Translating birdsong: songbirds as a model for basic and applied medical research.

Songbirds, long of interest to basic neuroscience, have great potential as a model system for translational neuroscience. Songbirds learn their complex vocal behavior in a manner that exemplifies general processes of perceptual and motor skill learning and, more specifically, resembles human speech learning. Song is subserved by circuitry that is specialized for vocal learning and production but that has strong similarities to mammalian brain pathways. The combination of highly quantifiable behavior and discrete neural substrates facilitates understanding links between brain and behavior, both in normal states and in disease. Here we highlight (a) behavioral and mechanistic parallels between birdsong and aspects of speech and social communication, including insights into mirror neurons, the function of auditory feedback, and genes underlying social communication disorders, and (b) contributions of songbirds to understanding cortical-basal ganglia circuit function and dysfunction, including the possibility of harnessing adult neurogenesis for brain repair.

The Avian Basal Ganglia Are a Source of Rapid Behavioral Variation That Enables Vocal Motor Exploration.

The basal ganglia (BG) participate in aspects of reinforcement learning that require evaluation and selection of motor programs associated with improved performance. However, whether the BG additionally contribute to behavioral variation ("motor exploration") that forms the substrate for such learning remains unclear. In songbirds, a tractable system for studying BG-dependent skill learning, a role for the BG in generating exploratory variability, has been challenged by the finding that lesions of Area X, the song-specific component of the BG, have no lasting effects on several forms of vocal variability that have been studied. Here we demonstrate that lesions of Area X in adult male zebra finches (Taeniopygia gutatta) permanently eliminate rapid within-syllable variation in fundamental frequency (FF), which can act as motor exploration to enable reinforcement-driven song learning. In addition, we found that this within-syllable variation is elevated in juveniles and in adults singing alone, conditions that have been linked to enhanced song plasticity and elevated neural variability in Area X. Consistent with a model that variability is relayed from Area X, via its cortical target, the lateral magnocellular nucleus of the anterior nidopallium (LMAN), to influence song motor circuitry, we found that lesions of LMAN also eliminate within-syllable variability. Moreover, we found that electrical perturbation of LMAN can drive fluctuations in FF that mimic naturally occurring within-syllable variability. Together, these results demonstrate that the BG are a central source of rapid behavioral variation that can serve as motor exploration for vocal learning.SIGNIFICANCE STATEMENT Many complex motor skills, such as speech, are not innately programmed but are learned gradually through trial and error. Learning involves generating exploratory variability in action ("motor exploration") and evaluating subsequent performance to acquire motor programs that lead to improved performance. Although it is well established that the basal ganglia (BG) process signals relating to action evaluation and selection, whether and how the BG promote exploratory motor variability remain unclear. We investigated this question in songbirds, which learn to produce complex vocalizations through trial and error. In contrast with previous studies that did not find effects of BG lesions on vocal motor variability, we demonstrate that the BG are an essential source of rapid behavioral variation linked to vocal learning.

Zebra finches are sensitive to combinations of temporally distributed features in a model of word recognition.

Discrimination between spoken words composed of overlapping elements, such as "captain" and "captive," relies on sensitivity to unique combinations of prefix and suffix elements that span a "uniqueness point" where the word candidates diverge. To model such combinatorial processing, adult female zebra finches were trained to discriminate between target and distractor syllable sequences that shared overlapping "contextual" prefixes and differed only in their "informative" suffixes. The transition from contextual to informative syllables thus created a uniqueness point analogous to that present between overlapping word candidates, where targets and distractors diverged. It was found that target recognition depended not only on informative syllables, but also on contextual syllables that were shared with distractors. Moreover, the influence of each syllable depended on proximity to the uniqueness point. Birds were then trained birds with targets and distractors that shared both prefix and suffix sequences and could only be discriminated by recognizing unique combinations of those sequences. Birds learned to robustly discriminate target and distractor combinations and maintained significant discrimination when the local transitions from prefix to suffix were disrupted. These findings indicate that birds, like humans, combine information across temporally distributed features, spanning contextual and informative elements, in recognizing and discriminating word-like stimuli.

Task-related "cortical" bursting depends critically on basal ganglia input and is linked to vocal plasticity.

Basal ganglia-thalamocortical circuits are critical for motor control and motor learning. Classically, basal ganglia nuclei are thought to regulate motor behavior by increasing or decreasing cortical firing rates, and basal ganglia diseases are assumed to reflect abnormal overall activity levels. More recent studies suggest instead that motor disorders derive from abnormal firing patterns, and have led to the hypothesis that surgical treatments, such as pallidotomy, act primarily by eliminating pathological firing patterns. Surprisingly little is known, however, about how the basal ganglia normally influence task-related cortical activity to regulate motor behavior, and how lesions of the basal ganglia influence cortical firing properties. Here, we investigated these questions in a songbird circuit that has striking homologies to mammalian basal ganglia-thalamocortical circuits but is specialized for singing. The "cortical" outflow nucleus of this circuit is required for song plasticity and normally exhibits increased firing during singing and song-locked burst firing. We found that lesions of the striato-pallidal nucleus in this circuit prevented hearing-dependent song changes. These basal ganglia lesions also stripped the cortical outflow neurons of their patterned burst firing during singing, without changing their spontaneous or singing-related firing rates. Taken together, these results suggest that the basal ganglia are essential not for normal cortical firing rates but for driving task-specific cortical firing patterns, including bursts. Moreover, such patterned bursting appears critical for motor plasticity. Our findings thus provide support for therapies that aim to treat basal ganglia movement disorders by normalizing firing patterns.

Dopamine physiology in the basal ganglia of male zebra finches during social stimulation.

Accumulating evidence suggests that dopamine (DA) is involved in altering neural activity and gene expression in a zebra finch cortical-basal ganglia circuit specialized for singing, upon the shift between solitary singing and singing as a part of courtship. Our objective here was to sample changes in the extracellular concentrations of DA in Area X of adult and juvenile birds, to test the hypothesis that DA levels would change similarly during presentation of a socially salient stimulus in both age groups. We used microdialysis to sample the extracellular milieu of Area X in awake, behaving adult and juvenile male zebra finches, and analysed the dialysate using high-performance liquid chromatography coupled with electrochemical detection. The extracellular levels of DA in Area X increased significantly during both female presentation to adult males and tutor presentation to juvenile males. DA levels were not correlated with the time spent singing. We also reverse-dialysed Area X with pharmacologic agents that act either on DA systems directly or on norepinephrine, and found that all of these agents significantly increased DA levels (3- to 10-fold) in Area X. These findings suggest that changes in extracellular DA levels can be stimulated similarly by very different social contexts (courtship and interaction with tutor), and influenced potently by dopaminergic and noradrenergic drugs. These results raise the possibility that the arousal level or attentional state of the subject (rather than singing behavior) is the common feature eliciting changes in extracellular DA concentration.

Social performance reveals unexpected vocal competency in young songbirds.

Vocal ontogeny in songbirds provides a good model for understanding how complex motor behavior, including speech, is learned. For birdsong, as for other motor learning, it has generally been assumed that a subject's motor output at any point during learning represents what the subject has learned to produce by that time. Here, we show, however, that juvenile zebra finches partway through song learning, singing immature song, are capable of producing song with much more mature properties, depending on the behavioral context. In these birds, we were able to elicit courtship (female-directed) song, which young birds normally sing infrequently, and to compare it with the alone or "undirected" song (Undir) predominantly produced during learning as well as with the same bird's subsequent adult song. We found that the juvenile courtship song was much less variable than the immature Undir and as stereotyped as the adult song produced after a further month of practice. More strikingly, the juvenile courtship song was also acoustically much more similar than Undir to the adult song. This finding demonstrates that the Undir that juvenile birds usually produce underestimates the extent of learning and that song structure is learned faster than previously thought. Moreover, the rapid improvement in song quality in response to external social cues supports the idea that courtship singing is a state of motor "performance," in which the bird selects the best variants of the song learned during singing alone, and suggests that such performance states can reveal unappreciated progression of learning.

Two-dimensional adaptation in the auditory forebrain.

Sensory neurons exhibit two universal properties: sensitivity to multiple stimulus dimensions, and adaptation to stimulus statistics. How adaptation affects encoding along primary dimensions is well characterized for most sensory pathways, but if and how it affects secondary dimensions is less clear. We studied these effects for neurons in the avian equivalent of primary auditory cortex, responding to temporally modulated sounds. We showed that the firing rate of single neurons in field L was affected by at least two components of the time-varying sound log-amplitude. When overall sound amplitude was low, neural responses were based on nonlinear combinations of the mean log-amplitude and its rate of change (first time differential). At high mean sound amplitude, the two relevant stimulus features became the first and second time derivatives of the sound log-amplitude. Thus a strikingly systematic relationship between dimensions was conserved across changes in stimulus intensity, whereby one of the relevant dimensions approximated the time differential of the other dimension. In contrast to stimulus mean, increases in stimulus variance did not change relevant dimensions, but selectively increased the contribution of the second dimension to neural firing, illustrating a new adaptive behavior enabled by multidimensional encoding. Finally, we demonstrated theoretically that inclusion of time differentials as additional stimulus features, as seen so prominently in the single-neuron responses studied here, is a useful strategy for encoding naturalistic stimuli, because it can lower the necessary sampling rate while maintaining the robustness of stimulus reconstruction to correlated noise.

Social context-induced song variation affects female behavior and gene expression.

Social cues modulate the performance of communicative behaviors in a range of species, including humans, and such changes can make the communication signal more salient. In songbirds, males use song to attract females, and song organization can differ depending on the audience to which a male sings. For example, male zebra finches (Taeniopygia guttata) change their songs in subtle ways when singing to a female (directed song) compared with when they sing in isolation (undirected song), and some of these changes depend on altered neural activity from a specialized forebrain-basal ganglia circuit, the anterior forebrain pathway (AFP). In particular, variable activity in the AFP during undirected song is thought to actively enable syllable variability, whereas the lower and less-variable AFP firing during directed singing is associated with more stereotyped song. Consequently, directed song has been suggested to reflect a "performance" state, and undirected song a form of vocal motor "exploration." However, this hypothesis predicts that directed-undirected song differences, despite their subtlety, should matter to female zebra finches, which is a question that has not been investigated. We tested female preferences for this natural variation in song in a behavioral approach assay, and we found that both mated and socially naive females could discriminate between directed and undirected song-and strongly preferred directed song. These preferences, which appeared to reflect attention especially to aspects of song variability controlled by the AFP, were enhanced by experience, as they were strongest for mated females responding to their mate's directed songs. We then measured neural activity using expression of the immediate early gene product ZENK, and found that social context and song familiarity differentially modulated the number of ZENK-expressing cells in telencephalic auditory areas. Specifically, the number of ZENK-expressing cells in the caudomedial mesopallium (CMM) was most affected by whether a song was directed or undirected, whereas the caudomedial nidopallium (NCM) was most affected by whether a song was familiar or unfamiliar. Together these data demonstrate that females detect and prefer the features of directed song and suggest that high-level auditory areas including the CMM are involved in this social perception.

Contributions of an avian basal ganglia-forebrain circuit to real-time modulation of song.

Cortical-basal ganglia circuits have a critical role in motor control and motor learning. In songbirds, the anterior forebrain pathway (AFP) is a basal ganglia-forebrain circuit required for song learning and adult vocal plasticity but not for production of learned song. Here, we investigate functional contributions of this circuit to the control of song, a complex, learned motor skill. We test the hypothesis that neural activity in the AFP of adult birds can direct moment-by-moment changes in the primary motor areas responsible for generating song. We show that song-triggered microstimulation in the output nucleus of the AFP induces acute and specific changes in learned parameters of song. Moreover, under both natural and experimental conditions, variability in the pattern of AFP activity is associated with variability in song structure. Finally, lesions of the output nucleus of the AFP prevent naturally occurring modulation of song variability. These findings demonstrate a previously unappreciated capacity of the AFP to direct real-time changes in song. More generally, they suggest that frontal cortical and basal ganglia areas may contribute to motor learning by biasing motor output towards desired targets or by introducing stochastic variability required for reinforcement learning.

Temporal processing and adaptation in the songbird auditory forebrain.

Songbird auditory neurons must encode the dynamics of natural sounds at many volumes. We investigated how neural coding depends on the distribution of stimulus intensities. Using reverse-correlation, we modeled responses to amplitude-modulated sounds as the output of a linear filter and a nonlinear gain function, then asked how filters and nonlinearities depend on the stimulus mean and variance. Filter shape depended strongly on mean amplitude (volume): at low mean, most neurons integrated sound over many milliseconds, while at high mean, neurons responded more to local changes in amplitude. Increasing the variance (contrast) of amplitude modulations had less effect on filter shape but decreased the gain of firing in most cells. Both filter and gain changes occurred rapidly after a change in statistics, suggesting that they represent nonlinearities in processing. These changes may permit neurons to signal effectively over a wider dynamic range and are reminiscent of findings in other sensory systems.

Activity propagation in an avian basal ganglia-thalamocortical circuit essential for vocal learning.

In mammalian basal ganglia-thalamocortical circuits, GABAergic pallidal neurons are thought to "gate" or modulate excitation in thalamus with their strong inhibitory inputs and thus signal to cortex by pausing and permitting thalamic neurons to fire in response to excitatory drive. In contrast, in a homologous circuit specialized for vocal learning in songbirds, evidence suggests that pallidal neurons signal by eliciting postinhibitory rebound spikes in thalamus, which could occur even without any excitatory drive to thalamic neurons. To test whether songbird pallidal neurons can also communicate with thalamus by gating excitatory drive, as well as by postinhibitory rebound, we examined the activity of thalamic relay neurons in response to acute inactivation of the basal ganglia structure Area X; Area X contains the pallidal neurons that project to thalamus. Although inactivation of Area X should eliminate rebound-mediated spiking in thalamus, this manipulation tonically increased the firing rate of thalamic relay neurons, providing evidence that songbird pallidal neurons can gate tonic thalamic excitatory drive. We also found that the increased thalamic activity was fed forward to its target in the avian equivalent of cortex, which includes neurons that project to the vocal premotor area. These data raise the possibility that basal ganglia circuits can signal to cortex through thalamus both by generating postinhibitory rebound and by gating excitatory drive and may switch between these modes depending on the statistics of pallidal firing. Moreover, these findings provide insight into the strikingly different disruptive effects of basal ganglia and cortical lesions on songbird vocal learning.

Activity in a cortical-basal ganglia circuit for song is required for social context-dependent vocal variability.

Variability in adult motor output is important for enabling animals to respond to changing external conditions. Songbirds are useful for studying variability because they alter the amount of variation in their song depending on social context. When an adult zebra finch male sings to a female ("directed"), his song is highly stereotyped, but when he sings alone ("undirected"), his song varies across renditions. Lesions of the lateral magnocellular nucleus of the anterior nidopallium (LMAN), the output nucleus of a cortical-basal ganglia circuit for song, reduce song variability to that of the stereotyped "performance" state. However, such lesions not only eliminate LMAN's synaptic input to its targets, but can also cause structural or physiological changes in connected brain regions, and thus cannot assess whether the acute activity of LMAN is important for social modulation of adult song variability. To evaluate the effects of ongoing LMAN activity, we reversibly silenced LMAN in singing zebra finches by bilateral reverse microdialysis of the GABA(A) receptor agonist muscimol. We found that LMAN inactivation acutely reduced undirected song variability, both across and even within syllable renditions, to the level of directed song variability in all birds examined. Song variability returned to pre-muscimol inactivation levels after drug washout. However, unlike LMAN lesions, LMAN inactivation did not eliminate social context effects on song tempo in adult birds. These results indicate that the activity of LMAN neurons acutely and actively generates social context-dependent increases in adult song variability but that social regulation of tempo is more complex.

Differential influence of frequency, timing, and intensity cues in a complex acoustic categorization task.

Songbirds, which, like humans, learn complex vocalizations, provide an excellent model for the study of acoustic pattern recognition. Here we examined the role of three basic acoustic parameters in an ethologically relevant categorization task. Female zebra finches were first trained to classify songs as belonging to one of two males and then asked whether they could generalize this knowledge to songs systematically altered with respect to frequency, timing, or intensity. Birds' performance on song categorization fell off rapidly when songs were altered in frequency or intensity, but they generalized well to songs that were changed in duration by >25%. Birds were not deaf to timing changes, however; they detected these tempo alterations when asked to discriminate between the same song played back at two different speeds. In addition, when birds were retrained with songs at many intensities, they could correctly categorize songs over a wide range of volumes. Thus although they can detect all these cues, birds attend less to tempo than to frequency or intensity cues during song categorization. These results are unexpected for several reasons: zebra finches normally encounter a wide range of song volumes but most failed to generalize across volumes in this task; males produce only slight variations in tempo, but females generalized widely over changes in song duration; and all three acoustic parameters are critical for auditory neurons. Thus behavioral data place surprising constraints on the relationship between previous experience, behavioral task, neural responses, and perception. We discuss implications for models of auditory pattern recognition.

Neurons in a forebrain nucleus required for vocal plasticity rapidly switch between precise firing and variable bursting depending on social context.

Song is a learned vocal behavior influenced by social interactions. Prior work has suggested that the anterior forebrain pathway (AFP), a specialized pallial-basal ganglia circuit critical for vocal plasticity, mediates the influence of social signals on song. Here, we investigate the signals the AFP sends to song motor areas and their dependence on social context by characterizing singing-related activity of single neurons in the AFP output nucleus LMAN (lateral magnocellular nucleus of the anterior nidopallium). We show that interaction with females causes marked, real-time changes in firing properties of individual LMAN neurons. When males sing to females ("directed"), LMAN neurons exhibit reliable firing of single spikes precisely locked to song. In contrast, when males sing alone ("undirected"), the same LMAN neurons exhibit prominent burst firing and trial-by-trial variability. Burst structure and timing vary substantially across repeated undirected trials. Despite context-dependent differences in firing statistics, the average pattern of song-locked firing for an individual neuron is similar across behavioral contexts, suggesting a common underlying signal. Different LMAN neurons in the same bird, however, exhibit distinct firing patterns, suggesting that subsets of neurons jointly encode song features. Together, our findings demonstrate that behavioral interactions reversibly transform the signaling mode of LMAN neurons. Such changes may contribute to rapid switching of motor activity between variable and precise states. More generally, our results suggest that pallial-basal ganglia circuits contribute to motor learning and production through multiple mechanisms: patterned signals could guide changes in motor output while state-dependent variability could subserve motor exploration.

Neural encoding of auditory temporal context in a songbird basal ganglia nucleus, and its independence of birds' song experience.

Some of the most complex auditory neurons known are found in the songbird forebrain, throughout the 'song system', including its basal ganglia nucleus Area X. These cells are selective for the temporal order of the bird's own song (BOS): they typically respond strongly to BOS, but more weakly when the syllable sequence of BOS is played in reverse order (roBOS), indicating that they integrate auditory information over more than single syllables. Here, studying the zebra finch Area X, we found that order selectivity strongly depends on the mean syllable duration of individual songs, decreasing markedly as this duration approaches 150-200 ms. Simply segmenting the same songs differently, creating an increase in average syllable length towards 150 ms, caused a similar decrease in order selectivity. This suggests that song neurons integrate acoustic information over a relatively limited time window, predominantly less than 150 ms. We provided further support for this by showing that a significant fraction of Area X order selectivity was accounted for by the acoustic similarity between each BOS and roBOS, measured using cross-correlation with fixed window sizes, but only when the correlation windows were at least 50 ms and no more than 200 ms long. All the same findings were evident in birds raised without tutor exposure, indicating that tutor learning has little effect on neural mechanisms underlying song temporal selectivity. Our results suggest that song-selective neurons encode much of the temporal context of song using a short, constant time window that is conserved across differences in songs, birds and learning.

Birdbrains could teach basal ganglia research a new song.

Recent advances in anatomical, physiological and histochemical characterization of avian basal ganglia neurons and circuitry have revealed remarkable similarities to mammalian basal ganglia. A modern revision of the avian anatomical nomenclature has now provided a common language for studying the function of the cortical-basal-ganglia-cortical loop, enabling neuroscientists to take advantage of the specialization of basal ganglia areas in various avian species. For instance, songbirds, which learn their vocal motor behavior using sensory feedback, have specialized a portion of their cortical-basal ganglia circuitry for song learning and production. This discrete circuit dedicated to a specific sensorimotor task could be especially tractable for elucidating the interwoven sensory, motor and reward signals carried by basal ganglia, and the function of these signals in task learning and execution.

Propagation of correlated activity through multiple stages of a neural circuit.

The timing of spikes can carry information, for instance, when the temporal pattern of firing across neurons results in correlated activity. However, in part because central synapses are unreliable, correlated activity has not been observed to propagate through multiple subsequent stages in neural circuits, although such propagation has frequently been used in theoretical models. Using simultaneous single-unit and multiunit recordings from two or three vocal control nuclei of songbirds, measurement of coherency and time delays, and manipulation of neural activity, we provide evidence here for preserved correlation of activity through multiple steps of the neural circuit for song, including a basal ganglia circuit and its target vocal motor pathway. This suggests that these pathways contain highly functionally interconnected neurons and represent a neural architecture that can preserve information about the timing of firing of groups of neurons. Because the interaction of these song pathways is critical to vocal learning, the preserved correlation of activity may be important to the learning and production of sequenced motor acts and could be a general feature of basal ganglia-cortical interaction.

Naturalistic stimulation drives opposing heterosynaptic plasticity at two inputs to songbird cortex.

Songbirds learn precisely sequenced motor skills (songs) subserved by distinct brain areas, including the premotor cortical analog HVC, which is essential for producing learned song, and a 'cortical'-basal ganglia loop required for song plasticity. Inputs from these nuclei converge in RA (robust nucleus of the arcopallium), making it a likely locus for song learning. However, activity-dependent synaptic plasticity has never been described in either input. Using a slice preparation, we found that stimulation patterns based on singing-related activity were able to drive opposing changes in the strength of RA's inputs: when one input was potentiated, the other was depressed, with the direction and magnitude of changes depending on the relative timing of stimulation of the inputs. Moreover, pharmacological manipulations that blocked synaptic plasticity in vitro also prevented reinforcement-driven changes to song in vivo. Together, these findings highlight the importance of precise timing in the basal ganglia-motor cortical interactions subserving adaptive motor skills.

Is the songbird Area X striatal, pallidal, or both? An anatomical study.

Anatomical and neurophysiological studies have established that Area X, a songbird nucleus essential for vocal learning, is a basal ganglia structure, with mammalian striatal properties. However, Area X also sends a gamma-aminobutyric acid (GABA)ergic projection to the medial portion of the dorsolateral thalamus (DLM), a projection characteristic of the pallidum. These findings suggested that Area X contains both striatal and pallidal neurons. To test this hypothesis further, we investigated the neurochemistry and connectivity of Area X and its projections by using neurotransmitter antibodies, in combination with tracing studies. Like the mammalian striatum, Area X contains small enkephalin- and substance P-immunopositive neurons. Choline acetyltransferase-positive cells of Area X do not retrogradely label from DLM and are probably cholinergic interneurons similar to those in mammals. Like pallidal cells, large GABAergic cells project from Area X to the thalamus, but they also contain enkephalin, a characteristic of striatal neurons projecting to indirect pathway pallidal neurons. Moreover, many Area X cells are labeled with the pallidal marker Nkx2.1, but these do not include any thalamus-projecting neurons, suggesting that the projection cells are not of pallidal embryonic origin. Thus, although Area X combines both striatal and pallidal features, it is not a simple recapitulation of the mammalian circuit or of the avian lateral striatopallidal pathway: some individual Area X neurons may function as pallidal-like projection neurons but have striatal characteristics as well. Such heterogeneity of basal ganglia circuitry, both within and across species, may be facilitated by the developmental history of basal ganglia, which involves extensive migration and cellular intermixing.

Cellular, circuit, and synaptic mechanisms in song learning.

Songbirds, much like humans, learn their vocal behavior, and must be able to hear both themselves and others to do so. Studies of the brain areas involved in singing and song learning could reveal the underlying neural mechanisms. Here we describe experiments that explore the properties of the songbird anterior forebrain pathway (AFP), a basal ganglia-forebrain circuit known to be critical for song learning and for adult modification of vocal output. First, neural recordings in anesthetized, juvenile birds show that auditory AFP neurons become selectively responsive to the song stimuli that are compared during sensorimotor learning. Individual AFP neurons develop tuning to the bird's own song (BOS), and in many cases to the tutor song as well, even when these stimuli are manipulated to be very different from each other. Such dual selectivity could be useful in the BOS-tutor song comparison critical to song learning. Second, simultaneous neural recordings from the AFP and its target nucleus in the song motor pathway in anesthetized adult birds reveal correlated activity that is preserved through multiple steps of the circuits for song, including the AFP. This suggests that the AFP contains highly functionally interconnected neurons, an architecture that can preserve information about the timing of firing of groups of neurons. Finally, in vitro studies show that recurrent synapses between neurons in the AFP outflow nucleus, which are expected to contribute importantly to AFP correlation, can undergo activity-dependent and timing-sensitive strengthening. This synaptic enhancement appears to be restricted to birds in the sensory critical and early sensorimotor phases of learning. Together, these studies show that the AFP contains cells that reflect learning of both BOS and tutor song, as well as developmentally regulated synaptic and circuit mechanisms well-suited to create temporally organized assemblies of such cells. Such experience-dependent sensorimotor assemblies are likely to be critical to the AFP's role in song learning. Moreover, studies of such mechanisms in this basal ganglia circuit specialized for song may shed light more generally on how basal ganglia circuits function in guiding motor learning using sensory feedback signals.