Cortical inter-hemispheric circuits for multimodal vocal learning in songbirds

A neuroethological view of the multifaceted sensory influences on birdsong

Learning and execution of complex motor skills are often modulated by sensory feedback and contextual cues arriving across multiple sensory modalities. Vocal motor behaviors, in particular, are primarily influenced by auditory inputs, both during learning and mature vocal production. The importance of auditory input in shaping vocal output has been investigated in several songbird species that acquire their adult song based on auditory exposure to a tutor during development. Recent studies have highlighted the influences of stimuli arriving through other sensory channels in juvenile song learning and in adult song production. Here, we review changes induced by diverse sensory stimuli during the song learning process and the production of adult song, considering the neuroethological significance of sensory channels in different species of songbirds. Additionally, we highlight advances, open questions, and possible future approaches for understanding the neural circuits that enable the multimodal shaping of singing behavior.

Changes in the dopaminergic circuitry and adult neurogenesis linked to reinforcement learning in corvids

The caudolateral nidopallium (NCL, an analog of the prefrontal cortex) is known to be involved in learning, memory, and discrimination in corvids (a songbird), whereas the involvement of other brain regions in these phenomena is not well explored. We used house crows (Corvus splendens) to explore the neural correlates of learning and decision-making by initially training them on a shape discrimination task followed by immunohistochemistry to study the immediate early gene expression (Arc), a dopaminoceptive neuronal marker (DARPP-32, Dopamine- and cAMP-regulated phosphoprotein, Mr 32 kDa) to understand the involvement of the reward pathway and an immature neuronal marker (DCX, doublecortin) to detect learning-induced changes in adult neurogenesis. We performed neuronal counts and neuronal tracing, followed by morphometric analyses. Our present results have demonstrated that besides NCL, other parts of the caudal nidopallium (NC), avian basal ganglia, and intriguingly, vocal control regions in house crows are involved in visual discrimination. We have also found that training on the visual discrimination task can be correlated with neurite pruning in mature dopaminoceptive neurons and immature DCX-positive neurons in the NC of house crows. Furthermore, there is an increase in the incorporation of new neurons throughout NC and the medial striatum which can also be linked to learning. For the first time, our results demonstrate that a combination of structural changes in mature and immature neurons and adult neurogenesis are linked to learning in corvids.

Parallel executive pallio-motor loops in the pigeon brain

A core component of the avian pallial cognitive network is the multimodal nidopallium caudolaterale (NCL) that is considered to be analogous to the mammalian prefrontal cortex (PFC). The NCL plays a key role in a multitude of executive tasks such as working memory, decision-making during navigation, and extinction learning in complex learning environments. Like the PFC, the NCL is positioned at the transition from ascending sensory to descending motor systems. For the latter, it sends descending premotor projections to the intermediate arcopallium (AI) and the medial striatum (MSt). To gain detailed insight into the organization of these projections, we conducted several retrograde and anterograde tracing experiments. First, we tested whether NCL neurons projecting to AI (NCLarco neurons) and MSt (NCLMSt neurons) are constituted by a single neuronal population with bifurcating neurons, or whether they form two distinct populations. Here, we found two distinct projection patterns to both target areas that were associated with different morphologies. Second, we revealed a weak topographic projection toward the medial and lateral striatum and a strong topographic projection toward AI with clearly distinguishable sensory termination fields. Third, we investigated the relationship between the descending NCL pathways to the arcopallium with those from the hyperpallium apicale, which harbors a second major descending pathway of the avian pallium. We embed our findings within a system of parallel pallio-motor loops that carry information from separate sensory modalities to different subpallial systems. Our results also provide insights into the evolution of the avian motor system from which, possibly, the song system has emerged.

Input and Output Connections of the Crow Nidopallium Caudolaterale

The avian telencephalic structure nidopallium caudolaterale (NCL) functions as an analog to the mammalian prefrontal cortex. In crows, corvid songbirds, it plays a crucial role in higher cognitive and executive functions. These functions rely on the NCL's extensive telencephalic connections. However, systematic investigations into the brain-wide connectivity of the NCL in crows or other songbirds are lacking. Here, we studied its input and output connections by injecting retrograde and anterograde tracers into the carrion crow NCL. Our results, mapped onto a published carrion crow brain atlas, confirm NCL multisensory connections and extend prior pigeon findings by identifying a novel input from the hippocampal formation. Furthermore, we analyze crow NCL efferent projections to the arcopallium and report newly identified arcopallial neurons projecting bilaterally to the NCL. These findings help to clarify the role of the NCL as central executive hub in the corvid songbird brain.

Uncoordinated sleep replay across hemispheres in the zebra finch

Bilaterally organized brain regions are often simultaneously active in both humans1,2,3 and animal models,4,5,6,7,8,9 but the extent to which the temporal progression of internally generated dynamics is coordinated across hemispheres and how this coordination changes with brain state remain poorly understood. To address these issues, we investigated the zebra finch courtship song (duration: 0.5-1.0 s), a highly stereotyped complex behavior10,11 produced by a set of bilaterally organized nuclei.12,13,14 Unilateral lesions to these structures can eliminate or degrade singing,13,15,16,17 indicating that both hemispheres are required for song production.18 Additionally, previous work demonstrated broadly coherent and symmetric bilateral premotor signals during song.9 To precisely track the temporal evolution of activity in each hemisphere, we recorded bilaterally in the song production pathway. We targeted the robust nucleus of the arcopallium (RA) in the zebra finch, where population activity reflects the moment-to-moment progression of the courtship song during awake vocalizations19,20,21,22,23,24 and sleep, where song-related network dynamics reemerge in "replay" events.24,25 We found that activity in the left and right RA is synchronized within a fraction of a millisecond throughout song. In stark contrast, the two hemispheres displayed largely independent replay activity during sleep, despite shared interhemispheric arousal levels. These findings demonstrate that the degree of bilateral coordination in the zebra finch song system is dynamically modulated by behavioral state.

Multi-channel recordings reveal age-related differences in the sleep of juvenile and adult zebra finches

Despite their phylogenetic differences and distinct pallial structures, mammals and birds show similar electroencephalography (EEG) traces during sleep, consisting of distinct rapid eye movement (REM) sleep and slow wave sleep (SWS) stages. Studies in human and a limited number of other mammalian species show that this organization of sleep into interleaving stages undergoes radical changes during lifetime. Do these age-dependent variations in sleep patterns also occur in the avian brain? Does vocal learning have an effect on sleep patterns in birds? To answer these questions, we recorded multi-channel sleep EEG from juvenile and adult zebra finches for several nights. Whereas adults spent more time in SWS and REM sleep, juveniles spent more time in intermediate sleep (IS). The amount of IS was significantly larger in male juveniles engaged in vocal learning compared to female juveniles, which suggests that IS could be important for vocal learning. In addition, we observed that functional connectivity increased rapidly during maturation of young juveniles, and was stable or declined at older ages. Synchronous activity during sleep was larger for recording sites in the left hemisphere for both juveniles and adults, and generally intra-hemispheric synchrony was larger than inter-hemispheric synchrony during sleep. A graph theory analysis revealed that in adults, highly correlated EEG activity tended to be distributed across fewer networks that were spread across a wider area of the brain, whereas in juveniles, highly correlated EEG activity was distributed across more numerous, albeit smaller, networks in the brain. Overall, our results reveal that significant changes occur in the neural signatures of sleep during maturation in an avian brain.

Neural correlates of cognitively controlled vocalizations in a corvid songbird

The neuronal basis of the songbird's song system is well understood. However, little is known about the neuronal correlates of the executive control of songbird vocalizations. Here, we record single-unit activity from the pallial endbrain region "nidopallium caudolaterale" (NCL) of crows that vocalize to the presentation of a visual go-cue but refrain from vocalizing during trials without a go-cue. We find that the preparatory activity of single vocalization-correlated neurons, but also of the entire population of NCL neurons, before vocal onset predicts whether or not the crows will produce an instructed vocalization. Fluctuations in baseline neuronal activity prior to the go-cue influence the premotor activity of such vocalization-correlated neurons and seemingly bias the crows' decision to vocalize. Neuronal response modulation significantly differs between volitional and task-unrelated vocalizations. This suggests that the NCL can take control over the vocal motor network during the production of volitional vocalizations in a corvid songbird.

Shared mechanisms of auditory and non-auditory vocal learning in the songbird brain

Songbirds and humans share the ability to adaptively modify their vocalizations based on sensory feedback. Prior studies have focused primarily on the role that auditory feedback plays in shaping vocal output throughout life. In contrast, it is unclear how non-auditory information drives vocal plasticity. Here, we first used a reinforcement learning paradigm to establish that somatosensory feedback (cutaneous electrical stimulation) can drive vocal learning in adult songbirds. We then assessed the role of a songbird basal ganglia thalamocortical pathway critical to auditory vocal learning in this novel form of vocal plasticity. We found that both this circuit and its dopaminergic inputs are necessary for non-auditory vocal learning, demonstrating that this pathway is critical for guiding adaptive vocal changes based on both auditory and somatosensory signals. The ability of this circuit to use both auditory and somatosensory information to guide vocal learning may reflect a general principle for the neural systems that support vocal plasticity across species.

A brain atlas of the carrion crow (Corvus corone)

Corvidae, passerine songbirds such as jays, crows, and ravens known as corvids, have become model systems for the study of avian cognition. The superior cognitive capabilities of corvids mainly emerge from a disproportionally large telencephalon found in these species. However, a systematic mapping of the neuroanatomy of the corvid brain, and the telencephalon in particular, is lacking so far. Here, we present a brain atlas of the carrion crow, Corvus corone, with special emphasis on the telencephalic pallium. We applied four staining techniques to brain slices (Nissl, myelin, combination of Nissl and myelin, and tyrosine hydroxylase targeting catecholaminergic neurons). This allowed us to identify brain nuclei throughout the brain and delineate the known pallial subdivisions termed hyperpallium, entopallium, mesopallium, nidopallium, arcopallium, and hippocampal complex. The extent of these subdivisions and brain nuclei are described according to stereotaxic coordinates. In addition, 3D depictions of pallial regions were reconstructed from these slices. While the overall organization of the carrion crow's brain matches other songbird brains, the relative proportions and expansions of associative pallial areas differ considerably in agreement with enhanced cognitive skills found in corvids. The presented global organization of the crow brain in stereotaxic coordinates will help to guide future neurobiological studies in corvids.

Responses to Song Playback Differ in Sleeping versus Anesthetized Songbirds

Vocal learning in songbirds is mediated by a highly localized system of interconnected forebrain regions, including recurrent loops that traverse the cortex, basal ganglia, and thalamus. This brain-behavior system provides a powerful model for elucidating mechanisms of vocal learning, with implications for learning speech in human infants, as well as for advancing our understanding of skill learning in general. A long history of experiments in this area has tested neural responses to playback of different song stimuli in anesthetized birds at different stages of vocal development. These studies have demonstrated selectivity for different song types that provide neural signatures of learning. In contrast to the ease of obtaining responses to song playback in anesthetized birds, song-evoked responses in awake birds are greatly reduced or absent, indicating that behavioral state is an important determinant of neural responsivity. Song-evoked responses can be elicited during sleep as well as anesthesia, and the selectivity of responses to song playback in adult birds is highly similar between anesthetized and sleeping states, encouraging the idea that anesthesia and sleep are similar. In contrast to that idea, we report evidence that cortical responses to song playback in juvenile zebra finches (Taeniopygia guttata) differ greatly between sleep and urethane anesthesia. This finding indicates that behavioral states differ in sleep versus anesthesia and raises questions about relationships between developmental changes in sleep activity, selectivity for different song types, and the neural substrate for vocal learning.

Multidimensional Tuning in Motor Cortical Neurons during Active Behavior

A region within songbird cortex, dorsal intermediate arcopallium (AId), is functionally analogous to motor cortex in mammals and has been implicated in song learning during development. Non-vocal factors such as visual and social cues are known to mediate song learning and performance, yet previous chronic-recording studies of regions important for song behavior have focused exclusively on neural activity in relation to song production. Thus, we have little understanding of the range of non-vocal information that single neurons may encode. We made chronic recordings in AId of freely behaving juvenile zebra finches and evaluated neural activity during diverse motor behaviors throughout entire recording sessions, including song production as well as hopping, pecking, preening, fluff-ups, beak interactions, scratching, and stretching. These movements are part of natural behavioral repertoires and are important components of both song learning and courtship behavior. A large population of AId neurons showed significant modulation of activity during singing. In addition, single neurons demonstrated heterogeneous response patterns during multiple movements (including excitation during one movement type and suppression during another), and some neurons showed differential activity depending on the context in which movements occurred. Moreover, we found evidence of neurons that did not respond during discrete movements but were nonetheless modulated during active behavioral states compared with quiescence. Our results suggest that AId neurons process both vocal and non-vocal information, highlighting the importance of considering the variety of multimodal factors that can contribute to vocal motor learning during development.

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.

Interhemispheric functional connectivity in the zebra finch brain, absent the corpus callosum in normal ontogeny

Bilaterally symmetric intrinsic brain activity (homotopic functional connectivity; FC) is a fundamental feature of the mammalian brain's functional architecture. In mammals, homotopic FC is primarily mediated by the corpus callosum (CC), a large interhemispheric white matter tract thought to balance the bilateral coordination and hemispheric specialization critical for many complex brain functions, including human language. The CC first emerged with the Eutherian (placental) mammals ∼160 MYA and is not found among other vertebrates. Despite this, other vertebrates also exhibit complex brain functions requiring hemispheric specialization and coordination. For example, the zebra finch (Taeniopygia guttata) songbird learns to sing from tutors much as humans acquire speech and must balance hemispheric specialization and coordination to successfully learn and produce song. We therefore tested whether the zebra finch also exhibits homotopic FC, despite lacking the CC. Resting-state fMRI analyses demonstrated widespread homotopic FC throughout the zebra finch brain across development, including within a network required for learned song that lacks direct interhemispheric structural connectivity. The presence of homotopic FC in a non-Eutherian suggests that ancestral pathways, potentially including indirect connectivity via the anterior commissure, are sufficient for maintaining a homotopic functional architecture, an insight with broad implications for understanding interhemispheric coordination across phylogeny.

Molecular architecture of the zebra finch arcopallium

The arcopallium, a key avian forebrain region, receives inputs from numerous brain areas and is a major source of descending sensory and motor projections. While there is evidence of arcopallial subdivisions, the internal organization or the arcopallium is not well understood. The arcopallium is also considered the avian homologue of mammalian deep cortical layers and/or amygdalar subdivisions, but one-to-one correspondences are controversial. Here we present a molecular characterization of the arcopallium in the zebra finch, a passerine songbird species and a major model organism for vocal learning studies. Based on in situ hybridization for arcopallial-expressed transcripts (AQP1, C1QL3, CBLN2, CNTN4, CYP19A1, ESR1/2, FEZF2, MGP, NECAB2, PCP4, PVALB, SCN3B, SCUBE1, ZBTB20, and others) in comparison with cytoarchitectonic features, we have defined 20 distinct regions that can be grouped into six major domains (anterior, posterior, dorsal, ventral, medial, and intermediate arcopallium, respectively; AA, AP, AD, AV, AM, and AI). The data also help to establish the arcopallium as primarily pallial, support a unique topography of the arcopallium in passerines, highlight similarities between the vocal robust nucleus of the arcopallium (RA) and AI, and provide insights into the similarities and differences of cortical and amygdalar regions between birds and mammals. We also propose the use of AMV (instead of nucleus taenia/TnA), AMD, AD, and AI as initial steps toward a universal arcopallial nomenclature. Besides clarifying the internal organization of the arcopallium, the data provide a coherent basis for further functional and comparative studies of this complex avian brain region.

The expression of tyrosine hydroxylase and DARPP-32 in the house crow (Corvus splendens) brain

Birds of the family Corvidae which includes diverse species such as crows, rooks, ravens, magpies, jays, and jackdaws are known for their amazing abilities at problem-solving. Since the catecholaminergic system, especially the neurotransmitter dopamine, plays a role in cognition, we decided to study the distribution of tyrosine hydroxylase (TH), the rate-limiting enzyme in the synthesis of catecholamines in the brain of house crows (Corvus splendens). We also studied the expression of DARPP-32 (dopamine and cAMP-regulated phosphoprotein), which is expressed in dopaminoceptive neurons. Our results demonstrated that as in other avian species, the expression of both TH and DARPP-32 was highest in the house crow striatum. The caudolateral nidopallium (NCL, the avian analogue of the mammalian prefrontal cortex) could be differentiated from the surrounding pallial regions based on a larger number of TH-positive "baskets" of fibers around neurons in this region and greater intensity of DARPP-32 staining in the neuropil in this region. House crows also possessed distinct nuclei in their brains which corresponded to song control regions in other songbirds. Whereas immunoreactivity for TH was higher in the vocal control region Area X compared to the surrounding MSt (medial striatum) in house crows, staining in RA and HVC was not as prominent. Furthermore, the arcopallial song control regions RA (nucleus robustus arcopallialis) and AId (intermediate arcopallium) were strikingly negative for DARPP-32 staining, in contrast to the surrounding arcopallium. Patterns of immunoreactivity for TH and DARPP-32 in "limbic" areas such as the hippocampus, septum, and extended amygdala have also been described.

Differential developmental changes in cortical representations of auditory-vocal stimuli in songbirds

Procedural skill learning requires iterative comparisons between feedback of self-generated motor output and a goal sensorimotor pattern. In juvenile songbirds, neural representations of both self-generated behaviors (each bird's own immature song) and the goal motor pattern (each bird's adult tutor song) are essential for vocal learning, yet little is known about how these behaviorally relevant stimuli are encoded. We made extracellular recordings during song playback in anesthetized juvenile and adult zebra finches ( Taeniopygia guttata) in adjacent cortical regions RA (robust nucleus of the arcopallium), AId (dorsal intermediate arcopallium), and RA cup, each of which is well situated to integrate auditory-vocal information: RA is a motor cortical region that drives vocal output, AId is an adjoining cortical region whose projections converge with basal ganglia loops for song learning in the dorsal thalamus, and RA cup surrounds RA and receives inputs from primary and secondary auditory cortex. We found strong developmental differences in neural selectivity within RA, but not in AId or RA cup. Juvenile RA neurons were broadly responsive to multiple songs but preferred juvenile over adult vocal sounds; in addition, spiking responses lacked consistent temporal patterning. By adulthood, RA neurons responded most strongly to each bird's own song with precisely timed spiking activity. In contrast, we observed a complete lack of song responsivity in both juvenile and adult AId, even though this region receives song-responsive inputs. A surprisingly large proportion of sites in RA cup of both juveniles and adults did not respond to song playback, and responsive sites showed little evidence of song selectivity. NEW & NOTEWORTHY Motor skill learning entails changes in selectivity for behaviorally relevant stimuli across cortical regions, yet the neural representation of these stimuli remains understudied. We investigated how information important for vocal learning in zebra finches is represented in regions analogous to infragranular layers of motor and auditory cortices during vs. after the developmentally regulated learning period. The results provide insight into how neurons in higher level stages of cortical processing represent stimuli important for motor skill learning.

A pan-mammalian map of interhemispheric brain connections predates the evolution of the corpus callosum

The brain of mammals differs from that of all other vertebrates, in having a six-layered neocortex that is extensively interconnected within and between hemispheres. Interhemispheric connections are conveyed through the anterior commissure in egg-laying monotremes and marsupials, whereas eutherians evolved a separate commissural tract, the corpus callosum. Although the pattern of interhemispheric connectivity via the corpus callosum is broadly shared across eutherian species, it is not known whether this pattern arose as a consequence of callosal evolution or instead corresponds to a more ancient feature of mammalian brain organization. Here we show that, despite cortical axons using an ancestral commissural route, monotremes and marsupials share features of interhemispheric connectivity with eutherians that likely predate the origin of the corpus callosum. Based on ex vivo magnetic resonance imaging and tractography, we found that connections through the anterior commissure in both fat-tailed dunnarts (Marsupialia) and duck-billed platypus (Monotremata) are spatially segregated according to cortical area topography. Moreover, cell-resolution retrograde and anterograde interhemispheric circuit mapping in dunnarts revealed several features shared with callosal circuits of eutherians. These include the layered organization of commissural neurons and terminals, a broad map of connections between similar (homotopic) regions of each hemisphere, and regions connected to different areas (heterotopic), including hyperconnected hubs along the medial and lateral borders of the cortex, such as the cingulate/motor cortex and claustrum/insula. We therefore propose that an interhemispheric connectome originated in early mammalian ancestors, predating the evolution of the corpus callosum. Because these features have been conserved throughout mammalian evolution, they likely represent key aspects of neocortical organization.

Neuroplasticity in the cerebello-thalamo-basal ganglia pathway: A longitudinal in vivo MRI study in male songbirds

Similar to human speech, bird song is controlled by several pathways including a cortico-basal ganglia-thalamo-cortical (C-BG-T-C) loop. Neurotoxic disengagement of the basal ganglia component, i.e. Area X, induces long-term changes in song performance, while most of the lesioned area regenerates within the first months. Importantly however, the timing and spatial extent of structural neuroplastic events potentially affecting other constituents of the C-BG-T-C loop is not clear. We designed a longitudinal MRI study where changes in brain structure were evaluated relative to the time after neurotoxic lesioning or to vocal performance. By acquiring both Diffusion Tensor Imaging and 3-dimensional anatomical scans, we were able to track alterations in respectively intrinsic tissue properties and local volume. Voxel-based statistical analyses revealed structural remodeling remote to the lesion, i.e. in the thalamus and, surprisingly, the cerebellum, both peaking within the first two months after lesioning Area X. Voxel-wise correlations between song performance and MRI parameters uncovered intriguing brain-behavior relationships in several brain areas pertaining to the C-BG-T-C loop supervising vocal motor control. Our results clearly point to structural neuroplasticity in the cerebellum induced by basal ganglia (striatal) damage and might point to the existence of a human-like cerebello-thalamic-basal ganglia pathway capable of modifying vocal motor output.

Neural activity in cortico-basal ganglia circuits of juvenile songbirds encodes performance during goal-directed learning

Cortico-basal ganglia circuits are thought to mediate goal-directed learning by a process of outcome evaluation to gradually select appropriate motor actions. We investigated spiking activity in core and shell subregions of the cortical nucleus LMAN during development as juvenile zebra finches are actively engaged in evaluating feedback of self-generated behavior in relation to their memorized tutor song (the goal). Spiking patterns of single neurons in both core and shell subregions during singing correlated with acoustic similarity to tutor syllables, suggesting a process of outcome evaluation. Both core and shell neurons encoded tutor similarity via either increases or decreases in firing rate, although only shell neurons showed a significant association at the population level. Tutor similarity predicted firing rates most strongly during early stages of learning, and shell but not core neurons showed decreases in response variability across development, suggesting that the activity of shell neurons reflects the progression of learning.