Vocal exploration is locally regulated during song learning

Dynamic top-down biasing implements rapid adaptive changes to individual movements

Complex behaviors depend on the coordinated activity of neural ensembles in interconnected brain areas. The behavioral function of such coordination, often measured as co-fluctuations in neural activity across areas, is poorly understood. One hypothesis is that rapidly varying co-fluctuations may be a signature of moment-by-moment task-relevant influences of one area on another. We tested this possibility for error-corrective adaptation of birdsong, a form of motor learning which has been hypothesized to depend on the top-down influence of a higher-order area, LMAN (lateral magnocellular nucleus of the anterior nidopallium), in shaping moment-by-moment output from a primary motor area, RA (robust nucleus of the arcopallium). In paired recordings of LMAN and RA in singing birds, we discovered a neural signature of a top-down influence of LMAN on RA, quantified as an LMAN-leading co-fluctuation in activity between these areas. During learning, this co-fluctuation strengthened in a premotor temporal window linked to the specific movement, sequential context, and acoustic modification associated with learning. Moreover, transient perturbation of LMAN activity specifically within this premotor window caused rapid occlusion of pitch modifications, consistent with LMAN conveying a temporally localized motor-biasing signal. Combined, our results reveal a dynamic top-down influence of LMAN on RA that varies on the rapid timescale of individual movements and is flexibly linked to contexts associated with learning. This finding indicates that inter-area co-fluctuations can be a signature of dynamic top-down influences that support complex behavior and its adaptation.

Generative models of birdsong learning link circadian fluctuations in song variability to changes in performance

Learning skilled behaviors requires intensive practice over days, months, or years. Behavioral hallmarks of practice include exploratory variation and long-term improvements, both of which can be impacted by circadian processes. During weeks of vocal practice, the juvenile male zebra finch transforms highly variable and simple song into a stable and precise copy of an adult tutor's complex song. Song variability and performance in juvenile finches also exhibit circadian structure that could influence this long-term learning process. In fact, one influential study reported juvenile song regresses towards immature performance overnight, while another suggested a more complex pattern of overnight change. However, neither of these studies thoroughly examined how circadian patterns of variability may structure the production of more or less mature songs. Here we relate the circadian dynamics of song maturation to circadian patterns of song variation, leveraging a combination of data-driven approaches. In particular we analyze juvenile singing in learned feature space that supports both data-driven measures of song maturity and generative developmental models of song production. These models reveal that circadian fluctuations in variability lead to especially regressive morning variants even without overall overnight regression, and highlight the utility of data-driven generative models for untangling these contributions.

Dopamine neurons evaluate natural fluctuations in performance quality

Many motor skills are learned by comparing ongoing behavior to internal performance benchmarks. Dopamine neurons encode performance error in behavioral paradigms where error is externally induced, but it remains unknown whether dopamine also signals the quality of natural performance fluctuations. Here, we record dopamine neurons in singing birds and examine how spontaneous dopamine spiking activity correlates with natural fluctuations in ongoing song. Antidromically identified basal ganglia-projecting dopamine neurons correlate with recent, and not future, song variations, consistent with a role in evaluation, not production. Furthermore, maximal dopamine spiking occurs at a single vocal target, consistent with either actively maintaining the existing song or shifting the song to a nearby form. These data show that spontaneous dopamine spiking can evaluate natural behavioral fluctuations unperturbed by experimental events such as cues or rewards.

Learning and producing skilled behavior requires an internal measure of performance. Duffy et al. examine dopamine neurons’ relationship to natural song in singing birds. Spontaneous dopamine activity correlates with song fluctuations in a manner consistent with evaluation of natural behavioral variations, independent of external perturbations, cues, or rewards.

Vocal Learning and Behaviors in Birds and Human Bilinguals: Parallels, Divergences and Directions for Research

Comparisons between the communication systems of humans and animals are instrumental in contextualizing speech and language into an evolutionary and biological framework and for illuminating mechanisms of human communication. As a complement to previous work that compares developmental vocal learning and use among humans and songbirds, in this article we highlight phenomena associated with vocal learning subsequent to the development of primary vocalizations (i.e., the primary language (L1) in humans and the primary song (S1) in songbirds). By framing avian “second-song” (S2) learning and use within the human second-language (L2) context, we lay the groundwork for a scientifically-rich dialogue between disciplines. We begin by summarizing basic birdsong research, focusing on how songs are learned and on constraints on learning. We then consider commonalities in vocal learning across humans and birds, in particular the timing and neural mechanisms of learning, variability of input, and variability of outcomes. For S2 and L2 learning outcomes, we address the respective roles of age, entrenchment, and social interactions. We proceed to orient current and future birdsong inquiry around foundational features of human bilingualism: L1 effects on the L2, L1 attrition, and L1<–>L2 switching. Throughout, we highlight characteristics that are shared across species as well as the need for caution in interpreting birdsong research. Thus, from multiple instructive perspectives, our interdisciplinary dialogue sheds light on biological and experiential principles of L2 acquisition that are informed by birdsong research, and leverages well-studied characteristics of bilingualism in order to clarify, contextualize, and further explore S2 learning and use in songbirds.

What Is the Role of Thalamostriatal Circuits in Learning Vocal Sequences?

Basal ganglia (BG) circuits integrate sensory and motor-related information from the cortex, thalamus, and midbrain to guide learning and production of motor sequences. Birdsong, like speech, is comprised of precisely sequenced vocal elements. Learning song sequences during development relies on Area X, a vocalization related region in the medial striatum of the songbird BG. Area X receives inputs from cortical-like pallial song circuits and midbrain dopaminergic circuits and sends projections to the thalamus. It has recently been shown that thalamic circuits also send substantial projections back to Area X. Here, we outline a gated-reinforcement learning model for how Area X may use signals conveyed by thalamostriatal inputs to direct song learning. Integrating conceptual advances from recent mammalian and songbird literature, we hypothesize that thalamostriatal pathways convey signals linked to song syllable onsets and offsets and influence striatal circuit plasticity via regulation of cholinergic interneurons (ChIs). We suggest that syllable sequence associated vocal-motor information from the thalamus drive precisely timed pauses in ChIs activity in Area X. When integrated with concurrent corticostriatal and dopaminergic input, this circuit helps regulate plasticity on medium spiny neurons (MSNs) and the learning of syllable sequences. We discuss new approaches that can be applied to test core ideas of this model and how associated insights may provide a framework for understanding the function of BG circuits in learning motor sequences.

A memory-driven auditory program ensures selective and precise vocal imitation in zebra finches

In the vocal learning model, the juvenile first memorizes a model sound, and the imprinted memory gradually converts into vocal-motor output during the sensorimotor integration. However, early acquired memory may not precisely represent the fine structures of a model sound. How do juveniles ensure precise model imitation? Here we show that juvenile songbirds develop an auditory learning program by actively and attentively engaging with tutor’s singing during the sensorimotor phase. The listening/approaching behavior requires previously acquired model memory and the individual variability of approaching behavior correlates with the precision of tutor song imitation. Moreover, it is modulated by dopamine and associated with forebrain regions for sensory processing. Overall, precise vocal learning may involve two steps of auditory processing: a passive imprinting of model memory occurs during the early sensory period; the previously acquired memory then guides an active and selective engagement of the re-exposed model to fine tune model imitation.

Wan-Chun Liu et al. demonstrate that the sensory phase of vocal learning in zebra finches is split across two stages: (1) passive listening and formation of a memory, and (2) active listening and behavioral engagement of juveniles with adult tutors. Furthermore, they show that approach behavior is correlated with song imitation quality, and immediate early gene expression in the caudal medial nidopallium linked to auditory behavior.

Inception of memories that guide vocal learning in the songbird

Animals learn many complex behaviors by emulating the behavior of more experienced individuals. This essential, yet still poorly understood, form of learning relies on the ability to encode lasting memories of observed behaviors. We identified a vocal-motor pathway in the zebra finch where memories that guide learning of song-element durations can be implanted. Activation of synapses in this pathway seeds memories that guide learning of song-element duration and can override learning from social interactions with other individuals. Genetic lesions of this circuit after memory formation, however, do not disrupt subsequent song imitation, which suggests that these memories are stored at downstream synapses. Thus, activity at these sensorimotor synapses can bypass learning from auditory and social experience and embed memories that guide learning of song timing.

An automatic behavior recognition system classifies animal behaviors using movements and their temporal context

Animals can perform complex and purposeful behaviors by executing simpler movements in flexible sequences. It is particularly challenging to analyze behavior sequences when they are highly variable, as is the case in language production, certain types of birdsong and, as in our experiments, flies grooming. High sequence variability necessitates rigorous quantification of large amounts of data to identify organizational principles and temporal structure of such behavior. To cope with large amounts of data, and minimize human effort and subjective bias, researchers often use automatic behavior recognition software. Our standard grooming assay involves coating flies in dust and videotaping them as they groom to remove it. The flies move freely and so perform the same movements in various orientations. As the dust is removed, their appearance changes. These conditions make it difficult to rely on precise body alignment and anatomical landmarks such as eyes or legs and thus present challenges to existing behavior classification software. Human observers use speed, location, and shape of the movements as the diagnostic features of particular grooming actions. We applied this intuition to design a new automatic behavior recognition system (ABRS) based on spatiotemporal features in the video data, heavily weighted for temporal dynamics and invariant to the animal's position and orientation in the scene. We use these spatiotemporal features in two steps of supervised classification that reflect two time-scales at which the behavior is structured. As a proof of principle, we show results from quantification and analysis of a large data set of stimulus-induced fly grooming behaviors that would have been difficult to assess in a smaller dataset of human-annotated ethograms. While we developed and validated this approach to analyze fly grooming behavior, we propose that the strategy of combining alignment-invariant features and multi-timescale analysis may be generally useful for movement-based classification of behavior from video data.

Language Origins Viewed in Spontaneous and Interactive Vocal Rates of Human and Bonobo Infants

From the first months of life, human infants produce "protophones," speech-like, non-cry sounds, presumed absent, or only minimally present in other apes. But there have been no direct quantitative comparisons to support this presumption. In addition, by 2 months, human infants show sustained face-to-face interaction using protophones, a pattern thought also absent or very limited in other apes, but again, without quantitative comparison. Such comparison should provide evidence relevant to determining foundations of language, since substantially flexible vocalization, the inclination to explore vocalization, and the ability to interact socially by means of vocalization are foundations for language. Here we quantitatively compare data on vocalization rates in three captive bonobo (Pan paniscus) mother-infant pairs with various sources of data from our laboratories on human infant vocalization. Both humans and bonobos produced distress sounds (cries/screams) and laughter. The bonobo infants also produced sounds that were neither screams nor laughs and that showed acoustic similarities to the human protophones. These protophone-like sounds confirm that bonobo infants share with humans the capacity to produce vocalizations that appear foundational for language. Still, there were dramatic differences between the species in both quantity and function of the protophone and protophone-like sounds. The bonobo protophone-like sounds were far less frequent than the human protophones, and the human protophones were far less likely to be interpreted as complaints and more likely as vocal play. Moreover, we found extensive vocal interaction between human infants and mothers, but no vocal interaction in the bonobo mother-infant pairs-while bonobo mothers were physically responsive to their infants, we observed no case of a bonobo mother vocalization directed to her infant. Our cross-species comparison focuses on low- and moderate-arousal circumstances because we reason the roots of language entail vocalization not triggered by excitement, for example, during fighting or intense play. Language appears to be founded in flexible vocalization, used to regulate comfortable social interaction, to share variable affective states at various levels of arousal, and to explore vocalization itself.

The Development of Structured Vocalizations in Songbirds and Humans: A Comparative Analysis

Humans and songbirds face a common challenge: acquiring the complex vocal repertoire of their social group. Although humans are thought to be unique in their ability to convey symbolic meaning through speech, speech and birdsong are comparable in their acoustic complexity and the mastery with which the vocalizations of adults are acquired by young individuals. In this review, we focus on recent advances in the study of vocal development in humans and songbirds that shed new light on the emergence of distinct structural levels of vocal behavior and point to new possible parallels between both groups.

Capacities and neural mechanisms for auditory statistical learning across species

Statistical learning has been proposed as a possible mechanism by which individuals can become sensitive to the structures of language fundamental for speech perception. Since its description in human infants, statistical learning has been described in human adults and several non-human species as a general process by which animals learn about stimulus-relevant statistics. The neurobiology of statistical learning is beginning to be understood, but many questions remain about the underlying mechanisms. Why is the developing brain particularly sensitive to stimulus and environmental statistics, and what neural processes are engaged in the adult brain to enable learning from statistical regularities in the absence of external reward or instruction? This review will survey the statistical learning abilities of humans and non-human animals with a particular focus on communicative vocalizations. We discuss the neurobiological basis of statistical learning, and specifically what can be learned by exploring this process in both humans and laboratory animals. Finally, we describe advantages of studying vocal communication in rodents as a means to further our understanding of the cortical plasticity mechanisms engaged during statistical learning. We examine the use of rodents in the context of pup retrieval, which is an auditory-based and experience-dependent form of maternal behavior.

Experience-Dependent Intrinsic Plasticity During Auditory Learning

Song learning in zebra finches (Taeniopygia guttata) requires exposure to the song of a tutor, resulting in an auditory memory. This memory is the foundation for later sensorimotor learning, resulting in the production of a copy of the tutor's song. The cortical premotor nucleus HVC (proper name) is necessary for auditory and sensorimotor learning as well as the eventual production of adult song. We recently discovered that the intrinsic physiology of HVC neurons changes across stages of song learning, but are those changes the result of learning or are they experience-independent developmental changes? To test the role of auditory experience in driving intrinsic changes, patch-clamp experiments were performed comparing HVC neurons in juvenile birds with varying amounts of tutor exposure. The intrinsic physiology of HVC neurons changed as a function of tutor exposure. Counterintuitively, tutor deprivation resulted in juvenile HVC neurons showing an adult-like phenotype not present in tutor-exposed juveniles. Biophysical models were developed to predict which ion channels were modulated by experience. The models indicate that tutor exposure transiently suppressed the I _h and T-type Ca²⁺ currents in HVC neurons that target the basal ganglia, whereas tutor exposure increased the resting membrane potential and decreased the spike amplitude in HVC neurons that drive singing. Our findings suggest that intrinsic plasticity may be part of the mechanism for auditory learning in the HVC. More broadly, models of learning and memory should consider intrinsic plasticity as a possible mechanism by which the nervous system encodes the lasting effects of experience.SIGNIFICANCE STATEMENT It is well established that learning involves plasticity of the synapses between neurons. However, the activity of a neural circuit can also be dramatically altered by changes in the intrinsic properties (ion channels) of the component neurons. The present experiments show experience-dependent changes in the intrinsic physiology of neurons in the cortical premotor nucleus HVC (proper name) in juvenile zebra finches (Taeniopygia guttata) during auditory learning of a tutor's song. Tutor deprivation does not "arrest" development of intrinsic properties, but rather results in neurons with a premature adult-like physiological phenotype. It is possible that auditory learning involves a form of nonsynaptic plasticity and that experience-dependent suppression of specific ion channels may work in concert with synaptic plasticity to promote vocal learning.

Songbirds work around computational complexity by learning song vocabulary independently of sequence

While acquiring motor skills, animals transform their plastic motor sequences to match desired targets. However, because both the structure and temporal position of individual gestures are adjustable, the number of possible motor transformations increases exponentially with sequence length. Identifying the optimal transformation towards a given target is therefore a computationally intractable problem. Here we show an evolutionary workaround for reducing the computational complexity of song learning in zebra finches. We prompt juveniles to modify syllable phonology and sequence in a learned song to match a newly introduced target song. Surprisingly, juveniles match each syllable to the most spectrally similar sound in the target, regardless of its temporal position, resulting in unnecessary sequence errors, that they later try to correct. Thus, zebra finches prioritize efficient learning of syllable vocabulary, at the cost of inefficient syntax learning. This strategy provides a non-optimal but computationally manageable solution to the task of vocal sequence learning.

Rhythmic Continuous-Time Coding in the Songbird Analog of Vocal Motor Cortex

Songbirds learn and produce complex sequences of vocal gestures. Adult birdsong requires premotor nucleus HVC, in which projection neurons (PNs) burst sparsely at stereotyped times in the song. It has been hypothesized that PN bursts, as a population, form a continuous sequence, while a different model of HVC function proposes that both HVC PN and interneuron activity is tightly organized around motor gestures. Using a large dataset of PNs and interneurons recorded in singing birds, we test several predictions of these models. We find that PN bursts in adult birds are continuously and nearly uniformly distributed throughout song. However, we also find that PN and interneuron firing rates exhibit significant 10-Hz rhythmicity locked to song syllables, peaking prior to syllable onsets and suppressed prior to offsets-a pattern that predominates PN and interneuron activity in HVC during early stages of vocal learning.

The Role of Variability in Motor Learning

Trial-to-trial variability in the execution of movements and motor skills is ubiquitous and widely considered to be the unwanted consequence of a noisy nervous system. However, recent studies have suggested that motor variability may also be a feature of how sensorimotor systems operate and learn. This view, rooted in reinforcement learning theory, equates motor variability with purposeful exploration of motor space that, when coupled with reinforcement, can drive motor learning. Here we review studies that explore the relationship between motor variability and motor learning in both humans and animal models. We discuss neural circuit mechanisms that underlie the generation and regulation of motor variability and consider the implications that this work has for our understanding of motor learning.

Maturation-dependent control of vocal temporal plasticity in a songbird

Birdsong is a unique model to address learning mechanisms of the timing control of sequential behaviors, with characteristic temporal structures consisting of serial sequences of brief vocal elements (syllables) and silent intervals (gaps). Understanding the neural mechanisms for plasticity of such sequential behavior should be aided by characterization of its developmental changes. Here, we assessed the level of acute vocal plasticity between young and adult Bengalese finches, and also quantified developmental change in variability of temporal structure. Acute plasticity was tested by delivering aversive noise bursts contingent on duration of a target gap, such that birds could avoid the noise by modifying their song. We found that temporal variability of song features decreased with birds' maturation. Noise-avoidance experiments demonstrated that maximal changes of gap durations were larger in young that in adult birds. After these young birds matured, the maximal change decreased to a similar level as adults. The variability of these target gaps also decreased as the birds matured. Such parallel changes suggest that the level of acute temporal plasticity could be predicted from ongoing temporal variability. Further, we found that young birds gradually began to stop their song at the target gap and restart from the introductory part of song, whereas adults did not. According to a synaptic chain model for timing sequence generation in premotor nuclei, adult learning would be interpreted as adaptive changes in conduction delays between chain-to-chain connections, whereas the learning of young birds could mainly depend on changes of the connections. © 2017 Wiley Periodicals, Inc. Develop Neurobiol 77: 995-1006, 2017.

A Distributed Recurrent Network Contributes to Temporally Precise Vocalizations

How do forebrain and brainstem circuits interact to produce temporally precise and reproducible behaviors? Birdsong is an elaborate, temporally precise, and stereotyped vocal behavior controlled by a network of forebrain and brainstem nuclei. An influential idea is that song premotor neurons in a forebrain nucleus (HVC) form a synaptic chain that dictates song timing in a top-down manner. Here we combine physiological, dynamical, and computational methods to show that song timing is not generated solely by a mechanism localized to HVC but instead is the product of a distributed and recurrent synaptic network spanning the forebrain and brainstem, of which HVC is a component.

A quantitative method for analyzing species-specific vocal sequence pattern and its developmental dynamics

BACKGROUND

Songbirds are a preeminent animal model for understanding the neural basis underlying the development and evolution of a complex learned behavior, bird song. However, only a few quantitative methods exist to analyze these species-specific sequential behaviors in multiple species using the same calculation method.

NEW METHOD

We report a method of analysis that focuses on calculating the frequency of characteristic syllable transitions in songs. This method comprises two steps: The first step involves forming correlation matrices of syllable similarity scores, named syllable similarity matrices (SSMs); these are obtained by calculating the round-robin comparison of all the syllables in two songs, while maintaining the sequential order of syllables in the songs. In the second step, each occurrence rate of three patterns of binarized "2 rows×2 columns" cells in the SSMs is calculated to extract information on the characteristic syllable transitions.

RESULTS

The SSM analysis method allowed obtaining species-specific features of song patterns and intraspecies individual variability simultaneously. Furthermore, it enabled quantitative tracking of the developmental trajectory of the syllable sequence patterns.

COMPARISON WITH EXISTING METHOD

This method enables us to extract the species-specific song patterns and dissect the regulation of song syntax development without human-biased procedures for syllable identification. This method can be adapted to study the acoustic communication systems in several animal species, such as insects and mammals.

CONCLUSIONS

This present method provides a comprehensive qualitative approach for understanding the regulation of species specificity and its development in vocal learning.

Neural circuits. Inhibition protects acquired song segments during vocal learning in zebra finches

Vocal imitation involves incorporating instructive auditory information into relevant motor circuits through processes that are poorly understood. In zebra finches, we found that exposure to a tutor's song drives spiking activity within premotor neurons in the juvenile, whereas inhibition suppresses such responses upon learning in adulthood. We measured inhibitory currents evoked by the tutor song throughout development while simultaneously quantifying each bird's learning trajectory. Surprisingly, we found that the maturation of synaptic inhibition onto premotor neurons is correlated with learning but not age. We used synthetic tutoring to demonstrate that inhibition is selective for specific song elements that have already been learned and not those still in refinement. Our results suggest that structured inhibition plays a crucial role during song acquisition, enabling a piece-by-piece mastery of complex tasks.

Acute off-target effects of neural circuit manipulations

Rapid and reversible manipulations of neural activity in behaving animals are transforming our understanding of brain function. An important assumption underlying much of this work is that evoked behavioural changes reflect the function of the manipulated circuits. We show that this assumption is problematic because it disregards indirect effects on the independent functions of downstream circuits. Transient inactivations of motor cortex in rats and nucleus interface (Nif) in songbirds severely degraded task-specific movement patterns and courtship songs, respectively, which are learned skills that recover spontaneously after permanent lesions of the same areas. We resolve this discrepancy in songbirds, showing that Nif silencing acutely affects the function of HVC, a downstream song control nucleus. Paralleling song recovery, the off-target effects resolved within days of Nif lesions, a recovery consistent with homeostatic regulation of neural activity in HVC. These results have implications for interpreting transient circuit manipulations and for understanding recovery after brain lesions.

Growth and splitting of neural sequences in songbird vocal development

Neural sequences are a fundamental feature of brain dynamics underlying diverse behaviors, but the mechanisms by which they develop during learning remain unknown. Songbirds learn vocalizations composed of syllables; in adult birds, each syllable is produced by a different sequence of action potential bursts in the premotor cortical area HVC. Here we carried out recordings of large populations of HVC neurons in singing juvenile birds throughout learning to examine the emergence of neural sequences. Early in vocal development, HVC neurons begin producing rhythmic bursts, temporally locked to a ‘prototype’ syllable. Different neurons are active at different latencies relative to syllable onset to form a continuous sequence. Through development, as new syllables emerge from the prototype syllable, initially highly overlapping burst sequences become increasingly distinct. We propose a mechanistic model in which multiple neural sequences can emerge from the growth and splitting of a common precursor sequence.

A hypothesis on improving foreign accents by optimizing variability in vocal learning brain circuits

Rapid vocal motor learning is observed when acquiring a language in early childhood, or learning to speak another language later in life. Accurate pronunciation is one of the hardest things for late learners to master and they are almost always left with a non-native accent. Here, I propose a novel hypothesis that this accent could be improved by optimizing variability in vocal learning brain circuits during learning. Much of the neurobiology of human vocal motor learning has been inferred from studies on songbirds. Jarvis (2004) proposed the hypothesis that as in songbirds there are two pathways in humans: one for learning speech (the striatal vocal learning pathway), and one for production of previously learnt speech (the motor pathway). Learning new motor sequences necessary for accurate non-native pronunciation is challenging and I argue that in late learners of a foreign language the vocal learning pathway becomes inactive prematurely. The motor pathway is engaged once again and learners maintain their original native motor patterns for producing speech, resulting in speaking with a foreign accent. Further, I argue that variability in neural activity within vocal motor circuitry generates vocal variability that supports accurate non-native pronunciation. Recent theoretical and experimental work on motor learning suggests that variability in the motor movement is necessary for the development of expertise. I propose that there is little trial-by-trial variability when using the motor pathway. When using the vocal learning pathway variability gradually increases, reflecting an exploratory phase in which learners try out different ways of pronouncing words, before decreasing and stabilizing once the “best” performance has been identified. The hypothesis proposed here could be tested using behavioral interventions that optimize variability and engage the vocal learning pathway for longer, with the prediction that this would allow learners to develop new motor patterns that result in more native-like pronunciation.

Behavior-linked FoxP2 regulation enables zebra finch vocal learning

Mutations in the FOXP2 transcription factor cause an inherited speech and language disorder, but how FoxP2 contributes to learning of these vocal communication signals remains unclear. FoxP2 is enriched in corticostriatal circuits of both human and songbird brains. Experimental knockdown of this enrichment in song control neurons of the zebra finch basal ganglia impairs tutor song imitation, indicating that adequate FoxP2 levels are necessary for normal vocal learning. In unmanipulated birds, vocal practice acutely downregulates FoxP2, leading to increased vocal variability and dynamic regulation of FoxP2 target genes. To determine whether this behavioral regulation is important for song learning, here, we used viral-driven overexpression of FoxP2 to counteract its downregulation. This manipulation disrupted the acute effects of song practice on vocal variability and caused inaccurate song imitation. Together, these findings indicate that dynamic behavior-linked regulation of FoxP2, rather than absolute levels, is critical for vocal learning.

Vocal learning in a social mammal: Demonstrated by isolation and playback experiments in bats

The evolution of human language is shrouded in mystery as it is unparalleled in the animal kingdom. Whereas vocal learning is crucial for the development of speech in humans, it seems rare among nonhuman animals. Songbirds often serve as a model for vocal learning, but the lack of a mammalian model hinders our quest for the origin of this capability. We report the influence of both isolation and playback experiments on the vocal development of a mammal, the Egyptian fruit bat. We continuously recorded pups from birth to adulthood and found that, when raised in a colony, pups acquired the adult repertoire, whereas when acoustically isolated, they exhibited underdeveloped vocalizations. Isolated pups that heard bat recordings exhibited a repertoire that replicated the playbacks they were exposed to. These findings demonstrate vocal learning in a social mammal, and suggest bats as a model for language acquisition.

A neural circuit mechanism for regulating vocal variability during song learning in zebra finches

Motor skill learning is characterized by improved performance and reduced motor variability. The neural mechanisms that couple skill level and variability, however, are not known. The zebra finch, a songbird, presents a unique opportunity to address this question because production of learned song and induction of vocal variability are instantiated in distinct circuits that converge on a motor cortex analogue controlling vocal output. To probe the interplay between learning and variability, we made intracellular recordings from neurons in this area, characterizing how their inputs from the functionally distinct pathways change throughout song development. We found that inputs that drive stereotyped song-patterns are strengthened and pruned, while inputs that induce variability remain unchanged. A simple network model showed that strengthening and pruning of action-specific connections reduces the sensitivity of motor control circuits to variable input and neural ‘noise’. This identifies a simple and general mechanism for learning-related regulation of motor variability.

DOI:http://dx.doi.org/10.7554/eLife.03697.001

‘Practice makes perfect’ captures the essence of how we learn new skills. When learning to play a musical instrument, for example, it often takes hours of practice before we can play a single piece of music properly for the first time. And as we get better, the variability in our performance—which is an advantage during the early stages of learning—becomes less. Likewise, songbirds need lots of practice in order to master the intricate songs they need to sing to attract mates.

Studies in songbirds show that the neural circuits in the brain that are responsible for producing song and for generating vocal variability both converge on a motor control region called the robust nucleus of the arcopallium (or RA for short). However, the details of how learning a song leads to reduced variability in vocal performance are poorly understood.

Now Garst-Orozco et al. have investigated the relationship between learning and variability by studying brain slices of zebra finches. Their experiments reveal that the inputs received by RA neurons from a higher-order brain region that controls song change with practice, with some inputs becoming stronger and others being eliminated as the birds' singing ability improves. However, inputs received by RA neurons from the circuit that generates vocal variability do not change despite the song becoming increasingly precise.

Using a computer simulation, Garst-Orozco et al. show that the sensitivity of RA neurons to variable or ‘noisy’ input is reduced when inputs from the brain region that controls song are adaptively strengthened and eliminated. This ensures that when the notes and syllables that make up the bird's song have finally been learned, they will be uttered with high fidelity and precision. Intriguingly, motor skill learning in mammals have been associated with neural connectivity changes very similar to those described by Garst-Orozco et al., suggesting that insights from songbirds may lead to a better understanding of how ‘practice makes perfect’ also works in humans.

DOI:http://dx.doi.org/10.7554/eLife.03697.002

Vocal learning beyond imitation: mechanisms of adaptive vocal development in songbirds and human infants

Studies of vocal learning in songbirds typically focus on the acquisition of sensory templates for song imitation and on the consequent process of matching song production to templates. However, functional vocal development also requires the capacity to adaptively diverge from sensory templates, and to flexibly assemble vocal units. Examples of adaptive divergence include the corrective imitation of abnormal songs, and the decreased tendency to copy over-abundant syllables. Such frequency-dependent effects might mirror tradeoffs between the assimilation of group identity (culture) while establishing individual and flexibly expressive songs. Intriguingly, although the requirements for vocal plasticity vary across songbirds, and more so between birdsong and language, the capacity to flexibly assemble vocal sounds develops in a similar, stepwise manner across species. Therefore, universal features of vocal learning go well beyond the capacity to imitate.

Two neural streams, one voice: pathways for theme and variation in the songbird brain

Birdsong offers a unique model system to understand how a developing brain - once given a set of purely acoustic targets - teaches itself the vocal-tract gestures necessary to imitate those sounds. Like human infants, to juvenile male zebra finches (Taeniopygia guttata) falls the burden of initiating the vocal-motor learning of adult sounds. In both species, adult caregivers provide only a set of sounds to be imitated, with little or no information about the vocal-tract gestures used to produce the sounds. Here, we focus on the central control of birdsong and review the recent discovery that zebra finch song is under dual premotor control. Distinct forebrain pathways for structured (theme) and unstructured (variation) singing not only raise new questions about mechanisms of sensory-motor integration, but also provide a fascinating new research opportunity. A cortical locus for a motor memory of the learned song is now firmly established, meaning that anatomical, physiological, and computational approaches are poised to reveal the neural mechanisms used by the brain to compose the songs of birds.

Vocal motor changes beyond the sensitive period for song plasticity

Behavior is critically shaped during sensitive periods in development. Birdsong is a learned vocal behavior that undergoes dramatic plasticity during a sensitive period of sensorimotor learning. During this period, juvenile songbirds engage in vocal practice to shape their vocalizations into relatively stereotyped songs. By the time songbirds reach adulthood, their songs are relatively stable and thought to be "crystallized." Recent studies, however, highlight the potential for adult song plasticity and suggest that adult song could naturally change over time. As such, we investigated the degree to which temporal and spectral features of song changed over time in adult Bengalese finches. We observed that the sequencing and timing of song syllables became more stereotyped over time. Increases in the stereotypy of syllable sequencing were due to the pruning of infrequently produced transitions and, to a lesser extent, increases in the prevalence of frequently produced transitions. Changes in song tempo were driven by decreases in the duration and variability of intersyllable gaps. In contrast to significant changes to temporal song features, we found little evidence that the spectral structure of adult song syllables changed over time. These data highlight differences in the degree to which temporal and spectral features of adult song change over time and support evidence for distinct mechanisms underlying the control of syllable sequencing, timing, and structure. Furthermore, the observed changes to temporal song features are consistent with a Hebbian framework of behavioral plasticity and support the notion that adult song should be considered a form of vocal practice.

Models of vocal learning in the songbird: Historical frameworks and the stabilizing critic

Birdsong is a form of sensorimotor learning that involves a mirror-like system that activates with both song hearing and production. Early models of song learning, based on behavioral measures, identified key features of vocal plasticity, such as the requirements for memorization of a tutor song and auditory feedback during song practice. The concept of a comparator, which compares the memory of the tutor song to auditory feedback, featured prominently. Later models focused on linking anatomically-defined neural modules to behavioral concepts, such as the comparator. Exploiting the anatomical modularity of the songbird brain, localized lesions illuminated mechanisms of the neural song system. More recent models have integrated neuronal mechanisms identified in other systems with observations in songbirds. While these models explain multiple aspects of song learning, they must incorporate computational elements based on unknown biological mechanisms to bridge the motor-to-sensory delay and/or transform motor signals into the sensory domain. Here, I introduce the stabilizing critic hypothesis, which enables sensorimotor learning by (1) placing a purely sensory comparator afferent of the song system and (2) endowing song system disinhibitory interneuron networks with the capacity both to bridge the motor-sensory delay through prolonged bursting and to stabilize song segments selectively based on the comparator signal. These proposed networks stabilize an otherwise variable signal generated by both putative mirror neurons and a cortical-basal ganglia-thalamic loop. This stabilized signal then temporally converges with a matched premotor signal in the efferent song motor cortex, promoting spike-timing-dependent plasticity in the premotor circuitry and behavioral song learning.

An automated procedure for evaluating song imitation

Songbirds have emerged as an excellent model system to understand the neural basis of vocal and motor learning. Like humans, songbirds learn to imitate the vocalizations of their parents or other conspecific “tutors.” Young songbirds learn by comparing their own vocalizations to the memory of their tutor song, slowly improving until over the course of several weeks they can achieve an excellent imitation of the tutor. Because of the slow progression of vocal learning, and the large amounts of singing generated, automated algorithms for quantifying vocal imitation have become increasingly important for studying the mechanisms underlying this process. However, methodologies for quantifying song imitation are complicated by the highly variable songs of either juvenile birds or those that learn poorly because of experimental manipulations. Here we present a method for the evaluation of song imitation that incorporates two innovations: First, an automated procedure for selecting pupil song segments, and, second, a new algorithm, implemented in Matlab, for computing both song acoustic and sequence similarity. We tested our procedure using zebra finch song and determined a set of acoustic features for which the algorithm optimally differentiates between similar and non-similar songs.

The putative pigeon homologue to song bird LMAN does not modulate behavioral variability

The active generation of behavioral variability is thought to be a pivotal element in reinforcement based learning. One example for this principle is song learning in oscine birds. Oscines possess a highly specialized set of brain areas that compose the song system. It is yet unclear how the song system evolved. One important hypothesis assumes a motor origin of the song system, i.e. the song system may have developed from motor pathways that were present in an early ancestor of extant birds. Indeed, in pigeons neural pathways are present that parallel the song system. We examined whether one component of these pathways, a forebrain area termed nidopallium intermedium medialis pars laterale (NIML), is functionally comparable to its putative homologue, the lateral magnocellular nucleus of the anterior nidopallium (LMAN) of the song system. LMAN conveys variability into the motor output during singing; a function crucial for song learning and maintenance. We tested if NIML is likewise associated with the generation of variability. We used a behavioral paradigm in which pigeons had to find hidden target areas on a touch screen to gain food rewards. Alterations in pecking variability would result in changes of performance levels in this search paradigm. We found that pharmacological inactivation of NIML did not reduce pecking variability contrasting increases of song stereotypy observed after LMAN inactivation.

Neural representation of a target auditory memory in a cortico-basal ganglia pathway

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: (1) one in which neurons are selectively tuned to memorized sounds and (2) 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: after 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, whereas sensorimotor circuits support motor aspects of learning.

Rhythmic cortical neurons increase their oscillations and sculpt basal ganglia signaling during motor learning

The function and modulation of neural circuits underlying motor skill may involve rhythmic oscillations (Feller, 1999; Marder and Goaillard, 2006; Churchland et al., 2012). In the proposed pattern generator for birdsong, the cortical nucleus HVC, the frequency and power of oscillatory bursting during singing increases with development (Crandall et al., 2007; Day et al., 2009). We examined the maturation of cellular activity patterns that underlie these changes. Single unit ensemble recording combined with antidromic identification (Day et al., 2011) was used to study network development in anesthetized zebra finches. Autocovariance quantified oscillations within single units. A subset of neurons oscillated in the theta/alpha/mu/beta range (8-20 Hz), with greater power in adults compared to juveniles. Across the network, the normalized oscillatory power in the 8-20 Hz range was greater in adults than juveniles. In addition, the correlated activity between rhythmic neuron pairs increased with development. We next examined the functional impact of the oscillators on the output neurons of HVC. We found that the firing of oscillatory neurons negatively correlated with the activity of cortico-basal ganglia neurons (HVC(X)s), which project to Area X (the song basal ganglia). If groups of oscillators work together to tonically inhibit and precisely control the spike timing of adult HVC(X)s with coordinated release from inhibition, then the activity of HVC(X)s in juveniles should be decreased relative to adults due to uncorrelated, tonic inhibition. Consistent with this hypothesis, HVC(X)s had lower activity in juveniles. These data reveal network changes that shape cortical-to-basal ganglia signaling during motor learning.

Vocal learning is constrained by the statistics of sensorimotor experience

The brain uses sensory feedback to correct behavioral errors. Larger errors by definition require greater corrections, and many models of learning assume that larger sensory feedback errors drive larger motor changes. However, an alternative perspective is that larger errors drive learning less effectively because such errors fall outside the range of errors normally experienced and are therefore unlikely to reflect accurate feedback. This is especially crucial in vocal control because auditory feedback can be contaminated by environmental noise or sensory processing errors. A successful control strategy must therefore rely on feedback to correct errors while disregarding aberrant auditory signals that would lead to maladaptive vocal corrections. We hypothesized that these constraints result in compensation that is greatest for smaller imposed errors and least for larger errors. To test this hypothesis, we manipulated the pitch of auditory feedback in singing Bengalese finches. We found that learning driven by larger sensory errors was both slower than that resulting from smaller errors and showed less complete compensation for the imposed error. Additionally, we found that a simple principle could account for these data: the amount of compensation was proportional to the overlap between the baseline distribution of pitch production and the distribution experienced during the shift. Correspondingly, the fraction of compensation approached zero when pitch was shifted outside of the song's baseline pitch distribution. Our data demonstrate that sensory errors drive learning best when they fall within the range of production variability, suggesting that learning is constrained by the statistics of sensorimotor experience.

Development of temporal structure in zebra finch song

Zebra finch song has provided an excellent case study in the neural basis of sequence learning, with a high degree of temporal precision and tight links with precisely timed bursting in forebrain neurons. To examine the development of song timing, we measured the following four aspects of song temporal structure at four age ranges between 65 and 375 days posthatch: the mean durations of song syllables and the silent gaps between them, timing variability linked to song tempo, timing variability expressed independently across syllables and gaps, and transition probabilities between consecutive syllable pairs. We found substantial increases in song tempo between 65 and 85 days posthatch, due almost entirely to a shortening of gaps. We also found a decrease in tempo variability, also specific to gaps. Both the magnitude of the increase in tempo and the decrease in tempo variability were correlated on gap-by-gap basis with increases in the reliability of corresponding syllable transitions. Syllables had no systematic increase in tempo or decrease in tempo variability. In contrast to tempo parameters, both syllables and gaps showed an early sharp reduction in independent variability followed by continued reductions over the first year. The data suggest that links between syllable-based representations are strengthened during the later parts of the traditional period of song learning and that song rhythm continues to become more regular throughout the first year of life. Similar learning patterns have been identified in human sequence learning, suggesting a potentially rich area of comparative research.

Directed functional connectivity matures with motor learning in a cortical pattern generator

Sequential motor skills may be encoded by feedforward networks that consist of groups of neurons that fire in sequence (Abeles 1991; Long et al. 2010). However, there has been no evidence of an anatomic map of activation sequence in motor control circuits, which would be potentially detectable as directed functional connectivity of coactive neuron groups. The proposed pattern generator for birdsong, the HVC (Long and Fee 2008; Vu et al. 1994), contains axons that are preferentially oriented in the rostrocaudal axis (Nottebohm et al. 1982; Stauffer et al. 2012). We used four-tetrode recordings to assess the activity of ensembles of single neurons along the rostrocaudal HVC axis in anesthetized zebra finches. We found an axial, polarized neural network in which sequential activity is directionally organized along the rostrocaudal axis in adult males, who produce a stereotyped song. Principal neurons fired in rostrocaudal order and with interneurons that were rostral to them, suggesting that groups of excitatory neurons fire at the leading edge of travelling waves of inhibition. Consistent with the synchronization of neurons by caudally travelling waves of inhibition, the activity of interneurons was more coherent in the orthogonal mediolateral axis than in the rostrocaudal axis. If directed functional connectivity within the HVC is important for stereotyped, learned song, then it may be lacking in juveniles, which sing a highly variable song. Indeed, we found little evidence for network directionality in juveniles. These data indicate that a functionally directed network within the HVC matures during sensorimotor learning and may underlie vocal patterning.