An automated procedure for evaluating song imitation

Self-organization of songbird neural sequences during social isolation

Behaviors emerge via a combination of experience and innate predispositions. As the brain matures, it undergoes major changes in cellular, network, and functional properties that can be due to sensory experience as well as developmental processes. In normal birdsong learning, neural sequences emerge to control song syllables learned from a tutor. Here, we disambiguate the role of tutor experience and development in neural sequence formation by delaying exposure to a tutor. Using functional calcium imaging, we observe neural sequences in the absence of tutoring, demonstrating that tutor experience is not necessary for the formation of sequences. However, after exposure to a tutor, pre-existing sequences can become tightly associated with new song syllables. Since we delayed tutoring, only half our birds learned new syllables following tutor exposure. The birds that failed to learn were the birds in which pre-tutoring neural sequences were most 'crystallized,' that is, already tightly associated with their (untutored) song.

Machine learning and statistical classification of birdsong link vocal acoustic features with phylogeny

Birdsong is a longstanding model system for studying evolution and biodiversity. Here, we collected and analyzed high quality song recordings from seven species in the family Estrildidae. We measured the acoustic features of syllables and then used dimensionality reduction and machine learning classifiers to identify features that accurately assigned syllables to species. Species differences were captured by the first 3 principal components, corresponding to basic frequency, power distribution, and spectrotemporal features. We then identified the measured features underlying classification accuracy. We found that fundamental frequency, mean frequency, spectral flatness, and syllable duration were the most informative features for species identification. Next, we tested whether specific acoustic features of species' songs predicted phylogenetic distance. We found significant phylogenetic signal in syllable frequency features, but not in power distribution or spectrotemporal features. Results suggest that frequency features are more constrained by species' genetics than are other features, and are the best signal features for identifying species from song recordings. The absence of phylogenetic signal in power distribution and spectrotemporal features suggests that these song features are labile, reflecting learning processes and individual recognition.

A researcher's guide to the comparative assessment of vocal production learning

Vocal production learning (VPL) is the capacity to learn to produce new vocalizations, which is a rare ability in the animal kingdom and thus far has only been identified in a handful of mammalian taxa and three groups of birds. Over the last few decades, approaches to the demonstration of VPL have varied among taxa, sound production systems and functions. These discrepancies strongly impede direct comparisons between studies. In the light of the growing number of experimental studies reporting VPL, the need for comparability is becoming more and more pressing. The comparative evaluation of VPL across studies would be facilitated by unified and generalized reporting standards, which would allow a better positioning of species on any proposed VPL continuum. In this paper, we specifically highlight five factors influencing the comparability of VPL assessments: (i) comparison to an acoustic baseline, (ii) comprehensive reporting of acoustic parameters, (iii) extended reporting of training conditions and durations, (iv) investigating VPL function via behavioural, perception-based experiments and (v) validation of findings on a neuronal level. These guidelines emphasize the importance of comparability between studies in order to unify the field of vocal learning. This article is part of the theme issue 'Vocal learning in animals and humans'.

Low-dimensional learned feature spaces quantify individual and group differences in vocal repertoires

Increases in the scale and complexity of behavioral data pose an increasing challenge for data analysis. A common strategy involves replacing entire behaviors with small numbers of handpicked, domain-specific features, but this approach suffers from several crucial limitations. For example, handpicked features may miss important dimensions of variability, and correlations among them complicate statistical testing. Here, by contrast, we apply the variational autoencoder (VAE), an unsupervised learning method, to learn features directly from data and quantify the vocal behavior of two model species: the laboratory mouse and the zebra finch. The VAE converges on a parsimonious representation that outperforms handpicked features on a variety of common analysis tasks, enables the measurement of moment-by-moment vocal variability on the timescale of tens of milliseconds in the zebra finch, provides strong evidence that mouse ultrasonic vocalizations do not cluster as is commonly believed, and captures the similarity of tutor and pupil birdsong with qualitatively higher fidelity than previous approaches. In all, we demonstrate the utility of modern unsupervised learning approaches to the quantification of complex and high-dimensional vocal behavior.

An avian cortical circuit for chunking tutor song syllables into simple vocal-motor units

How are brain circuits constructed to achieve complex goals? The brains of young songbirds develop motor circuits that achieve the goal of imitating a specific tutor song to which they are exposed. Here, we set out to examine how song-generating circuits may be influenced early in song learning by a cortical region (NIf) at the interface between auditory and motor systems. Single-unit recordings reveal that, during juvenile babbling, NIf neurons burst at syllable onsets, with some neurons exhibiting selectivity for particular emerging syllable types. When juvenile birds listen to their tutor, NIf neurons are also activated at tutor syllable onsets, and are often selective for particular syllable types. We examine a simple computational model in which tutor exposure imprints the correct number of syllable patterns as ensembles in an interconnected NIf network. These ensembles are then reactivated during singing to train a set of syllable sequences in the motor network.

Young songbirds learn to imitate their parents’ songs. Here, the authors find that, in baby birds, neurons in a brain region at the interface of auditory and motor circuits signal the onsets of song syllables during both tutoring and babbling, suggesting a specific neural mechanism for vocal imitation.

A comparative analysis of the dopaminergic innervation of the executive caudal nidopallium in pigeon, chicken, zebra finch, and carrion crow

Despite the long, separate evolutionary history of birds and mammals, both lineages developed a rich behavioral repertoire of remarkably similar executive control generated by distinctly different brains. The seat for executive functioning in birds is the nidopallium caudolaterale (NCL) and the mammalian equivalent is known as the prefrontal cortex (PFC). Both are densely innervated by dopaminergic fibers, and are an integration center of sensory input and motor output. Whereas the variation of the PFC has been well documented in different mammalian orders, we know very little about the NCL across the avian clade. In order to investigate whether this structure adheres to species-specific variations, this study aimed to describe the trajectory of the NCL in pigeon, chicken, carrion crow and zebra finch. We employed immunohistochemistry to map dopaminergic innervation, and executed a Gallyas stain to visualize the dorsal arcopallial tract that runs between the NCL and the arcopallium. Our analysis showed that whereas the trajectory of the NCL in the chicken is highly comparable to the pigeon, the two Passeriformes show a strikingly different pattern. In both carrion crow and zebra finch, we identified four different subareas of high dopaminergic innervation that span the entire caudal forebrain. Based on their sensory input, motor output, and involvement in dopamine-related cognitive control of the delineated areas here, we propose that at least three morphologically different subareas constitute the NCL in these songbirds. Thus, our study shows that comparable to the PFC in mammals, the NCL in birds varies considerably across species.

Learning is enhanced by tailoring instruction to individual genetic differences

It is widely argued that personalized instruction based on individual differences in learning styles or genetic predispositions could improve learning outcomes. However, this proposition has resisted clear demonstration in human studies, where it is difficult to control experience and quantify outcomes. Here, we take advantage of the tractable nature of vocal learning in songbirds (Lonchura striata domestica) to test the idea that matching instruction to individual genetic predispositions can enhance learning. We use both cross-fostering and computerized instruction with synthetic songs to demonstrate that matching the tutor song to individual predispositions can improve learning across genetic backgrounds. Moreover, we find that optimizing instruction in this fashion can equalize learning differences across individuals that might otherwise be construed as genetically determined. Our results demonstrate potent, synergistic interactions between experience and genetics in shaping song, and indicate the likely importance of such interactions for other complex learned behaviors.

Emergent tuning for learned vocalizations in auditory cortex

Vocal learners use early social experience to develop auditory skills specialized for communication. However, it is unknown where in the auditory pathway neural responses become selective for vocalizations or how the underlying encoding mechanisms change with experience. We used a vocal tutoring manipulation in two species of songbird to reveal that tuning for conspecific song arises within the primary auditory cortical circuit. Neurons in the deep region of primary auditory cortex responded more to conspecific songs than to other species' songs and more to species-typical spectrotemporal modulations, but neurons in the intermediate (thalamorecipient) region did not. Moreover, birds that learned song from another species exhibited parallel shifts in selectivity and tuning toward the tutor species' songs in the deep but not the intermediate region. Our results locate a region in the auditory processing hierarchy where an experience-dependent coding mechanism aligns auditory responses with the output of a learned vocal motor behavior.

Songbird Ventral Pallidum Sends Diverse Performance Error Signals to Dopaminergic Midbrain

Motor skills improve with practice, requiring outcomes to be evaluated against ever-changing performance benchmarks, yet it remains unclear how performance error signals are computed. Here, we show that the songbird ventral pallidum (VP) is required for song learning and sends diverse song timing and performance error signals to the ventral tegmental area (VTA). Viral tracing revealed inputs to VP from auditory and vocal motor thalamus, auditory and vocal motor cortex, and VTA. Our findings show that VP circuits, commonly associated with hedonic functions, signal performance error during motor sequence learning.

Beyond Critical Period Learning: Striatal FoxP2 Affects the Active Maintenance of Learned Vocalizations in Adulthood

In humans, mutations in the transcription factor forkhead box P2 (FOXP2) result in language disorders associated with altered striatal structure. Like speech, birdsong is learned through social interactions during maturational critical periods, and it relies on auditory feedback during initial learning and on-going maintenance. Hearing loss causes learned vocalizations to deteriorate in adult humans and songbirds. In the adult songbird brain, most FoxP2-enriched regions (e.g., cortex, thalamus) show a static expression level, but in the striatal song control nucleus, area X, FoxP2 is regulated by singing and social context: when juveniles and adults sing alone, its levels drop, and songs are more variable. When males sing to females, FoxP2 levels remain high, and songs are relatively stable: this “on-line” regulation implicates FoxP2 in ongoing vocal processes, but its role in the auditory-based maintenance of learned vocalization has not been examined. To test this, we overexpressed FoxP2 in both hearing and deafened adult zebra finches and assessed effects on song sung alone versus songs directed to females. In intact birds singing alone, no changes were detected between songs of males expressing FoxP2 or a GFP construct in area X, consistent with the marked stability of mature song in this species. In contrast, songs of males overexpressing FoxP2 became more variable and were less preferable to females, unlike responses to songs of GFP-expressing control males. In deafened birds, song deteriorated more rapidly following FoxP2 overexpression relative to GFP controls. Together, these experiments suggest that behavior-driven FoxP2 expression and auditory feedback interact to precisely maintain learned vocalizations.

An automated approach to the quantitation of vocalizations and vocal learning in the songbird

Studies of learning mechanisms critically depend on the ability to accurately assess learning outcomes. This assessment can be impeded by the often complex, multidimensional nature of behavior. We present a novel, automated approach to evaluating imitative learning. Conceptually, our approach estimates how much of the content present in a reference behavior is absent from the learned behavior. We validate our approach through examination of songbird vocalizations, complex learned behaviors the study of which has provided many insights into sensory-motor learning in general and vocal learning in particular. Historically, learning has been holistically assessed by human inspection or through comparison of specific song features selected by experimenters (e.g. fundamental frequency, spectral entropy). In contrast, our approach uses statistical models to broadly capture the structure of each song, and then estimates the divergence between the two models. We show that our measure of song learning (the Kullback-Leibler divergence between two distributions corresponding to specific song data, or, Song D_KL) is well correlated with human evaluation of song learning. We then expand the analysis beyond learning and show that Song D_KL also detects the typical song deterioration that occurs following deafening. Finally, we illustrate how this measure can be extended to quantify differences in other complex behaviors such as human speech and handwriting. This approach potentially provides a framework for assessing learning across a broad range of behaviors like song that can be described as a set of discrete and repeated motor actions.

Measuring learning outcomes is a critical objective of research into the mechanisms that support learning. Demonstration that a given manipulation results in better or worse learning outcomes requires an accurate and consistent measurement of learning quality. However, many behaviors (e.g. speaking, walking, and writing) are complex and multidimensional, confounding the assessment of learning. One behavior subject to such confounds, vocal learning in Estrildid finches, has emerged as a vital model for sensory motor learning broadly and human speech learning in particular. Here, we demonstrate a new approach to the assessment of learning for complex high dimensional behaviors. Conceptually, we estimate the amount of content present in a reference behavior that is absent in the resultant learned behavior. We show that this measure provides a holistic and automated assessment of vocal learning in Estrildid finches that is consistent with human assessment. We then illustrate how this measure can be used to quantify changes in other complex behaviors such as human speech. We conclude that this approach could be useful in assessing shared content in other similarly structured behaviors composed of a set of discrete and repeated motor actions.

A subcortical circuit linking the cerebellum to the basal ganglia engaged in vocal learning

Speech is a complex sensorimotor skill, and vocal learning involves both the basal ganglia and the cerebellum. These subcortical structures interact indirectly through their respective loops with thalamo-cortical and brainstem networks, and directly via subcortical pathways, but the role of their interaction during sensorimotor learning remains undetermined. While songbirds and their song-dedicated basal ganglia-thalamo-cortical circuitry offer a unique opportunity to study subcortical circuits involved in vocal learning, the cerebellar contribution to avian song learning remains unknown. We demonstrate that the cerebellum provides a strong input to the song-related basal ganglia nucleus in zebra finches. Cerebellar signals are transmitted to the basal ganglia via a disynaptic connection through the thalamus and then conveyed to their cortical target and to the premotor nucleus controlling song production. Finally, cerebellar lesions impair juvenile song learning, opening new opportunities to investigate how subcortical interactions between the cerebellum and basal ganglia contribute to sensorimotor learning.

Vocal learning in songbirds requires cholinergic signaling in a motor cortex-like nucleus

Cholinergic inputs to cortex modulate plasticity and sensory processing, yet little is known about their role in motor control. Here, we show that cholinergic signaling in a songbird vocal motor cortical area, the robust nucleus of the arcopallium (RA), is required for song learning. Reverse microdialysis of nicotinic and muscarinic receptor antagonists into RA in juvenile birds did not significantly affect syllable timing or acoustic structure during vocal babbling. However, chronic blockade over weeks reduced singing quantity and impaired learning, resulting in an impoverished song with excess variability, abnormal acoustic features, and reduced similarity to tutor song. The demonstration that cholinergic signaling in a motor cortical area is required for song learning motivates the songbird as a tractable model system to identify roles of the basal forebrain cholinergic system in motor control. NEW & NOTEWORTHY Cholinergic inputs to cortex are evolutionarily conserved and implicated in sensory processing and synaptic plasticity. However, functions of cholinergic signals in motor areas are understudied and poorly understood. Here, we show that cholinergic signaling in a songbird vocal motor cortical area is not required for normal vocal variability during babbling but is essential for developmental song learning. Cholinergic modulation of motor cortex is thus required for learning but not for the ability to sing.

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.

VoICE: A semi-automated pipeline for standardizing vocal analysis across models

The study of vocal communication in animal models provides key insight to the neurogenetic basis for speech and communication disorders. Current methods for vocal analysis suffer from a lack of standardization, creating ambiguity in cross-laboratory and cross-species comparisons. Here, we present VoICE (Vocal Inventory Clustering Engine), an approach to grouping vocal elements by creating a high dimensionality dataset through scoring spectral similarity between all vocalizations within a recording session. This dataset is then subjected to hierarchical clustering, generating a dendrogram that is pruned into meaningful vocalization “types” by an automated algorithm. When applied to birdsong, a key model for vocal learning, VoICE captures the known deterioration in acoustic properties that follows deafening, including altered sequencing. In a mammalian neurodevelopmental model, we uncover a reduced vocal repertoire of mice lacking the autism susceptibility gene, Cntnap2. VoICE will be useful to the scientific community as it can standardize vocalization analyses across species and laboratories.

A role for descending auditory cortical projections in songbird vocal learning

Many learned motor behaviors are acquired by comparing ongoing behavior with an internal representation of correct performance, rather than using an explicit external reward. For example, juvenile songbirds learn to sing by comparing their song with the memory of a tutor song. At present, the brain regions subserving song evaluation are not known. In this study, we report several findings suggesting that song evaluation involves an avian 'cortical' area previously shown to project to the dopaminergic midbrain and other downstream targets. We find that this ventral portion of the intermediate arcopallium (AIV) receives inputs from auditory cortical areas, and that lesions of AIV result in significant deficits in vocal learning. Additionally, AIV neurons exhibit fast responses to disruptive auditory feedback presented during singing, but not during nonsinging periods. Our findings suggest that auditory cortical areas may guide learning by transmitting song evaluation signals to the dopaminergic midbrain and/or other subcortical targets.