Neurobiological adaptations supporting vocal plasticity have accumulated in the marine carnivores

Neurobiological adaptations supporting vocal plasticity have accumulated in the marine carnivores | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Biological Sciences - Article Neurobiological adaptations supporting vocal plasticity have accumulated in the marine carnivores Peter Cook, Andrew Rouse, Eva Sawyer, Karla Miller, Gregory Berns This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6222519/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 11 Mar, 2026 Read the published version in Science → Version 1 posted You are reading this latest preprint version Abstract The evolutionary neurobiology of mammalian vocal learning is poorly understood. Pinnipeds are among the most promising model clades for addressing this knowledge gap 1–4 . The whole clade has been adaptively endowed with exquisite volitional breathing control to which the monachinae seals add developmental call plasticity and the phocinae add the capability for formant and frequency modulation and potentially mimicry 5–10 . Until now, there were no comparative neurobiological data on vocal behavior in this clade. Here, using histology and ex vivo dMRI tractography, we provide strong first evidence for a phylogenetic spectrum of accumulative neural adaptations supporting aspects of volitional vocal control across pinniped species. Otariids and phocid seals, but not coyotes, showed robust direct cortical vocal motor connectivity to the brainstem nucleus ambiguus. This pathway may have evolved to facilitate volitional breathing, submerged prey consumption, and underwater call production for an amphibious lifestyle in a basal pinniped prior to exaptation for vocal production learning in select clades. Phocid seals, but not otariids, showed a robust arcuate-like auditory-premotor cortical pathway potentially related to developmental call learning. Harbor seals (branch phocinae ) showed hypertrophic connectivity in the pathway between anterior ventrolateral thalamus and vocal premotor cortex, one part of the striatal-thalamocortical anterior forebrain circuit related to vocal motor reward learning in birds 11,12 . Thalamic-premotor connectivity is specifically implicated in human language production and vocal mimicry in parrots 13,14 . Biological sciences/Neuroscience/Motor control Biological sciences/Neuroscience/Neural circuits Biological sciences/Neuroscience/Sensorimotor processing Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Vocal behavior in the vast majority of animal species has been considered inflexible or obligate 1 , and is governed by interactions between an intricate network of midbrain and brainstem nuclei 15–17 that predominantly respond to environmental and affective conditions. The ability to vocalize under flexible, volitional control is distributed patchily among clades 18 . Non-obligate vocal behavior is typically attributed to vocal learning, but there are multiple dissociable cognitive and physiological mechanisms involved 3 . Vocal learning is increasingly referred to as a “spectrum” of interrelated capabilities. A growing number of species has been demonstrated to deploy species-typical calls in novel contexts 2 . Truly rare are species whose members can learn to produce novel calls with frequency and formant characteristics outside of the inherited repertoire. These “vocal production learners” are most common in avian lineages, where the behavioral and neurobiological mechanisms of vocal learning have been extensively studied in both field and laboratory 19 . Such studies have indicated that the ability to alter both filter (formants, via the mouth) and source phonation (frequency, via the syrinx) components of species-typical calls relies on elaboration or doubling of pathways in two forebrain circuits. The “anterior” circuit connects vocal premotor regions with anterior striatum and anterior ventrolateral thalamus 20 , and allows reinforcement learning to alter vocal motor production. A “posterior” circuit directly connects vocal motor cortex with brainstem phonatory neurons, allowing volitional control of the syrinx and vocal tract. The study of vocal learning in birds has exemplified the benefit of a “model clade” 21 approach to comparative behavioral neuroscience. Because scientists have been able to study closely related bird species that vary in systematic ways in terms of vocal flexibility (e.g., oscine and suboscine song birds and parrots), the neurobiological underpinnings of different aspects of call production and learning in birds have been elucidated. Heretofore we have lacked a comparable model clade in the mammalian order. Comparative Studies of Vocal Learning in Humans and Other Mammals Humans are the best studied vocally flexible mammal. They demonstrate high degrees of vocal plasticity and associated nervous system adaptations 22,23 , including developmental auditory-motor learning; contextual deployment of calls; and flexible breathing, phonation, and filter control throughout the lifespan, most fully exemplified by highly articulated vocal mimicry. Comparisons of human and avian vocal neurobiology have been productive, and have led to a number of hypotheses about neurobiological adaptations supporting vocal plasticity and learning 24,25 . However, the wide evolutionary gulf between the mammal and avian nervous systems and respective sound production mechanisms underscores the need for comparisons within the mammalian lineage. Behavioral vocal flexibility is not systematically distributed across non-human primate species, most of which show vocal stereotypy across measurable dimensions 26,27 , and human vocal behavior differs from that of other primates across too many dimensions to isolate specific neurobiological differences directly supporting different behavioral capabilities. In comparison to other primates, humans show enhanced connectivity in circuits analogous to the anterior and posterior vocal learning pathways in birds 24,25,28 . Striatal thalamocortical connectivity is implicated in language learning and use, but has not been systematically assessed across primates. Monosynaptic connections between laryngeal motor cortex and the nucleus ambiguus, theoretically supporting fine-grained volitional control of phonation, appear to be unique to humans among primates 29 . Accordingly, the influential “Kuyper-Jürgens hypothesis” 30 suggests the presence of direct connection between vocal motor cortex and brainstem phonatory nuclei such as the nucleus ambiguus is a precondition for flexible vocal control in any species. In addition, although it has been deemphasized in studies of bird vocal learning, there is extensive evidence that, in humans, auditory-vocal premotor pathways (e.g., the arcuate fasciculus connecting Wernicke’s and Broca’s areas) support language learning 31 . These pathways are less robust in non-human primates 32 . Mammalian comparisons would be most productive within clades demonstrating graded variability in production flexibility, which would elucidate the potential role of striatal thalamocortical and auditory-premotor adaptations in vocal learning, and would allow direct tests of the Kuyper-Jürgens hypothesis. Cetaceans are believed to be highly vocally flexible 33 , but are highly derived anatomically and logistically difficult to study. Bats have shown promise as a model clade, and are accessible to laboratory behavioral neuroscience 34,35 . However, it appears that even the most vocally flexible bat has limited volitional phonatory and filter control throughout the lifespan 36 . The Behavioral Case for Graded Vocal Flexibility in Pinnipeds There is one extant mammalian clade with clear behavioral evidence of differentially distributed components of vocal flexibility ranging from onset-offset control of stereotypical calls through developmental learning and, potentially, mimicry: the pinnipeds 7,37 . As such, they are a promising model clade for exploring the full spectrum of vocal learning adaptations in the mammalian brain. Pinnipeds are a lineage of marine-adapted Carnivores comprising odobenids (walruses), otariids (eared seals), and phocids (true seals), that have diverged and radiated over the last 25 million years from an amphibious canid carnivore ancestor 38 . In comparison to terrestrial carnivores, pinnipeds have experienced strong adaptive pressures to control breathing, swallowing, and sound production for a semi-aquatic lifestyle. Carnivores, including pinnipeds’ closest mustelid relatives, have not been found to express high vocal flexibility 1 . In strong contrast, all extant pinnipeds show some evidence of enhanced vocal flexibility, if only in onset/offset control 7 . Extensive field work shows that otariids, the fur seals and sea lions, produce highly stereotypic calls with no obvious ability to learn filter and phonation modulation 39 . However, there is strong experimental evidence that otariids easily learn to produce and inhibit species-typical calls under contextual control 40 , an ability that is difficult for many terrestrial species. Onset/offset control has been linked with cingulate-midbrain periaqueductal grey (PAG) connectivity 2 , but has not been explicitly linked to monosynaptic connections between laryngeal motor cortex and brainstem phonatory nuclei or gross adaptations to striatal thalamocortical vocal reinforcement learning circuits. The otariidae and phocidae (true seals) branches of pinnipeds split at least 24 million years ago, and unlike otariids, a number of phocid species including both monachinae (including elephant and monk seals) and phocinae (including harbor and grey seals) show some evidence of vocal learning over development 6,7,8,41,42 . This type of learning involves developmental modulation of phonation based on auditory input, and, in birds, requires cortical-brainstem phonatory connections and striatal thalamocortical adaptations 43 . In humans, developmental vocal learning also relies on an adapted circuit connecting auditory with vocal premotor regions (including the arcuate, mentioned above). At the far end of the pinniped flexibility spectrum is the phocinae branch of the phocid family, which split from the monachinae over 15 mya. Strikingly, some phocids show evidence of highly flexible vocal learning and mimicry, which may not be fully constrained to a strict developmental period. Evidence of mimicry has been collected in two harbor seals raised in captivity 44,45 , who learned to replicate human words and phrases. An older harbor seal in captivity learned to produce novel formants through operant techniques 10 . More recently, in an experimental context, grey seals learned to mimic novel tonal sequences and vowel-like formants 46 . This behavior requires fine-grained volitional control of the parts of the vocal apparatus responsible for filter (mouth, formants) and phonation (larynx, frequency) control. The neurobiology of mimicry is not fully understood, but in parrots seems to be related to further elaboration of thalamocortical circuits beyond the adaptations to the anterior pathway found in song birds 13 . Finally, among the pinnipeds, odobenids (walruses) are less widely studied, but have produced some of the most compelling experimental evidence of vocal flexibility 47,48 . Across the pinnipeds this range of capabilities, in conjunction with comparisons to terrestrial carnivore cousins, spans the spectrum from relatively fixed vocal stereotypy through mimicry in a set of closely related species with established phylogenetic branch points. Determining covariance between neurobiological and behavioral traits across the pinnipeds will thus yield insight into how changes to mammalian brains support vocal flexibility and learning, and will support new neurobehavioral hypotheses regarding the underpinnings of human language. In the current study, we directly assessed neural pathways potentially related to vocal plasticity and learning in several pinniped species and a terrestrial carnivore and compared these to available behavioral vocal data for these species. This allowed us to assemble a preliminary neurobehavioral phylogeny of vocal learning in the canid carnivore line. The advent of high signal-to-noise-ratio, high-resolution post-mortem diffusion MRI and tractography 49-51 provides a new opportunity for matching neurobiological adaptations to behavioral capabilities in previously inaccessible species. We used a novel post-mortem-optimized imaging sequence to scan opportunistically obtained brains of four California sea lions ( Zalophus californianus, otariidae ), three northern elephant seals ( Mirounga angustirostris, monachinae ), four harbor seals ( Phoca vitulina, phocinae ), and four coyotes ( Canis latrans, terrestrial canids ) (Subject details in Extended Data Table 1). We then conducted multi-region probabilistic tractography in right and left hemisphere of each subject of the following pathways (Extended Data Table 2). First, we examined nucleus ambiguus to vocal motor cortex (VMC[a] ) to test the Kuyper-Jürgens hypothesis. Next, we examined two-way connections between the anterior ventrolateral thalamus, anterior striatum, and pre-VMC to assess the role of the anterior striatal thalamocortical circuit in mammalian vocal production flexibility. To look for an arcuate-like connection potentially related to developmental vocal learning, we examined auditory cortex to pre-VMC. Finally, we assessed anterior cingulate to periaqueductal grey (PAG), a connection related to volitional deployment of species-typical calls in primates 2 , To establish seeds and regions for probabilistic tracing, we manually segmented all regions (25 per subject, 375 total, Figure 1, Extended Data Figure 1) using multiple methodologies: brainstem histology (nucleus ambiguus), prior cortical electrophysiological data (motor cortex/VMC/premotor cortex, A1 in coyotes), anatomical criteria (PAG, caudal colliculi, medial geniculate nucleus, anterior ventrolateral thalamus, and anterior cingulate), and tract tracing (A1 in pinnipeds, based on cortical projection from MGN, and pre VMC via tract tracing from VMC (Extended Data Table 3) [a] n.b, the mammalian vocal learning neurobiology literature typically refers to “laryngeal motor cortex,” but we broadly targeted brain regions involved in volitional motor control of the neck and face that would contribute to vocal motor control. Results We found clear evidence of brain circuit differences corresponding to differential behavioral evidence for vocal flexibility in each species. There was robust, bilateral, direct cortical connectivity between the putative vocal motor cortex and the nucleus ambiguus in the brainstem in all pinnipeds and in none of the coyotes (Figure 2). Importantly, there were no significant differences between the coyotes and any of the other species in the two control tracts, the whole-brainstem and corticospinal pathway (Extended Data Figure 2). Elephant seals and harbor seals, but not sea lions, showed robust arcuate-like connectivity between putative auditory cortex and vocal premotor cortex. The harbor seal brains showed exceptionally robust anterior cingulate to periaqueductal grey connectivity in comparison to all other brains evaluated. Finally, harbor seals showed enhanced connectivity between the anterior ventrolateral thalamus and pre-VMC, one pathway in the putative forebrain anterior vocal learning circuit (Figure 3, Extended Data Figure 3). Discussion This is the first evidence of a neurobehavioral phylogeny of vocal flexibility in closely related mammalian species with highly variable vocal capability, and it validates the pinnipeds as a model clade for studying the neurobiology of vocal production and learning (Figure 4). A parsimonious interpretation of our findings suggests that an early ancestral pinniped evolved a direct cortical pathway between vocal motor cortex and phonatory brainstem nuclei. This preceded adaptations supporting vocal production learning that emerged selectively in other pinniped branches, but did not, by itself, allow for high degrees of vocal plasticity. Enhanced auditory-motor cortical pathways supporting developmental vocal learning may have emerged in phocidae after the split from otariidae > 24 mya. Robust connections between vocal premotor cortex and anterior ventrolateral thalamus as well as enhanced anterior cingulate to PAG connections, both potentially related to enhanced vocal flexibility throughout the lifespan, may have emerged in phocinae after the split with monachinae . Our findings support the Kuyper-Jürgens hypothesis 30 that direct synaptic connections between cortical vocal regions and brainstem phonatory neurons are an evolutionary precondition for vocal production flexibility. They also emphasize the importance of direct cortical auditory-motor pathways to vocal learning in mammals. Our results do not as clearly implicate anterior striatal thalamocortical circuitry in mammalian vocal plasticity, but do underline the potential contribution of the anterior ventrolateral thalamus-pre-VMC pathway to vocal mimicry. Given categorical differences in these tracts between species, our results are cautionary for over-broad claims regarding vocal production learning across all pinnipeds 4 . Neurobiological and behavioral data should be acquired on a broader range of pinnipeds to examine evolutionary addition and potentially subtraction of these traits. Vocal Motor Cortex to Nucleus Ambiguus Pathway All pinnipeds tested showed robust direct connections between cortical vocal motor regions and nucleus ambiguus without transit through the PAG, which interposes between these regions in vocally inflexible species. This pathway was not found in any of the coyotes, despite robust primary motor corticospinal connectivity through the brainstem in all brains. The cortical-ambiguus pathway we found in the pinnipeds matches the putative posterior forebrain circuit for vocal learning proposed by Jarvis 11 . Behaviorally, compared to terrestrial carnivores, pinnipeds do show increased vocal motor control. However, there are no data of vocal production learning (filter/phonation control) in the otariidae . While the cortical-ambiguus pathway in the current study was most robust in harbor seals, all four sea lions in this study had clear bilateral cortical-nucleus ambiguus connections. Nucleus ambiguus is part of a network of medullary nuclei controlling throat, mouth, and chest musculature involved in breathing and swallowing 54,55 . In most terrestrial mammals, breathing is obligate based on peripheral physiological signals like blood-gas concentration and lung tension 56 . When adapting for amphibious life, pinnipeds would have faced intense pressure to develop breath control during submersion. In addition, in most terrestrial mammals studied, swallowing is controlled by central pattern generators in the brainstem that interface with medullary motor neurons controlling the throat and pharynx, and is wholly obligate and automatic once triggered 57 . Early pinnipeds must have evolved mechanisms to avoid ingesting sea water when consuming prey underwater. We suggest that the apparent vocal motor cortical-ambiguus pathway we identified here evolved primarily to bring pinniped breathing, and potentially swallowing, under volitional control (as suggested in Ravignani et al., 2016 37 ). This would, in turn, make control of laryngeal muscles and thoracic muscles involved in phonation subject to broadly conserved corticostriatal and corticocerebellar general motor learning mechanisms. The integration of breathing and swallowing motor control with subcortical learning mechanisms may have paved the way for subsequent vocal adaptations. However, in the otariids, this appears to be limited to context learning, not alteration of frequency and formants. Our current findings suggest that direct cortical control of medullary phonatory nuclei may be necessary, but not sufficient, for vocal production learning. Subsequent adaptations may be required to drive learning to alter vocal behavior based on auditory input. This is in line with theories suggesting that, in humans, vocal learning emerged subsequently to adaptations for breathing and swallowing control to avoid choking/food inhalation 45,58,59,60 . To disentangle breathing and vocal control, future studies should assess voluntary control of breathing with and without vocalization in otariids and terrestrial carnivores. Auditory Cortex to Vocal Premotor Pathway Our findings further implicate arcuate-like auditory to premotor cortical pathways in vocal production learning in the pinnipeds. Phocids in both the monachinae and phocinae line, but not otariids, show some evidence of vocal learning and flexibility over development. All phocids in the current study showed clear, direct auditory to premotor connectivity, while the sea lions did not. Auditory to motor connections in birds have been observed in vocal learners and non-learners, and so have not been the focus of neurobiological assessment of vocal flexibility in avians 11 , although see Roberts et al. (2017) 61 . These pathways do show evolutionary variability in the primates 32 . Further, when the arcuate is severed or disrupted in humans, the ability to learn to alter vocal output based on auditory input is greatly reduced, or potentially removed entirely. These longitudinal white matter tracts have been shown to be essential to developmental learning of speech 62 , both through integration of the auditory perception of conspecific vocalization with the motor pattern for one’s own vocal output and for refinement of one’s own vocal output based on the auditory feedback it produces. The presence of an arcuate-like pathway in all phocids but none of the otariids in the present study underlines its potential broad relevance to mammalian vocal plasticity and learning. Despite some evidence that otariid male calls may be distinct between colonies 63 , there is no evidence in these species of complex vocal production learning. The auditory-motor circuit described in phocids in the current study may have emerged or been strengthened in an early phocid to support ecologically relevant developmental vocal learning. Notably, most phocids rely far more heavily on vocal signaling in breeding than do otariids, suggesting parallel evolutionary pressures to those experienced by songbird males 64 . Elephant seals, a phocid species that mate on land, potentially rely on learned vocal cues for signaling to and tracking rival males during breeding season 6,65 . Most phocids, including harbor seals, breed in the water, where calls may serve male-male competition or potentially signaling to females 66,67 . These seals tend to have larger and more complex vocal repertoires that are produced seasonally 68 . While evidence of phocid developmental vocal learning from field studies is rare 8 , there is a small but growing body of laboratory studies showing some degree of developmental vocal plasticity in phocid species 42,69 . Some have even gone so far as to dub harbor seal vocalizations analogous to bird or whale song 70 . However, it must be emphasized that in the wild harbor seal vocalizations are among the most stereotypic of the phocids—the case for song awaits further data, and may be stronger in other phocids and walruses 68,71 . More data from the field and more controlled laboratory studies are required to fully characterize phocid vocal repertoires and determine the potential influence of developmental learning on call characteristics. It will be particularly important to rigorously quantify dimensions of vocal plasticity in captive animals as well as to determine when and where phocids are exposed to adult calls, and how these potential exposures relate to any subsequent vocal plasticity in the listeners. Because most phocinae pups stay on the beach to nurse during the time that adult males are producing breeding calls in offshore water, it is not clear when they might be exposed to auditory models related to call learning. In addition, there is some evidence to suggest that some seals raised in isolation from conspecifics produce species-typical calls as adults 41 . Despite these complexities, our current finding of an arcuate-like pathway in all phocids, but not in sea lions or reliably in coyotes, supports an argument for breeding-related vocal plasticity in the phocids. In anecdotal support of this interpretation, the arcuate-like pathway we described was notably stronger bilaterally in the two male harbor seal brains in comparison to the two female harbor seal brains, mirroring findings from seminal bird studies 72 . In addition, all phocid subjects in the current study were juveniles, suggesting that the auditory premotor pathway is present early in brain development, potentially supporting learning from very early auditory exposures neonatally or even in utero. As male seals are far more vocal than females, future work should seek to compare auditory-premotor connections in male and female brains from a broader range of phocid species, and across a wider range of ages. Pre- and post-pubertal comparisons are of particular interest. Our findings also suggest future comparisons between terrestrial canids—only two of our four coyotes showed notable auditory-premotor connectivity, and it was right lateralized. New domestic dog data indicate a robust left-lateralized auditory-premotor pathway 73 . Anterior Striatal Thalamocortical Circuit Despite its centrality in neurobiological accounts of vocal learning in birds, there were limited interspecies differences across the anterior striatal thalamocortical circuit in the present study. The cortical-brainstem phonatory and auditory-premotor pathways may be more relevant to vocal plasticity in the pinniped line. Notably, harbor seals did show increased connectivity between the anterior ventrolateral thalamus and pre-VMC. The ventrolateral thalamus in humans, but not other primates, is essential for vocal production 14 , serving as a hub connecting striatal and cortical regions involved in vocal release and inhibition, learning, and vocal working memory. Thalamic-pre motor connections facilitate transition between call sub-components in finches 74 , and elaboration of these circuits in parrots appears to be related to their capability for vocal mimicry throughout the lifespan 13 . Given this, and the current findings, the anterior ventrolateral thalamus-pre-VMC pathway may play a particular role in putative vocal mimicry in the phocinae line. The phocinae branch of phocids are the only pinnipeds so far demonstrated capable of mimicry, shown anecdotally in harbor seals and experimentally in grey seals. Importantly, these animals have never been observed to produce vocal mimicry in the wild, where harbor seal calls are fairly stereotypic 68 . More work is required to chart their abilities over the course of development. Intriguingly, despite the anecdotal and laboratory evidence indicating the importance of developmental exposure to phocinae mimicry, an adult male harbor seal was also able to learn to modulate his vocal motor output to produce novel formants, and without an external auditory model 10 . In birds, the thalamus to pre-VMC pathway relates to learned control of phonatory musculature, not exclusively to vocal learning based on auditory input and models 75 . Due to their benthic feeding and suction feeding and nursing strategies, phocid seals may have enhanced mouth motor control in comparison to otariids and terrestrial carnivores. As suggested by Schusterman (2008) 40 , this control might also serve fine-grained vocal production flexibility, with or without auditory models. Future behavioral research should seek to examine dimensions of vocal plasticity in phocids at different developmental time points, with and without auditory models. Anterior Cingulate to PAG Pathway In primates, stronger direct connections between the anterior cingulate and midbrain PAG are linked with increased vocal context learning—the ability not to produce novel vocalizations, but to learn to vocalize or not in specific situations and in response to specific cues 2 . Given the need to coordinate vocal behavior above and beneath the surface of the water, one might expect this pathway to show increased strength in pinnipeds compared to terrestrial carnivores. We did not find this to be the case in the current study; coyotes, sea lions, and elephant seals were indistinguishable in terms of path strength. Harbor seals, however, showed extremely robust anterior-cingulate to PAG connectivity. The lack of adaptation in this pathway in the sea lions and elephant seals, despite their demonstrated ability to acquire contextual control over vocal behavior, suggests that other neural circuits serve to control and organize phonation above and beneath the water—perhaps the direct cortical-brainstem phonatory pathway lacking in the coyotes. The role of this pathway in harbor seals remains to be elucidated. Limitations Our study’s limitations were primarily due to sample size–more brains should be collected and analyzed, balanced for age and sex. This is a slow, arduous, necessarily collaborative, and potentially expensive process with marine mammals. Lack of common neuroanatomical atlases for these species and limited data on functional division of the cortex are further limitations–these were overcome in the current study through painstaking sample-by-sample manual segmentation based on multiple lines of prior evidence and cross-validation. More baseline work on pinniped neuroanatomy would be useful. Examination of putative anterior and posterior forebrain vocal learning pathways in the walrus should also be undertaken, as they represent a distinct evolutionary branch of the pinnipeds, and both laboratory and field studies suggest they have substantial vocal flexibility 47,48 . On the technical side, probabilistic tractography is not as reliable as chemical tracing for determining brain connectivity. However, it has been repeatedly validated, and our specialized post-mortem technique yields particularly high SNR and high resolution, without the detriments of biological noise from in vivo scanning. In addition, the primary limitation of current diffusion approaches is resolving crossing fibers–our methodology did not depend on digitally dissecting potentially crossing tracts but rather searching for direct pathways between delineated structures. In terms of those structures, vocal motor cortex and nucleus ambiguus were particularly central to the current study. We confidently identified and validated VMC via prior electrophysiology and histology studies cross-referenced to anatomical landmarks. However, it is important to emphasize that our cortical motor regions are likely involved with volitional control of the throat and face—without invasive work or in vivo imaging in awake animals, we cannot be certain which, if any, sub-components of this system are specific to vocalization. For localization of the nucleus ambiguus, we were able to rely on species-specific histology in sea lions, elephant seals, and harbor seals, but had to use domestic dog data to identify this structure in coyotes. Importantly, we found no evidence of VMC projections in the vicinity of putative nucleus ambiguus in coyotes (Extended Data Figure 9). In addition, we were not able to stain and image brains from the same individuals in each species. While we cross-validated all regions used in this study in multiple ways (see method and Extended Data Table 3), future work could use histology-to-imaging registration 76 , ideally within the same samples. This will require very high-resolution structural imaging to achieve necessary contrast in the brainstem. The nucleus ambiguus has been considered particularly difficult to find in MR datasets. Our localization was based on clear, species-specific histology, and was further validated by connectivity–when traced separately, each putative pinniped nucleus ambiguus reached frontal cortex in the vicinity of putative VMC. In addition, all nucleus ambiguus regions, bilaterally, in all subjects, had robust PAG connectivity, bilateral connectivity with the other putative nucleus ambiguus, and lateral connectivity to putative vagal nerve root 55 . Finally, although nucleus ambiguus is in the vicinity of the reticular formation, which also has frontal projections, those transit the thalamus, which was not along the primary pathway to frontal cortex for any of the pinnipeds in the current study (Extended Data Figure 4). Conclusion In providing the first mammalian cladistic framework for assessing specific neurobiological adaptations supporting different aspects of vocal flexibility, our study validates and qualifies emerging behavioral evidence of vocal plasticity and learning in phocids, and opens up horizons for new research. One particularly pressing question: given strong evidence for vocal mimicry, and now, neural adaptations supporting vocal plasticity, why is harbor seal vocal behavior in the wild so apparently stereotypic? In addition to the already mentioned need to better characterize vocal and breathing behavior and control in the laboratory and the field across phocid species, it would be informative to examine whether otariids can be trained to express aspects of vocal flexibility in captivity. Their cortical-brainstem phonatory pathway might theoretically support frequency plasticity. In addition, grey seals, a vocally flexible phocid relative of harbor seals have been shown to express FOXP2 in cortex 77 , a gene linked with motor learning and vocal learning specifically 78 . Given our current findings, it would be particularly interesting to look for this gene in elephant seals, harbor seals, and sea lions. Otariid vocal plasticity directly relates to another comparative research literature featuring pinnipeds and other marine mammals: rhythm 79 . Much has been made of the idea that adaptations supporting vocal learning also allow flexible sensorimotor synchrony 80,81 . However, a California sea lion is currently the most reliable, precise, and flexible non-human beat keeper 82,83 . In the current study, sea lions appear to lack the auditory-premotor adaptations of the other pinnipeds, and don’t have clear elaboration of the striatal thalamocortical circuit. Does the cortical vocal motor-brainstem phonatory pathway we described support flexible sensorimotor synchrony, or is synchrony largely unrelated to neural adaptations for vocal learning? Our findings also pave the way for parallel neurobiological studies in other species. If the correct brainstem nuclei can be precisely segmented, we could replicate our approach in toothed and baleen whales, both of which demonstrate vocal flexibility and physiological adaptations related to sound production 84,85 . Circuit-specific comparisons to bats, which, as mentioned previously seem to show more limited vocal plasticity, would also be informative. Finally, although our findings generally support the Kuyper-Jürgens hypothesis of the necessity of direct cortical connectivity between cortical vocal motor regions and brainstem phonatory nuclei, they also suggest that this pathway may not confer high degrees of vocal plasticity without other parallel adaptations. Given elaboration of the anterior ventrolateral thalamus to vocal motor cortex pathway in harbor seals and parrots, more work should examine development of and dynamic interactions between these regions in relation to parrot mimicry and human speech behavior. In addition, if the distinct neural circuits we examined in the current study evolved separately to subserve different aspects of vocal plasticity, then they may also have done so in early hominids before coming together into the multicomponent, integrated language learning system of extant humans 86 . Declarations Acknowledgements We thank the volunteers and staff at The Marine Mammal Center in Sausalito, CA and, the USDA Wildlife Research Center in Logan, Utah for help in obtaining samples. Colleagues at the Yerkes Primate Center in Atlanta, GA assisted with coyote brain fixation and extraction. Orion Keifer at Emory University assisted with methodology for setting brains for imaging. We also thank numerous colleagues who have provided preliminary feedback on this manuscript, most notably C. Casey and P. Tyack. A special thanks to C. Reichmuth for detailed feedback on interpretation of vocal behavior data in pinnipeds. Pinniped tissues were obtained under NMFS permit 18786. Funding for P.F. Cook was provided by Human Frontiers Science Program grant RGP0019. Author Contributions P.F.C. and G.B. conceived of the study and P.F.C., G.B., and E.S. collected data. The imaging method was developed by K.M. P.F.C. and A.R. developed and implemented the criteria and methodology for region of interest definition and segmentation. P.F.C. and A.R. analyzed the data and created figures. P.F.C. wrote the manuscript with input from all other authors. Data Availability Structural and dMRI data for all subjects are featured on University of Oxford’s Digital Brain Zoo, and are available by request: https://open.win.ox.ac.uk/DigitalBrainBank/#/datasets/zoo ROIs used in this study were uploaded for submission to Nature’s figshare service, and are available for reviewers. References Janik, V. M. & Slater, P. J. B. Vocal Learning in Mammals. in Advances in the Study of Behavior vol. 26 59–99 (Elsevier, 1997). Tyack, P. L. A taxonomy for vocal learning. Trans. R. Soc. B Biol. Sci. 375 , 20180406 (2020). Vernes, S. C. et al. The multi-dimensional nature of vocal learning. Trans. R. Soc. B Biol. Sci. 376 , 20200236 (2021). Wirthlin, M. E. et al. Vocal learning–associated convergent evolution in mammalian proteins and regulatory elements. Science 383 , eabn3263 (2024). Collins, K. T., Terhune, J. M., Rogers, T. L., Wheatley, K. E. & Harcourt, R. G. Vocal individuality of in-air Weddell seal ( Leptonychotes weddellii ) pup “primary” calls. Mammal Sci. 22 , 933–951 (2006). Sanvito, S., Galimberti, F. & Miller, E. H. Observational Evidences of Vocal Learning in Southern Elephant Seals: a Longitudinal Study. Ethology 113 , 137–146 (2007). Reichmuth, C. J. & Casey, C. Vocal learning in seals, sea lions, and walruses. Opin. Neurobiol. 28C , 66–71 (2014). Stansbury, A. L. & Janik, V. M. The role of vocal learning in call acquisition of wild grey seal pups. Trans. R. Soc. B Biol. Sci. 376 , 20200251 (2021). Duengen, D., Fitch, W. T. & Ravignani, A. Hoover the talking seal. Biol. 33 , R50–R52 (2023). Goncharova, M., Jadoul, Y., Reichmuth, C., Fitch, W. T. & Ravignani, A. Vocal tract dynamics shape the formant structure of conditioned vocalizations in a harbor seal. N. Y. Acad. Sci. 1538 , 107–116 (2024). Jarvis, E. D. Evolution of vocal learning and spoken language. Science 366 , 50–54 (2019). Jarvis, E. D. et al. Behaviourally driven gene expression reveals song nuclei in hummingbird brain. Nature 406 , 628–632 (2000). Zhao, Z. et al. Anterior forebrain pathway in parrots is necessary for producing learned vocalizations with individual signatures. Biol. 33 , 5415-5426.e4 (2023). Jürgens, U. Neuronal Control of Vocal Production in Non-Human and Human Primates. in Current Topics in Primate Vocal Communication (eds. Zimmermann, E., Newman, J. D. & Jürgens, U.) 199–206 (Springer US, Boston, MA, 1995). doi:10.1007/978-1-4757-9930-9_10. Jürgens, U. The Neural Control of Vocalization in Mammals: A Review. Voice 23 , 1–10 (2009). Briefer, E. F. Vocal expression of emotions in mammals: mechanisms of production and evidence. Zool. 288 , 1–20 (2012). Park, J. et al. Brainstem control of vocalization and its coordination with respiration. Science 383 , eadi8081 (2024). Janik, V. M. & Knörnschild, M. Vocal production learning in mammals revisited. Trans. R. Soc. B Biol. Sci. 376 , 20200244 (2021). Beecher, M. D. & Brenowitz, E. A. Functional aspects of song learning in songbirds. Trends Ecol. Evol. 20 , 143–149 (2005). Jarvis, E. D. Neural systems for vocal learning in birds and humans: a synopsis. Ornithol. 148 , 35–44 (2007). Jourjine, N. & Hoekstra, H. E. Expanding evolutionary neuroscience: insights from comparing variation in behavior. Neuron 109 , 1084–1099 (2021). Jürgens, U. Neural pathways underlying vocal control. Biobehav. Rev. 26 , 235–258 (2002). Ghazanfar, A. A. & Rendall, D. Evolution of human vocal production. Biol. 18 , R457–R460 (2008). Jarvis, E. D. Learned Birdsong and the Neurobiology of Human Language. N. Y. Acad. Sci. 1016 , 749–777 (2004). Kuhl, P. K. Birds and babies: Ontogeny of vocal learning. Natl. Acad. Sci. 121 , e2405626121 (2024). Hammerschmidt, K. & Fischer, J. Constraints in Primate Vocal Production. in Evolution of Communicative Flexibility (eds. Oller, D. K. & Griebel, U.) 92–119 (The MIT Press, 2008). doi:10.7551/mitpress/9780262151214.003.0005. Ghazanfar, A. A. & Liao, D. A. Constraints and flexibility during vocal development: insights from marmoset monkeys. Opin. Behav. Sci. 21 , 27–32 (2018). Hebb, A. O. & Ojemann, G. A. The thalamus and language revisited. Brain Lang. 126 , 99–108 (2013). Simonyan, K. The laryngeal motor cortex: its organization and connectivity. Opin. Neurobiol. 28 , 15–21 (2014). Fitch, W. T. The Evolution of Syntax: An Exaptationist Perspective. Evol. Neurosci. 3 , (2011). Bernal, B. & Ardila, A. The role of the arcuate fasciculus in conduction aphasia. Brain 132 , 2309–2316 (2009). Rilling, J. K. et al. The evolution of the arcuate fasciculus revealed with comparative DTI. Neurosci. 11 , 426–428 (2008). Janik, V. M. Cetacean vocal learning and communication. Opin. Neurobiol. 28 , 60–65 (2014). Prat, Y., Taub, M. & Yovel, Y. Vocal learning in a social mammal: Demonstrated by isolation and playback experiments in bats. Adv. 1 , e1500019 (2015). Genzel, D., Desai, J., Paras, E. & Yartsev, M. M. Long-term and persistent vocal plasticity in adult bats. Commun. 10 , 3372 (2019). Vernes, S. C. & Wilkinson, G. S. Behaviour, biology and evolution of vocal learning in bats. Trans. R. Soc. B Biol. Sci. 375 , 20190061 (2019). Ravignani, A. et al. What Pinnipeds Have to Say about Human Speech, Music, and the Evolution of Rhythm. Neurosci. 10 , (2016). Berta, A., Churchill, M. & Boessenecker, R. W. The Origin and Evolutionary Biology of Pinnipeds: Seals, Sea Lions, and Walruses. Rev. Earth Planet. Sci. 46 , 203–228 (2018). Charrier, I. Vocal Communication in Otariids and Odobenids. in Ethology and Behavioral Ecology of Otariids and the Odobenid (eds. Campagna, C. & Harcourt, R.) 265–289 (Springer International Publishing, Cham, 2021). doi:10.1007/978-3-030-59184-7_14. Schusterman, R. J. Vocal learning in mammals with special emphasis on pinnipeds. in The Evolution of Communicative Flexibility: Complexity, Creativity, and Adaptability in Human and Animal Communication (eds. Oller, D. K. & Gribel, U.) 41–70 (MIT Press, Cambridge, MA, 2008). Casey, C., Sills, J. M., Knaub, S., Sotolotto, K. & Reichmuth, C. Lifelong Patterns of Sound Production in Two Seals. Mamm. 47 , 499–514 (2021). Duengen, D., Jadoul, Y. & Ravignani, A. Vocal usage learning and vocal comprehension learning in harbor seals. BMC Neurosci. 25 , 48 (2024). Brainard, M. S. & Doupe, A. J. Auditory feedback in learning and maintenance of vocal behaviour. Rev. Neurosci. 1 , 31–40 (2000). Ralls, K., Fiorelli, P. & Gish, S. Vocalizations and vocal mimicry in captive harbor seals, Phoca vitulina. J. Zool. 63 , 1050–1056 (1985). Deacon, T. W. The Symbolic Species: The Co-Evolution of Language and the Brain . (Norton, New York, NY, 1998). Stansbury, A. L. & Janik, V. M. Formant Modification through Vocal Production Learning in Gray Seals. Biol. 29 , 2244-2249.e4 (2019). Schusterman, R. J. & Reichmuth, C. J. Novel sound production through contingency learning in the Pacific walrus (Odobenus rosmarus divergens). Cogn. 11 , 319–27 (2008). McNab, J. A. et al. High resolution diffusion-weighted imaging in fixed human brain using diffusion-weighted steady state free precession. NeuroImage 46 , 775–785 (2009). Berns, G. S. et al. Diffusion tensor imaging of dolphin brains reveals direct auditory pathway to temporal lobe. R. Soc. B Biol. Sci. 282 , 20151203 (2015). Tendler, B. C. et al. The Digital Brain Bank, an open access platform for post-mortem imaging datasets. eLife 11 , e73153 (2022). Sawyer, E. K., Turner, E. C. & Kaas, J. H. Somatosensory brainstem, thalamus, and cortex of the California sea lion ( Zalophus californianus) . Comp. Neurol. 524 , 1957–1975 (2016). Rioch, D. McK. A physiological and histological study of the frontal cortex of the seal (Phoca vitulina). Bull. 73 , 591–602 (1937). Chiao, G. Z., Larson, C. R., Yajima, Y., Ko, P. & Kahrilas, P. J. Neuronal activity in nucleus ambiguus during deglutition and vocalization in conscious monkeys. Brain Res. 100 , (1994). Akunna, G. G. & Abah, E. D. Deciphering the Nucleus Ambiguus: Anatomy, Functions and Clinical Implications Explored. Clin. Med. Sci. 7 , (2023). Nattie, E. CO2, brainstem chemoreceptors and breathing. Neurobiol. 59 , 299–331 (1999). Jean, A. Brain Stem Control of Swallowing: Neuronal Network and Cellular Mechanisms. Rev. 81 , 929–969 (2001). Herrero, J. L., Khuvis, S., Yeagle, E., Cerf, M. & Mehta, A. D. Breathing above the brain stem: volitional control and attentional modulation in humans. Neurophysiol. 119 , 145–159 (2018). Belyk, M. & Brown, S. The origins of the vocal brain in humans. Biobehav. Rev. 77 , 177–193 (2017). Fitch, W. T. The evolution of speech: a comparative review. Trends Cogn. Sci. 4 , 258–267 (2000). Roberts, T. F. et al. Identification of a motor-to-auditory pathway important for vocal learning. Neurosci. 20 , 978–986 (2017). López-Barroso, D. et al. Word learning is mediated by the left arcuate fasciculus. Natl. Acad. Sci. 110 , 13168–13173 (2013). Attard, M. R. G., Pitcher, B. J., Charrier, I., Ahonen, H. & Harcourt, R. G. Vocal Discrimination in Mate Guarding Male Australian Sea Lions: Familiarity Breeds Contempt. Ethology 116 , 704–712 (2010). Nottebohm, F. The Origins of Vocal Learning. Nat. 106 , 116–140 (1972). Charrier, I. & Casey, C. Social Communication in Phocids. in Ethology and Behavioral Ecology of Phocids (eds. Costa, D. P. & McHuron, E. A.) 69–100 (Springer International Publishing, Cham, 2022). doi:10.1007/978-3-030-88923-4_3. Stirling, I. & Thomas, J. A. Relationships between underwater vocalizations and mating systems in phocid seals. Mamm. 29 , 227–246 (2003). Cassini, M. H. The evolution of reproductive systems in pinnipeds. Ecol. 10 , 612–616 (1999). Terhune, J. M. The underwater vocal complexity of seals (Phocidae) is not related to their phylogeny. Canadian Journal of Zoology , 97 (3), 232-240 (2019). Ravignani, A. et al. Ontogeny of vocal rhythms in harbor seal pups: an exploratory study. Zool. 65 , 107–120 (2019). Duengen, D. & Ravignani, A. The paradox of learned song in a semi‐solitary mammal. Ethology 129 , 445–453 (2023). Nottebohm, F. & Arnold, A. P. Sexual Dimorphism in Vocal Control Areas of the Songbird Brain. Science 194 , 211–213 (1976). Levin, I., Sinha, M., Barton, S. & Hecht, E. A left-lateralized white matter tract associated with communication in domestic dogs. Biol. 34 , R1069–R1070 (2024). Moll, F. W. et al. Thalamus drives vocal onsets in the zebra finch courtship song. Nature 616 , 132–136 (2023). McGregor, J. N. et al. Shared mechanisms of auditory and non-auditory vocal learning in the songbird brain. eLife 11 , e75691 (2022). Huszar, I. N. et al. Tensor image registration library: Deformable registration of stand‐alone histology images to whole‐brain post‐mortem MRI data. NeuroImage 265 , 119792 (2023). Hoeksema, N. et al. Neuroanatomy of the grey seal brain: bringing pinnipeds into the neurobiological study of vocal learning. Trans. R. Soc. B Biol. Sci. 376 , 20200252 (2021). Vargha-Khadem, F., Gadian, D. G., Copp, A. & Mishkin, M. FOXP2 and the neuroanatomy of speech and language. Rev. Neurosci. 6 , 131–138 (2005). Hersh, T. A., Ravignani, A. & Whitehead, H. Cetaceans are the next frontier for vocal rhythm research. Natl. Acad. Sci. 121 , e2313093121 (2024). Rouse, A. A., Patel, A. D. & Kao, M. H. Vocal learning and flexible rhythm pattern perception are linked: Evidence from songbirds. Natl. Acad. Sci. 118 , e2026130118 (2021). Patel, A. D. Vocal learning as a preadaptation for the evolution of human beat perception and synchronization. Trans. R. Soc. B Biol. Sci. 376 , 20200326 (2021). Cook, P. F., Rouse, A. A., Wilson, M. & Reichmuth, C. A California sea lion (Zalophus californianus) can keep the beat: motor entrainment to rhythmic auditory stimuli in a non vocal mimic. Comp. Psychol. 127 , 412–27 (2013). Cook, P. F., Hood, C., Rouse, A. A. & Reichmuth, C. Sensorimotor Synchronization to Rhythm in an Experienced Sea Lion Rivals that of Human Subjects. Rep. (In review). Elemans, C. P. H. et al. Evolutionary novelties underlie sound production in baleen whales. Nature 627 , 123–129 (2024). Madsen, P. T., Siebert, U. & Elemans, C. P. H. Toothed whales use distinct vocal registers for echolocation and communication. Science 379 , 928–933 (2023). Fedorenko, E., Ivanova, A. A. & Regev, T. I. The language network as a natural kind within the broader landscape of the human brain. Rev. Neurosci. 25 , 289–312 (2024). Methods Materials This study included four California sea lion brains, three elephant seal brains, four harbor seal brains, and four coyote brains. For each species there were two male and two female brains except for elephant seals, which were all male. Two sea lions were juveniles, and two were adults. All harbor seals and elephant seals were juveniles. All coyotes were adults. All brains for this study were obtained opportunistically from rehabilitation facilities and were held and imaged under appropriate permits (Extended Data Table 1). Brains were obtained opportunistically from animals euthanized for veterinary cause, and were extracted by experienced veterinarians and necropsy techs. Marine mammal brains used in this study were collected under NMFS permit 18786, and researchers conducting histology and MRI were permitted by NMFS to hold those tissues. Pinniped brains were obtained from The Marine Mammal Center in Sausalito, CA, and were extracted immediately after veterinary euthanasia. They were placed in cold 10% buffered formalin and fixed in a cold room (~4.5 degrees C) with regular formalin refreshes for at least two weeks before being shipped to Emory University for imaging at FERN. The coyote brains were obtained from the USDA Wildlife Research Center in Logan, Utah. Following veterinary euthanasia, heads were removed and shipped on ice to Emory University. There, the skull was cracked and the entire heads were placed in buckets of cold 10% buffered formalin and kept in a cold room for a week. At that point, the brains were removed from the skull and continued fixation for a month before imaging. All pinnipeds and coyotes had a veterinary diagnosis prior to euthanasia (Extended Data Table 1). Brains were selected from subjects with no evidence of neurological disease. Following imaging all brains were also assessed for any evidence of gross neurological disease–none was found. Three pinniped brains were used for histological assessment of the nucleus ambiguus. The sea lion and harbor seal brain used for histology were acquired from The Marine Mammal Center under the same permit as the brains used for MRI. The harbor seal brain was archival and acquired over 40 years ago from an NIH brain collection. Brain Imaging All brains for this study were imaged at Emory University’s FERN imaging center on a Siemen’s Trio 3T magnet with standard gradients and a 32-channel head receive coil. Protocol details are in Cook et al. (2018) 87 and Cook et al. (2022) 88 . In brief, for imaging pinniped brains were set in 2% agarose (Phenix Research Products Low EEO Molecular Biology Grade Agarose) doped with 2 mM gadolinium (III) oxide (Acros Organics, Fisher Scientific) to provide stability, a low-signal environment, and minimize magnetic field homogeneity. Coyote brains were placed in Fluorinert FC-3283 (3M) for the same purposes. High-resolution (0.6 x 0.6 x 0.5 mm) structural images were obtained for the pinnipeds using balanced steady-state free precession (SSFP) sequences (TR = 7.09 ms, TE = 3.55 ms, flip angle = 37 degrees) that were averaged together 89 . High-resolution (0.5 mm isotropic) T2 structural images were obtained for the coyotes using a standard turbo spin echo sequence (TR = 1000 ms, TE = 47 ms, flip angle = 180 degrees). High resolution (1 mm isotropic) diffusion images were obtained for all subjects using a specialized diffusion-weighted SSFP sequence developed specifically for post-mortem diffusion imaging 90 . Fifty-two directions were acquired for each subject (FOV = 128 mm, TR = 31 ms, TE = 24 ms, flip angle = 35 degrees, bandwidth = 159 Hz/pixel, q = 255 cm -1 , Gmax = 38.0 mT/m, gradient duration = 15.76ms, effective b value = 3500). For each subject a reference image analogous to b=0 was obtained using the same parameters but with q = 10 cm -1 . Diffusion images were registered to the higher-resolution structural image space using FSL’s FLIRT to allow use of seeds defined on the structural images to be used for tract tracing and for display of tracing results in high resolution space. Tractography Tractography for this study was probabilistic and used a modified version of BEDPOSTX adjusted for the SSFP sequence, and modeled primary and secondary diffusion direction at each voxel to account for crossing fibers 91 . All tracts were based on manually defined seeds based on anatomical criteria (Extended Data Table 3). For quantitative analysis, seed to target region and seed to target region with exclusion region traces were used (Extended Data Table 2). Single seed traces were used for validation and identification of certain seeds and regions. An anatomically defined brainstem cross-section to brainstem cross-section tract was used for a tracing quality check for each brain. For all tracts, default values were used in FSL’s PROBTRACKX 91 (5,000 streamlines, step length = 0.5 mm, curvature threshold = 80 degrees). To control for differences in subject and species brain, and thus region of interest, size, quantitative values for multi-region traces were computed as a percentage of total streamlines that made it from the seed region to the target region. We took the waytotal (total number of probabilistic streamlines generated from the seed region that reached the target region) and divided by the number of voxels in the seed region. The resultant value represents the percentage of generated streamlines that reached the target and instantiates a correction for differences in size of anatomical regions between samples. All quantitative tract tracing results in the current study are drawn from two-region traces, with a seed and target region. During region segmentation and validation, a number of single-seed-region traces were conducted to cross-validate segmentations (see Extended Data Table 3). These were displayed with streamline thresholds selected to produce a coherent, dense pathway. For visualization of multi-region tracts, fsleyes render was used with the same parameters for each sample and tract (high resolution structural underlay: alpha = 75, brightness = 45, contrast = 50, display range = 80-200, gamma = 50; tracts: alpha = 100, brightness = 63, contrast = 85, gamma = 75). ROI and seed definition All tracing in this study used seeds and targets that were either hand drawn voxel by voxel or constituted 3D cubes generated at a hand-selected center voxel. The criteria for drawing and generating each ROI differed by region (Figure 1, Extended Data Figure 1, Extended Data Table 3). Each region was delineated separately in the left and right hemisphere, except for PAG and the brainstem sections. Brainstem control tract regions Whole cross sections of brainstem were segmented caudally, just caudal to the caudal extent of the cerebellar peduncles, and rostrally, just caudal to the caudal extent of the pons (Extended Data Figure 1). Pyramids The descending primary motor pathway (corticospinal tract) in mammals goes through the pyramids, large structures ventromedial in the brainstem running longitudinally down the brainstem just dorsal to the pons. We overlaid primary diffusion vector directions with the high-resolution structural images and segmented the pyramids in one transverse slice just dorsal to the caudal boundary of the pons. All medial voxels with rostrocaudal directionality were included (Extended Data Figure 1) Internal capsule The corticospinal tract descends through the internal capsule in mammals. We placed three-dimensional cube ROIs bilaterally by hand in the rostral internal capsule for each subject, ventral to the lateral projection of the head of the caudate and dorsal to the putamen. Locations were validated by direct connections to the putamen. Cube size was determined by internal capsule thickness, and was 5 mm 3 for coyotes, 7 mm 3 for CSLs and harbor seals, and 9 mm 3 for elephant seals (Extended Data Figure 1). Nucleus ambiguus An anatomist with expertise in pinniped neurobiology identified N. Amb histologically in the brains of a juvenile female sea lion, juvenile female elephant seal, and juvenile female harbor seal, all euthanized for veterinary cause at TMMC (Extended Data Figure 5). Sections were cut through the brainstem at 50 micrometer thickness and stained with cresyl violet. The N. Amb was identified bilaterally as a cluster of large multipolar neurons, rostrocaudally close to the level of the obex, dorsal to the inferior olive, and rostromedial to the spinal trigeminal nucleus 92 . High resolution histology slides were overlaid at low opacity using the program pureref (https://www.pureref.com) with high resolution trufi images for each pinniped (Figure 1, Extended Data Figure 6). The central histological slice for each animal was matched with an axial brainstem slice just rostral to the opening of the obex. The overlay was scaled and rotated to best match external brainstem countours. Because some brainstems were distorted and non-symmetrical along the right-left axis, the histological overlay was matched separately to each half (right/left) of the brainstem for segmentation. All voxels that even partially overlapped the stained cells identified in N. Amb were included in right and left segmentations. Full histological series of N. Amb rostral to caudal were available in sea lion and elephant seal (Extended Data Figure 7, Extended Data Figure 8), and were used in series to segment ambiguus in the MR data. In harbor seals, a histological slice was only available in the middle of the ambiguus. This was used to segment the corresponding brainstem section in the MR Data and on the slices immediately rostral and caudal to that slice. In coyotes, dog histology sections of brainstem were used ( https://vanat.ahc.umn.edu/brainsect/showAtlasFrames.html ) with the same method as the harbor seal brains. All segmentations were confirmed to be interposed between IO and SPVI, overlapping neither. Single seed traces were conducted with each segmentation for validation, looking for A. direct rostral connection to the PAG, B. bilateral connection between left and right N. Amb, and C. lateral connection to apparent vagus nerve (Extended Data Figure 4). Because N. Amb is close to the reticular formation, which connects directly to thalamus, segmentations were further validated by lack of direct connection to thalamus in single seed traces. Because no projections of the putative N. Amb past the midbrain were identified in single-mask tractography in any of the coyotes, we also performed single seed traces from the coyote vocal motor cortex regions and confirmed that there were no coherent pathways dorsal to the pyramidal tract at the level of the obex (Extended Data Figure 4). Vocal motor cortex In older physiological studies, motor cortex related to production of vocal behavior has been identified in ventrolateral precuciate gyrus in harbor seal and dog and cat brains 53,93,94 . This location has been incidentally corroborated by more recent physiological and histological work identifying primary somatosensory cortex in sea lions between the coronal (rostral) and pseudosylvian (caudal) sulci 52 , and the afferent targets for mouth and neck just caudal to the caudal facing bend in the middle of the coronal sulcus. This location is, in line with conserved relationships between S1 and M1 in mammalian brains, just caudal to the putative vocal motor cortex in the current study. Of note, although this brain region is frequently referred to as “laryngeal motor cortex” in mammalian studies, we could not confirm that our ROIs were specific to the laryngeal representation as opposed to also including surrounding neck, mouth, and vocal tract representations. Therefore, we refer to our region as “vocal motor cortex” (VMC) throughout this paper. In the current study, three-dimensional cube ROIs were placed by hand for each subject on the ventral bank of the precruciate gyrus, with the center of the cube in white matter just caudal to the ventrorostral bank of the precruciate dorsolateral to the presylvian sulcus and ventrolateral to the coronal sulcus (Figure 1). Because of the more dorsoventrally constrained (“flatter”) shape of the phocid brains this location was lower by absolute coordinates in the elephant seal and harbor seal brains than in the CSL and coyote brains. To account for gross differences in brain volume, cubes for elephant seal brains were 13 voxels isotropic, 11 voxels isotropic for sea lions, 9 voxels isotropic for harbor seals, and 7 voxels isotropic for coyotes. All VMC ROIs were validated to be part of motor cortex with single mask traces looking for direct connections to the globus pallidus and, with distance correction, passage through the pyramids in the brainstem (Extended Data Figure 9). Periaqueductal Grey PAG is a bilateral band of grey matter situated around the mesencephalic aqueduct, and its location and function is conserved in mammalian brains. In the current study, the PAG was segmented in the transverse plane based on anatomical criteria (Extended Data Figure 1). The grey matter/white matter boundary of the periphery of the PAG was clearly visible in the structural images for all subjects, and the structure was segmented from the rostral boundary of the posterior commissure to the caudal extent of the enclosed portion of the mesencephalic aqueduct. The dorsal raphe nucleus was also included ventrally in all segmentations. Pre vocal motor cortex Premotor cortex has been identified in carnivores with chemical tracing 95 and is dorsorostral to the motor cortex in the presylvian gyrus. In this study we created pre-VMC ROIs by centering three dimensional cubes (13, 11, 9, 7 mm as with VMC) in the rostral most portion of the white matter underlying presylvian gyrus along maximum projection from single-seed traces from each subjects’ VMC ROIs bilaterally (Figure 1). ROIs were further validated with single seed tracings looking for direct connections to the ventrorostral thalamus and ventral striatum. Medial geniculate nucleus MGN has been identified in previous work with canid carnivores, including sea lions 96 . In the current study It was segmented anatomically in the dorsal plane for all subjects. MGN presents as a small protuberance of grey matter on the caudal ventrolateral boundary of the thalamus, just medial to the medial parahippocampal cortex. MGN was validated for each segmentation by confirming that it was along the direct rostral projection pathway of the ipsilateral inferior colliculus (Figure 1). Inferior colliculus IC was segmented in the dorsal plane using anatomical criteria. In all subjects it clearly presented as a dome of grey matter protruding caudally from the midbrain ventral to the superior colliculus. It was bounded rostrally by the caudal edge of the PAG (Figure 1). It was validated in each subject by projection to putative MGN. Auditory cortex Primary auditory cortex’s location is broadly conserved in mammal brains 97 , and was identified in the current study from distance-corrected single seed tracings from MGN bilaterally. Prior physiological work has identified A1 in dog brains 98 in the rostrocaudal middle of the ectosylvian gyrus, just dorsal to the dorsolateral extent of the pseudosylvian gyrus, which roughly corresponds to the same position of the suprasylvian gyrus in acrtoid carnivores 99 . We situated three dimensional cubes bilaterally in these gyri along the strongest direct pathways from MGN for each subject (Figure 1). Anterior ventrolateral thalamus AVT has not been explicitly delineated in pinniped brains previously, but has been in other carnivores, including dogs 100 . AVT is at the ventrorostral portion of the thalamus, situated dorsal to the subthalamic nucleus and just inside the zona incerta. It is lateral to the rostral lamina and the rostral-most thalamic nucleus which connects to prefrontal cortex. In the current study, AVT was segmented in each subject in the transverse plane based on anatomical criteria (Extended Data Figure 1). Lamina and other intrathalamic boundaries are more visible in the B0 than trufi and T2 images, so tracing was aided with overlays of B0 images registered to high resolution structural images. VRT was segmented by including the ventrorostral-most portion of the thalamus just lateral to the lamina. The ventral most boundary was the ventral portion of the thalamus. The dorsal boundary was arbitrarily but consistently set as the rostral-most extension of the rostral portion of the thalamus. Segmentations were validated by single seed traces looking for direct rostral projection through the ventral striatum. Ventral striatum Ventral striatum has been widely studied in carnivores 101 and was segmented in the transverse plane in the current study based on explicit anatomical criteria (Extended Data Figure 1). The entire rostral portion of the putamen was segmented and the head of the caudate nucleus was segmented in the same dorsoventral extent. Both regions were included in left and right striatal masks. In carnivores, the head of the caudate is enlarged and clearly visible as a bulb of grey matter caudal to the olfactory probuterances and rostral to the thalamus 101 . The putamen is notably smaller, but can be viewed clearly in the transverse plane as a separate island of grey matter medial to the medial most grey matter of the frontal and temporal lobe. The white matter between the caudate and the putamen and the striations interdigitating it were not included in segmentations, which were kept to the smooth coherent bodies of the basal ganglia structures. Anterior cingulate cortex The cingulate cortex has been identified and functionally assessed in primates, rodents, and carnivores 102 . It is a bundle of fibers running rostrocaudally just dorsal to the corpus callosum, with a ventral turn at the rostral extent. In the current study the anterior cingulate was identified by corroborating anatomical criteria with diffusion direction. The V1 voxel-wise diffusion maps from dtifit were overlayed on the high resolution structural images (Extended Data Figure 1). Then, cingulate was segmented in the sagittal view by following the most ventral gyri of the longitudinal fissure, including all grey matter voxels with caudal-rostral directionality immediately dorsal to the middle portion of the corpus callosum and dorsoventral directionality rostral to the rostral dip (genu) of the corpus callosum. The caudal boundary was arbitrarily but consistently defined as rostral to the rostral boundary of the thalamus. Statistical method For the eight region-to-region traces in this study, tracts were assessed quantitatively. Each pathway was assessed in each of the brains. The waytotal generated by PROBTRACKX (number of probabilistic streamlines that successfully traversed from the seed to the target region) was taken as a percentage of total streamlines generated (5000 x the number of voxels in the seed mask). This allowed direct comparison of tract strength accounting for differences in brain and region size between subjects. Because of the small sample size, comparisons were made between subjects’ tract strength nonparametrically. A two-sided Kruskal-Wallis test was run via MATLAB for each tract by species to determine if any of the median species values differed significantly, followed by Bonferroni-corrected multiple comparisons by species. Methods References 87. Cook, P. F., Berns, G. S., Colegrove, K., Johnson, S. & Gulland, F. Postmortem DTI reveals altered hippocampal connectivity in wild sea lions diagnosed with chronic toxicosis from algal exposure. J. Comp. Neurol. 526, 216–228 (2018). 88. Cook, P. F. & Berns, G. Volumetric and connectivity assessment of the caudate nucleus in California sea lions and coyotes. Anim. Cogn. 25, 1231–1240 (2022). 89. Miller, K. L. et al. Diffusion imaging of whole, post-mortem human brains on a clinical MRI scanner. NeuroImage 57, 167–181 (2011). 90. Miller, K. L., McNab, J. A., Jbabdi, S. & Douaud, G. Diffusion tractography of post-mortem human brains: Optimization and comparison of spin echo and steady-state free precession techniques. NeuroImage 59, 2284–2297 (2012). 91. Behrens, T. E. J., Berg, H. J., Jbabdi, S., Rushworth, M. F. S. & Woolrich, M. W. Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? NeuroImage 34, 144–155 (2007). 92. Berman, Alvin L. The Brain Stem of the Cat; a Cytoarchitectonic Atlas with Stereotaxic Coordinates. (University of Wisconsin Press, Madison, 1968). 93. Milojevic, B. & Hast, M. H. Cortical motor centers of the laryngeal muscles in the cat and dog. Ann. Otol. Rhinol. Laryngol. 73, 979–988 (1964). 94. De Lanerolle, N. C. & Lang, F. F. Functional Neural Pathways for Vocalization in the Domestic Cat. in The Physiological Control of Mammalian Vocalization (ed. Newman, J. D.) 21–41 (Springer US, Boston, MA, 1988). doi:10.1007/978-1-4613-1051-8_3. 95. Radtke-Schuller, S. et al. Dorsal prefrontal and premotor cortex of the ferret as defined by distinctive patterns of thalamo-cortical projections. Brain Struct. Funct. 225, 1643–1667 (2020). 96. Montie, E. W. et al. Neuroanatomy and Volumes of Brain Structures of a Live California Sea Lion ( Zalophus californianus ) From Magnetic Resonance Images. Anat. Rec. 292, 1523–1547 (2009). 97. Krubitzer, L. The Magnificent Compromise: Cortical Field Evolution in Mammals. Neuron 56, 201–208 (2007). 98. Andics, A., Gácsi, M., Faragó, T., Kis, A. & Miklósi, Á. Voice-Sensitive Regions in the Dog and Human Brain Are Revealed by Comparative fMRI. Curr. Biol. 24, 574–578 (2014). 99. Boch, M. et al. Comparative neuroimaging of the carnivoran brain: Neocortical sulcal anatomy. Preprint at https://doi.org/10.1101/2024.06.03.597118 (2024). 100. Salazar, I., Ruiz Pesini, P., Fernández Alvarez, P. & Cifuentes, J. M. The thalamus of the dog: a tridimensional and cytoarchitectonic study. Anat. Anz. 169, 101–113 (1989). 101. Baer, E. et al. Predictive Methods and Probabilistic Mapping of Subcortical Brain Components in Fossil Carnivora. J. Comp. Neurol. 533, e70014 (2025). 102. Devinsky, O., Morrell, M. J. & Vogt, B. A. Contributions of anterior cingulate cortex to behaviour. Brain 118, 279–306 (1995). Additional Declarations There is NO Competing Interest. Supplementary Files ExtendedData.docx SupplementaryTable1.docx Cite Share Download PDF Status: Published Journal Publication published 11 Mar, 2026 Read the published version in Science → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6222519","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Biological Sciences - Article","associatedPublications":[],"authors":[{"id":428983577,"identity":"ba82bc77-142e-4ead-a58f-d172dfec57a9","order_by":0,"name":"Peter Cook","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAnklEQVRIiWNgGAWjYJACxoYKBgb2BgYGCRK0nGFg4DlAkpbGNlK0mLMfPvhw5rw78jwMzAdv8xCjxbInLdlw47Znhj0MbMnWRGkxuMFjJvlw2+EEewYeM2kStMw5nMDDwP+NBC0bG0BaeNiI0wL2y4xjhw17mNmMLecQowUcYj01h+V52Jsf3nhDlMPgLGZilKNqGQWjYBSMglGACwAAkeMtkE0JoAEAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-8064-5217","institution":"New College of Florida","correspondingAuthor":true,"prefix":"","firstName":"Peter","middleName":"","lastName":"Cook","suffix":""},{"id":428983578,"identity":"44b9a4b6-cc86-4ed5-9c7e-ed25c089b695","order_by":1,"name":"Andrew Rouse","email":"","orcid":"https://orcid.org/0000-0003-3504-5906","institution":"Long Marine Laboratory, University of California Santa Cruz","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"","lastName":"Rouse","suffix":""},{"id":428983579,"identity":"1799fe3e-d622-4da2-9efd-40dbeb911115","order_by":2,"name":"Eva Sawyer","email":"","orcid":"","institution":"Vanderbilt University","correspondingAuthor":false,"prefix":"","firstName":"Eva","middleName":"","lastName":"Sawyer","suffix":""},{"id":428983580,"identity":"b4bb1267-6813-432c-95dc-cdc9f8306dc4","order_by":3,"name":"Karla Miller","email":"","orcid":"https://orcid.org/0000-0002-2511-3189","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Karla","middleName":"","lastName":"Miller","suffix":""},{"id":428983581,"identity":"f55b075b-6e35-4b94-8fea-c8c17baa529f","order_by":4,"name":"Gregory Berns","email":"","orcid":"","institution":"Emory University","correspondingAuthor":false,"prefix":"","firstName":"Gregory","middleName":"","lastName":"Berns","suffix":""}],"badges":[],"createdAt":"2025-03-13 20:45:44","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6222519/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6222519/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1126/science.adx9367","type":"published","date":"2026-03-12T00:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":78721810,"identity":"89251faf-7959-43f4-b8e1-03b396126098","added_by":"auto","created_at":"2025-03-18 05:04:23","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1367897,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eLocation and identification methodology for the nucleus ambiguus (N.Amb) and cortical regions (A1, VMC, and pre-VMC) used in tract tracing for this study. a: cresyl violet Nissl stain of the brainstem of a CSL in an axial section, localizing N. Amb in the brainstem with an expanded view below. b: Histology slices were overlaid over high-resolution MR images of each species based on anatomical landmarks, and all voxels overlapping with N. Amb were included in MR segmentations. Brainstem of CSL in axial (above) and sagittal (below) plane. c: 3D render of a California sea lion brain with cortical ROIs overlaid and validations of auditory (green), sensory (orange), motor (yellow), and VMC and pre-VMC (purple) surrounding. Frontal sulci are indicated: the pseudosylivan fissure (PSF), presylvian sulcus (PRS), Coronal sulcus (CoS), suprasylvian sulcus (SSA). Left: A1 was identified via direct probabilistic traces from the MGN, identified anatomically and validated by tracing from the inferior colliculus. An isotropic cube was defined in the principle white matter lateral projection site of MGN for each subject. Top right: Pre-VMC was identified by probabilistic tracing from VMC, and validated by comparison with electrophysiological studies in terrestrial carnivores (ferret, dog). In the current study, an isotropic cube ROI was placed for pre-VMC along the maximum trace projection from VMC just rostral to the dorsal portion of the presylvian sulcus. Bottom right: representation of somatosensory cortex in grey, with mouth and neck cortex in red in the California sea lion, and primary motor cortex in grey with vocal motor cortex in yellow for the harbor seal. These regions were identified with histology and invasive electrophysiology\u003c/em\u003e\u003csup\u003e52,53\u003c/sup\u003e\u003cem\u003e, figures adapted by H. Cook. In the current study, an isotropic cube ROI was placed for VMC caudal to the ventral presylvian sulcus. Identification of subcortical regions are in Extended Data Figure 1. Anatomical orientation noted as: L: left, R: rostral, D: dorsal.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/38dc6af3e81b8f784bbb4be2.png"},{"id":78722414,"identity":"e23e7eb9-ac07-4058-8fa5-45ca0bf890e4","added_by":"auto","created_at":"2025-03-18 05:12:23","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":735223,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e3D renderings of left (blue) and right (red) nucleus ambiguus (N. Amb) to VMC and A1 to pre-VMC pathways for all subjects. Pathways are from probabilistic traces conducted in FSL’s PROBTRACKX using standard settings, with a minimum threshold of 400 streamlines. a: N. Amb to VMC pathways shown in a California sea lion brain in the dorsal (left) and sagittal (right) view. Left and right N. Amb were seeded for all subjects separately with left and right VMC as waypoints and the bilateral PAG as an exclusion mask. b: A1 to pre-VMC pathways shown in a harbor seal brain in the dorsal (left) and sagittal (right) view. Left and right A1 were seeded for all subjects separately with left and right pre-VMC as waypoints and left and right MGN as exclusion masks. c: results for N. Amb to VMC tracings in all subjects (top to bottom: Coyote, California sea lion, northern elephant seal, harbor seal), with a minimum threshold of 400 streamlines. d: results for A1 to pre-VMC tracings for all subjects. Anatomical orientation noted as: L: left, R: rostral, D: dorsal.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/35f068b38293683cd55e355f.png"},{"id":78721808,"identity":"a50897a6-4325-4256-88c8-6c335628ed16","added_by":"auto","created_at":"2025-03-18 05:04:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":518508,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eCorrected tract strength for multi-region probabilistic traces (for brainstem and corticospinal control tracts and comparisons see Extended Data Figure 3). The six pathways are shown in the sagittal view with seeds and waypoint ROIs in red, exclusion masks (where present) in green, and paths in yellow. Each plot shows the percent of total streamlines for paths on the Y axis and the species in the X axis (see Extended Data Table 4 and Supplementary Table 1). 5000 probabilistic streamlines were generated from each voxel of the seed regions for each trace. Paths that make it from the seed region to the waypoint without transiting any exclusion masks constitute the final probabilistic trace, and the number of such paths is the “waytotal.” Each subject’s corrected waytotal for each trace is shown for left hemisphere in blue and right hemisphere in yellow, with a dotted line connecting the values for each subject. Significant differences between species values are indicated by brackets and asterisks (Kruskal-Wallis test with Bonferroni-corrected pairwise comparisons, *p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001) a: N. Amb to VMC traces b: A1 to VMC traces c: Anterior cingulate to PAG d: Anterior ventrolateral thalamus to pre-VMC (HS tract strength remains significantly higher than California sea lion and coyotes’ if potential outlier value removed). e: Striatum to pre-VMC (CSL tract strength does not remain significantly higher than that of coyotes if potential outlier value removed). F: Anterior ventrolateral thalamus to striatum. Notably, both the vocal motor-ambiguus connections and the auditory-pre vocal motor connections were stronger in two male harbor seal brains than the two female harbor seal brains.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/c8bf1b2cc43878997a2be520.png"},{"id":78778952,"identity":"f63c288c-3910-47cd-a21b-67f49d1a0b2e","added_by":"auto","created_at":"2025-03-18 19:02:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1075656,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eNeurobiological phylogeny of vocal flexibility in pinnipeds. a: Simplified pinniped phylogenetic tree, representing branch points for the split between the arctoidae (in which branch the pinnipeds are found) and the canidae, the otariidae and phocidae, and the monachinae and phocinae. Potential emergence points for different vocal capabilities are indicated on the left by color. b: Schematic representations of four distinct brain pathways traced in the present study that showed significant difference between species and likely bear on vocal flexibility. c: Schematic representations of mean path strength for each species in the current study. All pinnipeds tested have a robust N Amb to VMC pathway, while this pathway does not appear in the coyotes. This path is linked with volitional control of the larynx. Elephant seals and harbor seals have robust connections between auditory and pre VMC–in humans, this pathway is linked to developmental vocal learning. Harbor seals add to this robust anterior cingulate to PAG and anterior ventrolateral thalamus to Pre-VMC pathways. Anterior cingulate to PAG is associated with control of species typical vocalizations, while the thalamus to pre VMC path is linked with associative learning of vocal behavior and potentially mimicry in parrots.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/be734588ac8c78f480375571.png"},{"id":104740923,"identity":"d3f04afc-6208-451d-8646-27560619ddb4","added_by":"auto","created_at":"2026-03-16 16:19:42","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4251430,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/0e0d9bd3-72cd-4c41-ba22-f20862afc25f.pdf"},{"id":78721807,"identity":"bd1cb8f2-0421-4ecd-8d51-dd90a135e2e4","added_by":"auto","created_at":"2025-03-18 05:04:23","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":4448980,"visible":true,"origin":"","legend":"","description":"","filename":"ExtendedData.docx","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/72ecc6988f33337fa18d64b9.docx"},{"id":78721806,"identity":"3feadfdc-a047-415e-a7d2-45163d73adce","added_by":"auto","created_at":"2025-03-18 05:04:23","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":24776,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-6222519/v1/bf6cf563f60216f758ac536a.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Neurobiological adaptations supporting vocal plasticity have accumulated in the marine carnivores","fulltext":[{"header":"Introduction","content":"\u003cp\u003eVocal behavior in the vast majority of animal species has been considered inflexible or obligate\u003csup\u003e1\u003c/sup\u003e, and is governed by interactions between an intricate network of midbrain and brainstem nuclei\u003csup\u003e15\u0026ndash;17\u003c/sup\u003e that predominantly respond to environmental and affective conditions. The ability to vocalize under flexible, volitional control is distributed patchily among clades\u003csup\u003e18\u003c/sup\u003e. Non-obligate vocal behavior is typically attributed to vocal learning, but there are multiple dissociable cognitive and physiological mechanisms involved\u003csup\u003e3\u003c/sup\u003e. Vocal learning is increasingly referred to as a \u0026ldquo;spectrum\u0026rdquo; of interrelated capabilities. A growing number of species has been demonstrated to deploy species-typical calls in novel contexts\u003csup\u003e2\u003c/sup\u003e. Truly rare are species whose members can learn to produce novel calls with frequency and formant characteristics outside of the inherited repertoire. These \u0026ldquo;vocal production learners\u0026rdquo; are most common in avian lineages, where the behavioral and neurobiological mechanisms of vocal learning have been extensively studied in both field and laboratory\u003csup\u003e19\u003c/sup\u003e. Such studies have indicated that the ability to alter both filter (formants, via the mouth) and source phonation (frequency, via the syrinx) components of species-typical calls relies on elaboration or doubling of pathways in two forebrain circuits. The \u0026ldquo;anterior\u0026rdquo; circuit connects vocal premotor regions with anterior striatum and anterior ventrolateral thalamus\u003csup\u003e20\u003c/sup\u003e, and allows reinforcement learning to alter vocal motor production. A \u0026ldquo;posterior\u0026rdquo; circuit directly connects vocal motor cortex with brainstem phonatory neurons, allowing volitional control of the syrinx and vocal tract. The study of vocal learning in birds has exemplified the benefit of a \u0026ldquo;model clade\u0026rdquo;\u003csup\u003e21\u003c/sup\u003e approach to comparative behavioral neuroscience. Because scientists have been able to study closely related bird species that vary in systematic ways in terms of vocal flexibility (e.g., oscine and suboscine song birds and parrots), the neurobiological underpinnings of different aspects of call production and learning in birds have been elucidated. Heretofore we have lacked a comparable model clade in the mammalian order.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparative Studies of Vocal Learning in Humans and Other Mammals\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHumans are the best studied vocally flexible mammal. They demonstrate high degrees of vocal plasticity and associated nervous system adaptations\u003csup\u003e22,23\u003c/sup\u003e, including developmental auditory-motor learning; contextual deployment of calls; and flexible breathing, phonation, and filter control throughout the lifespan, most fully exemplified by highly articulated vocal mimicry. Comparisons of human and avian vocal neurobiology have been productive, and have led to a number of hypotheses about neurobiological adaptations supporting vocal plasticity and learning\u003csup\u003e24,25\u003c/sup\u003e. However, the wide evolutionary gulf between the mammal and avian nervous systems and respective sound production mechanisms underscores the need for comparisons within the mammalian lineage.\u003c/p\u003e\n\u003cp\u003eBehavioral vocal flexibility is not systematically distributed across non-human primate species, most of which show vocal stereotypy across measurable dimensions\u003csup\u003e26,27\u003c/sup\u003e, and human vocal behavior differs from that of other primates across too many dimensions to isolate specific neurobiological differences directly supporting different behavioral capabilities. In comparison to other primates, humans show enhanced connectivity in circuits analogous to the anterior and posterior vocal learning pathways in birds\u003csup\u003e24,25,28\u003c/sup\u003e. Striatal thalamocortical connectivity is implicated in language learning and use, but has not been systematically assessed across primates. \u0026nbsp;Monosynaptic connections between laryngeal motor cortex and the nucleus ambiguus, theoretically supporting fine-grained volitional control of phonation, appear to be unique to humans among primates\u003csup\u003e29\u003c/sup\u003e. Accordingly, the influential \u0026ldquo;Kuyper-J\u0026uuml;rgens hypothesis\u0026rdquo;\u003csup\u003e30\u003c/sup\u003e suggests the presence of direct connection between vocal motor cortex and brainstem phonatory nuclei such as the nucleus ambiguus is a precondition for flexible vocal control in any species. In addition, although it has been deemphasized in studies of bird vocal learning, there is extensive evidence that, in humans, auditory-vocal premotor pathways (e.g., the arcuate fasciculus connecting Wernicke\u0026rsquo;s and Broca\u0026rsquo;s areas) support language learning\u003csup\u003e31\u003c/sup\u003e. These pathways are less robust in non-human primates\u003csup\u003e32\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMammalian comparisons would be most productive within clades demonstrating graded variability in production flexibility, which would elucidate the potential role of striatal thalamocortical and auditory-premotor adaptations in vocal learning, and would allow direct tests of the Kuyper-J\u0026uuml;rgens hypothesis. Cetaceans are believed to be highly vocally flexible\u003csup\u003e33\u003c/sup\u003e, but are highly derived anatomically and logistically difficult to study. Bats have shown promise as a model clade, and are accessible to laboratory behavioral neuroscience\u003csup\u003e34,35\u003c/sup\u003e. However, it appears that even the most vocally flexible bat has limited volitional phonatory and filter control throughout the lifespan\u003csup\u003e36\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eThe Behavioral Case for Graded Vocal Flexibility in Pinnipeds\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere is one extant mammalian clade with clear behavioral evidence of differentially distributed components of vocal flexibility ranging from onset-offset control of stereotypical calls through developmental learning and, potentially, mimicry: the pinnipeds\u003csup\u003e7,37\u003c/sup\u003e. As such, they are a promising model clade for exploring the full spectrum of vocal learning adaptations in the mammalian brain. Pinnipeds are a lineage of marine-adapted Carnivores comprising odobenids (walruses), otariids (eared seals), and phocids (true seals), that have diverged and radiated over the last 25 million years from an amphibious canid carnivore ancestor\u003csup\u003e38\u003c/sup\u003e. In comparison to terrestrial carnivores, pinnipeds have experienced strong adaptive pressures to control breathing, swallowing, and sound production for a semi-aquatic lifestyle. Carnivores, including pinnipeds\u0026rsquo; closest mustelid relatives, have not been found to express high vocal flexibility\u003csup\u003e1\u003c/sup\u003e. In strong contrast, all extant pinnipeds show some evidence of enhanced vocal flexibility, if only in onset/offset control\u003csup\u003e7\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eExtensive field work shows that otariids, the fur seals and sea lions, produce highly stereotypic calls with no obvious ability to learn filter and phonation modulation\u003csup\u003e39\u003c/sup\u003e. However, there is strong experimental evidence that otariids easily learn to produce and inhibit species-typical calls under contextual control\u003csup\u003e40\u003c/sup\u003e, an ability that is difficult for many terrestrial species. Onset/offset control has been linked with cingulate-midbrain periaqueductal grey (PAG) connectivity\u003csup\u003e2\u003c/sup\u003e, but has not been explicitly linked to monosynaptic connections between laryngeal motor cortex and brainstem phonatory nuclei or gross adaptations to striatal thalamocortical vocal reinforcement learning circuits. The \u003cem\u003eotariidae\u003c/em\u003e and \u003cem\u003ephocidae\u003c/em\u003e (true seals) branches of pinnipeds split at least 24 million years ago, and unlike otariids, a number of phocid species including both \u003cem\u003emonachinae\u003c/em\u003e (including elephant and monk seals) and \u003cem\u003ephocinae\u003c/em\u003e (including harbor and grey seals) show some evidence of vocal learning over development\u003csup\u003e6,7,8,41,42\u003c/sup\u003e. This type of learning involves developmental modulation of phonation based on auditory input, and, in birds, requires cortical-brainstem phonatory connections and striatal thalamocortical adaptations\u003csup\u003e43\u003c/sup\u003e. In humans, developmental vocal learning also relies on an adapted circuit connecting auditory with vocal premotor regions (including the arcuate, mentioned above).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAt the far end of the pinniped flexibility spectrum is the \u003cem\u003ephocinae\u003c/em\u003e branch of the phocid family, which split from the \u003cem\u003emonachinae\u003c/em\u003e over 15 mya. Strikingly, some phocids show evidence of highly flexible vocal learning and mimicry, which may not be fully constrained to a strict developmental period. Evidence of mimicry has been collected in two harbor seals raised in captivity\u003csup\u003e44,45\u003c/sup\u003e, who learned to replicate human words and phrases. An older harbor seal in captivity learned to produce novel formants through operant techniques\u003csup\u003e10\u003c/sup\u003e. More recently, in an experimental context, grey seals learned to mimic novel tonal sequences and vowel-like formants\u003csup\u003e46\u003c/sup\u003e. This behavior requires fine-grained volitional control of the parts of the vocal apparatus responsible for filter (mouth, formants) and phonation (larynx, frequency) control. The neurobiology of mimicry is not fully understood, but in parrots seems to be related to further elaboration of thalamocortical circuits beyond the adaptations to the anterior pathway found in song birds\u003csup\u003e13\u003c/sup\u003e. Finally, among the pinnipeds, odobenids (walruses) are less widely studied, but have produced some of the most compelling experimental evidence of vocal flexibility\u003csup\u003e47,48\u003c/sup\u003e. Across the pinnipeds this range of capabilities, in conjunction with comparisons to terrestrial carnivore cousins, spans the spectrum from relatively fixed vocal stereotypy through mimicry in a set of closely related species with established phylogenetic branch points. Determining covariance between neurobiological and behavioral traits across the pinnipeds will thus yield insight into how changes to mammalian brains support vocal flexibility and learning, and will support new neurobehavioral hypotheses regarding the underpinnings of human language.\u003c/p\u003e\n\u003cp\u003eIn the current study, we directly assessed neural pathways potentially related to vocal plasticity and learning in several pinniped species and a terrestrial carnivore and compared these to available behavioral vocal data for these species. This allowed us to assemble a preliminary neurobehavioral phylogeny of vocal learning in the canid carnivore line. The advent of high signal-to-noise-ratio, high-resolution post-mortem diffusion MRI and tractography\u003csup\u003e49-51\u003c/sup\u003e provides a new opportunity for matching neurobiological adaptations to behavioral capabilities in previously inaccessible species. We used a novel post-mortem-optimized imaging sequence to scan opportunistically obtained brains of four California sea lions (\u003cem\u003eZalophus californianus, otariidae\u003c/em\u003e), three northern elephant seals (\u003cem\u003eMirounga angustirostris, monachinae\u003c/em\u003e), four harbor seals (\u003cem\u003ePhoca vitulina, phocinae\u003c/em\u003e), and four coyotes (\u003cem\u003eCanis latrans, terrestrial canids\u003c/em\u003e) (Subject details in Extended Data Table 1). We then conducted multi-region probabilistic tractography in right and left hemisphere of each subject of the following pathways (Extended Data Table 2). First, we examined nucleus ambiguus to vocal motor cortex (VMC[a] ) to test the Kuyper-J\u0026uuml;rgens hypothesis. Next, we examined two-way connections between the anterior ventrolateral thalamus, anterior striatum, and pre-VMC to assess the role of the anterior striatal thalamocortical circuit in mammalian vocal production flexibility. To look for an arcuate-like connection potentially related to developmental vocal learning, we examined auditory cortex to pre-VMC. Finally, we assessed anterior cingulate to periaqueductal grey (PAG), a connection related to volitional deployment of species-typical calls in primates\u003csup\u003e2\u003c/sup\u003e, To establish seeds and regions for probabilistic tracing, we manually segmented all regions (25 per subject, 375 total, Figure 1, Extended Data Figure 1) using multiple methodologies: brainstem histology (nucleus ambiguus), prior cortical electrophysiological data (motor cortex/VMC/premotor cortex, A1 in coyotes), anatomical criteria (PAG, caudal colliculi, medial geniculate nucleus, anterior ventrolateral thalamus, and anterior cingulate), and tract tracing (A1 in pinnipeds, based on cortical projection from MGN, and pre VMC via tract tracing from VMC (Extended Data Table 3)\u003c/p\u003e\n\u003cp\u003e[a] n.b, the mammalian vocal learning neurobiology literature typically refers to \u0026ldquo;laryngeal motor cortex,\u0026rdquo; but we broadly targeted brain regions involved in volitional motor control of the neck and face that would contribute to vocal motor control.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eWe found clear evidence of brain circuit differences corresponding to differential behavioral evidence for vocal flexibility in each species. There was robust, bilateral, direct cortical connectivity between the putative vocal motor cortex and the nucleus ambiguus in the brainstem in all pinnipeds and in none of the coyotes (Figure 2). Importantly, there were no significant differences between the coyotes and any of the other species in the two control tracts, the whole-brainstem and corticospinal pathway (Extended Data Figure 2). Elephant seals and harbor seals, but not sea lions, showed robust arcuate-like connectivity between putative auditory cortex and vocal premotor cortex. The harbor seal brains showed exceptionally robust anterior cingulate to periaqueductal grey connectivity in comparison to all other brains evaluated. Finally, harbor seals showed enhanced connectivity between the anterior ventrolateral thalamus and pre-VMC, one pathway in the putative forebrain anterior vocal learning circuit (Figure 3, Extended Data Figure 3).\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis is the first evidence of a neurobehavioral phylogeny of vocal flexibility in closely related mammalian species with highly variable vocal capability, and it validates the pinnipeds as a model clade for studying the neurobiology of vocal production and learning (Figure 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA parsimonious interpretation of our findings suggests that an early ancestral pinniped evolved a direct cortical pathway between vocal motor cortex and phonatory brainstem nuclei. This preceded adaptations supporting vocal production learning that emerged selectively in other pinniped branches, but did not, by itself, allow for high degrees of vocal plasticity. Enhanced auditory-motor cortical pathways supporting developmental vocal learning may have emerged in \u003cem\u003ephocidae\u003c/em\u003e after the split from \u003cem\u003eotariidae \u0026gt;\u003c/em\u003e24 mya. Robust connections between vocal premotor cortex and anterior ventrolateral thalamus as well as enhanced anterior cingulate to PAG connections, both potentially related to enhanced vocal flexibility throughout the lifespan, may have emerged in \u003cem\u003ephocinae\u003c/em\u003e after the split with \u003cem\u003emonachinae\u003c/em\u003e. Our findings support the Kuyper-J\u0026uuml;rgens hypothesis\u003csup\u003e30\u003c/sup\u003e that direct synaptic connections between cortical vocal regions and brainstem phonatory neurons are an evolutionary precondition for vocal production flexibility. They also emphasize the importance of direct cortical auditory-motor pathways to vocal learning in mammals. Our results do not as clearly implicate anterior striatal thalamocortical circuitry in mammalian vocal plasticity, but do underline the potential contribution of the anterior ventrolateral thalamus-pre-VMC pathway to vocal mimicry. Given categorical differences in these tracts between species, our results are cautionary for over-broad claims regarding vocal production learning across all pinnipeds\u003csup\u003e4\u003c/sup\u003e. Neurobiological and behavioral data should be acquired on a broader range of pinnipeds to examine evolutionary addition and potentially subtraction of these traits.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVocal Motor Cortex to Nucleus Ambiguus Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll pinnipeds tested showed robust direct connections between cortical vocal motor regions and nucleus ambiguus without transit through the PAG, which interposes between these regions in vocally inflexible species. This pathway was not found in any of the coyotes, despite robust primary motor corticospinal connectivity through the brainstem in all brains. The cortical-ambiguus pathway we found in the pinnipeds matches the putative posterior forebrain circuit for vocal learning proposed by Jarvis\u003csup\u003e11\u003c/sup\u003e. Behaviorally, compared to terrestrial carnivores, pinnipeds do show increased vocal motor control. However, there are no data of vocal production learning (filter/phonation control) in the \u003cem\u003eotariidae\u003c/em\u003e. While the cortical-ambiguus pathway in the current study was most robust in harbor seals, all four sea lions in this study had clear bilateral cortical-nucleus ambiguus connections. Nucleus ambiguus is part of a network of medullary nuclei controlling throat, mouth, and chest musculature involved in breathing and swallowing\u003csup\u003e54,55\u003c/sup\u003e. In most terrestrial mammals, breathing is obligate based on peripheral physiological signals like blood-gas concentration and lung tension\u003csup\u003e56\u003c/sup\u003e. When adapting for amphibious life, pinnipeds would have faced intense pressure to develop breath control during submersion. In addition, in most terrestrial mammals studied, swallowing is controlled by central pattern generators in the brainstem that interface with medullary motor neurons controlling the throat and pharynx, and is wholly obligate and automatic once triggered\u003csup\u003e57\u003c/sup\u003e. Early pinnipeds must have evolved mechanisms to avoid ingesting sea water when consuming prey underwater.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe suggest that the apparent vocal motor cortical-ambiguus pathway we identified here evolved primarily to bring pinniped breathing, and potentially swallowing, under volitional control (as suggested in Ravignani et al., 2016\u003csup\u003e37\u003c/sup\u003e). This would, in turn, make control of laryngeal muscles and thoracic muscles involved in phonation subject to broadly conserved corticostriatal and corticocerebellar general motor learning mechanisms. The integration of breathing and swallowing motor control with subcortical learning mechanisms may have paved the way for subsequent vocal adaptations. However, in the otariids, this appears to be limited to context learning, not alteration of frequency and formants. Our current findings suggest that direct cortical control of medullary phonatory nuclei may be necessary, but not sufficient, for vocal production learning. Subsequent adaptations may be required to drive learning to alter vocal behavior based on auditory input. This is in line with theories suggesting that, in humans, vocal learning emerged subsequently to adaptations for breathing and swallowing control to avoid choking/food inhalation\u003csup\u003e45,58,59,60\u003c/sup\u003e. To disentangle breathing and vocal control, future studies should assess voluntary control of breathing with and without vocalization in otariids and terrestrial carnivores.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuditory Cortex to Vocal Premotor Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur findings further implicate arcuate-like auditory to premotor cortical pathways in vocal production learning in the pinnipeds. Phocids in both the \u003cem\u003emonachinae\u003c/em\u003e and \u003cem\u003ephocinae\u003c/em\u003e line, but not otariids, show some evidence of vocal learning and flexibility over development. All phocids in the current study showed clear, direct auditory to premotor connectivity, while the sea lions did not. Auditory to motor connections in birds have been observed in vocal learners and non-learners, and so have not been the focus of neurobiological assessment of vocal flexibility in avians\u003csup\u003e11\u003c/sup\u003e, although see Roberts et al. (2017)\u003csup\u003e61\u003c/sup\u003e. These pathways do show evolutionary variability in the primates\u003csup\u003e32\u003c/sup\u003e. Further, when the arcuate is severed or disrupted in humans, the ability to learn to alter vocal output based on auditory input is greatly reduced, or potentially removed entirely. These longitudinal white matter tracts have been shown to be essential to developmental learning of speech\u003csup\u003e62\u003c/sup\u003e, both through integration of the auditory perception of conspecific vocalization with the motor pattern for one\u0026rsquo;s own vocal output and for refinement of one\u0026rsquo;s own vocal output based on the auditory feedback it produces. The presence of an arcuate-like pathway in all phocids but none of the otariids in the present study underlines its potential broad relevance to mammalian vocal plasticity and learning.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDespite some evidence that otariid male calls may be distinct between colonies\u003csup\u003e63\u003c/sup\u003e, there is no evidence in these species of complex vocal production learning. The auditory-motor circuit described in phocids in the current study may have emerged or been strengthened in an early phocid to support ecologically relevant developmental vocal learning. Notably, most phocids rely far more heavily on vocal signaling in breeding than do otariids, suggesting parallel evolutionary pressures to those experienced by songbird males\u003csup\u003e64\u003c/sup\u003e. Elephant seals, a phocid species that mate on land, potentially rely on learned vocal cues for signaling to and tracking rival males during breeding season\u003csup\u003e6,65\u003c/sup\u003e. Most phocids, including harbor seals, breed in the water, where calls may serve male-male competition or potentially signaling to females\u003csup\u003e66,67\u003c/sup\u003e. These seals tend to have larger and more complex vocal repertoires that are produced seasonally\u003csup\u003e68\u003c/sup\u003e. While evidence of phocid developmental vocal learning from field studies is rare\u003csup\u003e8\u003c/sup\u003e, there is a small but growing body of laboratory studies showing some degree of developmental vocal plasticity in phocid species\u003csup\u003e42,69\u003c/sup\u003e. Some have even gone so far as to dub harbor seal vocalizations analogous to bird or whale song\u003csup\u003e70\u003c/sup\u003e. However, it must be emphasized that in the wild harbor seal vocalizations are among the most stereotypic of the phocids\u0026mdash;the case for song awaits further data, and may be stronger in other phocids and walruses\u003csup\u003e68,71\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMore data from the field and more controlled laboratory studies are required to fully characterize phocid vocal repertoires and determine the potential influence of developmental learning on call characteristics. It will be particularly important to rigorously quantify dimensions of vocal plasticity in captive animals as well as to determine when and where phocids are exposed to adult calls, and how these potential exposures relate to any subsequent vocal plasticity in the listeners. Because most \u003cem\u003ephocinae\u0026nbsp;\u003c/em\u003epups stay on the beach to nurse during the time that adult males are producing breeding calls in offshore water, it is not clear when they might be exposed to auditory models related to call learning. In addition, there is some evidence to suggest that some seals raised in isolation from conspecifics produce species-typical calls as adults\u003csup\u003e41\u003c/sup\u003e. Despite these complexities, our current finding of an arcuate-like pathway in all phocids, but not in sea lions or reliably in coyotes, supports an argument for breeding-related vocal plasticity in the phocids. In anecdotal support of this interpretation, the arcuate-like pathway we described was notably stronger bilaterally in the two male harbor seal brains in comparison to the two female harbor seal brains, mirroring findings from seminal bird studies\u003csup\u003e72\u003c/sup\u003e. In addition, all phocid subjects in the current study were juveniles, suggesting that the auditory premotor pathway is present early in brain development, potentially supporting learning from very early auditory exposures neonatally or even in utero. As male seals are far more vocal than females, future work should seek to compare auditory-premotor connections in male and female brains from a broader range of phocid species, and across a wider range of ages. Pre- and post-pubertal comparisons are of particular interest. Our findings also suggest future comparisons between terrestrial canids\u0026mdash;only two of our four coyotes showed notable auditory-premotor connectivity, and it was right lateralized. New domestic dog data indicate a robust left-lateralized auditory-premotor pathway\u003csup\u003e73\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnterior Striatal Thalamocortical Circuit\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDespite its centrality in neurobiological accounts of vocal learning in birds, there were limited interspecies differences across the anterior striatal thalamocortical circuit in the present study. The cortical-brainstem phonatory and auditory-premotor pathways may be more relevant to vocal plasticity in the pinniped line. Notably, harbor seals did show increased connectivity between the anterior ventrolateral thalamus and pre-VMC. The ventrolateral thalamus in humans, but not other primates, is essential for vocal production\u003csup\u003e14\u003c/sup\u003e, serving as a hub connecting striatal and cortical regions involved in vocal release and inhibition, learning, and vocal working memory. Thalamic-pre motor connections facilitate transition between call sub-components in finches\u003csup\u003e74\u003c/sup\u003e, and elaboration of these circuits in parrots appears to be related to their capability for vocal mimicry throughout the lifespan\u003csup\u003e13\u003c/sup\u003e. Given this, and the current findings, the anterior ventrolateral thalamus-pre-VMC pathway may play a particular role in putative vocal mimicry in the \u003cem\u003ephocinae\u003c/em\u003e line.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe \u003cem\u003ephocinae\u003c/em\u003e branch of phocids are the only pinnipeds so far demonstrated capable of mimicry, shown anecdotally in harbor seals and experimentally in grey seals. Importantly, these animals have never been observed to produce vocal mimicry in the wild, where harbor seal calls are fairly stereotypic\u003csup\u003e68\u003c/sup\u003e. More work is required to chart their abilities over the course of development. Intriguingly, despite the anecdotal and laboratory evidence indicating the importance of developmental exposure to \u003cem\u003ephocinae\u003c/em\u003e mimicry, an adult male harbor seal was also able to learn to modulate his vocal motor output to produce novel formants, and without an external auditory model\u003csup\u003e10\u003c/sup\u003e. In birds, the thalamus to pre-VMC pathway relates to learned control of phonatory musculature, not exclusively to vocal learning based on auditory input and models\u003csup\u003e75\u003c/sup\u003e. Due to their benthic feeding and suction feeding and nursing strategies, phocid seals may have enhanced mouth motor control in comparison to otariids and terrestrial carnivores. As suggested by Schusterman (2008)\u003csup\u003e40\u003c/sup\u003e, this control might also serve fine-grained vocal production flexibility, with or without auditory models. Future behavioral research should seek to examine dimensions of vocal plasticity in phocids at different developmental time points, with and without auditory models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAnterior Cingulate to PAG Pathway\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn primates, stronger direct connections between the anterior cingulate and midbrain PAG are linked with increased vocal context learning\u0026mdash;the ability not to produce novel vocalizations, but to learn to vocalize or not in specific situations and in response to specific cues\u003csup\u003e2\u003c/sup\u003e. Given the need to coordinate vocal behavior above and beneath the surface of the water, one might expect this pathway to show increased strength in pinnipeds compared to terrestrial carnivores. We did not find this to be the case in the current study; coyotes, sea lions, and elephant seals were indistinguishable in terms of path strength. Harbor seals, however, showed extremely robust anterior-cingulate to PAG connectivity. The lack of adaptation in this pathway in the sea lions and elephant seals, despite their demonstrated ability to acquire contextual control over vocal behavior, suggests that other neural circuits serve to control and organize phonation above and beneath the water\u0026mdash;perhaps the direct cortical-brainstem phonatory pathway lacking in the coyotes. The role of this pathway in harbor seals remains to be elucidated.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOur study\u0026rsquo;s limitations were primarily due to sample size\u0026ndash;more brains should be collected and analyzed, balanced for age and sex. This is a slow, arduous, necessarily collaborative, and potentially expensive process with marine mammals. Lack of common neuroanatomical atlases for these species and limited data on functional division of the cortex are further limitations\u0026ndash;these were overcome in the current study through painstaking sample-by-sample manual segmentation based on multiple lines of prior evidence and cross-validation. More baseline work on pinniped neuroanatomy would be useful. Examination of putative anterior and posterior forebrain vocal learning pathways in the walrus should also be undertaken, as they represent a distinct evolutionary branch of the pinnipeds, and both laboratory and field studies suggest they have substantial vocal flexibility\u003csup\u003e47,48\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOn the technical side, probabilistic tractography is not as reliable as chemical tracing for determining brain connectivity. However, it has been repeatedly validated, and our specialized post-mortem technique yields particularly high SNR and high resolution, without the detriments of biological noise from \u003cem\u003ein vivo\u003c/em\u003e scanning. In addition, the primary limitation of current diffusion approaches is resolving crossing fibers\u0026ndash;our methodology did not depend on digitally dissecting potentially crossing tracts but rather searching for direct pathways between delineated structures. In terms of those structures, vocal motor cortex and nucleus ambiguus were particularly central to the current study. We confidently identified and validated VMC via prior electrophysiology and histology studies cross-referenced to anatomical landmarks. However, it is important to emphasize that our cortical motor regions are likely involved with volitional control of the throat and face\u0026mdash;without invasive work or \u003cem\u003ein vivo\u003c/em\u003e imaging in awake animals, we cannot be certain which, if any, sub-components of this system are specific to vocalization. For localization of the nucleus ambiguus, we were able to rely on species-specific histology in sea lions, elephant seals, and harbor seals, but had to use domestic dog data to identify this structure in coyotes. Importantly, we found no evidence of VMC projections in the vicinity of putative nucleus ambiguus in coyotes (Extended Data Figure 9). In addition, we were not able to stain and image brains from the same individuals in each species. While we cross-validated all regions used in this study in multiple ways (see method and Extended Data Table 3), future work could use histology-to-imaging registration\u003csup\u003e76\u003c/sup\u003e, ideally within the same samples. This will require very high-resolution structural imaging to achieve necessary contrast in the brainstem.\u003c/p\u003e\n\u003cp\u003eThe nucleus ambiguus has been considered particularly difficult to find in MR datasets. Our localization was based on clear, species-specific histology, and was further validated by connectivity\u0026ndash;when traced separately, each putative pinniped nucleus ambiguus reached frontal cortex in the vicinity of putative VMC. In addition, all nucleus ambiguus regions, bilaterally, in all subjects, had robust PAG connectivity, bilateral connectivity with the other putative nucleus ambiguus, and lateral connectivity to putative vagal nerve root\u003csup\u003e55\u003c/sup\u003e. Finally, although nucleus ambiguus is in the vicinity of the reticular formation, which also has frontal projections, those transit the thalamus, which was not along the primary pathway to frontal cortex for any of the pinnipeds in the current study (Extended Data Figure 4).\u0026nbsp;\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn providing the first mammalian cladistic framework for assessing specific neurobiological adaptations supporting different aspects of vocal flexibility, our study validates and qualifies emerging behavioral evidence of vocal plasticity and learning in phocids, and opens up horizons for new research. One particularly pressing question: given strong evidence for vocal mimicry, and now, neural adaptations supporting vocal plasticity, why is harbor seal vocal behavior in the wild so apparently stereotypic? In addition to the already mentioned need to better characterize vocal and breathing behavior and control in the laboratory and the field across phocid species, it would be informative to examine whether \u003cem\u003eotariids\u003c/em\u003e can be trained to express aspects of vocal flexibility in captivity. Their cortical-brainstem phonatory pathway might theoretically support frequency plasticity. In addition, grey seals, a vocally flexible phocid relative of harbor seals have been shown to express FOXP2 in cortex\u003csup\u003e77\u003c/sup\u003e, a gene linked with motor learning and vocal learning specifically\u003csup\u003e78\u003c/sup\u003e. Given our current findings, it would be particularly interesting to look for this gene in elephant seals, harbor seals, and sea lions. Otariid vocal plasticity directly relates to another comparative research literature featuring pinnipeds and other marine mammals: rhythm\u003csup\u003e79\u003c/sup\u003e. Much has been made of the idea that adaptations supporting vocal learning also allow flexible sensorimotor synchrony\u003csup\u003e80,81\u003c/sup\u003e. However, a California sea lion is currently the most reliable, precise, and flexible non-human beat keeper\u003csup\u003e82,83\u003c/sup\u003e. In the current study, sea lions appear to lack the auditory-premotor adaptations of the other pinnipeds, and don’t have clear elaboration of the striatal thalamocortical circuit. Does the cortical vocal motor-brainstem phonatory pathway we described support flexible sensorimotor synchrony, or is synchrony largely unrelated to neural adaptations for vocal learning?\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOur findings also pave the way for parallel neurobiological studies in other species. If the correct brainstem nuclei can be precisely segmented, we could replicate our approach in toothed and baleen whales, both of which demonstrate vocal flexibility and physiological adaptations related to sound production\u003csup\u003e84,85\u003c/sup\u003e. Circuit-specific comparisons to bats, which, as mentioned previously seem to show more limited vocal plasticity, would also be informative. Finally, although our findings generally support the Kuyper-Jürgens hypothesis of the necessity of direct cortical connectivity between cortical vocal motor regions and brainstem phonatory nuclei, they also suggest that this pathway may not confer high degrees of vocal plasticity without other parallel adaptations. Given elaboration of the anterior ventrolateral thalamus to vocal motor cortex pathway in harbor seals and parrots, more work should examine development of and dynamic interactions between these regions in relation to parrot mimicry and human speech behavior. In addition, if the distinct neural circuits we examined in the current study evolved separately to subserve different aspects of vocal plasticity, then they may also have done so in early hominids before coming together into the multicomponent, integrated language learning system of extant humans\u003csup\u003e86\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eWe thank the volunteers and staff at The Marine Mammal Center in Sausalito, CA and, the USDA Wildlife Research Center in Logan, Utah for help in obtaining samples. Colleagues at the Yerkes Primate Center in Atlanta, GA assisted with coyote brain fixation and extraction. Orion Keifer at Emory University assisted with methodology for setting brains for imaging. We also thank numerous colleagues who have provided preliminary feedback on this manuscript, most notably C. Casey and P. Tyack. A special thanks to C. Reichmuth for detailed feedback on interpretation of vocal behavior data in pinnipeds. Pinniped tissues were obtained under NMFS permit 18786. Funding for P.F. Cook was provided by\u0026nbsp;Human Frontiers Science Program grant RGP0019.\u003c/p\u003e\n\u003cp\u003eAuthor Contributions\u003c/p\u003e\n\u003cp\u003eP.F.C. and G.B. conceived of the study and P.F.C., G.B., and E.S. collected data. The imaging method was developed by K.M. P.F.C. and A.R. developed and implemented the criteria and methodology for region of interest definition and segmentation. P.F.C. and A.R. analyzed the data and created figures. P.F.C. wrote the manuscript with input from all other authors.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData Availability\u003c/p\u003e\n\u003cp\u003eStructural and dMRI data for all subjects are featured on University of Oxford’s Digital Brain Zoo, and are available by request: https://open.win.ox.ac.uk/DigitalBrainBank/#/datasets/zoo\u003c/p\u003e\n\u003cp\u003eROIs used in this study were uploaded for submission to Nature’s figshare service, and are available for reviewers.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eJanik, V. M. \u0026amp; Slater, P. J. B. Vocal Learning in Mammals. in \u003cem\u003eAdvances in the Study of Behavior\u003c/em\u003e vol. 26 59\u0026ndash;99 (Elsevier, 1997).\u003c/li\u003e\n\u003cli\u003eTyack, P. L. A taxonomy for vocal learning. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e375\u003c/strong\u003e, 20180406 (2020).\u003c/li\u003e\n\u003cli\u003eVernes, S. C. \u003cem\u003eet al.\u003c/em\u003e The multi-dimensional nature of vocal learning. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e376\u003c/strong\u003e, 20200236 (2021).\u003c/li\u003e\n\u003cli\u003eWirthlin, M. E. \u003cem\u003eet al.\u003c/em\u003e Vocal learning\u0026ndash;associated convergent evolution in mammalian proteins and regulatory elements. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e383\u003c/strong\u003e, eabn3263 (2024).\u003c/li\u003e\n\u003cli\u003eCollins, K. T., Terhune, J. M., Rogers, T. L., Wheatley, K. E. \u0026amp; Harcourt, R. G. Vocal individuality of in-air Weddell seal (\u003cem\u003eLeptonychotes weddellii\u003c/em\u003e ) pup \u0026ldquo;primary\u0026rdquo; calls. \u003cem\u003e Mammal Sci.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 933\u0026ndash;951 (2006).\u003c/li\u003e\n\u003cli\u003eSanvito, S., Galimberti, F. \u0026amp; Miller, E. H. Observational Evidences of Vocal Learning in Southern Elephant Seals: a Longitudinal Study. \u003cem\u003eEthology\u003c/em\u003e \u003cstrong\u003e113\u003c/strong\u003e, 137\u0026ndash;146 (2007).\u003c/li\u003e\n\u003cli\u003eReichmuth, C. J. \u0026amp; Casey, C. Vocal learning in seals, sea lions, and walruses. \u003cem\u003e Opin. Neurobiol.\u003c/em\u003e \u003cstrong\u003e28C\u003c/strong\u003e, 66\u0026ndash;71 (2014).\u003c/li\u003e\n\u003cli\u003eStansbury, A. L. \u0026amp; Janik, V. M. The role of vocal learning in call acquisition of wild grey seal pups. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e376\u003c/strong\u003e, 20200251 (2021).\u003c/li\u003e\n\u003cli\u003eDuengen, D., Fitch, W. T. \u0026amp; Ravignani, A. Hoover the talking seal. \u003cem\u003e Biol.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, R50\u0026ndash;R52 (2023).\u003c/li\u003e\n\u003cli\u003eGoncharova, M., Jadoul, Y., Reichmuth, C., Fitch, W. T. \u0026amp; Ravignani, A. Vocal tract dynamics shape the formant structure of conditioned vocalizations in a harbor seal. \u003cem\u003e N. Y. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e1538\u003c/strong\u003e, 107\u0026ndash;116 (2024).\u003c/li\u003e\n\u003cli\u003eJarvis, E. D. Evolution of vocal learning and spoken language. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e366\u003c/strong\u003e, 50\u0026ndash;54 (2019).\u003c/li\u003e\n\u003cli\u003eJarvis, E. D. \u003cem\u003eet al.\u003c/em\u003e Behaviourally driven gene expression reveals song nuclei in hummingbird brain. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e406\u003c/strong\u003e, 628\u0026ndash;632 (2000).\u003c/li\u003e\n\u003cli\u003eZhao, Z. \u003cem\u003eet al.\u003c/em\u003e Anterior forebrain pathway in parrots is necessary for producing learned vocalizations with individual signatures. \u003cem\u003e Biol.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 5415-5426.e4 (2023).\u003c/li\u003e\n\u003cli\u003eJ\u0026uuml;rgens, U. Neuronal Control of Vocal Production in Non-Human and Human Primates. in \u003cem\u003eCurrent Topics in Primate Vocal Communication\u003c/em\u003e (eds. Zimmermann, E., Newman, J. D. \u0026amp; J\u0026uuml;rgens, U.) 199\u0026ndash;206 (Springer US, Boston, MA, 1995). doi:10.1007/978-1-4757-9930-9_10.\u003c/li\u003e\n\u003cli\u003eJ\u0026uuml;rgens, U. The Neural Control of Vocalization in Mammals: A Review. \u003cem\u003e Voice\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1\u0026ndash;10 (2009).\u003c/li\u003e\n\u003cli\u003eBriefer, E. F. Vocal expression of emotions in mammals: mechanisms of production and evidence. \u003cem\u003e Zool.\u003c/em\u003e \u003cstrong\u003e288\u003c/strong\u003e, 1\u0026ndash;20 (2012).\u003c/li\u003e\n\u003cli\u003ePark, J. \u003cem\u003eet al.\u003c/em\u003e Brainstem control of vocalization and its coordination with respiration. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e383\u003c/strong\u003e, eadi8081 (2024).\u003c/li\u003e\n\u003cli\u003eJanik, V. M. \u0026amp; Kn\u0026ouml;rnschild, M. Vocal production learning in mammals revisited. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e376\u003c/strong\u003e, 20200244 (2021).\u003c/li\u003e\n\u003cli\u003eBeecher, M. D. \u0026amp; Brenowitz, E. A. Functional aspects of song learning in songbirds. \u003cem\u003eTrends Ecol. Evol.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 143\u0026ndash;149 (2005).\u003c/li\u003e\n\u003cli\u003eJarvis, E. D. Neural systems for vocal learning in birds and humans: a synopsis. \u003cem\u003e Ornithol.\u003c/em\u003e \u003cstrong\u003e148\u003c/strong\u003e, 35\u0026ndash;44 (2007).\u003c/li\u003e\n\u003cli\u003eJourjine, N. \u0026amp; Hoekstra, H. E. Expanding evolutionary neuroscience: insights from comparing variation in behavior. \u003cem\u003eNeuron\u003c/em\u003e \u003cstrong\u003e109\u003c/strong\u003e, 1084\u0026ndash;1099 (2021).\u003c/li\u003e\n\u003cli\u003eJ\u0026uuml;rgens, U. Neural pathways underlying vocal control. \u003cem\u003e Biobehav. Rev.\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e, 235\u0026ndash;258 (2002).\u003c/li\u003e\n\u003cli\u003eGhazanfar, A. A. \u0026amp; Rendall, D. Evolution of human vocal production. \u003cem\u003e Biol.\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, R457\u0026ndash;R460 (2008).\u003c/li\u003e\n\u003cli\u003eJarvis, E. D. Learned Birdsong and the Neurobiology of Human Language. \u003cem\u003e N. Y. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e1016\u003c/strong\u003e, 749\u0026ndash;777 (2004).\u003c/li\u003e\n\u003cli\u003eKuhl, P. K. Birds and babies: Ontogeny of vocal learning. \u003cem\u003e Natl. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e, e2405626121 (2024).\u003c/li\u003e\n\u003cli\u003eHammerschmidt, K. \u0026amp; Fischer, J. Constraints in Primate Vocal Production. in \u003cem\u003eEvolution of Communicative Flexibility\u003c/em\u003e (eds. Oller, D. K. \u0026amp; Griebel, U.) 92\u0026ndash;119 (The MIT Press, 2008). doi:10.7551/mitpress/9780262151214.003.0005.\u003c/li\u003e\n\u003cli\u003eGhazanfar, A. A. \u0026amp; Liao, D. A. Constraints and flexibility during vocal development: insights from marmoset monkeys. \u003cem\u003e Opin. Behav. Sci.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 27\u0026ndash;32 (2018).\u003c/li\u003e\n\u003cli\u003eHebb, A. O. \u0026amp; Ojemann, G. A. The thalamus and language revisited. \u003cem\u003eBrain Lang.\u003c/em\u003e \u003cstrong\u003e126\u003c/strong\u003e, 99\u0026ndash;108 (2013).\u003c/li\u003e\n\u003cli\u003eSimonyan, K. The laryngeal motor cortex: its organization and connectivity. \u003cem\u003e Opin. Neurobiol.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 15\u0026ndash;21 (2014).\u003c/li\u003e\n\u003cli\u003eFitch, W. T. The Evolution of Syntax: An Exaptationist Perspective. \u003cem\u003e Evol. Neurosci.\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, (2011).\u003c/li\u003e\n\u003cli\u003eBernal, B. \u0026amp; Ardila, A. The role of the arcuate fasciculus in conduction aphasia. \u003cem\u003eBrain\u003c/em\u003e \u003cstrong\u003e132\u003c/strong\u003e, 2309\u0026ndash;2316 (2009).\u003c/li\u003e\n\u003cli\u003eRilling, J. K. \u003cem\u003eet al.\u003c/em\u003e The evolution of the arcuate fasciculus revealed with comparative DTI. \u003cem\u003e Neurosci.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 426\u0026ndash;428 (2008).\u003c/li\u003e\n\u003cli\u003eJanik, V. M. Cetacean vocal learning and communication. \u003cem\u003e Opin. Neurobiol.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 60\u0026ndash;65 (2014).\u003c/li\u003e\n\u003cli\u003ePrat, Y., Taub, M. \u0026amp; Yovel, Y. Vocal learning in a social mammal: Demonstrated by isolation and playback experiments in bats. \u003cem\u003e Adv.\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, e1500019 (2015).\u003c/li\u003e\n\u003cli\u003eGenzel, D., Desai, J., Paras, E. \u0026amp; Yartsev, M. M. Long-term and persistent vocal plasticity in adult bats. \u003cem\u003e Commun.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 3372 (2019).\u003c/li\u003e\n\u003cli\u003eVernes, S. C. \u0026amp; Wilkinson, G. S. Behaviour, biology and evolution of vocal learning in bats. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e375\u003c/strong\u003e, 20190061 (2019).\u003c/li\u003e\n\u003cli\u003eRavignani, A. \u003cem\u003eet al.\u003c/em\u003e What Pinnipeds Have to Say about Human Speech, Music, and the Evolution of Rhythm. \u003cem\u003e Neurosci.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, (2016).\u003c/li\u003e\n\u003cli\u003eBerta, A., Churchill, M. \u0026amp; Boessenecker, R. W. The Origin and Evolutionary Biology of Pinnipeds: Seals, Sea Lions, and Walruses. \u003cem\u003e Rev. Earth Planet. Sci.\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 203\u0026ndash;228 (2018).\u003c/li\u003e\n\u003cli\u003eCharrier, I. Vocal Communication in Otariids and Odobenids. in \u003cem\u003eEthology and Behavioral Ecology of Otariids and the Odobenid\u003c/em\u003e (eds. Campagna, C. \u0026amp; Harcourt, R.) 265\u0026ndash;289 (Springer International Publishing, Cham, 2021). doi:10.1007/978-3-030-59184-7_14.\u003c/li\u003e\n\u003cli\u003eSchusterman, R. J. Vocal learning in mammals with special emphasis on pinnipeds. in \u003cem\u003eThe Evolution of Communicative Flexibility: Complexity, Creativity, and Adaptability in Human and Animal Communication\u003c/em\u003e (eds. Oller, D. K. \u0026amp; Gribel, U.) 41\u0026ndash;70 (MIT Press, Cambridge, MA, 2008).\u003c/li\u003e\n\u003cli\u003eCasey, C., Sills, J. M., Knaub, S., Sotolotto, K. \u0026amp; Reichmuth, C. Lifelong Patterns of Sound Production in Two Seals. \u003cem\u003e Mamm.\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 499\u0026ndash;514 (2021).\u003c/li\u003e\n\u003cli\u003eDuengen, D., Jadoul, Y. \u0026amp; Ravignani, A. Vocal usage learning and vocal comprehension learning in harbor seals. \u003cem\u003eBMC Neurosci.\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 48 (2024).\u003c/li\u003e\n\u003cli\u003eBrainard, M. S. \u0026amp; Doupe, A. J. Auditory feedback in learning and maintenance of vocal behaviour. \u003cem\u003e Rev. Neurosci.\u003c/em\u003e \u003cstrong\u003e1\u003c/strong\u003e, 31\u0026ndash;40 (2000).\u003c/li\u003e\n\u003cli\u003eRalls, K., Fiorelli, P. \u0026amp; Gish, S. Vocalizations and vocal mimicry in captive harbor seals, Phoca vitulina. \u003cem\u003e J. Zool.\u003c/em\u003e \u003cstrong\u003e63\u003c/strong\u003e, 1050\u0026ndash;1056 (1985).\u003c/li\u003e\n\u003cli\u003eDeacon, T. W. \u003cem\u003eThe Symbolic Species: The Co-Evolution of Language and the Brain\u003c/em\u003e. (Norton, New York, NY, 1998).\u003c/li\u003e\n\u003cli\u003eStansbury, A. L. \u0026amp; Janik, V. M. Formant Modification through Vocal Production Learning in Gray Seals. \u003cem\u003e Biol.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 2244-2249.e4 (2019).\u003c/li\u003e\n\u003cli\u003eSchusterman, R. J. \u0026amp; Reichmuth, C. J. Novel sound production through contingency learning in the Pacific walrus (Odobenus rosmarus divergens). \u003cem\u003e Cogn.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 319\u0026ndash;27 (2008).\u003c/li\u003e\n\u003cli\u003eMcNab, J. A. \u003cem\u003eet al.\u003c/em\u003e High resolution diffusion-weighted imaging in fixed human brain using diffusion-weighted steady state free precession. \u003cem\u003eNeuroImage\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e, 775\u0026ndash;785 (2009).\u003c/li\u003e\n\u003cli\u003eBerns, G. S. \u003cem\u003eet al.\u003c/em\u003e Diffusion tensor imaging of dolphin brains reveals direct auditory pathway to temporal lobe. \u003cem\u003e R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e282\u003c/strong\u003e, 20151203 (2015).\u003c/li\u003e\n\u003cli\u003eTendler, B. C. \u003cem\u003eet al.\u003c/em\u003e The Digital Brain Bank, an open access platform for post-mortem imaging datasets. \u003cem\u003eeLife\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, e73153 (2022).\u003c/li\u003e\n\u003cli\u003eSawyer, E. K., Turner, E. C. \u0026amp; Kaas, J. H. Somatosensory brainstem, thalamus, and cortex of the California sea lion ( \u003cem\u003eZalophus californianus)\u003c/em\u003e. \u003cem\u003e Comp. Neurol.\u003c/em\u003e \u003cstrong\u003e524\u003c/strong\u003e, 1957\u0026ndash;1975 (2016).\u003c/li\u003e\n\u003cli\u003eRioch, D. McK. A physiological and histological study of the frontal cortex of the seal (Phoca vitulina). \u003cem\u003e Bull.\u003c/em\u003e \u003cstrong\u003e73\u003c/strong\u003e, 591\u0026ndash;602 (1937).\u003c/li\u003e\n\u003cli\u003eChiao, G. Z., Larson, C. R., Yajima, Y., Ko, P. \u0026amp; Kahrilas, P. J. Neuronal activity in nucleus ambiguus during deglutition and vocalization in conscious monkeys. \u003cem\u003e Brain Res.\u003c/em\u003e \u003cstrong\u003e100\u003c/strong\u003e, (1994).\u003c/li\u003e\n\u003cli\u003eAkunna, G. G. \u0026amp; Abah, E. D. Deciphering the Nucleus Ambiguus: Anatomy, Functions and Clinical Implications Explored. \u003cem\u003e Clin. Med. Sci.\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eNattie, E. CO2, brainstem chemoreceptors and breathing. \u003cem\u003e Neurobiol.\u003c/em\u003e \u003cstrong\u003e59\u003c/strong\u003e, 299\u0026ndash;331 (1999).\u003c/li\u003e\n\u003cli\u003eJean, A. Brain Stem Control of Swallowing: Neuronal Network and Cellular Mechanisms. \u003cem\u003e Rev.\u003c/em\u003e \u003cstrong\u003e81\u003c/strong\u003e, 929\u0026ndash;969 (2001).\u003c/li\u003e\n\u003cli\u003eHerrero, J. L., Khuvis, S., Yeagle, E., Cerf, M. \u0026amp; Mehta, A. D. Breathing above the brain stem: volitional control and attentional modulation in humans. \u003cem\u003e Neurophysiol.\u003c/em\u003e \u003cstrong\u003e119\u003c/strong\u003e, 145\u0026ndash;159 (2018).\u003c/li\u003e\n\u003cli\u003eBelyk, M. \u0026amp; Brown, S. The origins of the vocal brain in humans. \u003cem\u003e Biobehav. Rev.\u003c/em\u003e \u003cstrong\u003e77\u003c/strong\u003e, 177\u0026ndash;193 (2017).\u003c/li\u003e\n\u003cli\u003eFitch, W. T. The evolution of speech: a comparative review. \u003cem\u003eTrends Cogn. Sci.\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 258\u0026ndash;267 (2000).\u003c/li\u003e\n\u003cli\u003eRoberts, T. F. \u003cem\u003eet al.\u003c/em\u003e Identification of a motor-to-auditory pathway important for vocal learning. \u003cem\u003e Neurosci.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 978\u0026ndash;986 (2017).\u003c/li\u003e\n\u003cli\u003eL\u0026oacute;pez-Barroso, D. \u003cem\u003eet al.\u003c/em\u003e Word learning is mediated by the left arcuate fasciculus. \u003cem\u003e Natl. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e110\u003c/strong\u003e, 13168\u0026ndash;13173 (2013).\u003c/li\u003e\n\u003cli\u003eAttard, M. R. G., Pitcher, B. J., Charrier, I., Ahonen, H. \u0026amp; Harcourt, R. G. Vocal Discrimination in Mate Guarding Male Australian Sea Lions: Familiarity Breeds Contempt. \u003cem\u003eEthology\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e, 704\u0026ndash;712 (2010).\u003c/li\u003e\n\u003cli\u003eNottebohm, F. The Origins of Vocal Learning. \u003cem\u003e Nat.\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 116\u0026ndash;140 (1972).\u003c/li\u003e\n\u003cli\u003eCharrier, I. \u0026amp; Casey, C. Social Communication in Phocids. in \u003cem\u003eEthology and Behavioral Ecology of Phocids\u003c/em\u003e (eds. Costa, D. P. \u0026amp; McHuron, E. A.) 69\u0026ndash;100 (Springer International Publishing, Cham, 2022). doi:10.1007/978-3-030-88923-4_3.\u003c/li\u003e\n\u003cli\u003eStirling, I. \u0026amp; Thomas, J. A. Relationships between underwater vocalizations and mating systems in phocid seals. \u003cem\u003e Mamm.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 227\u0026ndash;246 (2003).\u003c/li\u003e\n\u003cli\u003eCassini, M. H. The evolution of reproductive systems in pinnipeds. \u003cem\u003e Ecol.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 612\u0026ndash;616 (1999).\u003c/li\u003e\n\u003cli\u003eTerhune, J. M. The underwater vocal complexity of seals (Phocidae) is not related to their phylogeny.\u0026nbsp;\u003cem\u003eCanadian Journal of Zoology\u003c/em\u003e,\u0026nbsp;\u003cstrong\u003e\u003cem\u003e97\u003c/em\u003e\u003c/strong\u003e(3), 232-240 (2019).\u003c/li\u003e\n\u003cli\u003eRavignani, A. \u003cem\u003eet al.\u003c/em\u003e Ontogeny of vocal rhythms in harbor seal pups: an exploratory study. \u003cem\u003e Zool.\u003c/em\u003e \u003cstrong\u003e65\u003c/strong\u003e, 107\u0026ndash;120 (2019).\u003c/li\u003e\n\u003cli\u003eDuengen, D. \u0026amp; Ravignani, A. The paradox of learned song in a semi‐solitary mammal. \u003cem\u003eEthology\u003c/em\u003e \u003cstrong\u003e129\u003c/strong\u003e, 445\u0026ndash;453 (2023).\u003c/li\u003e\n\u003cli\u003eNottebohm, F. \u0026amp; Arnold, A. P. Sexual Dimorphism in Vocal Control Areas of the Songbird Brain. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e194\u003c/strong\u003e, 211\u0026ndash;213 (1976).\u003c/li\u003e\n\u003cli\u003eLevin, I., Sinha, M., Barton, S. \u0026amp; Hecht, E. A left-lateralized white matter tract associated with communication in domestic dogs. \u003cem\u003e Biol.\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e, R1069\u0026ndash;R1070 (2024).\u003c/li\u003e\n\u003cli\u003eMoll, F. W. \u003cem\u003eet al.\u003c/em\u003e Thalamus drives vocal onsets in the zebra finch courtship song. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e616\u003c/strong\u003e, 132\u0026ndash;136 (2023).\u003c/li\u003e\n\u003cli\u003eMcGregor, J. N. \u003cem\u003eet al.\u003c/em\u003e Shared mechanisms of auditory and non-auditory vocal learning in the songbird brain. \u003cem\u003eeLife\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, e75691 (2022).\u003c/li\u003e\n\u003cli\u003eHuszar, I. N. \u003cem\u003eet al.\u003c/em\u003e Tensor image registration library: Deformable registration of stand‐alone histology images to whole‐brain post‐mortem MRI data. \u003cem\u003eNeuroImage\u003c/em\u003e \u003cstrong\u003e265\u003c/strong\u003e, 119792 (2023).\u003c/li\u003e\n\u003cli\u003eHoeksema, N. \u003cem\u003eet al.\u003c/em\u003e Neuroanatomy of the grey seal brain: bringing pinnipeds into the neurobiological study of vocal learning. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e376\u003c/strong\u003e, 20200252 (2021).\u003c/li\u003e\n\u003cli\u003eVargha-Khadem, F., Gadian, D. G., Copp, A. \u0026amp; Mishkin, M. FOXP2 and the neuroanatomy of speech and language. \u003cem\u003e Rev. Neurosci.\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 131\u0026ndash;138 (2005).\u003c/li\u003e\n\u003cli\u003eHersh, T. A., Ravignani, A. \u0026amp; Whitehead, H. Cetaceans are the next frontier for vocal rhythm research. \u003cem\u003e Natl. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e, e2313093121 (2024).\u003c/li\u003e\n\u003cli\u003eRouse, A. A., Patel, A. D. \u0026amp; Kao, M. H. Vocal learning and flexible rhythm pattern perception are linked: Evidence from songbirds. \u003cem\u003e Natl. Acad. Sci.\u003c/em\u003e \u003cstrong\u003e118\u003c/strong\u003e, e2026130118 (2021).\u003c/li\u003e\n\u003cli\u003ePatel, A. D. Vocal learning as a preadaptation for the evolution of human beat perception and synchronization. \u003cem\u003e Trans. R. Soc. B Biol. Sci.\u003c/em\u003e \u003cstrong\u003e376\u003c/strong\u003e, 20200326 (2021).\u003c/li\u003e\n\u003cli\u003eCook, P. F., Rouse, A. A., Wilson, M. \u0026amp; Reichmuth, C. A California sea lion (Zalophus californianus) can keep the beat: motor entrainment to rhythmic auditory stimuli in a non vocal mimic. \u003cem\u003e Comp. Psychol.\u003c/em\u003e \u003cstrong\u003e127\u003c/strong\u003e, 412\u0026ndash;27 (2013).\u003c/li\u003e\n\u003cli\u003eCook, P. F., Hood, C., Rouse, A. A. \u0026amp; Reichmuth, C. Sensorimotor Synchronization to Rhythm in an Experienced Sea Lion Rivals that of Human Subjects. \u003cem\u003e Rep.\u003c/em\u003e (In review).\u003c/li\u003e\n\u003cli\u003eElemans, C. P. H. \u003cem\u003eet al.\u003c/em\u003e Evolutionary novelties underlie sound production in baleen whales. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e627\u003c/strong\u003e, 123\u0026ndash;129 (2024).\u003c/li\u003e\n\u003cli\u003eMadsen, P. T., Siebert, U. \u0026amp; Elemans, C. P. H. Toothed whales use distinct vocal registers for echolocation and communication. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e379\u003c/strong\u003e, 928\u0026ndash;933 (2023).\u003c/li\u003e\n\u003cli\u003eFedorenko, E., Ivanova, A. A. \u0026amp; Regev, T. I. The language network as a natural kind within the broader landscape of the human brain. \u003cem\u003e Rev. Neurosci.\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 289\u0026ndash;312 (2024).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Methods","content":"\u003cp\u003eMaterials\u003c/p\u003e\n\u003cp\u003eThis study included four California sea lion brains, three elephant seal brains, four harbor seal brains, and four coyote brains. For each species there were two male and two female brains except for elephant seals, which were all male. Two sea lions were juveniles, and two were adults. All harbor seals and elephant seals were juveniles. All coyotes were adults. All brains for this study were obtained opportunistically from rehabilitation facilities and were held and imaged under appropriate permits (Extended Data Table 1). Brains were obtained opportunistically from animals euthanized for veterinary cause, and were extracted by experienced veterinarians and necropsy techs. Marine mammal brains used in this study were collected under NMFS permit 18786, and researchers conducting histology and MRI were permitted by NMFS to hold those tissues.\u0026nbsp;Pinniped brains were obtained from The Marine Mammal Center in Sausalito, CA, and were extracted immediately after veterinary euthanasia. They were placed in cold 10% buffered formalin and fixed in a cold room (~4.5 degrees C) with regular formalin refreshes for at least two weeks before being shipped to Emory University for imaging at FERN. The coyote brains were obtained from the USDA Wildlife Research Center in Logan, Utah. Following veterinary euthanasia, heads were removed and shipped on ice to Emory University. There, the skull was cracked and the entire heads were placed in buckets of cold 10% buffered formalin and kept in a cold room for a week. At that point, the brains were removed from the skull and continued fixation for a month before imaging.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll pinnipeds and coyotes had a veterinary diagnosis prior to euthanasia (Extended Data Table 1). Brains were selected from subjects with no evidence of neurological disease. Following imaging all brains were also assessed for any evidence of gross neurological disease–none was found.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThree pinniped brains were used for histological assessment of the nucleus ambiguus. The sea lion and harbor seal brain used for histology were acquired from The Marine Mammal Center under the same permit as the brains used for MRI. The harbor seal brain was archival and acquired over 40 years ago from an NIH brain collection.\u003c/p\u003e\n\u003cp\u003eBrain Imaging\u003c/p\u003e\n\u003cp\u003eAll brains for this study were imaged at Emory University’s FERN imaging center on a Siemen’s Trio 3T magnet with standard gradients and a 32-channel head receive coil. Protocol details are in Cook et al. (2018)\u003csup\u003e87\u003c/sup\u003e and Cook et al. (2022)\u003csup\u003e88\u003c/sup\u003e. In brief, for imaging pinniped brains were set in 2% agarose (Phenix Research Products Low EEO Molecular Biology Grade Agarose) doped with 2 mM gadolinium (III) oxide (Acros Organics, Fisher Scientific) to provide stability, a low-signal environment, and minimize magnetic field homogeneity. Coyote brains were placed in Fluorinert FC-3283 (3M) for the same purposes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHigh-resolution (0.6 x 0.6 x 0.5 mm) structural images were obtained for the pinnipeds using balanced steady-state free precession (SSFP) sequences (TR = 7.09 ms, TE = 3.55 ms, flip angle = 37 degrees) that were averaged together\u003csup\u003e89\u003c/sup\u003e. High-resolution (0.5 mm isotropic) T2 structural images were obtained for the coyotes using a standard turbo spin echo sequence (TR = 1000 ms, TE = 47 ms, flip angle = 180 degrees). High resolution (1 mm isotropic) diffusion images were obtained for all subjects using a specialized diffusion-weighted SSFP sequence developed specifically for post-mortem diffusion imaging\u003csup\u003e90\u003c/sup\u003e. Fifty-two directions were acquired for each subject (FOV = 128 mm, TR = 31 ms, TE = 24 ms, flip angle = 35 degrees, bandwidth = 159 Hz/pixel, q = 255 cm\u003csup\u003e-1\u003c/sup\u003e, Gmax = 38.0 mT/m, gradient duration = 15.76ms, effective b value = 3500). For each subject a reference image analogous to b=0 was obtained using the same parameters but with q = 10 cm\u003csup\u003e\u0026nbsp;-1\u003c/sup\u003e. Diffusion images were registered to the higher-resolution structural image space using FSL’s FLIRT to allow use of seeds defined on the structural images to be used for tract tracing and for display of tracing results in high resolution space.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTractography\u003c/p\u003e\n\u003cp\u003eTractography for this study was probabilistic and used a modified version of BEDPOSTX adjusted for the SSFP sequence, and modeled primary and secondary diffusion direction at each voxel to account for crossing fibers\u003csup\u003e91\u003c/sup\u003e. All tracts were based on manually defined seeds based on anatomical criteria (Extended Data Table 3). For quantitative analysis, seed to target region and seed to target region with exclusion region traces were used (Extended Data Table 2). Single seed traces were used for validation and identification of certain seeds and regions. An anatomically defined brainstem cross-section to brainstem cross-section tract was used for a tracing quality check for each brain.\u003c/p\u003e\n\u003cp\u003eFor all tracts, default values were used in FSL’s PROBTRACKX\u003csup\u003e91\u003c/sup\u003e (5,000 streamlines, step length = 0.5 mm, curvature threshold = 80 degrees). To control for differences in subject and species brain, and thus region of interest, size, quantitative values for multi-region traces were computed as a percentage of total streamlines that made it from the seed region to the target region. We took the waytotal (total number of probabilistic streamlines generated from the seed region that reached the target region) and divided by the number of voxels in the seed region. The resultant value represents the percentage of generated streamlines that reached the target and instantiates a correction for differences in size of anatomical regions between samples. All quantitative tract tracing results in the current study are drawn from two-region traces, with a seed and target region. During region segmentation and validation, a number of single-seed-region traces were conducted to cross-validate segmentations (see Extended Data Table 3). These were displayed with streamline thresholds selected to produce a coherent, dense pathway. For visualization of multi-region tracts, fsleyes render was used with the same parameters for each sample and tract (high resolution structural underlay: alpha = 75, brightness = 45, contrast = 50, display range = 80-200, gamma = 50; tracts: alpha = 100, brightness = 63, contrast = 85, gamma = 75).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eROI and seed definition\u003c/p\u003e\n\u003cp\u003eAll tracing in this study used seeds and targets that were either hand drawn voxel by voxel or constituted 3D cubes generated at a hand-selected center voxel. The criteria for drawing and generating each ROI differed by region (Figure 1, Extended Data Figure 1, Extended Data Table 3). Each region was delineated separately in the left and right hemisphere, except for PAG and the brainstem sections.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBrainstem control tract regions\u003c/p\u003e\n\u003cp\u003eWhole cross sections of brainstem were segmented caudally, just caudal to the caudal extent of the cerebellar peduncles, and rostrally, just caudal to the caudal extent of the pons (Extended Data Figure 1).\u003c/p\u003e\n\u003cp\u003ePyramids\u003c/p\u003e\n\u003cp\u003eThe descending primary motor pathway (corticospinal tract) in mammals goes through the pyramids, large structures ventromedial in the brainstem running longitudinally down the brainstem just dorsal to the pons. We overlaid primary diffusion vector directions with the high-resolution structural images and segmented the pyramids in one transverse slice just dorsal to the caudal boundary of the pons. All medial voxels with rostrocaudal directionality were included (Extended Data Figure 1)\u003c/p\u003e\n\u003cp\u003eInternal capsule\u003c/p\u003e\n\u003cp\u003eThe corticospinal tract descends through the internal capsule in mammals. We placed three-dimensional cube ROIs bilaterally by hand in the rostral internal capsule for each subject, ventral to the lateral projection of the head of the caudate and dorsal to the putamen. Locations were validated by direct connections to the putamen. Cube size was determined by internal capsule thickness, and was 5 mm\u003csup\u003e3\u003c/sup\u003e for coyotes, 7 mm\u003csup\u003e3\u003c/sup\u003e for CSLs and harbor seals, and 9 mm\u003csup\u003e3\u003c/sup\u003e for elephant seals (Extended Data Figure 1).\u003c/p\u003e\n\u003cp\u003eNucleus ambiguus\u003c/p\u003e\n\u003cp\u003eAn anatomist with expertise in pinniped neurobiology identified N. Amb histologically in the brains of a juvenile female sea lion, juvenile female elephant seal, and juvenile female harbor seal, all euthanized for veterinary cause at TMMC (Extended Data Figure 5). Sections were cut through the brainstem at 50 micrometer thickness and stained with cresyl violet. The N. Amb was identified bilaterally as a cluster of large multipolar neurons, rostrocaudally close to the level of the obex, dorsal to the inferior olive, and rostromedial to the spinal trigeminal nucleus\u003csup\u003e92\u003c/sup\u003e. High resolution histology slides were overlaid at low opacity using the program pureref (https://www.pureref.com) with high resolution trufi images for each pinniped (Figure 1, Extended Data Figure 6). The central histological slice for each animal was matched with an axial brainstem slice just rostral to the opening of the obex. The overlay was scaled and rotated to best match external brainstem countours. Because some brainstems were distorted and non-symmetrical along the right-left axis, the histological overlay was matched separately to each half (right/left) of the brainstem for segmentation. All voxels that even partially overlapped the stained cells identified in N. Amb were included in right and left segmentations. Full histological series of N. Amb rostral to caudal were available in sea lion and elephant seal (Extended Data Figure 7, Extended Data Figure 8), and were used in series to segment ambiguus in the MR data. In harbor seals, a histological slice was only available in the middle of the ambiguus. This was used to segment the corresponding brainstem section in the MR Data and on the slices immediately rostral and caudal to that slice. In coyotes, dog histology sections of brainstem were used (\u003cu\u003ehttps://vanat.ahc.umn.edu/brainsect/showAtlasFrames.html\u003c/u\u003e) with the same method as the harbor seal brains.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll segmentations were confirmed to be interposed between IO and SPVI, overlapping neither. Single seed traces were conducted with each segmentation for validation, looking for A. direct rostral connection to the PAG, B. bilateral connection between left and right N. Amb, and C. lateral connection to apparent vagus nerve (Extended Data Figure 4). Because N. Amb is close to the reticular formation, which connects directly to thalamus, segmentations were further validated by lack of direct connection to thalamus in single seed traces. Because no projections of the putative N. Amb past the midbrain were identified in single-mask tractography in any of the coyotes, we also performed single seed traces from the coyote vocal motor cortex regions and confirmed that there were no coherent pathways dorsal to the pyramidal tract at the level of the obex (Extended Data Figure 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eVocal motor cortex\u003c/p\u003e\n\u003cp\u003eIn older physiological studies, motor cortex related to production of vocal behavior has been identified in ventrolateral precuciate gyrus in harbor seal and dog and cat brains\u003csup\u003e53,93,94\u003c/sup\u003e. This location has been incidentally corroborated by more recent physiological and histological work identifying primary somatosensory cortex in sea lions between the coronal (rostral) and pseudosylvian (caudal) sulci\u003csup\u003e52\u003c/sup\u003e, and the afferent targets for mouth and neck just caudal to the caudal facing bend in the middle of the coronal sulcus. This location is, in line with conserved relationships between S1 and M1 in mammalian brains, just caudal to the putative vocal motor cortex in the current study. Of note, although this brain region is frequently referred to as “laryngeal motor cortex” in mammalian studies, we could not confirm that our ROIs were specific to the laryngeal representation as opposed to also including surrounding neck, mouth, and vocal tract representations. Therefore, we refer to our region as “vocal motor cortex” (VMC) throughout this paper.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the current study, three-dimensional cube ROIs were placed by hand for each subject on the ventral bank of the precruciate gyrus, with the center of the cube in white matter just caudal to the ventrorostral bank of the precruciate dorsolateral to the presylvian sulcus and ventrolateral to the coronal sulcus (Figure 1). Because of the more dorsoventrally constrained (“flatter”) shape of the phocid brains this location was lower by absolute coordinates in the elephant seal and harbor seal brains than in the CSL and coyote brains. To account for gross differences in brain volume, cubes for elephant seal brains were 13 voxels isotropic, 11 voxels isotropic for sea lions, 9 voxels isotropic for harbor seals, and 7 voxels isotropic for coyotes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll VMC ROIs were validated to be part of motor cortex with single mask traces looking for direct connections to the globus pallidus and, with distance correction, passage through the pyramids in the brainstem (Extended Data Figure 9).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePeriaqueductal Grey\u003c/p\u003e\n\u003cp\u003ePAG is a bilateral band of grey matter situated around the mesencephalic aqueduct, and its location and function is conserved in mammalian brains. In the current study, the PAG was segmented in the transverse plane based on anatomical criteria (Extended Data Figure 1). The grey matter/white matter boundary of the periphery of the PAG was clearly visible in the structural images for all subjects, and the structure was segmented from the rostral boundary of the posterior commissure to the caudal extent of the enclosed portion of the mesencephalic aqueduct. The dorsal raphe nucleus was also included ventrally in all segmentations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePre vocal motor cortex\u003c/p\u003e\n\u003cp\u003ePremotor cortex has been identified in carnivores with chemical tracing\u003csup\u003e95\u003c/sup\u003e and is dorsorostral to the motor cortex in the presylvian gyrus. In this study we created pre-VMC ROIs by centering three dimensional cubes (13, 11, 9, 7 mm as with VMC) in the rostral most portion of the white matter underlying presylvian gyrus along maximum projection from single-seed traces from each subjects’ VMC ROIs bilaterally (Figure 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eROIs were further validated with single seed tracings looking for direct connections to the ventrorostral thalamus and ventral striatum.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMedial geniculate nucleus\u003c/p\u003e\n\u003cp\u003eMGN has been identified in previous work with canid carnivores, including sea lions\u003csup\u003e96\u003c/sup\u003e. In the current study It was segmented anatomically in the dorsal plane for all subjects. MGN presents as a small protuberance of grey matter on the caudal ventrolateral boundary of the thalamus, just medial to the medial parahippocampal cortex.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMGN was validated for each segmentation by confirming that it was along the direct rostral projection pathway of the ipsilateral inferior colliculus (Figure 1).\u003c/p\u003e\n\u003cp\u003eInferior colliculus\u003c/p\u003e\n\u003cp\u003eIC was segmented in the dorsal plane using anatomical criteria. In all subjects it clearly presented as a dome of grey matter protruding caudally from the midbrain ventral to the superior colliculus. It was bounded rostrally by the caudal edge of the PAG (Figure 1). It was validated in each subject by projection to putative MGN.\u003c/p\u003e\n\u003cp\u003eAuditory cortex\u003c/p\u003e\n\u003cp\u003ePrimary auditory cortex’s location is broadly conserved in mammal brains\u003csup\u003e97\u003c/sup\u003e, and was identified in the current study from distance-corrected single seed tracings from MGN bilaterally. Prior physiological work has identified A1 in dog brains\u003csup\u003e98\u003c/sup\u003e in the rostrocaudal middle of the ectosylvian gyrus, just dorsal to the dorsolateral extent of the pseudosylvian gyrus, which roughly corresponds to the same position of the suprasylvian gyrus in acrtoid carnivores\u003csup\u003e99\u003c/sup\u003e. We situated three dimensional cubes bilaterally in these gyri along the strongest direct pathways from MGN for each subject (Figure 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnterior ventrolateral thalamus\u003c/p\u003e\n\u003cp\u003eAVT has not been explicitly delineated in pinniped brains previously, but has been in other carnivores, including dogs\u003csup\u003e100\u003c/sup\u003e. AVT is at the ventrorostral portion of the thalamus, situated dorsal to the subthalamic nucleus and just inside the zona incerta. It is lateral to the rostral lamina and the rostral-most thalamic nucleus which connects to prefrontal cortex. In the current study, AVT was segmented in each subject in the transverse plane based on anatomical criteria (Extended Data Figure 1). Lamina and other intrathalamic boundaries are more visible in the B0 than trufi and T2 images, so tracing was aided with overlays of B0 images registered to high resolution structural images. VRT was segmented by including the ventrorostral-most portion of the thalamus just lateral to the lamina. The ventral most boundary was the ventral portion of the thalamus. The dorsal boundary was arbitrarily but consistently set as the rostral-most extension of the rostral portion of the thalamus.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSegmentations were validated by single seed traces looking for direct rostral projection through the ventral striatum.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eVentral striatum\u003c/p\u003e\n\u003cp\u003eVentral striatum has been widely studied in carnivores\u003csup\u003e101\u003c/sup\u003e and was segmented in the transverse plane in the current study based on explicit anatomical criteria (Extended Data Figure 1). The entire rostral portion of the putamen was segmented and the head of the caudate nucleus was segmented in the same dorsoventral extent. Both regions were included in left and right striatal masks. In carnivores, the head of the caudate is enlarged and clearly visible as a bulb of grey matter caudal to the olfactory probuterances and rostral to the thalamus\u003csup\u003e101\u003c/sup\u003e. The putamen is notably smaller, but can be viewed clearly in the transverse plane as a separate island of grey matter medial to the medial most grey matter of the frontal and temporal lobe. The white matter between the caudate and the putamen and the striations interdigitating it were not included in segmentations, which were kept to the smooth coherent bodies of the basal ganglia structures.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnterior cingulate cortex\u003c/p\u003e\n\u003cp\u003eThe cingulate cortex has been identified and functionally assessed in primates, rodents, and carnivores\u003csup\u003e102\u003c/sup\u003e. It is a bundle of fibers running rostrocaudally just dorsal to the corpus callosum, with a ventral turn at the rostral extent. In the current study the anterior cingulate was identified by corroborating anatomical criteria with diffusion direction. The V1 voxel-wise diffusion maps from dtifit were overlayed on the high resolution structural images (Extended Data Figure 1). Then, cingulate was segmented in the sagittal view by following the most ventral gyri of the longitudinal fissure, including all grey matter voxels with caudal-rostral directionality immediately dorsal to the middle portion of the corpus callosum and dorsoventral directionality rostral to the rostral dip (genu) of the corpus callosum. The caudal boundary was arbitrarily but consistently defined as rostral to the rostral boundary of the thalamus.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eStatistical method\u003c/p\u003e\n\u003cp\u003eFor the eight region-to-region traces in this study, tracts were assessed quantitatively. Each pathway was assessed in each of the brains. The waytotal generated by PROBTRACKX (number of probabilistic streamlines that successfully traversed from the seed to the target region) was taken as a percentage of total streamlines generated (5000 x the number of voxels in the seed mask). This allowed direct comparison of tract strength accounting for differences in brain and region size between subjects. Because of the small sample size, comparisons were made between subjects’ tract strength nonparametrically.\u003c/p\u003e\n\u003cp\u003eA two-sided Kruskal-Wallis test was run via MATLAB for each tract by species to determine if any of the median species values differed significantly, followed by Bonferroni-corrected multiple comparisons by species.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMethods References\u003c/p\u003e\n\u003cp\u003e87.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Cook, P. F., Berns, G. S., Colegrove, K., Johnson, S. \u0026amp; Gulland, F. Postmortem DTI reveals altered hippocampal connectivity in wild sea lions diagnosed with chronic toxicosis from algal exposure. J. Comp. Neurol. 526, 216–228 (2018).\u003c/p\u003e\n\u003cp\u003e88.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Cook, P. F. \u0026amp; Berns, G. Volumetric and connectivity assessment of the caudate nucleus in California sea lions and coyotes. Anim. Cogn. 25, 1231–1240 (2022).\u003c/p\u003e\n\u003cp\u003e89.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Miller, K. L. et al. Diffusion imaging of whole, post-mortem human brains on a clinical MRI scanner. NeuroImage 57, 167–181 (2011).\u003c/p\u003e\n\u003cp\u003e90.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Miller, K. L., McNab, J. A., Jbabdi, S. \u0026amp; Douaud, G. Diffusion tractography of post-mortem human brains: Optimization and comparison of spin echo and steady-state free precession techniques. NeuroImage 59, 2284–2297 (2012).\u003c/p\u003e\n\u003cp\u003e91.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Behrens, T. E. J., Berg, H. J., Jbabdi, S., Rushworth, M. F. S. \u0026amp; Woolrich, M. W. Probabilistic diffusion tractography with multiple fibre orientations: What can we gain? NeuroImage 34, 144–155 (2007).\u003c/p\u003e\n\u003cp\u003e92.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Berman, Alvin L. The Brain Stem of the Cat; a Cytoarchitectonic Atlas with Stereotaxic Coordinates. (University of Wisconsin Press, Madison, 1968).\u003c/p\u003e\n\u003cp\u003e93.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Milojevic, B. \u0026amp; Hast, M. H. Cortical motor centers of the laryngeal muscles in the cat and dog. Ann. Otol. Rhinol. Laryngol. 73, 979–988 (1964).\u003c/p\u003e\n\u003cp\u003e94.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;De Lanerolle, N. C. \u0026amp; Lang, F. F. Functional Neural Pathways for Vocalization in the Domestic Cat. in The Physiological Control of Mammalian Vocalization (ed. Newman, J. D.) 21–41 (Springer US, Boston, MA, 1988). doi:10.1007/978-1-4613-1051-8_3.\u003c/p\u003e\n\u003cp\u003e95.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Radtke-Schuller, S. et al. Dorsal prefrontal and premotor cortex of the ferret as defined by distinctive patterns of thalamo-cortical projections. Brain Struct. Funct. 225, 1643–1667 (2020).\u003c/p\u003e\n\u003cp\u003e96.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Montie, E. W. et al. Neuroanatomy and Volumes of Brain Structures of a Live California Sea Lion ( Zalophus californianus ) From Magnetic Resonance Images. Anat. Rec. 292, 1523–1547 (2009).\u003c/p\u003e\n\u003cp\u003e97.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Krubitzer, L. The Magnificent Compromise: Cortical Field Evolution in Mammals. Neuron 56, 201–208 (2007).\u003c/p\u003e\n\u003cp\u003e98.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Andics, A., Gácsi, M., Faragó, T., Kis, A. \u0026amp; Miklósi, Á. Voice-Sensitive Regions in the Dog and Human Brain Are Revealed by Comparative fMRI. Curr. Biol. 24, 574–578 (2014).\u003c/p\u003e\n\u003cp\u003e99.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Boch, M. et al. Comparative neuroimaging of the carnivoran brain: Neocortical sulcal anatomy. Preprint at https://doi.org/10.1101/2024.06.03.597118 (2024).\u003c/p\u003e\n\u003cp\u003e100.\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Salazar, I., Ruiz Pesini, P., Fernández Alvarez, P. \u0026amp; Cifuentes, J. M. The thalamus of the dog: a tridimensional and cytoarchitectonic study. Anat. Anz. 169, 101–113 (1989).\u003c/p\u003e\n\u003cp\u003e101.\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Baer, E. et al. Predictive Methods and Probabilistic Mapping of Subcortical Brain Components in Fossil Carnivora. J. Comp. Neurol. 533, e70014 (2025).\u003c/p\u003e\n\u003cp\u003e102.\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;Devinsky, O., Morrell, M. J. \u0026amp; Vogt, B. A. Contributions of anterior cingulate cortex to behaviour. Brain 118, 279–306 (1995).\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6222519/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6222519/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe evolutionary neurobiology of mammalian vocal learning is poorly understood. Pinnipeds are among the most promising model clades for addressing this knowledge gap\u003csup\u003e1–4\u003c/sup\u003e. The whole clade has been adaptively endowed with exquisite volitional breathing control to which the \u003cem\u003emonachinae \u003c/em\u003eseals add developmental call plasticity and the \u003cem\u003ephocinae\u003c/em\u003e add the capability for formant and frequency modulation and potentially mimicry\u003csup\u003e5–10\u003c/sup\u003e. Until now, there were no comparative neurobiological data on vocal behavior in this clade. Here, using histology and \u003cem\u003eex vivo\u003c/em\u003e dMRI tractography, we provide strong first evidence for a phylogenetic spectrum of accumulative neural adaptations supporting aspects of volitional vocal control across pinniped species. Otariids and phocid seals, but not coyotes, showed robust direct cortical vocal motor connectivity to the brainstem nucleus ambiguus. This pathway may have evolved to facilitate volitional breathing, submerged prey consumption, and underwater call production for an amphibious lifestyle in a basal pinniped prior to exaptation for vocal production learning in select clades. Phocid seals, but not otariids, showed a robust arcuate-like auditory-premotor cortical pathway potentially related to developmental call learning. Harbor seals (branch \u003cem\u003ephocinae\u003c/em\u003e) showed hypertrophic connectivity in the pathway between anterior ventrolateral thalamus and vocal premotor cortex, one part of the striatal-thalamocortical anterior forebrain circuit related to vocal motor reward learning in birds\u003csup\u003e11,12\u003c/sup\u003e. Thalamic-premotor connectivity is specifically implicated in human language production and vocal mimicry in parrots\u003csup\u003e13,14\u003c/sup\u003e.\u003c/p\u003e","manuscriptTitle":"Neurobiological adaptations supporting vocal plasticity have accumulated in the marine carnivores","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-18 05:04:19","doi":"10.21203/rs.3.rs-6222519/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a192246d-fd08-46f2-8727-9fc74f54ff40","owner":[],"postedDate":"March 18th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":45704043,"name":"Biological sciences/Neuroscience/Motor control"},{"id":45704044,"name":"Biological sciences/Neuroscience/Neural circuits"},{"id":45704045,"name":"Biological sciences/Neuroscience/Sensorimotor processing"}],"tags":[],"updatedAt":"2026-03-16T16:18:07+00:00","versionOfRecord":{"articleIdentity":"rs-6222519","link":"https://doi.org/10.1126/science.adx9367","journal":{"identity":"science","isVorOnly":true,"title":"Science"},"publishedOn":"2026-03-12 00:00:00","publishedOnDateReadable":"March 12th, 2026"},"versionCreatedAt":"2025-03-18 05:04:19","video":"","vorDoi":"10.1126/science.adx9367","vorDoiUrl":"https://doi.org/10.1126/science.adx9367","workflowStages":[]},"version":"v1","identity":"rs-6222519","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6222519","identity":"rs-6222519","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}