subregions is the result of weaker responses to non-
preferred stimuli.
Because relative levels of inhibition contribute to neural selectivity (Rosen and Mooney,
2000, Rosen and Mooney, 2003, Pinaud et al., 2008) as well as regulation of sensitive periods
(Hensch et al., 1998), we examined the proportion of all auditory neurons with enhanced versus
suppressed responses throughout development. LMAN
CORE
is known to include song-suppressed
neurons in adult but not juvenile birds (Doupe, 1997). In accord with this fact, we found that the
proportion of CORE neurons that showed a suppressed response to at least one song stimulus
increased throughout development (gray bars, Fig. 9A; 0.01 at 45 dph, 0.13 at 60 dph, 0.25 in
adults; p < 0.001, three-way chi-square). Among BOS-responsive neurons, the percentage in
CORE that were suppressed by BOS also increased significantly from 0% at 45 dph to 8% at 60
dph and 12% in adults (gray bars Fig. 9B; p = 0.016, three-way chi-square). In contrast to CORE,
the proportion of SHELL neurons that showed suppression to at least one song stimulus did not
change during development (red bars Fig. 9A; 0.26 in 45 dph birds, 0.33 in 60 dph, 0.18 in
adults; p = 0.12, three-way chi-square). Moreover, the percentage of BOS-responsive neurons in
SHELL that were suppressed by BOS did not change during development (red bars Fig. 9B; 19%
at 45 dph, 19% at 60 dph, 13% in adults; p = 0.55, three-way chi-square). Thus, suppressive
responses to song in LMAN
CORE
mature throughout development, and gradually come to match
70
the level of song suppression seen in LMAN
SHELL
. In addition, these results suggest that
maturation of inhibitory circuits may underlie increased selective tuning in LMAN and represent
a correlate of learning.

71

Figure 9. Developmental changes in song-suppressive responses in LMAN
CORE
and LMAN
SHELL

neurons. A, Proportion of auditory neurons which responded with suppression to any song
stimulus in CORE (gray) and SHELL (red) at 45 dph (CORE n = 83; SHELL n = 77), 60 dph (CORE n
= 76; SHELL n = 46) and in adult birds (CORE n = 120; SHELL n = 154). B, Proportion of BOS-
responsive neurons which were suppressed by BOS in CORE and SHELL at 45 dph (CORE n = 65;
SHELL n = 36), 60 dph (CORE n = 53; SHELL n = 26) and in adult birds (CORE n = 82; SHELL n =
100). * and *** indicate p < 0.05 and p < 0.001, respectively.

72
DISCUSSION
These results demonstrate that LMAN
SHELL
contains a population of neurons that encodes
a neural representation of learned tutor sounds in juvenile songbirds during early stages of
sensorimotor integration for vocal learning (45 dph). This discovery represents the first instance
in which a substantial population of vocal-control neurons responds selectively to tutor song in
birds that are acquiring learned vocal behavior. The TUT-tuned neurons are present at a time
when lesions of AI
d
(dorsal intermediate arcopallium) which interrupt the trans-cortical recurrent
pathway from LMAN
SHELL
through dorsal thalamus, prevent both imitation of the tutor song as
well as the development of a stable sequence of syllables (Bottjer and Altenau, 2010).
Furthermore, the large proportion of neurons which responds only to tutor song is not present in
LMAN
SHELL
of adult birds that have finished learning to produce a stereotyped imitation of that
song. Tutor-tuned neurons in 45 dph LMAN
SHELL
co-exist with a separate population that
responds selectively to the bird’s own song, suggesting that discrete populations of SHELL
neurons receive signals from regions of higher-level auditory cortex that process either vocal-
related feedback (BOS tuning) or aspects of tutor song (TUT tuning) (Phan et al., 2006, London
and Clayton, 2008, Gobes et al., 2010, Hahnloser and Kotowicz, 2010). In contrast to
LMAN
SHELL
, most neurons in LMAN
CORE
of young birds responded to BOS or to BOS plus other
songs, in agreement with previous studies (Solis and Doupe, 1999, 2000, Yazaki-Sugiyama and
Mooney, 2004). The overall pattern of results indicates that distinct representations of both the
tutor song and the bird’s own song are available only in SHELL at 45 dph, and that neurons in
CORE do not develop a selective representation of self-produced vocalizations until late in
learning.
73
At 45 dph, LMAN
SHELL
contained TUT- and BOS-suppressed neurons, demonstrating that
SHELL neurons represent the tutor song and bird’s own song via either increases or decreases in
firing rate, similar to responses to learned sounds in prefrontal cortex of mammals (Suzuki et al.,
1997, Fritz et al., 2010). Thalamic input to LMAN

is excitatory (Livingston and Mooney, 1997,
Boettiger and Doupe, 1998, Bottjer et al., 1998, Pinaud et al., 2007), so suppression presumably
results from intrinsic inhibitory interneurons like those that contribute to selectivity in CORE
(Rosen and Mooney, 2000). Thus, TUT- and BOS-tuned excitatory neurons in DLM and
LMAN
SHELL
may drive intrinsic inhibitory activity onto specific subgroups of SHELL neurons
(BOS- and TUT-suppressed neurons) and contribute to greater selectivity (Blättler and
Hahnloser, 2011).
Comparisons of response strength and selectivity across all TUT-responsive neurons
showed that individual cells in 45 dph birds were more selectively tuned to TUT in SHELL than in
CORE. This selectivity was attributable to the fact that SHELL neurons showed a much weaker
response to non-tutor songs than did CORE neurons. In addition, BOS-responsive CORE neurons
at 45 dph were less selective for BOS than SHELL neurons. By the time birds reached an
intermediate stage of sensorimotor integration (60 dph), a robust representation of tutor song was
no longer present in SHELL. By this age, BOS-suppressed responses have emerged in CORE, and
CORE neurons have developed a selective representation of self-produced sounds (Eliades and
Wang, 2008). The earlier maturation of selective tuning in LMAN
SHELL
suggests that SHELL
neurons may contribute to emergent auditory tuning in CORE neurons between 45 and 60 dph.
60 dph corresponds to the earliest time when birds have formed a fairly stable sequence of song
syllables and lesions to LMAN (including both CORE and SHELL) have lost the ability to
substantially disrupt song behavior (Bottjer et al., 1984). Additionally, the volume of
74
LMAN
SHELL
increases dramatically during early sensorimotor integration and regresses by 60
dph (Johnson and Bottjer, 1992, Johnson et al., 1995). Thus, the suite of developmental changes
that occurs in LMAN by this age, including the loss of TUT-tuned SHELL neurons, may reflect a
diminished capacity for plasticity, and thereby contribute to the close of the sensitive period for
vocal learning.
A neural memory of the tutor song represents a necessary target to which feedback of
current song behavior can be compared (Thorpe, 1958, Konishi, 1965). Multiple sites are likely
to encode a representation of the tutor song (Nick and Konishi, 2005b, London and Clayton,
2008, Prather et al., 2010, Adret et al., 2012), some of which may convey information to
LMAN
SHELL
. Another source of tutor song information may come through a postsynaptic target
of LMAN
SHELL
, dNCL (dorsal caudolateral nidopallium; Fig. 10), which contributes to the trans-
cortical recurrent loop. Indeed, singing in combination with hearing tutor song strongly induces
immediate early gene expression in dNCL neurons (Bottjer et al., 2010).

75

Figure 10. Feedforward and feedback LMAN
SHELL
pathways. SHELL circuitry (red) contributes
both to feedback pathways through thalamus and through integrative feedforward pathways.
One feedforward pathway includes a projection from AI
d
(previously referred to as Ad) to
dopaminergic neurons in VTA, and then to basal ganglia of the CORE pathway (Area X, gray).
Feedforward pathways also project through a dorsal thalamic zone to MMAN and thence to
HVC which controls adult song production through its projection to vocal motor cortex (RA).
Main feedback pathways include those discussed in the text:
SHELL→dNCL→AI
d
→DTZ→SHELL and SHELL→Area X
shell
→DTZ→SHELL. Additional
recurrent loops may be made through AI
d
’s projection to VTA and lateral hypothalamus. In
addition, LMAN
SHELL
projects to ipsilateral posterior amygdala which sends axons to
contralateral LMAN
SHELL
, thereby constituting one pathway that could be used for
interhemispheric coordination (Johnson et al., 1995). Posterior pallial amygdala (PoA) was
referred to as Av (ventral arcopallium) in previous papers from our lab, but we have changed the
terminology here to conform to the nomenclature suggested by Reiner et al. 2004. Dashed line
indicates midline. Abbreviations: HVC (common name); RA (robust nucleus of arcopallium);
dNCL (dorsal caudolateral nidopallium); AI
d
(dorsal intermediate arcopallium); LMAN (lateral
magnocellular nucleus of the anterior nidopallium); MMAN (medial magnocellular nucleus of
the anterior nidopallium); DTZ (dorsal thalamic zone; includes DLM and DMP); Area X (basal
ganglia nucleus containing both striatal and pallidal neurons); LH (lateral hypothalamus); VTA
(ventral tegmental area); PoA (posterior pallial amygdala).

76
The presence of TUT-tuned neurons during a restricted interval of vocal learning
suggests that they may serve as a referent that enables a correct imitation of tutor song to
gradually emerge during sensorimotor integration. During vocal babbling, juvenile birds will
occasionally produce a sound that is similar enough to tutor song to activate TUT-selective
neurons; this would enable TUT-tuned neurons to act as a filter for correct performance.
Concurrent feedback of babbling sounds could drive BOS-selective neurons such that both TUT-
tuned and BOS-tuned neurons are activated simultaneously. Convergence of signals from TUT-
and BOS-tuned neurons in downstream circuitry would constitute a means of evaluating matches
between current and target behavior. Interestingly, the strong transient projection from juvenile
CORE to AI
d
(Miller-Sims and Bottjer, 2012) (Fig. 1B) during early stages of song development
indicates that SHELL circuitry receives an efference copy of the motor signal for vocal output. In
addition, AI
d
projects to the region of dopaminergic neurons in the ventral tegmental area that
projects to Area X (Bottjer et al., 2000), and hence could relay error/reinforcement signals into
the cortico-basal ganglia circuits that collectively regulate vocal learning.
Auditory responses are generally reduced during alert wakefulness in the song motor
system (Dave et al., 1998, Cardin and Schmidt, 2004), although BOS-selective responses have
been observed in LMAN
CORE
of awake birds (Hessler and Doupe, 1999). Our results were
obtained in anesthetized birds and the role of behavioral state in neural responses of SHELL
neurons is not known. It is possible that in addition to or instead of an online role, TUT-tuned
neurons in LMAN
SHELL
are important for offline learning that occurs during sleep (Shank and
Margoliash, 2009). Chronic recordings in CORE and SHELL of juvenile singing birds will be
needed to test directly the role of TUT- and BOS-selective neurons.
77
Models of reinforcement learning in which motor skills are acquired via evaluation of
goal-oriented outcomes suggest that parallel basal ganglia circuits like those of LMAN
CORE
and
LMAN
SHELL
mediate different aspects of learning (Nakahara et al., 2001, Samejima and Doya,
2007). In mammals, neurons in parallel cortico-basal ganglia loops process distinct aspects of
reward-based learning such as reward prediction error and selected motor actions (Hikosaka et
al., 1999, Miyachi et al., 2002, Tanaka et al., 2004, Yin and Knowlton, 2006, Atallah et al., 2007,
Samejima and Doya, 2007, Graybiel, 2008, Yin et al., 2009, Thorn et al., 2010, Redgrave et al.,
2011). Thorn et al. (2010) describe a model for cortico-basal ganglia circuitry in which
sensorimotor pathways regulate motor-related aspects of learning whereas associative-limbic
circuits process goal-directed responses and modulate the sensorimotor loop’s control over
action. Models based on reinforcement learning have suggested that the LMAN
CORE
circuit may
function as an “actor/experimenter” to drive variable vocal production (Sutton, 1998, Doya and
Sejnowski, 2000, Troyer and Doupe, 2000, Fiete et al., 2007, Miller et al., 2010, Fee and
Goldberg, 2011). The presence of neurons representing the target behavior—a memory of the
tutor song—in juvenile LMAN
SHELL
is consistent with a model in which SHELL participates in
evaluative goal-oriented responses (Abe and Watanabe, 2011). Like SHELL neurons, associative
cortico-basal ganglia neurons in prefrontal cortex of mammals show activity related to goal
states (Saito et al., 2005, Mushiake et al., 2006, Fritz et al., 2010). The LMAN
SHELL
circuit
contributes to both integrative feedforward and feedback pathways, including connections to
limbic and reward circuitry (Fig. 10), and so is well suited to integrate these signals with specific
song-related information.
The presence and nature of the auditory responses in LMAN
SHELL
, in addition to lesion
and anatomical evidence (Johnson and Bottjer, 1992, Johnson et al., 1995, Bottjer et al., 2010,
78
Bottjer and Altenau, 2010, Miller-Sims and Bottjer, 2012), challenge the idea of a single song-
control basal ganglia pathway. More generally, these findings show that two distinctly tuned
neuronal populations can represent current and target behaviors during learning. Therefore these
results raise the possibility that the division of learning-related functions into parallel cortico-
basal ganglia loops may be a conserved feature of goal-directed motor learning.

79
Chapter 3: Chronic recordings in LMAN CORE and SHELL in singing juvenile
birds
ABSTRACT
Early phases of procedural learning require attention and practice, during which sensory
feedback of a given motor action is compared to an internal sensory model, and the motor
program altered accordingly. Songbirds learn to copy the song of an adult tutor during a
sensitive period of development and therefore provide a valuable model of imitative procedural
learning. As in mammals, parallel cortico-basal ganglia loops in the songbird are required during
the motor learning period, however little is known about the dynamics of neural activity in these
pathways throughout the early learning stages. Here we show that during the initial and later
phases of learning, neurons in the CORE and SHELL cortico-basal ganglia loops exhibit firing rate
modulations during singing and during a pre-singing period prior to song syllable onsets. As the
CORE pathway outputs to motor circuitry that controls the vocal muscles, this activity in CORE is
thought to represent a premotor signal. However, pre-singing responses in the SHELL pathway,
which is not associated with downstream motor neurons, likely represent an efference copy of
the premotor signal. This is supported by the findings that neurons in SHELL exhibited lower
response strength and less bursting during singing than did CORE neurons, and that significant
pre-singing responses in SHELL lagged those in CORE by 12-14 ms. Additionally, between the
early and later phases of learning there was a shift in the singing responses of SHELL neurons
from mostly suppressed responses to mostly excited responses. One idea is that responses during
singing in SHELL reflect an evaluative signal used to construct a reinforcement signal through its
downstream projection to dopaminergic neurons in the ventral tegmental area. In this model, the
80
relative decrease in suppressed responses and increase in excited responses may reflect the
progression of learning and production of more reinforced vocalizations.
INTRODUCTION
Identifying mechanisms of sensorimotor integration and imitative motor learning in cortico-basal
ganglia pathways is critical for understanding normal procedural learning in humans as well as a
wide range of disease states, from autism spectrum disorder to Parkinson’s disease. Songbirds
provide an important model for the study of motor learning because they learn to produce a
specific vocal pattern early in life by listening to the song of an adult male tutor (Doupe and
Kuhl, 1999, Brainard and Doupe, 2013). Like humans, songbirds learn to imitate those
communication sounds by vocalizing and using sensory feedback to compare their own incipient
babbling sounds to the memory of tutor sounds (Konishi, 1965, Marler, 1970). In zebra finches
(Taeniopygia guttata), this period of sensorimotor integration begins when young males produce
variable babbling sounds (subsong) starting at approximately 35 days post hatch (dph), and
continues until the birds reach adulthood (90 dph) and are able to produce a stereotyped imitation
of the tutor song.
Parallel cortico-basal ganglia loops in the songbird are required during the learning
period for accurate imitation of the tutor song (Bottjer et al., 1984, Sohrabji and Nordeen, 1990,
Scharff and Nottebohm, 1991, Bottjer and Altenau, 2010). The output of these pathways is
provided by the cortical nucleus, LMAN (lateral magnocellular nucleus of the anterior
nidopallium); LMAN is composed of two sub-regions, a core and a surrounding shell, which
give rise to independent topographically organized pathways (Johnson and Bottjer, 1992, Iyengar
et al., 1999, Pinaud and Mello, 2007, Person et al., 2008) (Fig.1). LMAN
CORE
projects to a motor
81
cortical region, RA (robust nucleus of the arcopallium), which innervates hindbrain motor
neurons that control the vocal musculature (Nottebohm et al., 1976, Vicario, 1991, Wild, 1993);
LMAN
SHELL
projects to a region adjacent to RA within motor cortex, AI
d
(dorsal intermediate
arcopallium), which forms recurrent loops including projections through limbic and reward
circuitry (Johnson et al., 1995, Bottjer et al., 2000).
The CORE cortico-basal ganglia pathway functions as a premotor driver of early
vocalizations (Aronov et al., 2008, Aronov et al., 2011, Goldberg and Fee, 2012). In contrast to
the motor role of the CORE pathway during subsong, the SHELL pathway has no direct motor role.
Lesions of AI
d
of the SHELL pathway in juveniles produce no an immediate disruption of song or
any long-term effect on syllable phonology. However, AI
d
-lesioned birds were eventually
unable to copy tutor song syllables or produce stable song sequences (Bottjer and Altenau,
2009). The lack of immediate effect followed by an eventual impairment of imitative learning
suggests that the SHELL pathway is involved in evaluating motor learning. In addition, a
population of neurons in LMAN
SHELL
of juvenile birds respond selectively only to their tutor song

(Achiro and Bottjer, 2013), indicating that the SHELL pathway has access to a representation of
the tutor song. In addition, juvenile RA-projecting LMAN
CORE
neurons send a transient
projection into AI
d
, such that a copy of the subsong motor signal is conveyed to the SHELL
pathway

(Miller-Sims and Bottjer, 2012). Together, a representation of the tutor song and a copy
of the motor signal could be used for vocal evaluation during learning.
In mammals, distinct sensorimotor and associative cortico-basal ganglia loops contribute
to different aspects of motor learning (Samejima and Doya, 2007, Graybiel, 2008, Yin et al.,
2008, Yin et al., 2009, Ashby et al., 2010, Redgrave et al., 2010, Thorn et al., 2010, Gremel and
Costa, 2013). My hypothesis is that like in mammalian pathways, neurons in both songbird
82
parallel cortico-basal ganglia circuits are active during learning, however the activity is
differentially modulated in the two pathways such that CORE neurons participate in motor related
actions and SHELL neurons participate in evaluating motor performance (Histed et al., 2009, Yin
et al., 2009, Thorn et al., 2010, Gremel and Costa, 2013, Kim et al., 2013).
83

Figure 1. Parallel cortico-basal ganglia pathways in songbirds. LMAN CORE (gray) and SHELL
(red) regions form parallel pathways through the basal ganglia (Area X, contains both striatal and
pallidal neurons) and thalamus (DLM). LMAN projects to the analog of mammalian motor
cortex: RA and AI
d
. One additional pathway includes a projection from AI
d
to dopaminergic
neurons in VTA which project to the basal ganglia. Abbreviations: LMAN, lateral
magnocellular nucleus of the anterior nidopallium; AI
d
, dorsal intermediate arcopallium; RA,
robust nucleus of the arcopallium; DLM, dorsolateral medial thalamus, DL, dorsolateral, VM,
ventromedial; VTA, ventral tegmental area.

84
MATERIALS AND METHODS
Subjects
All procedures were performed in accordance with protocols approved by the University of
Southern California Animal Care and Use Committee. Nine juvenile male zebra finches
(Taeniopygia guttata) were used, 43-60 days post hatch (dph) at the time of neural recordings.
Birds were birds raised under naturalistic conditions (by their parents within group aviaries) until
they were at least 38 dph, ensuring normal tutor song exposure and social experience
(Immelmann, 1969, Böhner, 1983, Eales, 1985, Clayton, 1987, Böhner, 1990, Catchpole and
Slater, 1995, Mann and Slater, 1995, Roper and Zann, 2006).
Electrophysiology
Birds were anesthetized with isoflurane (1.5% inhalation) and an electrode assembly consisting
of seven tungsten wire (diameter 25 µm) stereotrodes was implanted into LMAN CORE and
SHELL. One stereotrode was implanted 1.5 mm dorsal to LMAN to serve as a reference
electrode, and silver wire (diameter 250 µm) was placed between the skull and skin for the
animal ground. Signals were acquired through a unity gain headstage (Neuralynx) with a
flexible cable that connected the electrode assembly to a commutator (Neuralynx) and the
amplifiers (Lynx8, Neuralynx). The audio signal was acquired through a Sanken COS-11D
microphone. Spike data was bandpassed between 300-6000 Hz and digitized at 32 kHz using
Spike 2 software (Cambridge Electronic Design). At the end of each experiment, small
electrolytic lesions (7 µA for 20 seconds) were made to confirm recording locations. Coronal
sections (50-µm thick) were Nissl-stained to reveal LMAN
CORE
borders, and an alternate series
85
was stained for calbindin using standard immunohistochemical procedures to confirm
LMAN
SHELL
borders (Pinaud et al., 2007). Isolated units were clustered automatically as
described previously (Achiro and Bottjer, 2013), using KlustaKwik (Ken Harris, Rutgers
University) and refined manually using MClust (A. David Redish, University of Minnesota) in
MATLAB (Mathworks). Single units were considered for analysis if signal to noise ratio was >
3 and if less than 1% of spikes had an interspike interval < 2 ms.
Song analysis
Song analysis was conducted with custom MATLAB software, applying many features and
methods originally used in Sound Analysis Pro (Tchernichovski et al., 2000). Song bouts and
syllables were automatically detected using amplitude threshold crossings and manually checked
to remove cage noise. Because juvenile birds in early stages of song learning do not produce
stereotyped song motifs, episodes of singing were defined as periods of continuous singing
separated by gaps of at least 300 ms. For analysis at the level of individual syllables,
automatically parsed syllables were verified manually and boundaries were adjusted where
needed.
In order to empirically determine whether a bird was in the subsong or plastic song stage
each day, we utilized methods previously described to define song stage (Aronov et al., 2011).
Juvenile birds in the subsong stage of sensorimotor integration produce syllables of variable
lengths, the distribution of which is well-fitted by an exponential function. As birds progress to
the plastic song phase, they produce more regular syllable types which begin to appear as peaks
in the distribution of syllable durations, and therefore are no longer well-fit by an exponential
(Tchernichovski et al., 2004, Aronov et al., 2011) . Based on this evidence, to assign the stage of
86
subsong or plastic song, we fit an exponential to all syllable duration distributions for each bird,
for each day of singing (500-8,000 syllables). We used the lilliefors test (function in MATLAB,
lilliefors) to quantify the goodness-of-fit, and the resulting goodness-of-fit coefficient was
scaled by the number of syllables. We found that distributions which were well-fit by the
exponential had goodness-of-fit coefficients less than 12 and therefore we determined
distributions with coefficients less than 12 to come from subsong vocalizations, and those more
than 12 were determined to be from plastic song vocalizations.
Analysis of neural activity
For the analysis of neural activity throughout singing episodes, singing was defined as the time
during an episode of singing plus the preceding 100 ms to include pre-episode related activity.
Baseline periods were defined as periods of silence lasting at least two seconds, which were at
least two seconds away from singing to avoid any pre- or post-singing responses. We
determined whether each single unit was responsive by testing for a significant change in firing
rate from baseline during singing (independent t-test, due to differing number of baseline and
singing episodes, p < 0.05). For response strength measures, the average of the two baseline
periods nearest in time to the singing event was used as a local baseline. To compare responses
in neurons with differing firing rates, standardized responses (SR) were calculated:
𝑆 𝑅 =
𝑆 − 𝐵 � 𝑉 𝑎𝑟 ( 𝑆 ) + 𝑉 𝑎𝑟 ( 𝐵 ) − 2 ∗ 𝐶 𝑜𝑣 𝑎 𝑟 ( 𝑆 , 𝐵 )
∗ √ 𝑛
where S is the mean firing rate during singing, B is the mean firing rate during baseline, and n is
the number of singing/baseline pairs. Therefore, a positive value indicates an increased rate
during singing and a negative value indicates a decreased rate during singing. To measure the
87
incidence of bursts during singing and baseline, we calculated a burst fraction by taking the
fraction of spike events with an interspike interval of less than 10 ms.
We also analyzed pre-singing related activity on a syllable by syllable level. We tested
for a significant change in firing rate between baseline periods and the 50 ms prior to each
syllable onset (independent t-test, p < 0.05). To analyze the temporal pattern of firing leading up
to syllable onsets, histograms of population activity were made by taking the mean subtracted
firing rate for each neuron and then averaging across neurons and smoothing with a Gaussian (2
ms bin size, 40 ms smoothing).
To determine responses at the syllable level, we calculated firing rates during the syllable
and the 50 ms prior to syllable onset to include pre-syllable related activity. In juveniles, the
average gap between syllables is ̴ 60 ms (Aronov et al., 2011, Glaze and Troyer, 2013), therefore
it is unlikely that activity during a previous syllable was included in these firing rates. To
determine selectivity of responses to specific syllable types (methods for syllable labeling
described below), we calculated an activity fraction (AF) (Vinje and Gallant, 2000, Meliza and
Margoliash, 2012) :
AF =
1 − � �
∑ 𝑟 𝑖 𝑛 �
�
2
∑ �
𝑟 𝑖 2
𝑛 �
� � �
1 −
1
𝑛 �

where r is the firing rate to the ith syllable type and n is the number of syllable types. Thus a
score of 0 indicates no selectivity for a specific syllable type and 1 indicates maximum
selectivity. In order to align syllables to construct peristimulus time histograms, we linearly time
warped each spike train to the average length for that syllable type.
88
Syllable labeling
As indicated above, juvenile birds produce highly variable sequences of syllables, therefore it is
not possible to align singing episodes to analyze the temporal pattern of neuronal activity across
songs (Tchernichovski et al., 2001). However, one can align multiple renditions of syllables of
the same type in birds in the plastic song stage which produce some recognizable syllables.
Because of the high variability of juvenile syllables, we created custom software in MATLAB
(using some features created for Sound Analysis Pro (Tchernichovski et al., 2000)) in order to
semi-automatically label syllables in birds during the plastic song stage. To assign syllables into
groups/types, we employed a combination of two measures of the acoustic distance between
syllables which were then used to cluster the syllables.
For the first distance measure we used the following syllable features: fundamental
frequency, center frequency, frequency modulation, pitch goodness (estimate of the amount of
periodic energy), Wiener entropy (estimate of spectral disorder), total amplitude, derivative of
total amplitude, amplitude modulation and the maximum temporal derivative across each
frequency’s power. We calculated the following summary statistics for each feature over the
length of the syllable: mean, standard deviation, maximum, minimum, onset (average of samples
1-3%), middle (average of samples 49-51%), offset (average of samples 97-100%), correlation
coefficient of the linear trend, time of maximum, and time of minimum. Each syllable was thus
represented as a point in high-dimensional space where each dimension was a summary statistic
of one of the features. We then calculated the Euclidian distances between each point.
For the second distance measure, we calculated the following syllable features for each
time point in the syllable (9.27 ms window size, 7.91 ms overlap): Wiener entropy, frequency
89
modulation, amplitude modulation, fundamental frequency and goodness of pitch
(Tchernichovski et al., 2000). Syllables were represented as multidimensional feature vectors in
time. The distance between the vectors was calculated with a dynamic time-warping algorithm
including a warping penalty (Vintsyuk, 1968). The time-warping algorithm tolerates small
perturbations in the timing of sub-syllabic events common to juvenile syllables.
In order to normalize the measures, we converted both distance measures to percentile
scores based on the empirical distribution (computed over all pairs of syllables). We averaged
the two scores to calculate the final combined distance score which ranged between 0 (perfect
similarity) and 1 (no similarity). Syllables were clustered into 4-25 types using hierarchical
agglomerative clustering with complete linkage (functions provided by MATLAB, cluster
and linkage). We then manually selected the number of types for each bird for each day,
confirmed cluster quality, and merged clusters which were similar. Some clusters were manually
rejected due to insufficient consistency.
Statistics
All continuous data were tested for normality using Kolmogorov-Smirnov and Shapiro-Wilk
tests. t-tests were used to compare means for normally distributed data, and Mann-Whitney tests
were used to compare medians for non-normally distributed data. Differences in proportions
were tested using chi-square tests, and ANOVAs were used to compare differences across
subsong and plastic song in addition to those between LMAN
CORE
and LMAN
SHELL
. Bonferroni
corrections were applied wherever multiple comparisons were made.

90
RESULTS
Singing-modulated neurons in LMAN CORE and SHELL of juvenile birds
Across all neurons recorded in juvenile LMAN (43-60 dph), the majority in LMAN
CORE
showed
significant firing rate modulation (either excitation or suppression) during singing compared with
baseline periods (86%; 101/118 neurons, independent t-tests). Similarly, the majority of
LMAN
SHELL
neurons showed significant modulation of firing rates during singing (73%; 109/149
neurons), however the proportion was smaller than that in CORE (p = 0.01, chi-square). Juveniles
produce syllable sequences which are highly inconsistent, making it unfeasible to align song
motifs in order to get firing rate histograms across renditions (Tchernichovski et al., 2001).
Across multiple examples of singing periods in a 53 dph bird (Fig. 2), the activity of a single
neuron in LMAN
CORE
was variable across and within song bouts, however peaks in firing often
occurred at syllable onsets, as shown previously (Aronov et al., 2008).

The activity of a single
SHELL neuron in a 50 dph bird (Fig. 3) also showed variable firing during singing with firing rate
peaks at some syllable onsets. We also examined changes in activity prior to each syllable onset.
Across the population of juvenile CORE neurons, 64% (75/118 neurons) showed a significant pre-
singing response during the 50 ms before syllable onsets. A somewhat smaller percentage of
SHELL neurons showed significant pre-singing activity (52%; 78/149 neurons; p = 0.07, chi-
square).
91

Figure 2. Example of a single CORE neuron during singing in a juvenile bird (53 dph). Six
examples of song behavior, shown as spectrograms (bottom portion of each example). Above
each spectrogram is the raster of the CORE neuron firing during that period and its instantaneous
firing rate (spikes/s). Bin size for firing rate = 10 ms.

92

Figure 3. Example of a single SHELL neuron during singing in a juvenile bird (50 dph). Six
examples of song behavior, shown as spectrograms (bottom portion of each example). Above
each spectrogram is the raster of the SHELL neuron firing during that period and its instantaneous
firing rate (spikes/s). Bin size for firing rate = 10 ms.

93
Classification of subsong and plastic song behavior
For each day, the singing behavior of each bird was empirically determined to be in the subsong
or plastic song stage based on syllable duration distributions (Fig. 4; see Materials and Methods).
Juveniles in the subsong stage of sensorimotor integration produce largely unrecognizable
syllables of various durations, and the distribution of syllable durations are well-fit by an
exponential function. As birds progress to the plastic song stage of sensorimotor integration,
they begin to produce recognizable syllables that appear as peaks in the distribution of syllable
durations, causing the distributions to no longer be approximated by an exponential function
(Tchernichovski et al., 2004, Aronov et al., 2011). Goodness-of-fit coefficients from an
exponential fit to the syllable duration distributions were used to assign birds to the subsong or
plastic song stage for each day of singing, where smaller coefficients indicate better fit to the
exponential and therefore more likely to result from subsong rather than plastic song syllables.
94

95
Figure 4. Progression of subsong to plastic song. The left panel shows example spectrograms of
syllables from the same bird at age 47, 49, 50 and 51 dph. The right panel shows the
distributions of syllable durations per day for this bird. The red line indicates the exponential fit
for the distributions. The coefficients shown for each day are the goodness-of-fit coefficients
from the lilliefors test for fit to an exponential function (lower numbers indicate better fit).
Based on the coefficients, this bird was considered in the subsong stage at 47 and 49 dph and in
the plastic song stage at 50 and 51 dph.

96
Proportions of excited and suppressed LMAN neurons during song development
Both CORE and SHELL neurons of birds in the subsong stage showed significant modulation of
response strength during singing: spiking activity was either excited or suppressed during singing
episodes. Table 1 shows that a large proportion of neurons in both subregions showed rate
suppression during singing in subsong-stage birds: 40% (23/57) of neurons in CORE and 58%
(23/40) of neurons in SHELL showed overall suppression in mean response strength across
singing episodes (p = 0.10, chi-square). However, relatively more neurons in SHELL were
suppressed during singing whereas relatively more neurons in CORE were excited during singing.
This pattern indicates that during subsong production, the majority of singing-modulated neurons
in LMAN
SHELL
are suppressed across the motif. The proportion of neurons that showed excitation
during singing increased in both CORE and SHELL by the plastic song stage (Table 1), although
this increase was significant only for SHELL (p = 0.02, chi-square). Thus, the relative increase in
the proportion of neurons that were singing-excited from subsong to plastic song was much
greater in SHELL than in CORE neurons (150% in SHELL compared to 80% in CORE). Despite this
fact, LMAN
CORE
contained a somewhat greater proportion of neurons that showed excitation
during singing than did LMAN
SHELL
during the plastic song stage (73%, 32/44 CORE neurons
compared to 65%, 41/63 SHELL neurons); this difference failed to reach significance (p = 0.40,
chi-square). In summary, the majority of neurons in LMAN
CORE
were singing-excited during
subsong production, and the proportion of excited neurons increased by the plastic song stage.
The majority of neurons in LMAN
SHELL
were singing-suppressed during subsong production, but
this pattern changed such that the majority of SHELL neurons were singing-excited by the plastic
song stage.
97

Table 1. Ratios of singing-excited to singing-suppressed neurons in juvenile CORE and SHELL.
Excitation and suppression is based on standardized response strengths during a singing episode
and the 50 ms preceding the episode compared to baseline activity.

Ratio excited:suppressed firing during singing
CORE
subsong
(n = 57)
1.48 (34/23)
plastic song
(n = 44)
2.67 (32/12)
SHELL
subsong
(n = 40)
0.74 (17/23)
plastic song
(n = 63)
1.86 (41/22)
98
Response strength of excited and suppressed LMAN neurons during song development
Baseline firing rates increased in both CORE and SHELL from the subsong to the plastic song
stage, although this increase was significant only in SHELL (SHELL 1.62 ± 0.26 spikes/s to 3.58 ±
0.90; mean ± s.e.m., p = 0.016, independent t-test; CORE 3.49 ± 0.80 to 5.07 ± 1.58, p = 0.37,
independent t-test). SHELL neurons had significantly lower baseline rates compared to CORE
during the subsong stage (p = 0.007, independent t-test). To compare firing rates across ages and
neurons with differing baseline rates, we calculated standardized response strengths for each
neuron (see Materials and Methods). Both CORE and SHELL neurons showed robust changes in
response strength during singing compared to baseline during subsong and plastic song stages; as
indicated above, neurons were either excited or suppressed (Fig. 5A, C). To compare CORE
versus SHELL neurons across song stages and excitation and suppression, we tested the absolute
value of the standardized response strengths using an ANOVA. This analysis revealed a main
effect of CORE versus SHELL (p = 0.018) and subsong versus plastic song (p < 0.001), reflecting
the lower response strength in SHELL compared to CORE and the increase in response strength
from subsong to plastic song in both regions. The absolute response strength of excitation versus
suppression was not significant (p = 0.62). Interestingly, response strength increased from the
subsong to the plastic song stage for singing-excited neurons in both CORE and SHELL (CORE p =
0.03, Mann-Whitney U; SHELL p < 0.001, independent t-test), but this trend was not significant
for singing-suppressed neurons (CORE p = 0.23; SHELL p = 0.30; Mann-Whitney U).
In order to determine if there was increased bursting during singing as has been reported
previously for LMAN
CORE
neurons in plastic song birds (Ölveczky et al., 2005), we analyzed the
fraction of spikes that occurred within a burst for each neuron (fraction of spikes with interspike
99
intervals < 10 ms). During subsong production, there was a small but non-significant increase in
mean burst fractions from baseline to singing in both CORE and SHELL (Fig. 5B; p = 0.32, no
effect of CORE versus SHELL, p = 0.50, ANOVA). During plastic song production, there was a
significant effect of both singing (p = 0.006) and CORE versus SHELL (p = 0.011 ANOVA; Fig.
5D) on burst fraction, showing that bursting increased significantly during production of plastic
song and that more bursting occurred in CORE than in SHELL neurons. A planned comparison of
mean burst fraction in SHELL neurons during baseline versus singing, revealed a significant
increase in bursts during the production of plastic song (p = 0.04, Mann-Whitney U). This
pattern results from the fact that in SHELL, there is a reduction in the burst fraction during
baseline periods from the subsong to plastic song stage (p = 0.006, Mann-Whitney U, Bonferroni
corrected). The planned comparison of burst fraction in CORE neurons during baseline versus
singing showed that the increase in bursts during the production of plastic song did not reach
significance (p = 0.58, Mann-Whitney U).
100

Figure 5. Firing characteristics during singing in juvenile CORE (gray) and SHELL (red). A,
Standardized response strength of singing-excited and singing-suppressed neurons in CORE and
SHELL of birds singing subsong. CORE n = 57, SHELL n = 40, see Table 1 for excitation to
suppression split. B, Fraction of spikes that occur in bursts (interspike intervals < 10 ms) during
baseline and singing subsong. CORE = 56, SHELL n= 40, neurons included whose firing rates
were significantly modulated during singing compared to baseline and had a firing rate < 0 Hz
101
during singing. C, Standardized response strength of singing-excited and singing-suppressed
neurons in CORE and SHELL of birds singing plastic song. CORE n = 44, SHELL n = 63. D, Fraction
of spikes that occur in bursts (ISIs < 10 ms) during baseline and singing plastic song. CORE = 40,
SHELL n= 62. * indicates p < 0.05.

102
Time course of excitatory pre-singing activity in CORE and SHELL
To investigate the time course of pre-singing related signals in CORE and SHELL, we constructed
population histograms that show the average mean-subtracted response of neurons which were
excited during the 50 ms prior to each syllable onset (Fig. 6). In both subsong and plastic song
stages, a higher proportion of neurons in CORE were pre-singing excited compared to SHELL,
although this difference was significant only if song stage was collapsed (subsong: CORE 43%, n
= 26/60 neurons; SHELL 29%, n = 20/68; plastic song: CORE 40%, n = 22/55 neurons, SHELL 25%,
n = 19/76; p = 0.013, chi-square). The population histogram shows that during subsong
production, both CORE and SHELL neurons showed significant pre-singing related activity which
peaked at approximately 10 ms prior to syllable onset. In CORE, pre-singing activity was
significant from 40 ms prior to syllable onset to 8 ms after syllable onset, whereas in SHELL,
significant responses began 12 ms after that in CORE (from -28 ms to +4 ms). There was also
significant transient suppression approximately 40 ms following syllable onset in CORE and
SHELL, although suppression lasted longer in CORE, to 82 ms after syllable onset. There was also
points of significant responses further from syllable onsets, however responses outside of ~100
ms likely reflect responses to preceding and succeeding syllables within the song (Aronov et al.,
2011, Glaze and Troyer, 2013).
During plastic song production, significant pre-singing activity in CORE ranged from 30
ms prior to syllable onset to 10 ms after syllable onset. In SHELL, pre-singing responses lagged
those in CORE by 14 ms (from -16 to 0 ms). Significant suppression occurred at approximately
the same time as suppression during subsong production (44-88 ms following syllable onset),
103
however this occurred in CORE but not SHELL for plastic song syllables. We found no significant
offset responses in either CORE or SHELL in juvenile birds (data not shown).
104

Figure 6. Syllable onset responses in CORE (gray) and SHELL (red) in birds singing subsong (top
panel) and plastic song (bottom panel), includes neurons significantly excited during the 50 ms
preceding syllable onsets. Solid lines show the smoothed average mean-subtracted rate
histograms aligned to syllable onsets (syllable onset at time 0), shading indicates s.e.m. Bars
above and below the traces indicate times at which the rate change is significant (95%
confidence interval outside of zero; located above traces for increased rates, below for decreased
rates). Subsong CORE n = 26, SHELL n = 20; Plastic song CORE n = 22, SHELL n = 19.

105
Neural responses during repeated renditions of emerging syllable types
Although juvenile song sequence is variable (Tchernichovski et al., 2001), we attempted to
analyze the temporal specificity of firing by aligning neural activity to multiple renditions of
plastic song syllables of the same type. To label syllables, we created custom software in
MATLAB (including features created for Sound Analysis Pro (Tchernichovski et al., 2000)) to
semi-automatically characterize syllables into labeled groups (Fig. 7A; see Materials and
Methods). Briefly, we calculated an acoustic similarity measure that was used to cluster
syllables into groups. This measure consisted of one quantity calculated from the similarity of
various feature statistics between syllables, and one quantity calculated from the similarity
between syllables of various features along the length of each syllable (Fig. 7B).

106

107
Figure 7. Examples of syllables assigned to different emerging syllable types, and four of the
features calculated across time in order to characterize syllables in a bird in the plastic song stage
(59 dph). A, Spectrograms for five exemplars of syllables for each syllable type, a-e, resulting
from semi-automatic labeling of syllables. B, Top row shows spectrograms for five exemplars of
syllable type e at an expanded time scale (as in A). Below are plots of four of the features
calculated across time for each syllable that are used for the distance measure needed for semi-
automatic syllable clustering (see Materials and Methods).

108
To determine if single neurons in LMAN CORE or SHELL exhibited temporal specificity in
firing either across renditions of the same syllable type or across syllable types, we constructed
peristimulus time histograms for each syllable type in one bird (Fig. 8). In order to time-align
these renditions, we linearly time warped each spike train to the average syllable length for that
syllable type. A single CORE neuron fired during four syllable types, however this neuron fired
variably from rendition to rendition, indicating no apparent temporal specificity for specific time
points in any syllable. Although it appears that this neuron fired most during the second syllable
type, this is due to the fact that the second syllable type is significantly longer than the other
types. When accounting for syllable length by measuring firing rate, this neuron showed similar
responses across the syllable types (types 1-4 firing rate (spikes/s): 7.88 ± 1.58; 10.5 ± 0.97; 10.5
± 0.97; 10.3 ± 1.63; p = 0.63 ANOVA for syllable type). A single SHELL neuron from the same
day in the same bird showed similar activity. Firing across renditions was not specific to time-
points in the syllables and this neuron also showed similar firing rates across syllable types
(types 1-4 firing rate: 11.4 ± 1.77; 10.4 ± 0.85; 10.2 ± 2.29; 13.9 ± 2.13; p = 0.45 ANOVA for
syllable type).
To calculate the selectivity of responses to specific syllable types of birds in plastic song
we computed an activity fraction for each neuron (AF; see Materials and Methods) (Vinje and
Gallant, 2000, Meliza and Margoliash, 2012), where a score of 0 indicates no selectivity for a
specific syllable type and 1 indicates maximum selectivity. On average, both CORE and SHELL
neurons displayed low selectivity for specific syllable types (CORE AF = 0.29 ± 0.07; SHELL AF =
0.29 ± 0.04). However, the cumulative distributions of AF scores revealed that in both regions,
some neurons show relatively high selectivity for specific syllable types (Fig. 9).
109

110
Figure 8. Examples of responses of single LMAN CORE and SHELL neurons to various syllable
types in a 53 dph bird. Top row shows spectrograms of syllable type. Below are the raster plots
and peristimulus time histograms for each syllable for a single CORE neuron and below that, for a
single SHELL neuron.

111

Figure 9. Quantification of plastic syllable type selectivity of LMAN CORE and SHELL neurons.
Cumulative distribution of AF scores for CORE (gray; n = 18) and SHELL (red; n = 32) neurons
from plastic song syllables.

112
DISCUSSION
In juvenile birds producing subsong or plastic song, neurons in both LMAN
CORE
and LMAN
SHELL

show significant singing-related activity, including both excitation and suppression. These data
represent the first recordings in LMAN
SHELL
during singing and suggest that SHELL plays a role in
the sensorimotor integration period in juveniles. Previous work has shown that in birds
producing subsong, neurons in LMAN
CORE
and its thalamic afferent exhibit premotor increases in
firing rate (Aronov et al., 2008, Goldberg and Fee, 2012). Our results from LMAN
CORE
confirm
these findings, and show that many SHELL neurons also display increases in firing prior to
syllable onset. LMAN
CORE
is required for the production of subsong behavior and drives vocal
output through its connection to the motor cortical region, RA, which innervates vocal motor
neurons in the hindbrain (Nottebohm et al., 1976, Vicario, 1991, Wild, 1993, Aronov et al.,
2008). Neurons in motor cortex receiving input from LMAN
SHELL
, however, have no direct
projections to any downstream motor circuits, and lesions of this region produce no immediate
changes in song production (Bottjer et al., 2000, Bottjer and Altenau, 2010). These anatomical
and behavioral results suggest that singing related activity in LMAN
SHELL
does not represent a
motor signal in that it is not likely generating motor action.
The idea that signals in SHELL during singing do not function to drive motor behavior is
reinforced by the finding that a smaller proportion of SHELL neurons showed activity which was
modulated during singing than did CORE neurons, and that those neurons which were singing-
modulated had lower response strength and exhibited less bursting during singing than did CORE
neurons. These results raise the question of what singing and pre-singing related activity in
LMAN
SHELL
signifies. One idea is that singing related activity in SHELL represents efference copy
113
signals from LMAN
CORE
. In juveniles but not adults, many RA-projecting LMAN
CORE
neurons
send a collateral axon into AI
d
(Miller-Sims and Bottjer, 2012), and therefore a copy of the
premotor signal generated by CORE neurons is routed into AI
d
. Through AI
d
’s input to
LMAN
SHELL
’s thalamic afferent, SHELL neurons could receive a copy of the CORE premotor signal.
Another possible source of an efference copy from CORE is through SHELL’s other afferent
projection, contralateral posterior amygdala, which also receives input from LMAN
CORE
(Johnson
et al., 1995, Bottjer et al., 2000). Our result that significant pre-singing activity in LMAN
SHELL

lagged the activity in CORE by 12-14 ms supports either of these possibilities. During singing,
efference copy could convey a prediction of auditory feedback that could be used for vocal
evaluation. Alternatively, such signals in SHELL may represent a sensory target rather than a
detailed sensory prediction, as in human speech production (Niziolek et al., 2013).
Both CORE and SHELL regions of LMAN contain a substantial proportion of neurons that
are suppressed during singing in juvenile birds. This suppression may arise from intrinsic
inhibitory networks (Bottjer et al., 1998, Rosen and Mooney, 2000), and the decrease in singing-
suppressed neurons from subsong to plastic song could represent a correlate of learning.
Singing-suppressed responses in CORE may reflect lateral inhibition of competing CORE neurons,
which could decrease as the circuit matures into the plastic song stage. Such suppression is not
likely to persist through the two or three excitatory synapses between CORE and SHELL neurons,
and therefore suppression in SHELL may signify a different function than that in CORE. One idea
is that these responses reflect a cancellation signal that is used to compare generated motor
activity with a goal signal. Determining the precise pattern of suppression would be important,
as birds evaluate song production on at least a syllable by syllable timescale (Sakata and
114
Brainard, 2006, Tumer and Brainard, 2007, Sober and Brainard, 2009, Warren et al., 2011,
Charlesworth et al., 2012).
During later stages of sensorimotor integration, when birds are producing plastic song,
evidence suggests that the motor driver of song gradually shifts from LMAN
CORE
to HVC (high
vocal center). In adult birds, the projection from HVC to RA is necessary and sufficient for the
production of song (Nottebohm et al., 1982, Simpson and Vicario, 1990, Yu and Margoliash,
1996, Hahnloser et al., 2002, Fujimoto et al., 2011). Lesions of LMAN at later stages of
sensorimotor integration produce little or no disruption of learning (Bottjer et al., 1984). In
contrast, lesions or inactivation of HVC disrupt plastic song such that vocalizations seem to
acquire the characteristics of subsong (Aronov et al., 2008, Aronov et al., 2011). It is interesting
therefore, that the proportion of singing-excited neurons and the strength of neural responses
during singing in LMAN CORE and SHELL are lower during the subsong stage than in the plastic
song stage of learning. One idea is that although the strength of the signal to motor cortex is
relatively low during subsong, slow NMDA (N-methyl-D-aspartate) receptor currents that are
typical in RA of birds in the subsong stage (White et al., 1999) allow substantial activation in
RA despite a relatively weak signal from LMAN.
Although LMAN
CORE
may no longer drive song production during the plastic song stage,
evidence indicates that CORE neurons contribute variability to song during plastic and adult song
production. Inactivation of LMAN during the plastic song phase results in a substantial
reduction in variability of repeated renditions of song (Ölveczky et al., 2005), and pauses in
firing of HVC neurons correlate with increased variability in plastic song (Day et al., 2008). Our
results showing a lack of temporal specificity in the activity of CORE and SHELL neurons across
115
renditions of the same syllable type, in addition to the finding that most CORE and SHELL neurons
fire during production of multiple syllable types support the idea these signals represent variable
motor perturbations (CORE) and the copy of these motor explorations (SHELL).
In contrast to LMAN, the activity of Area X striatal neurons correlates well to specific
(10 ms) time points in song during the plastic song stage of development (Goldberg and Fee,
2010), as does the activity of LMAN
CORE
-projecting thalamic neurons (Goldberg and Fee, 2012).
These findings suggest that timing information, presumably generated in HVC (Hahnloser et al.,
2002), is preserved though the basal ganglia and thalamus, but is not evident at the level of
LMAN
CORE
. Such findings may indicate that LMAN
CORE
receives input that provides information
regarding the time in the song, and can use that input to locally direct changes or variability to
specific song elements. Variability in behavior is important in reinforcement learning to explore
possible actions that could lead to reward, i.e. variable song patterns represent an exploration of
motor space so that some vocalizations will be a good match to the neural representation of the
tutor song and can be reinforced (Barto et al., 1983, Doya and Sejnowski, 2000, Fiete et al.,
2007). Supporting the idea such a stratagem is actually used, songbirds have the ability to
locally regulate variability in their songs during learning, producing more variable renditions of
only the song elements that are in the process of being learned (Ravbar, 2012).
Our working model is that the CORE pathway is important for exploring possible motor
actions and implementing changes in the song motor program, whereas the SHELL pathway is
important for monitoring song motor behavior and its relation to the goal outcome. In this
arrangement, the efference copy of motor signals used to alter song could be utilized in the
SHELL pathway to compare to the neural representation of the tutor song. In LMAN
SHELL
of
anesthetized juvenile birds, a population of neurons is selectively tuned to the tutor song (Achiro
116
and Bottjer, 2013). Such a functional segregation of motor learning into parallel cortico-basal
ganglia loops is reminiscent of corresponding roles during learning of sensorimotor and
associative cortico-basal ganglia loops of mammals (Alexander et al., 1986, Parent and Hazrati,
1995, Joel and Weiner, 1997, Brown et al., 1999, Nakahara et al., 2001, Histed et al., 2009, Yin
et al., 2009, Thorn et al., 2010, Gremel and Costa, 2013, Kim et al., 2013).
Cortico-basal ganglia loops are critical for normal speech production in humans
(Lieberman, 2001, Wildgruber et al., 2001) and speech disorders such as stuttering, and language
fluency problems due to Parkinson’s disease, are attributed to abnormalities in the basal ganglia
(Cooper et al., 1991, Giraud et al., 2008). Furthermore, autism, a disorder characterized by
deficient communication and speech acquisition, is linked to the FoxP2 gene which is highly
expressed in the birdsong basal ganglia pathway during the sensitive period for vocal learning
(Haesler et al., 2004, Haesler et al., 2007, Thompson et al., 2013). Therefore, connecting findings
in mammalian cortico-basal ganglia research with those in birdsong is a vital part of developing
hypotheses relating to sensorimotor integration and to deficits in motor behavior.
117
Chapter 4: Conclusions
Models of song learning postulate that young songbirds learn to imitate their tutor’s song by
using a reinforcement learning network with a controller and a critic (Doya and Sejnowski, 2000,
Troyer and Doupe, 2000, Fiete et al., 2007, Fee and Goldberg, 2011). Combining ideas from
these papers and the results of the work presented in this dissertation, I present the following
model of song learning.
In this model, the controller has essentially two roles: an experimenter and an actor.
Variability in motor behavior is required for reinforcement learning so that motor space is
explored for possible rewarding actions. Evidence suggests that the LMAN
CORE
pathway
functions as the experimenter, injecting variability into song behavior (Ölveczky et al., 2005,
Kao and Brainard, 2006, Aronov et al., 2008). The critic functions to evaluate motor output, and
I propose the LMAN
SHELL
pathway is important for this role. To avoid issues of feedback delay,
the critic may use an efference copy map of motor activity (Miall and Wolpert, 1996), and my
results from Chapter 3 support the idea that efference copy is conveyed to LMAN
SHELL
. The
motor output would then be compared to a neural representation of the tutor song. I showed in
the work presented in Chapter 2, that a population of neurons in LMAN
SHELL
of juvenile birds
contains a representation of the tutor song (Achiro and Bottjer, 2013). The reinforcement signal
could be generated by the input of dopaminergic neurons from the VTA into the basal ganglia
(Gale et al., 2008, Person et al., 2008), instructed through the SHELL pathway’s projection to the
VTA (Bottjer et al., 2000). The role of the actor would be to carry out the behavior, and the
HVC to RA circuit likely functions in this role (Nottebohm et al., 1982, Simpson and Vicario,
1990, Yu and Margoliash, 1996, Hahnloser et al., 2002, Aronov et al., 2008, Aronov et al., 2011,
118
Fujimoto et al., 2011, Warren et al., 2011, Charlesworth et al., 2012). Therefore, learning may
occur through the strengthening of HVC-RA synapses, although future work is required to
decipher how reinforcement signals may lead to such plasticity.
Other research I am performing, which was not included in this dissertation, aims to
uncover the mechanisms underlying an evaluative signal in LMAN
SHELL
of juvenile birds. As a
first step, I am testing if SHELL neural activity during singing is correlated to how well a
particular song element is matched to the tutor song. This and future studies are important to
discover specific mechanisms underlying procedural learning in order to address the numerous
disease states associated with deficits in motor learning and production.
119
References
Abe K, Watanabe D (2011) Songbirds possess the spontaneous ability to discriminate syntactic
rules. Nature Neuroscience 14:1067-1074.
Achiro JM, Bottjer SW (2013) Neural representation of a target auditory memory in a cortico-
Basal Ganglia pathway. The Journal of neuroscience : the official journal of the Society
for Neuroscience 33:14475-14488.
Adret P, Meliza CD, Margoliash D (2012) Song tutoring in presinging zebra finch juveniles
biases a small population of higher-order song-selective neurons toward the tutor song.
Journal of Neurophysiology 108:1977-1987.
Alexander GE, DeLong MR, Strick PL (1986) Parallel organization of functionally segregated
circuits linking basal ganglia and cortex. Annual Review of Neuroscience 9:357-381.
Aronov D, Andalman AS, Fee MS (2008) A specialized forebrain circuit for vocal babbling in
the juvenile songbird. Science 320:630-634.
Aronov D, Fee MS (2012) Natural changes in brain temperature underlie variations in song
tempo during a mating behavior. PLoS One 7:e47856.
Aronov D, Veit L, Goldberg JH, Fee MS (2011) Two distinct modes of forebrain circuit
dynamics underlie temporal patterning in the vocalizations of young songbirds. The
Journal of neuroscience : the official journal of the Society for Neuroscience 31:16353-
16368.
Ashby FG, Turner BO, Horvitz JC (2010) Cortical and basal ganglia contributions to habit
learning and automaticity. Trends in cognitive sciences 14:208-215.
Ashmore RC, Renk JA, Schmidt MF (2008) Bottom-up activation of the vocal motor forebrain
by the respiratory brainstem. Journal of Neuroscience 28:2613–2623.
Atallah HE, Lopez-Paniagua D, Rudy JW, O’Reilly RC (2007) Separate neural substrates for
skill learning and performance in the ventral and dorsal striatum. Nature Neuroscience
10:126-131.
Atoji Y, Karim MR (2012) Expression of the neocortical marker, RORβ, in the entopallium and
field L2 of adult chicken. Neuroscience letters 521:119-124.
Barto AG, Sutton RS, Anderson C (1983) Neuron-like adaptive elements that can solve difficult
learning control problems. IEEE Transactions on Systems, Man, and Cybernetics 834-
846.
Basham ME, Nordeen EJ, Nordeen KW (1996) Blockade of NMDA receptors in the anterior
forebrain impairs sensory acquisition in the zebra finch (Poephila guttata). Neurobiology
of Learning and Memory 66:295-304.
Blättler F, Hahnloser RHR (2011) An efficient coding hypothesis links sparsity and selectivity of
neural responses. PLoS ONE 6:e25506.
Boettiger CA, Doupe AJ (1998) Intrinsic and thalamic excitatory inputs onto songbird LMAN
neurons differ in their pharmacological and temporal properties. Journal of
Neurophysiology 79:2615-2628.
Böhner J (1983) Song learning in the zebra finch (taeniopygia guttata): Selectivity in the choice
of a tutor and accuracy of song copies. Animal Behaviour 31:231-237.
Böhner J (1990) Early acquisition of song in the zebra finch, Taeniopygia guttata. Animal
Behaviour 39:369-374.
Bolhuis JJ, Gahr M (2006) Neural mechanisms of birdsong memory. Nature reviews
Neuroscience 7:347-357.
120
Bottjer S, Glaessner S, Arnold A (1985) Ontogeny of brain nuclei controlling song learning and
behavior in zebra finches. The Journal of Neuroscience 5:1556-1562.
Bottjer SW (2004) Developmental regulation of basal ganglia circuitry during the sensitive
period for vocal learning in songbirds. Annals of the New York Academy of Sciences
1016:395–415.
Bottjer SW, Alderete TL, Chang D (2010) Conjunction of vocal production and perception
regulates expression of the immediate early gene ZENK in a novel cortical region of
songbirds. Journal of Neurophysiology 103:1833-1842.
Bottjer SW, Altenau B (2009) Parallel pathways for vocal learning in basal ganglia. Nature
Neuroscience 13:153-155.
Bottjer SW, Altenau B (2010) Parallel pathways for vocal learning in basal ganglia. Nature
Neuroscience 13:153-155.
Bottjer SW, Brady JD, Cribbs B (2000) Connections of a motor cortical region in zebra finches:
Relation to pathways for vocal learning. The Journal of Comparative Neurology 420:244
- 260.
Bottjer SW, Brady JD, Walsh JP (1998) Intrinsic and synaptic properties of neurons in the vocal-
control nucleus lMAN from in vitro slice preparations of juvenile and adult zebra finches.
Journal of neurobiology 37:642-658.
Bottjer SW, Miesner EA, Arnold AP (1984) Forebrain lesions disrupt development but not
maintenance of song in passerine birds. Science 224:901-903.
Bottjer SW, Miesner EA, Arnold AP (1986) Changes in neuronal number, density and size
account for increases in volume of song-control nuclei during song development in zebra
finches. Neuroscience letters 67:263-268.
Bottjer SW, To M (2012) Afferents from vocal motor and respiratory effectors are recruited
during vocal production in juvenile songbirds. The Journal of neuroscience : the official
journal of the Society for Neuroscience 32:10895-10906.
Brainard MS, Doupe AJ (2000) Interruption of a basal ganglia-forebrain circuit prevents
plasticity of learned vocalizations. Nature 404:762-766.
Brainard MS, Doupe AJ (2013) Translating birdsong: songbirds as a model for basic and applied
medical research. Annu Rev Neurosci 36:489-517.
Brown J, Bullock D, Grossberg S (1999) How the basal ganglia use parallel excitatory and
inhibitory learning pathways to selectively respond to unexpected rewarding cues. The
Journal of Neuroscience 19:10502–10511.
Cardin JA, Schmidt MF (2004) Auditory responses in multiple sensorimotor song system nuclei
are co-modulated by behavioral state. Journal of Neurophysiology 91:2148-2163.
Catchpole C, Slater PJB (1995) Bird song: Biological themes and variations. Cambridge:
Cambridge University Press.
Charlesworth JD, Warren TL, Brainard MS (2012) Covert skill learning in a cortical-basal
ganglia circuit. Nature 486:251-255.
Clayton NS (1987) Song tutor choice in zebra finches. Animal Behaviour 35:714-721.
Coleman, J. M, Mooney R (2004) Synaptic transformations underlying highly selective auditory
representations of learned birdsong. Journal of Neuroscience 24:7251-7265.
Cooper BG, Goller F (2006) Physiological insights into the social-context-dependent changes in
the rhythm of the song motor program. J Neurophysiol 95:3798-3809.
121
Cooper JA, Sagar HJ, Jordan N, Harvey NS, Sullivan EV (1991) Cognitive impairment in early,
untreated Parkinson's disease and its relationship to motor disability. Brain 114:2095-
2122.
Dave AS, Yu AC, Margoliash D (1998) Behavioral state modulation of auditory activity in a
vocal motor system. Science 282:2250-2254.
Derégnaucourt S, Mitra PP, Fehér O, Maul KK, Lints TJ, Tchernichovski O (2004) Song
development: in search of the error-signal. Annals of the New York Academy of Sciences
1016:364-376.
Doupe AJ (1997) Song- and order-selective neurons in the songbird anterior forebrain and their
emergence during vocal development. Journal of Neuroscience 17:1147-1167.
Doupe AJ, Kuhl PK (1999) Birdsong and human speech: common themes and mechanisms.
Annual Review of Neuroscience 22:567-631.
Doupe AJ, Solis MM (1997) Song- and order-selective neurons develop in the songbird anterior
forebrain during vocal learning. Journal of neurobiology 33:694-709.
Doupe AJ, Solis MM, Kimpo R, Boettiger CA (2004) Cellular, circuit, and synaptic mechanisms
in song learning. Annals of the New York Academy of Sciences 1016:495-523.
Doya K, Sejnowski TJ (2000) A computational model of avian song learning. In: The New
Cognitive Neurosciences, pp 469-488 Cambridge: MIT Press.
Eales LA (1985) Song learning in zebra finches: some effects of song model availability on what
is learnt and when. Animal Behaviour 33:1293-1300.
Eliades SJ, Wang X (2008) Neural substrates of vocalization feedback monitoring in primate
auditory cortex. Nature 453:1102-1106.
Fee MS, Goldberg JH (2011) A hypothesis for basal ganglia-dependent reinforcement learning in
the songbird. Neuroscience 198:152-170.
Fiete IR, Fee MS, Seung HS (2007) Model of birdsong learning based on gradient estimation by
dynamic perturbation of neural conductances. Journal of Neurophysiology 2038-2057.
Foster E, F., Bottjer SW (2001) Lesions of a telencephalic nucleus in male zebra finches:
influences on vocal behavior in juveniles and adults. Journal of neurobiology 46:142-165.
Foster EF, Bottjer SW (1998) Axonal connections of the high vocal center and surrounding
cortical regions in juvenile and adult male zebra finches. J Comp Neurol 397:118-138.
Fritz JB, David SV, Radtke-Schuller S, Yin P, Shamma S (2010) Adaptive, behaviorally gated,
persistent encoding of task-relevant auditory information in ferret frontal cortex. Nature
Neuroscience 13:1011-1019.
Fujimoto H, Hasegawa T, Watanabe D (2011) Neural Coding of Syntactic Structure in Learned
Vocalizations in the Songbird. The Journal of Neuroscience 31:10023-10033.
Fukushima Y, Aoki K (2002) Neural function of the mesencephalic dorsomedial nucleus (DM)
on distance call production in Bengalese finches. Zoological science 19:393-402.
Gale SD, Person AL, Perkel DJ (2008) A novel basal ganglia pathway forms a loop linking a
vocal learning circuit with its dopaminergic input. Journal of Comparative Neurology
508:824 - 839.
Giraud A-L, Neumann K, Bachoud-Levi A-C, von Gudenberg AW, Euler HA, Lanfermann H,
Preibisch C (2008) Severity of dysfluency correlates with basal ganglia activity in
persistent developmental stuttering. Brain and Language 104:190–199.
Glaze CM, Troyer TW (2013) Development of temporal structure in zebra finch song. Journal of
Neurophysiology 109:1025-1035.
122
Gobes SMH, Zandbergen MA, Bolhuis JJ (2010) Memory in the making: localized brain
activation related to song learning in young songbirds. Proceedings of the Royal Society
277:3343-3351.
Goldberg JH, Fee MS (2010) Singing-related neural activity distinguishes four classes of
putative striatal neurons in the songbird basal ganglia. J Neurophysiol 103:2002-2014.
Goldberg JH, Fee MS (2012) A cortical motor nucleus drives the basal ganglia-recipient
thalamus in singing birds. Nat Neurosci 15:620-627.
Graybiel AM (2008) Habits, rituals and the evaluative brain. Annual Review of Neuroscience
31:359-387.
Gremel CM, Costa RM (2013) Orbitofrontal and striatal circuits dynamically encode the shift
between goal-directed and habitual actions. Nature communications 4:2264.
Haesler S, Rochefort C, Georgi B, Licznerski P, Osten P, Scharff C (2007) Incomplete and
inaccurate vocal imitation after knockdown of FoxP2 in songbird basal ganglia nucleus
Area X. PLoS biology 5:e321.
Haesler S, Wada K, Nshdejan A, Morrisey EE, Lints T, Jarvis ED, Scharff C (2004) FoxP2
Expression in Avian Vocal Learners and Non-Learners. The Journal of Neuroscience
24:3164-3175.
Hahnloser R, Kozhevnikov A, Fee M (2002) An ultra-sparse code underlies the generation of
neural sequences in a songbird. Nature 65–70.
Hahnloser RHR, Kotowicz A (2010) Auditory representations and memory in birdsong learning.
Current Opinion in Neurobiology 20:332-339.
Hampton CM, Sakata JT, Brainard MS (2009) An avian basal ganglia-forebrain circuit
contributes differentially to syllable versus sequence variability of adult Bengalese finch
song. J Neurophysiol 101:3235-3245.
Harris KD, Henze DA, Csicsvari J, Hirase H, Buzsáki G (2000) Accuracy of tetrode spike
separation as determined by simultaneous intracellular and extracellular measurements.
Journal of Neurophysiology 84:401–414.
Hensch T, K., Fagiolini M, Mataga N, Stryker MP, Baekkeskov S, Kash SF (1998) Local GABA
circuit control of experience-dependent plasticity in developing visual cortex. Science
282:1504-1508.
Hessler NA, Doupe AJ (1999) Singing-related neural activity in a dorsal forebrain–basal ganglia
circuit of adult zebra finches. Journal of Neuroscience 19:10461–10481.
Hikosaka O, Nakahara H, Rand MK, Sakai K, Lu X, Nakamura K, Miyachi S, Doya K (1999)
Parallel neural networks for learning sequential procedures. Trends in Neurosciences
22:464-471.
Histed MH, Pasupathy A, Miller EK (2009) Learning substrates in the primate prefrontal cortex
and striatum: sustained activity related to successful actions. Neuron 63:244-253.
Immelmann K (1969) Song development in the zebra finch and other estrildid finches. In: Bird
vocalisations (Hinde, R. A., ed), pp 61-74 Cambridge: Cambridge University Press.
Iyengar S, Bottjer SW (2002a) Development of individual axon arbors in a thalamocortical
circuit necessary for song learning in zebra finches. Journal of Neuroscience 22:901-911.
Iyengar S, Bottjer SW (2002b) The Role of Auditory Experience in the Formation of Neural
Circuits Underlying Vocal Learning in Zebra Finches. The Journal of Neuroscience
22:946-958.
123
Iyengar S, Viswanathan SS, Bottjer SW (1999) Development of topography within song control
circuitry of zebra finches during the sensitive period for song learning. Journal of
Neuroscience 19:6037-6057.
Joel D, Weiner I (1997) The connections of the primate subthalamic nucleus: indirect pathways
and the open interconnected scheme of basal ganglia-thalamocortical circuitry. Brain
Research Reviews 23:62–78.
Johnson F, Bottjer SW (1992) Growth and regression of thalamic efferents in the song-control
system of male zebra finches. Journal of Comparative Neurology 326:442-450.
Johnson F, Sablan MM, Bottjer SW (1995) Topographic organization of a forebrain pathway
involved with vocal learning in zebra finches. Journal of Comparative Neurology
358:260-278.
Johnson F, Soderstrom K, Whitney O (2002) Quantifying song bout production during zebra
finch sensory-motor learning suggests a sensitive period for vocal practice. Behavioural
brain research 131:57-65.
Jurgens U (2002) Neural pathways underlying vocal control. Neuroscience and biobehavioral
reviews 26:235-258.
Kao M, Brainard M (2006) Lesions of an avian basal ganglia circuit prevent context-dependent
changes to song variability. Journal of Neurophysiology 96:1441-1455.
Kao MH, Doupe AJ, Brainard MS (2005) Contributions of an avian basal ganglia–forebrain
circuit to real-time modulation of song. Nature 433:638-643.
Kao MH, Wright BD, Doupe AJ (2008) Neurons in a forebrain nucleus required for vocal
plasticity rapidly switch between precise firing and variable bursting depending on social
context. Journal of Neuroscience 28:13232-13247.
Karten HJ (1997) Evolutionary developmental biology meets the brain: the origins of
mammalian cortex. Proceedings of the National Academy of Sciences 94:2800–2804.
Kim H, Lee D, Jung MW (2013) Signals for previous goal choice persist in the dorsomedial, but
not dorsolateral striatum of rats. Journal of Neuroscience 33:52-63.
Kojima S, Doupe AJ (2007) Song selectivity in the pallial-basal ganglia song circuit of zebra
finches raised without tutor song exposure. Journal of Neurophysiology 98:2099-2109.
Kojima S, Doupe AJ (2008) Neural encoding of auditory temporal context in a songbird basal
ganglia nucleus, and its independence of birds' song experience. European Journal of
Neuroscience 27:1231-1244.
Kojima S, Doupe AJ (2011) Social performance reveals unexpected vocal competency in young
songbirds. Proceedings of the National Academy of Sciences 108:1687-1692.
Konishi M (1965) The role of auditory feedback in the control of vocalization in the white
crowned sparrow. Zeitschrift fur Tierpsychologie 22:770-783.
Konishi M (2004) The Role of Auditory Feedback in Birdsong. Annals of the New York
Academy of Sciences 1016:463-475.
Konishi M, Akutagawa E (1985) Neuronal growth, atrophy and death in a sexually dimorphic
song nucleus in the zebra finch brain. Nature 315:145-147.
Kuhl PK (2004) Early language acquisition: cracking the speech code. Nature reviews
Neuroscience 5:831-843.
Lieberman P (2001) Human language and our reptilian brain: the subcortical basis of speech,
syntax, and thought. Perspectives in Biology and Medicine 44:32-51.
Livingston F, Mooney R (1997) Development of intrinsic and synaptic properties in a forebrain
nucleus essential to avian song learning. Journal of Neuroscience 17:8997-9009.
124
Lombardino AJ, Nottebohm F (2000) Age at deafening affects the stability of learned song in
adult male zebra finches. The Journal of neuroscience : the official journal of the Society
for Neuroscience 20:5054-5064.
London SE, Clayton DF (2008) Functional identification of sensory mechanisms required for
developmental song learning. Nature Neuroscience 11:579-586.
Long MA, Fee MS (2008) Using temperature to analyze temporal dynamics in the songbird
motor pathway. Nature 456:189-194.
Mann NI, Slater PJB (1995) Song tutor choice by zebra finches in aviaries. Animal Behaviour
49:811-820.
Margoliash DF, Eric S. (1992) Temporal and harmonic combination-sensitive neurons in the
zebra finch's HVC. Journal of Neuroscience 12:4309-4326.
Marler P (1956) The voice of the Chaffinch and its function as a language. Ibis: International
Journal of Avian Science 98:231-261.
Marler P (1970) A comparative approach to vocal learning: Song development in white-crowned
sparrows. Journal of Comparative and Physiological Psychology 71:1-25.
Meliza CD, Margoliash D (2012) Emergence of selectivity and tolerance in the avian auditory
cortex. Journal of Neuroscience 32:15158-15168.
Miall RC, Wolpert DM (1996) Forward models for physiological motor control. Neural
Networks 9:1265-1279.
Miller-Sims VC, Bottjer SW (2012) Auditory experience refines cortico-basal ganglia inputs to
motor cortex via re-mapping of single axons during vocal learning in zebra finches.
Journal of Neurophysiology 107:1142-1156.
Miller JE, Hilliard AT, White SA (2010) Song practice promotes acute vocal variability at a key
stage of sensorimotor learning. PLoS ONE 5:e8592.
Miyachi S, Hikosaka O, Lu X (2002) Differential activation of monkey striatal neurons in the
early and late stages of procedural learning. Experimental Brain Research 146:122-126.
Mooney R, Rao M (1994) Waiting periods versus early innervation: the development of axonal
connections in the zebra finch song system. The Journal of Neuroscience 14:6532-6543.
Mushiake H, Saito N, Sakamoto K, Itoyama Y, Tanji J (2006) Activity in the lateral prefrontal
cortex reflects multiple steps of future events in action plans. Neuron 50:631–641.
Nakahara H, Doya K, Hikosaka O (2001) Parallel cortico-basal ganglia mechanisms for
acquisition and execution of visuomotor sequences- a computational approach. Journal
of Cognitive Neuroscience 13:626-647.
Nick TA, Konishi M (2005a) Neural auditory selectivity develops in parallel with song. Journal
of neurobiology 62:469-481.
Nick TA, Konishi M (2005b) Neural song preference during vocal learning in the zebra finch
depends on age and state. Journal of neurobiology 62:231-242.
Niziolek CA, Nagarajan SS, Houde JF (2013) What Does Motor Efference Copy Represent?
Evidence from Speech Production. The Journal of Neuroscience 33:16110-16116.
Nordeen KW, Nordeen EJ (1988) Projection neurons within a vocal motor pathway are born
during song learning in zebra finches. Nature 334:149-151.
Nordeen KW, Nordeen EJ (1992) Auditory feedback is necessary for the maintenance of
stereotyped song in adult zebra finches. Behavior and Neural Biology 57:58-66.
Nordeen KW, Nordeen EJ (2010) Deafening-induced vocal deterioration in adult songbirds is
reversed by disrupting a basal ganglia-forebrain circuit. The Journal of neuroscience : the
official journal of the Society for Neuroscience 30:7392-7400.
125
Nottebohm F, Paton JA, Kelley DB (1982) Connections of vocal control nuclei in the canary
telencephalon. The Journal of Comparative Neurology 207:344-357.
Nottebohm F, Stokes TM, Leonard CM (1976) Central control of song in the canary, Serinus
canarius. J Comp Neurol 165:457-486.
O'Doherty J, Dayan P, Schultz J, Deichmann R, Friston K, Dolan RJ (2004) Dissociable roles of
ventral and dorsal striatum in instrumental conditioning. Science 304:452-454
Ölveczky BP, Andalman AS, Fee MS (2005) Vocal experimentation in the juvenile songbird
requires a basal ganglia circuit. PLoS biology 3:e153.
Ölveczky BP, Otchy TM, Goldberg JH, Aronov D, Fee MS (2011) Changes in the neural control
of a complex motor sequence during learning. Journal of Neurophysiology 106:386-397.
Parent A, Hazrati L-N (1995) Functional anatomy of the basal ganglia: I. The cortico-basal
ganglia-thalamo-cortical loop. Brain Research Reviews 20:91-127.
Pedreira C, Martinez J, Ison MJ, Quiroga RQ (2012) How many neurons can we see with current
spike sorting algorithms? Journal of Neuroscience Methods 211:58-65.
Person AL, Gale SD, Farries MA, Perkel DJ (2008) Organization of the songbird basal ganglia,
including Area X. The Journal of Comparative Neurology 508:840-866.
Phan ML, Pytte CL, Vicario DS (2006) Early auditory experience generates long-lasting
memories that may subserve vocal learning in songbirds. Proceedings of the National
Academy of Sciences 103:1088-1093.
Pinaud R, Mello C, V. (2007) GABA immunoreactivity in auditory and song control brain areas
of zebra finches. Journal of Chemical Neuroanatomy 34:1-21.
Pinaud R, Saldanha CJ, Wynne RD, Lovell PV, Mello CV (2007) The excitatory thalamo-
"cortical" projection within the song control system of zebra finches is formed by
calbindin-expressing neurons. The Journal of Comparative Neurology 504:601-618.
Pinaud R, Terleph TA, Tremere LA, Phan ML, Dagostin AA, Leão RM, Mello CV, Vicario DS
(2008) Inhibitory network interactions shape the auditory processing of natural
communication signals in the songbird auditory forebrain Journal of Neurophysiology
100:441-455.
Prather JF, Peters S, Nowicki S, Mooney R (2010) Persistent representation of juvenile
experience in the adult songbird brain. Journal of Neuroscience 30:10586-10598.
Ravbar PL, Dina; Parra, Lucas; Tchernichovski, Ofer (2012) Vocal exploration is locally
regulated during song learning. Journal of Neuroscience 32:3422-3432.
Redgrave P, Rodriguez M, Smith Y, Rodriguez-Oroz MC, Lehericy S, Bergman H, Agid Y,
DeLong MR, Obeso JA (2010) Goal-directed and habitual control in the basal ganglia:
implications for Parkinson's disease. Nature reviews Neuroscience 11:760-772.
Redgrave P, Vautrelle N, Reynolds JNJ (2011) Functional properties if the basal ganglia's re-
entrant loop architecture: selection and reinforcement. Neuroscience 198:138-151.
Reiner A, Perkel DJ, Bruce LL, Butler AB, Csillag A, Kuenzel W, Medina L, Paxinos G,
Shimizu T, Striedter G, Wild M, Ball GF, Durand S, Gütürkün O, Lee DW, Mello CV,
Powers A, White SA, Hough G, Kubikova L, Smulders TV, Wada K, Dugas-Ford J,
Husband S, Yamamoto K, Yu J, Siang C, Jarvis ED (2004) Revised nomenclature for
avian telencephalon and some related brainstem nuclei. Journal of Comparative
Neurology 473:377-414.
Reiner A, Yamamoto K, Karten HJ (2005) Organization and evolution of the avian forebrain.
Discoveries in molecular, cellular, and evolutionary biology 287:1080–1102.
126
Roper A, Zann R (2006) The onset of song learning and song tutor selection in fledgling zebra
finches. Ethology 112:458 - 470.
Rosen MJ, Mooney R (2000) Intrinsic and extrinsic contributions to auditory selectivity in a
song nucleus critical for vocal plasticity. Journal of Neuroscience 20:5437-5448.
Rosen MJ, Mooney R (2003) Inhibitory and excitatory mechanisms underlying auditory
responses to learned vocalizations in the songbird nucleus HVC. Neuron 39:177-194.
Saito N, Mushiake H, Sakamoto K, Itoyama Y, Tanji J (2005) Representation of immediate and
final behavioral goals in the monkey prefrontal cortex during an instructed delay period.
Cerebral Cortex 15:1535-1546.
Sakata JT, Brainard MS (2006) Real-time contributions of auditory feedback to avian. Journal of
Neuroscience 26:9619 –9628.
Sakata JT, Brainard MS (2009) Social context rapidly modulates the influence of auditory
feedback on avian vocal motor control. J Neurophysiol 102:2485-2497.
Samejima K, Doya K (2007) Multiple representations of belief states and action values in
corticobasal ganglia loops. Annals of the New York Academy of Sciences 1104:213-228.
Scharff C, Nottebohm F (1991) A comparative study of the behavioral deficits following lesions
of various parts of the zebra finch song system: implications for vocal learning. Journal
of Neuroscience 11:2898-2913.
Schmidt MF, Ashmore RC, Vu ET (2004) Bilateral control and interhemispheric coordination in
the avian song motor system. Annals of the New York Academy of Sciences 1016:171-
186.
Shank SS, Margoliash D (2009) Sleep and sensorimotor integration during early vocal learning
in a songbird. Nature 458:73-77.
Simpson HB, Vicario DS (1990) Brain pathways for learned and unlearned vocalizations differ
in zebra finches. Journal of Neuroscience 10:1541-1556.
Sober SJ, Brainard MS (2009) Adult birdsong is actively maintained by error correction. Nat
Neurosci 12:927-931.
Sohrabji F, Nordeen EJN, K.W. (1990) Selective impairment of song learning following lesions
of a forebrain nucleus in the juvenile zebra finch. Behavioral and Neural Biology 53:51-
63.
Solis MM, Doupe AJ (1997) Anterior forebrain neurons develop selectivity by an intermediate
stage of birdsong learning. Journal of Neuroscience 17:6447-6462.
Solis MM, Doupe AJ (1999) Contributions of tutor and bird’s own song experience to neural
selectivity in the songbird anterior forebrain. The Journal of Neuroscience 19:4559–4584.
Solis MM, Doupe AJ (2000) Compromised neural selectivity for song in birds with impaired
sensorimotor learning. Neuron 25:109–121.
Sossinka R, Böhner J (1980) Song Types in the Zebra Finch Poephila guttata castanotis.
Zeitschrift für Tierpsychologie 53:123-132.
Sutton RSB, Andrew G. (1998) Reinforcement learning: an introduction. Cambridge: MIT Press.
Suzuki WA, Miller EK, Desimone R (1997) Object and place memory in the macaque entorhinal
cortex. Journal of Neurophysiology 78:1062-1081.
Tanaka S, C., Doya K, Okada G, Ueda K, Okamoto Y, Yamawaki S (2004) Prediction of
immediate and future rewards differentially recruits cortico-basal ganglia loops. Nature
Neuroscience 7:887-893.
127
Tchernichovski O, Lints TJ, Deregnaucourt S, Cimenser A, Mitra PP (2004) Studying the song
development process: rationale and methods. Annals of the New York Academy of
Sciences 1016:348-363.
Tchernichovski O, Mitra PP, Lints T, Nottebohm F (2001) Dynamics of the vocal imitation
process: how a zebra finch learns its song. Science 291:2564 - 2569.
Tchernichovski O, Nottebohm F, Ho CE, Pesaran B, Mitra PP (2000) A procedure for an
automated measurement of song similarity. Animal Behaviour 59:1167-1176.
Terpstra NJ, Bolhuis JJ, den Boer-Visser AM (2004) An analysis of the neural representation of
birdsong memory. The Journal of neuroscience : the official journal of the Society for
Neuroscience 24:4971-4977.
Thompson CK, Schwabe F, Schoof A, Mendoza E, Gampe J, Rochefort C, Scharff C (2013)
Young and intense: FoxP2 immunoreactivity in Area X varies with age, song stereotypy,
and singing in male zebra finches. Frontiers in neural circuits 7:24.
Thorn C, A., Atallah H, Howe M, Graybiel AM (2010) Differential dynamics of activity changes
in dorsolateral and dorsomedial striatal loops during learning. Neuron 66:781-795.
Thorpe WH (1958) The learning of song patterns by birds, with especial reference to the song of
the Chaffinch Fringilla coelebs. Ibis: International Journal of Avian Science 100:535-570.
Troyer TW, Doupe AJ (2000) An associational model of birdsong sensorimotor learning I.
Efference copy and the learning of song syllables. Journal of Neurophysiology 1204-
1223.
Tumer EC, Brainard MS (2007) Performance variability enables adaptive plasticity of
'crystallized' adult birdsong. Nature 450:1240-1244.
Vicario DS (1991) Organization of the zebra finch song control system: II. Functional
organization of outputs from nucleus Robustus archistriatalis. J Comp Neurol 309:486-
494.
Vinje WE, Gallant JL (2000) Sparse Coding and Decorrelation in Primary Visual Cortex During
Natural Vision. Science 287:1273-1276.
Vintsyuk TK (1968) Speech discrimination by dynamic programming. Cybern Syst Anal 4:52-
57.
Vu ET, Schmidt MF, Mazurek ME (1998) Interhemispheric coordination of premotor neural
activity during singing in adult zebra finches. Journal of Neuroscience 18:9088-9098.
Wang Y, Brzozowska-Prechtl A, Karten H (2010) Laminar and columnar auditory cortex in
avian brain. Proceedings of the National Academy of Sciences 107:12676-12681.
Warren TL, Tumer EC, Charlesworth JD, Brainard MS (2011) Mechanisms and time course of
vocal learning and consolidation in the adult songbird. Journal of Neurophysiology
106:1806-1821.
White SA, Livingston FS, Mooney R (1999) Androgens modulate NMDA receptor-mediated
EPSCs in the zebra finch song system. J Neurophysiol 82:2221-2234.
Wild JM (1993) Descending projections of the songbird nucleus robustus archistriatalis. J Comp
Neurol 338:225-241.
Wild JM (2004) Functional neuroanatomy of the sensorimotor control of singing. Annals of the
New York Academy of Sciences 1016:438-462.
Wildgruber D, Ackermann H, Grodd W (2001) Differential contributions of motor cortex, basal
ganglia, and cerebellum to speech motor control: effects of syllable repetition rate
evaluated by fMRI. NeuroImage 13:101-109.
128
Williams H (1990) Models for song learning in the zebra finch: fathers or others? Animal
Behaviour 39:745-757.
Yazaki-Sugiyama Y, Mooney R (2004) Sequential learning from multiple tutors and serial
retuning of auditory neurons in a brain area important to birdsong learning. Journal of
Neurophysiology 92:2771-2788.
Yin HH, Knowlton BJ (2006) The role of the basal ganglia in habit formation. Nature Reviews
Neuroscience 7:464-476.
Yin HH, Mulcare SP, Hilário MRF, Clouse E, Holloway T, Davis MI, Hansson AC, Lovinger
DM, Costa RM (2009) Dynamic reorganization of striatal circuits during the acquisition
and consolidation of a skill. Nature Neuroscience 12:333-341.
Yin HH, Ostlund SB, Balleine BW (2008) Reward-guided learning beyond dopamine in the
nucleus accumbens: the integrative functions of cortico-basal ganglia networks. European
Journal of Neuroscience 28:1437-1448.
Yu AC, Margoliash D (1996) Temporal hierarchical control of singing in birds. Science
273:1871-1875.
Zann RA (1996) The Zebra Finch: A synthesis of field and laboratory studies. Oxford: Oxford
University Press.

129

Abstract (if available)

Abstract For over 70 years the study of vocal learning in songbirds, pioneered by William Thorpe and Peter Marler, has been transformative in the research of such areas as adult neurogenesis, the role of hormones in development and behavior, and genetic factors underlying complex behavior. Moreover, songbirds provide an ideal model for studying mechanisms of sensorimotor integration and imitative motor learning. In a process analogous to the acquisition of human speech, songbirds learn a specific vocal pattern by listening to adult male ""tutor"" song during a sensitive period early in life. Like humans, songbirds learn to imitate those communication sounds by vocalizing and using sensory feedback to compare incipient babbling to the memory of tutor sounds. This dissertation addresses one major unresolved question with regard to mechanisms of vocal learning: what neural circuits carry out comparisons of vocal feedback to the tutor sounds, and how is this comparison achieved?

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Motor cortical representations of sensorimotor information during skill learning

PDF

Neuroplasticity of the basal ganglia in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine mouse model of Parkinson's disease

PDF

Brain and behavior correlates of intrinsic motivation and skill learning

PDF

The behavioral and neural bases of tactile object localization

PDF

Computational principles in human motor adaptation: sources, memories, and variability

PDF

Synaptic mechanism underlying development and function of neural circuits in rat primary auditory cortex

PDF

Functional magnetic resonance imaging characterization of peripheral form vision

PDF

Synaptic integration in dendrites: theories and applications

PDF

Experimental and computational explorations of different forms of plasticity in motor learning and stroke recovery

PDF

Neural basis of number sense in zebrafish

PDF

The neural correlates of face recognition

PDF

Characterizing response and plasticity in sensory cortices of the Fmr1⁻⁄⁻ mouse

PDF

The role of the cofilin/Limk1 signaling pathway in axon growth during development and regeneration

PDF

Mechanisms of stress effects on learning and decision making in younger and older adults

PDF

Mechanisms of motor learning in immersive virtual reality and their influences on retention and context transfer

PDF

The role of Rho-GEF signaling in synapse function and autism-related disorders

PDF

Mapping multi-scale connectivity of the mouse posterior parietal cortex

PDF

Functional models of fMRI BOLD signal in the visual cortex

PDF

From sensory processing to behavior control: functions of the inferior colliculus

PDF

Synaptic circuits for information processing along the central auditory pathway

Asset Metadata

Creator Achiro, Jennifer McGrady (author)

Core Title Neural mechanisms of sensorimotor learning in cortico-basal ganglia pathways

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 05/08/2015

Defense Date 10/21/2013

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cortico-basal ganglia,motor learning,OAI-PMH Harvest,sensorimotor integration

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Bottjer, Sarah W. (committee chair), Dickman, Dion (committee member), Goldstein, Louis (committee member), Hirsch, Judith (committee member), Mel, Bartlett W. (committee member)

Creator Email jenniferachiro@gmail.com,mcgrady@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-345172

Unique identifier UC11296524

Identifier etd-AchiroJenn-2145.pdf (filename),usctheses-c3-345172 (legacy record id)

Legacy Identifier etd-AchiroJenn-2145.pdf

Dmrecord 345172

Document Type Dissertation

Format application/pdf (imt)

Rights Achiro, Jennifer McGrady

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA