Characterizing Vocal Repertoires--Hard vs. Soft Classification Approaches

Applying machine learning to primate bioacoustics: Review and perspectives

This paper provides a comprehensive review of the use of computational bioacoustics as well as signal and speech processing techniques in the analysis of primate vocal communication. We explore the potential implications of machine learning and deep learning methods, from the use of simple supervised algorithms to more recent self-supervised models, for processing and analyzing large data sets obtained within the emergence of passive acoustic monitoring approaches. In addition, we discuss the importance of automated primate vocalization analysis in tackling essential questions on animal communication and highlighting the role of comparative linguistics in bioacoustic research. We also examine the challenges associated with data collection and annotation and provide insights into potential solutions. Overall, this review paper runs through a set of common or innovative perspectives and applications of machine learning for primate vocal communication analysis and outlines opportunities for future research in this rapidly developing field.

Vocal repertoire and individuality in the plains zebra (Equus quagga)

Acoustic signals are vital in animal communication, and quantifying them is fundamental for understanding animal behaviour and ecology. Vocalizations can be classified into acoustically and functionally or contextually distinct categories, but establishing these categories can be challenging. Newly developed methods, such as machine learning, can provide solutions for classification tasks. The plains zebra is known for its loud and specific vocalizations, yet limited knowledge exists on the structure and information content of its vocalzations. In this study, we employed both feature-based and spectrogram-based algorithms, incorporating supervised and unsupervised machine learning methods to enhance robustness in categorizing zebra vocalization types. Additionally, we implemented a permuted discriminant function analysis to examine the individual identity information contained in the identified vocalization types. The findings revealed at least four distinct vocalization types-the 'snort', the 'soft snort', the 'squeal' and the 'quagga quagga'-with individual differences observed mostly in snorts, and to a lesser extent in squeals. Analyses based on acoustic features outperformed those based on spectrograms, but each excelled in characterizing different vocalization types. We thus recommend the combined use of these two approaches. This study offers valuable insights into plains zebra vocalization, with implications for future comprehensive explorations in animal communication.

Vocal complexity in the long calls of Bornean orangutans

Vocal complexity is central to many evolutionary hypotheses about animal communication. Yet, quantifying and comparing complexity remains a challenge, particularly when vocal types are highly graded. Male Bornean orangutans (Pongo pygmaeus wurmbii) produce complex and variable "long call" vocalizations comprising multiple sound types that vary within and among individuals. Previous studies described six distinct call (or pulse) types within these complex vocalizations, but none quantified their discreteness or the ability of human observers to reliably classify them. We studied the long calls of 13 individuals to: (1) evaluate and quantify the reliability of audio-visual classification by three well-trained observers, (2) distinguish among call types using supervised classification and unsupervised clustering, and (3) compare the performance of different feature sets. Using 46 acoustic features, we used machine learning (i.e., support vector machines, affinity propagation, and fuzzy c-means) to identify call types and assess their discreteness. We additionally used Uniform Manifold Approximation and Projection (UMAP) to visualize the separation of pulses using both extracted features and spectrogram representations. Supervised approaches showed low inter-observer reliability and poor classification accuracy, indicating that pulse types were not discrete. We propose an updated pulse classification approach that is highly reproducible across observers and exhibits strong classification accuracy using support vector machines. Although the low number of call types suggests long calls are fairly simple, the continuous gradation of sounds seems to greatly boost the complexity of this system. This work responds to calls for more quantitative research to define call types and quantify gradedness in animal vocal systems and highlights the need for a more comprehensive framework for studying vocal complexity vis-à-vis graded repertoires.

Many morphs: Parsing gesture signals from the noise

Parsing signals from noise is a general problem for signallers and recipients, and for researchers studying communicative systems. Substantial efforts have been invested in comparing how other species encode information and meaning, and how signalling is structured. However, research depends on identifying and discriminating signals that represent meaningful units of analysis. Early approaches to defining signal repertoires applied top-down approaches, classifying cases into predefined signal types. Recently, more labour-intensive methods have taken a bottom-up approach describing detailed features of each signal and clustering cases based on patterns of similarity in multi-dimensional feature-space that were previously undetectable. Nevertheless, it remains essential to assess whether the resulting repertoires are composed of relevant units from the perspective of the species using them, and redefining repertoires when additional data become available. In this paper we provide a framework that takes data from the largest set of wild chimpanzee (Pan troglodytes) gestures currently available, splitting gesture types at a fine scale based on modifying features of gesture expression using latent class analysis (a model-based cluster detection algorithm for categorical variables), and then determining whether this splitting process reduces uncertainty about the goal or community of the gesture. Our method allows different features of interest to be incorporated into the splitting process, providing substantial future flexibility across, for example, species, populations, and levels of signal granularity. Doing so, we provide a powerful tool allowing researchers interested in gestural communication to establish repertoires of relevant units for subsequent analyses within and between systems of communication.

GesturalOrigins: A bottom-up framework for establishing systematic gesture data across ape species

Current methodologies present significant hurdles to understanding patterns in the gestural communication of individuals, populations, and species. To address this issue, we present a bottom-up data collection framework for the study of gesture: GesturalOrigins. By "bottom-up", we mean that we minimise a priori structural choices, allowing researchers to define larger concepts (such as 'gesture types', 'response latencies', or 'gesture sequences') flexibly once coding is complete. Data can easily be re-organised to provide replication of, and comparison with, a wide range of datasets in published and planned analyses. We present packages, templates, and instructions for the complete data collection and coding process. We illustrate the flexibility that our methodological tool offers with worked examples of (great ape) gestural communication, demonstrating differences in the duration of action phases across distinct gesture action types and showing how species variation in the latency to respond to gestural requests may be revealed or masked by methodological choices. While GesturalOrigins is built from an ape-centred perspective, the basic framework can be adapted across a range of species and potentially to other communication systems. By making our gesture coding methods transparent and open access, we hope to enable a more direct comparison of findings across research groups, improve collaborations, and advance the field to tackle some of the long-standing questions in comparative gesture research.

Vocal complexity in a socially complex corvid: gradation, diversity and lack of common call repertoire in male rooks

Vocal communication is widespread in animals, with vocal repertoires of varying complexity. The social complexity hypothesis predicts that species may need high vocal complexity to deal with complex social organization (e.g. have a variety of different interindividual relations). We quantified the vocal complexity of two geographically distant captive colonies of rooks, a corvid species with complex social organization and cognitive performances, but understudied vocal abilities. We quantified the diversity and gradation of their repertoire, as well as the inter-individual similarity at the vocal unit level. We found that males produced call units with lower diversity and gradation than females, while song units did not differ between sexes. Surprisingly, while females produced highly similar call repertoires, even between colonies, each individual male produced almost completely different call repertoires from any other individual. These findings question the way male rooks communicate with their social partners. We suggest that each male may actively seek to remain vocally distinct, which could be an asset in their frequently changing social environment. We conclude that inter-individual similarity, an understudied aspect of vocal repertoires, should also be considered as a measure of vocal complexity.

Deep audio embeddings for vocalisation clustering

The study of non-human animals' communication systems generally relies on the transcription of vocal sequences using a finite set of discrete units. This set is referred to as a vocal repertoire, which is specific to a species or a sub-group of a species. When conducted by human experts, the formal description of vocal repertoires can be laborious and/or biased. This motivates computerised assistance for this procedure, for which machine learning algorithms represent a good opportunity. Unsupervised clustering algorithms are suited for grouping close points together, provided a relevant representation. This paper therefore studies a new method for encoding vocalisations, allowing for automatic clustering to alleviate vocal repertoire characterisation. Borrowing from deep representation learning, we use a convolutional auto-encoder network to learn an abstract representation of vocalisations. We report on the quality of the learnt representation, as well as of state of the art methods, by quantifying their agreement with expert labelled vocalisation types from 8 datasets of other studies across 6 species (birds and marine mammals). With this benchmark, we demonstrate that using auto-encoders improves the relevance of vocalisation representation which serves repertoire characterisation using a very limited number of settings. We also publish a Python package for the bioacoustic community to train their own vocalisation auto-encoders or use a pretrained encoder to browse vocal repertoires and ease unit wise annotation.

Improving the workflow to crack Small, Unbalanced, Noisy, but Genuine (SUNG) datasets in bioacoustics: The case of bonobo calls

Despite the accumulation of data and studies, deciphering animal vocal communication remains challenging. In most cases, researchers must deal with the sparse recordings composing Small, Unbalanced, Noisy, but Genuine (SUNG) datasets. SUNG datasets are characterized by a limited number of recordings, most often noisy, and unbalanced in number between the individuals or categories of vocalizations. SUNG datasets therefore offer a valuable but inevitably distorted vision of communication systems. Adopting the best practices in their analysis is essential to effectively extract the available information and draw reliable conclusions. Here we show that the most recent advances in machine learning applied to a SUNG dataset succeed in unraveling the complex vocal repertoire of the bonobo, and we propose a workflow that can be effective with other animal species. We implement acoustic parameterization in three feature spaces and run a Supervised Uniform Manifold Approximation and Projection (S-UMAP) to evaluate how call types and individual signatures cluster in the bonobo acoustic space. We then implement three classification algorithms (Support Vector Machine, xgboost, neural networks) and their combination to explore the structure and variability of bonobo calls, as well as the robustness of the individual signature they encode. We underscore how classification performance is affected by the feature set and identify the most informative features. In addition, we highlight the need to address data leakage in the evaluation of classification performance to avoid misleading interpretations. Our results lead to identifying several practical approaches that are generalizable to any other animal communication system. To improve the reliability and replicability of vocal communication studies with SUNG datasets, we thus recommend: i) comparing several acoustic parameterizations; ii) visualizing the dataset with supervised UMAP to examine the species acoustic space; iii) adopting Support Vector Machines as the baseline classification approach; iv) explicitly evaluating data leakage and possibly implementing a mitigation strategy.

Animal linguistics: a primer

The evolution of language has been investigated by several research communities, including biologists and linguists, striving to highlight similar linguistic capacities across species. To date, however, no consensus exists on the linguistic capacities of non-human species. Major controversies remain on the use of linguistic terminology, analysis methods and behavioural data collection. The field of 'animal linguistics' has emerged to overcome these difficulties and attempt to reach uniform methods and terminology. This primer is a tutorial review of 'animal linguistics'. It describes the linguistic concepts of semantics, pragmatics and syntax, and proposes minimal criteria to be fulfilled to claim that a given species displays a particular linguistic capacity. Second, it reviews relevant methods successfully applied to the study of communication in animals and proposes a list of useful references to detect and overcome major pitfalls commonly observed in the collection of animal behaviour data. This primer represents a step towards mutual understanding and fruitful collaborations between linguists and biologists.

The Ontogeny of Vocal Sequences: Insights from a Newborn Wild Chimpanzee (Pan troglodytes schweinfurthii)

Observations of early vocal behaviours in non-human primates (hereafter primates) are important for direct comparisons between human and primate vocal development. However, direct observations of births and perinatal behaviour in wild primates are rare, and the initial stages of behavioural ontogeny usually remain undocumented. Here, we report direct observations of the birth of a wild chimpanzee (Pan troglodytes schweinfurthii) in Budongo Forest, Uganda, including the behaviour of the mother and other group members. We monitored the newborn’s vocal behaviour for approximately 2 hours and recorded 70 calls. We categorised the vocalisations both qualitatively, using conventional call descriptions, and quantitatively, using cluster and discriminant acoustic analyses. We found evidence for acoustically distinct vocal units, produced both in isolation and in combination, including sequences akin to adult pant hoots, a vocal utterance regarded as the most complex vocal signal produced by this species. We concluded that chimpanzees possess the capacity to produce vocal sequences composed of different call types from birth, albeit in rudimentary forms. Our observations are in line with the idea that primate vocal repertoires are largely present from birth, with fine acoustic structures undergoing ontogenetic processes. Our study provides rare and valuable empirical data on perinatal behaviours in wild primates.

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Simple Summary

Unsupervised clustering algorithms are widely used in ecology and conservation to classify animal vocalizations, but also offer various advantages in basic research, contributing to the understanding of acoustic communication. Nevertheless, there are still some challenges to overcome. For instance, the quality of the clustering result depends on the audio transformation technique previously used to adjust the audio data. Moreover, it is difficult to verify the reliability of the clustering result. To analyze bioacoustic data using a clustering algorithm, it is, therefore, essential to select a reasonable algorithm from the many existing algorithms and prepare the recorded vocalizations so that the resulting values characterize a vocalization as accurately as possible. Frequency-modulated vocalizations, whose frequencies change over time, pose a particular problem. In this paper, we present the software CASE, which includes various clustering methods and provides an overview of their strengths and weaknesses concerning the classification of bioacoustic data. This software uses a multidimensional feature-extraction method to achieve better clustering results, especially for frequency-modulated vocalizations.

Abstract

Unsupervised clustering algorithms are widely used in ecology and conservation to classify animal sounds, but also offer several advantages in basic bioacoustics research. Consequently, it is important to overcome the existing challenges. A common practice is extracting the acoustic features of vocalizations one-dimensionally, only extracting an average value for a given feature for the entire vocalization. With frequency-modulated vocalizations, whose acoustic features can change over time, this can lead to insufficient characterization. Whether the necessary parameters have been set correctly and the obtained clustering result reliably classifies the vocalizations subsequently often remains unclear. The presented software, CASE, is intended to overcome these challenges. Established and new unsupervised clustering methods (community detection, affinity propagation, HDBSCAN, and fuzzy clustering) are tested in combination with various classifiers (k-nearest neighbor, dynamic time-warping, and cross-correlation) using differently transformed animal vocalizations. These methods are compared with predefined clusters to determine their strengths and weaknesses. In addition, a multidimensional data transformation procedure is presented that better represents the course of multiple acoustic features. The results suggest that, especially with frequency-modulated vocalizations, clustering is more applicable with multidimensional feature extraction compared with one-dimensional feature extraction. The characterization and clustering of vocalizations in multidimensional space offer great potential for future bioacoustic studies. The software CASE includes the developed method of multidimensional feature extraction, as well as all used clustering methods. It allows quickly applying several clustering algorithms to one data set to compare their results and to verify their reliability based on their consistency. Moreover, the software CASE determines the optimal values of most of the necessary parameters automatically. To take advantage of these benefits, the software CASE is provided for free download.

Investigating note repertoires and acoustic tradeoffs in the duet contributions of a basal haplorrhine primate

Acoustic communication serves a crucial role in the social interactions of vocal animals. Duetting—the coordinated singing among pairs of animals—has evolved independently multiple times across diverse taxonomic groups including insects, frogs, birds, and mammals. A crucial first step for understanding how information is encoded and transferred in duets is through quantifying the acoustic repertoire, which can reveal differences and similarities on multiple levels of analysis and provides the groundwork necessary for further studies of the vocal communication patterns of the focal species. Investigating acoustic tradeoffs, such as the tradeoff between the rate of syllable repetition and note bandwidth, can also provide important insights into the evolution of duets, as these tradeoffs may represent the physical and mechanical limits on signal design. In addition, identifying which sex initiates the duet can provide insights into the function of the duets. We have three main goals in the current study: (1) provide a descriptive, fine-scale analysis of Gursky’s spectral tarsier (Tarsius spectrumgurskyae) duets; (2) use unsupervised approaches to investigate sex-specific note repertoires; and (3) test for evidence of acoustic tradeoffs in the rate of note repetition and bandwidth of tarsier duet contributions. We found that both sexes were equally likely to initiate the duets and that pairs differed substantially in the duration of their duets. Our unsupervised clustering analyses indicate that both sexes have highly graded note repertoires. We also found evidence for acoustic tradeoffs in both male and female duet contributions, but the relationship in females was much more pronounced. The prevalence of this tradeoff across diverse taxonomic groups including birds, bats, and primates indicates the constraints that limit the production of rapidly repeating broadband notes may be one of the few ‘universals’ in vocal communication. Future carefully designed playback studies that investigate the behavioral response, and therefore potential information transmitted in duets to conspecifics, will be highly informative.

Vocal expression of emotional arousal across two call types in young rhesus macaques

As Darwin first recognized, the study of emotional communication has the potential to improve scientific understanding of the mechanisms of signal production as well as how signals evolve. We examined the relationships between emotional arousal and selected acoustic characteristics of coo and scream vocalizations produced by female rhesus macaques, Macaca mulatta, during development. For coos, arousal was assessed through measures of stress-induced elevations of plasma cortisol exhibited in response to the human intruder test. In the analysis of screams, arousal was evaluated from the intensity of aggression experienced by the vocalizer during natural social interactions. Both call types showed a positive relationship between arousal and overall fundamental frequency (F0, perceived as pitch in humans). In coos, this association was dampened over development from infancy (6 months) to the juvenile, prepubertal period (16 months) and further to menarche (21.3-31.3 months), perhaps reflecting developmental changes in physiology, anatomy and/or call function. Heightened arousal was also associated in coos with increases in an acoustic dimension related to F0 modulation and noisiness. As monkeys matured, coos showed decreases in overall F0 as well as increased noisiness and F0 modulation, likely reflecting growth of the vocal apparatus and changes in vocal fold oscillation. Within screams, only one acoustic dimension (related to F0 modulation) showed developmental change, and only within one subclass of screams within one behavioural context. Our results regarding the acoustic correlates of arousal in both call types are broadly consistent with findings in other species, supporting the hypothesis of evolutionary continuity in emotion expression. We discuss implications for broader theories of how vocal acoustics respond to selection pressures.

Comparing methodologies for classification of zebra finch distance calls

Bioacoustic analysis has been used for a variety of purposes including classifying vocalizations for biodiversity monitoring and understanding mechanisms of cognitive processes. A wide range of statistical methods, including various automated methods, have been used to successfully classify vocalizations based on species, sex, geography, and individual. A comprehensive approach focusing on identifying acoustic features putatively involved in classification is required for the prediction of features necessary for discrimination in the real world. Here, we used several classification techniques, namely discriminant function analyses (DFAs), support vector machines (SVMs), and artificial neural networks (ANNs), for sex-based classification of zebra finch (Taeniopygia guttata) distance calls using acoustic features measured from spectrograms. We found that all three methods (DFAs, SVMs, and ANNs) correctly classified the calls to respective sex-based categories with high accuracy between 92 and 96%. Frequency modulation of ascending frequency, total duration, and end frequency of the distance call were the most predictive features underlying this classification in all of our models. Our results corroborate evidence of the importance of total call duration and frequency modulation in the classification of male and female distance calls. Moreover, we provide a methodological approach for bioacoustic classification problems using multiple statistical analyses.

Parameterizing animal sounds and motion with animal-attached tags to study acoustic communication

Abstract

Stemming from the traditional use of field observers to score states and events, the study of animal behaviour often relies on analyses of discrete behavioural categories. Many studies of acoustic communication record sequences of animal sounds, classify vocalizations, and then examine how call categories are used relative to behavioural states and events. However, acoustic parameters can also convey information independent of call type, offering complementary study approaches to call classifications. Animal-attached tags can continuously sample high-resolution behavioural data on sounds and movements, which enables testing how acoustic parameters of signals relate to parameters of animal motion. Here, we present this approach through case studies on wild common bottlenose dolphins (Tursiops truncatus). Using data from sound-and-movement recording tags deployed in Sarasota (FL), we parameterized dolphin vocalizations and motion to investigate how senders and receivers modified movement parameters (including vectorial dynamic body acceleration, “VeDBA”, a proxy for activity intensity) as a function of signal parameters. We show that (1) VeDBA of one female during consortships had a negative relationship with centroid frequency of male calls, matching predictions about agonistic interactions based on motivation-structural rules; (2) VeDBA of four males had a positive relationship with modulation rate of their pulsed vocalizations, confirming predictions that click-repetition rate of these calls increases with agonism intensity. Tags offer opportunities to study animal behaviour through analyses of continuously sampled quantitative parameters, which can complement traditional methods and facilitate research replication. Our case studies illustrate the value of this approach to investigate communicative roles of acoustic parameter changes.

Significance statement

Studies of animal behaviour have traditionally relied on classification of behavioural patterns and analyses of discrete behavioural categories. Today, technologies such as animal-attached tags enable novel approaches, facilitating the use of quantitative metrics to characterize behaviour. In the field of acoustic communication, researchers typically classify vocalizations and examine usage of call categories. Through case studies of bottlenose dolphin social interactions, we present here a novel tag-based complementary approach. We used high-resolution tag data to parameterize dolphin sounds and motion, and we applied continuously sampled parameters to examine how individual dolphins responded to conspecifics’ signals and moved while producing sounds. Activity intensity of senders and receivers changed with specific call parameters, matching our predictions and illustrating the value of our approach to test communicative roles of acoustic parameter changes. Parametric approaches can complement traditional methods for animal behaviour and facilitate research replication.

Comparative Analysis of the Vocal Repertoires of the Indri (Indri indri) and the Diademed Sifaka (Propithecus diadema)

Strepsirrhine vocalisations are extraordinarily diverse and cross-species comparisons are needed to explore how this variability evolved. We contributed to the investigation of primate acoustic diversity by comparing the vocal repertoire of two sympatric lemur species, Propithecus diadema and Indri indri. These diurnal species belong to the same taxonomic family and have similar activity patterns but different social structures. These features make them excellent candidates for an investigation of the phylogenetic, environmental, and social influence on primate vocal behavior. We recorded 3 P. diadema groups in 2014 and 2016. From 1,872 recordings we selected and assigned 3814 calls to 9 a priori call types, on the basis of their acoustic structure. We implemented a reproducible technique performing an acoustic feature extraction relying on frequency bins, t-SNE data reduction, and a hard-clustering analysis. We first quantified the vocal repertoire of P. diadema, finding consistent results for the 9 putatively identified call types. When comparing this repertoire with a previously published repertoire of I. indri, we found highly species-specific repertoires, with only 2% of the calls misclassified by species identity. The loud calls of the two species were very distinct, while the low-frequency calls were more similar. Our results pinpoint the role of phylogenetic history, social and environmental features on the evolution of communicative systems and contribute to a deeper understanding of the evolutionary roots of primate vocal differentiation. We conclude by arguing that standardized and reproducible techniques, like the one we employed, allow robust comparisons and should be prioritized in the future.

A machine learning approach for classifying and quantifying acoustic diversity

1. Assessing diversity of discretely varying behavior is a classical ethological problem. In particular, the challenge of calculating an individuals' or species' vocal repertoire size is often an important step in ecological and behavioral studies, but a reproducible and broadly applicable method for accomplishing this task is not currently available. 2. We offer a generalizable method to automate the calculation and quantification of acoustic diversity using an unsupervised random forest framework. We tested our method using natural and synthetic datasets of known repertoire sizes that exhibit standardized variation in common acoustic features as well as in recording quality. We tested two approaches to estimate acoustic diversity using the output from unsupervised random forest analyses: (i) cluster analysis to estimate the number of discrete acoustic signals (e.g., repertoire size) and (ii) an estimation of acoustic area in acoustic feature space, as a proxy for repertoire size. 3. We find that our unsupervised analyses classify acoustic structure with high accuracy. Specifically, both approaches accurately estimate element diversity when repertoire size is small to intermediate (5-20 unique elements). However, for larger datasets (20-100 unique elements), we find that calculating the size of the area occupied in acoustic space is a more reliable proxy for estimating repertoire size. 4. We conclude that our implementation of unsupervised random forest analysis offers a generalizable tool that researchers can apply to classify acoustic structure of diverse datasets. Additionally, output from these analyses can be used to compare the distribution and diversity of signals in acoustic space, creating opportunities to quantify and compare the amount of acoustic variation among individuals, populations, or species in a standardized way. We provide R code and examples to aid researchers interested in using these techniques.

Discrimination of Acoustic Stimuli and Maintenance of Graded Alarm Call Structure in Captive Meerkats

Simple Summary

Preserving natural behaviors has many advantages for both research and animal welfare. Natural behaviors include producing vocalizations and responding to them. If it can be shown that the natural vocal repertoire is preserved in zoos, studies in zoos may help to expand the knowledge of acoustic behaviors and transfer it to animals in the wild. Once the meaning of diverse vocalizations is known, inferences can be made about an animal’s internal state in order to adapt and improve conditions for animals in zoos. In this paper, a natural and selective response of meerkats to potentially threatening acoustic signals such as the call of a predator is demonstrated. It can be shown that both the graded structure of meerkat alarm calls, which serves to convey the urgency of a dangerous situation, and the natural response to alarm calls are preserved. The obtained findings allow a continuation of the bioacoustic studies known for wild meerkats in zoos. The meerkat’s ability to already recognize acoustic signals as a potential threat may be crucial information for certain husbandry conditions. Vocalizing predators kept or naturally occurring near the meerkat enclosure form one example. The level of stress induced by potential threats and the associated alertness could be determined by using the graded alarm calls as a tool.

Abstract

Animals living in human care for several generations face the risk of losing natural behaviors, which can lead to reduced animal welfare. The goal of this study is to demonstrate that meerkats (Suricata suricatta) living in zoos can assess potential danger and respond naturally based on acoustic signals only. This includes that the graded information of urgency in alarm calls as well as a response to those alarm calls is retained in captivity. To test the response to acoustic signals with different threat potential, meerkats were played calls of various animals differing in size and threat (e.g., robin, raven, buzzard, jackal) while their behavior was observed. The emitted alarm calls were recorded and examined for their graded structure on the one hand and played back to them on the other hand by means of a playback experiment to see whether the animals react to their own alarm calls even in the absence of danger. A fuzzy clustering algorithm was used to analyze and classify the alarm calls. Subsequently, the features that best described the graded structure were isolated using the LASSO algorithm and compared to features already known from wild meerkats. The results show that the graded structure is maintained in captivity and can be described by features such as noise and duration. The animals respond to new threats and can distinguish animal calls that are dangerous to them from those that are not, indicating the preservation of natural cooperative behavior. In addition, the playback experiments show that the meerkats respond to their own alarm calls with vigilance and escape behavior. The findings can be used to draw conclusions about the intensity of alertness in captive meerkats and to adapt husbandry conditions to appropriate welfare.

All units are equal in humpback whale songs, but some are more equal than others

Flexible production and perception of vocalizations is linked to an impressive array of cognitive capacities including language acquisition by humans, song learning by birds, biosonar in bats, and vocal imitation by cetaceans. Here, we characterize a portion of the repertoire of one of the most impressive vocalizers in nature: the humpback whale. Qualitative and quantitative analyses of sounds (units) produced by humpback whales revealed that singers gradually morphed streams of units along multiple acoustic dimensions within songs, maintaining the continuity of spectral content across subjectively dissimilar unit "types." Singers consistently produced some unit forms more frequently and intensely than others, suggesting that units are functionally heterogeneous. The precision with which singing humpback whales continuously adjusted the acoustic characteristics of units shows that they possess exquisite vocal control mechanisms and vocal flexibility beyond what is seen in most animals other than humans. The gradual morphing of units within songs that we observed is inconsistent with past claims that humpback whales construct songs from a fixed repertoire of discrete unit types. These findings challenge the results of past studies based on fixed-unit classification methods and argue for the development of new metrics for characterizing the graded structure of units. The specific vocal variations that singers produced suggest that humpback whale songs are unlikely to provide detailed information about a singer's reproductive fitness, but can reveal the precise locations and movements of singers from long distances and may enhance the effectiveness of units as sonar signals.

Dominance style is a key predictor of vocal use and evolution across nonhuman primates

Animal communication has long been thought to be subject to pressures and constraints associated with social relationships. However, our understanding of how the nature and quality of social relationships relates to the use and evolution of communication is limited by a lack of directly comparable methods across multiple levels of analysis. Here, we analysed observational data from 111 wild groups belonging to 26 non-human primate species, to test how vocal communication relates to dominance style (the strictness with which a dominance hierarchy is enforced, ranging from 'despotic' to 'tolerant'). At the individual-level, we found that dominant individuals who were more tolerant vocalized at a higher rate than their despotic counterparts. This indicates that tolerance within a relationship may place pressure on the dominant partner to communicate more during social interactions. At the species-level, however, despotic species exhibited a larger repertoire of hierarchy-related vocalizations than their tolerant counterparts. Findings suggest primate signals are used and evolve in tandem with the nature of interactions that characterize individuals' social relationships.

Fuzzy clustering as a tool to differentiate between discrete and graded call types

Animals may communicate potential information to conspecifics using stereotyped "discrete" calls and variable "graded" calls. However, animal vocal research often centers on identifying the number of call types in a repertoire rather than quantifying the amount of gradation. Here, fuzzy clustering was applied to the social call repertoire of a species with a complex communication system, the humpback whale (Megaptera novaeangliae). Of 26 call types, 6 were classified as discrete, 7 as graded, and 13 as intermediate. These results indicate that humpback whales have a graded call repertoire, and fuzzy clustering may be a useful approach to investigate this variability.

Filling in the gaps: Acoustic gradation increases in the vocal ontogeny of chimpanzees (Pan troglodytes)

Human vocal ontogeny is considered to be a process whereby a large repertoire of discrete sounds seemingly emerges from a smaller number of acoustically graded vocalizations. While adult chimpanzee vocal behavior is highly graded, its developmental trajectory is poorly understood. In the present study, we therefore examined the size and structure of the chimpanzee vocal repertoire at different stages of ontogeny. Audio recordings were collected on infant (N = 13) and juvenile (N = 13) semi-wild chimpanzees at Chimfunshi Wildlife Orphanage, Zambia, using focal and ad libitum sampling. All observed call types were acoustically measured. These were predominantly grunts, whimpers, laughs, screams, hoos, and barks and squeaks. A range of spectral and temporal acoustic parameters were extracted, and fuzzy c-means clustering was used to quantify the size and structure of the repertoire. The infant and juvenile vocal repertoires were both best described with the same number of clusters. However, compared to infants, juvenile call clusters were less distinct from one another and could be extracted only when a low level of overlap between call clusters was permitted. Moreover, the acoustic overlap between call clusters was significantly higher for juveniles. Overall, this pattern shows greater acoustic overlap in juvenile vocalizations compared to infants, suggesting a trend toward increased acoustic gradation in chimpanzee vocal ontogeny. This may imply in contrast to humans, chimpanzees become increasingly proficient in using graded signals effectively rather than developing a larger repertoire of more discrete sounds in ontogeny.

Comparative bioacoustics: a roadmap for quantifying and comparing animal sounds across diverse taxa

Animals produce a wide array of sounds with highly variable acoustic structures. It is possible to understand the causes and consequences of this variation across taxa with phylogenetic comparative analyses. Acoustic and evolutionary analyses are rapidly increasing in sophistication such that choosing appropriate acoustic and evolutionary approaches is increasingly difficult. However, the correct choice of analysis can have profound effects on output and evolutionary inferences. Here, we identify and address some of the challenges for this growing field by providing a roadmap for quantifying and comparing sound in a phylogenetic context for researchers with a broad range of scientific backgrounds. Sound, as a continuous, multidimensional trait can be particularly challenging to measure because it can be hard to identify variables that can be compared across taxa and it is also no small feat to process and analyse the resulting high-dimensional acoustic data using approaches that are appropriate for subsequent evolutionary analysis. Additionally, terminological inconsistencies and the role of learning in the development of acoustic traits need to be considered. Phylogenetic comparative analyses also have their own sets of caveats to consider. We provide a set of recommendations for delimiting acoustic signals into discrete, comparable acoustic units. We also present a three-stage workflow for extracting relevant acoustic data, including options for multivariate analyses and dimensionality reduction that is compatible with phylogenetic comparative analysis. We then summarize available phylogenetic comparative approaches and how they have been used in comparative bioacoustics, and address the limitations of comparative analyses with behavioural data. Lastly, we recommend how to apply these methods to acoustic data across a range of study systems. In this way, we provide an integrated framework to aid in quantitative analysis of cross-taxa variation in animal sounds for comparative phylogenetic analysis. In addition, we advocate the standardization of acoustic terminology across disciplines and taxa, adoption of automated methods for acoustic feature extraction, and establishment of strong data archival practices for acoustic recordings and data analyses. Combining such practices with our proposed workflow will greatly advance the reproducibility, biological interpretation, and longevity of comparative bioacoustic studies.

Universal principles underlying segmental structures in parrot song and human speech

Despite the diversity of human languages, certain linguistic patterns are remarkably consistent across human populations. While syntactic universals receive more attention, there is stronger evidence for universal patterns in the inventory and organization of segments: units that are separated by rapid acoustic transitions which are used to build syllables, words, and phrases. Crucially, if an alien researcher investigated spoken human language how we analyze non-human communication systems, many of the phonological regularities would be overlooked, as the majority of analyses in non-humans treat breath groups, or "syllables" (units divided by silent inhalations), as the smallest unit. Here, we introduce a novel segment-based analysis that reveals patterns in the acoustic output of budgerigars, a vocal learning parrot species, that match universal phonological patterns well-documented in humans. We show that song in four independent budgerigar populations is comprised of consonant- and vowel-like segments. Furthermore, the organization of segments within syllables is not random. As in spoken human language, segments at the start of a vocalization are more likely to be consonant-like and segments at the end are more likely to be longer, quieter, and lower in fundamental frequency. These results provide a new foundation for empirical investigation of language-like abilities in other species.

Tolerant and intolerant macaques show different levels of structural complexity in their vocal communication

We tested the social complexity hypothesis which posits that animals living in complex social environments should use complex communication systems. We focused on two components of vocal complexity: diversity (number of categories of calls) and flexibility (degree of gradation between categories of calls). We compared the acoustic structure of vocal signals in groups of macaques belonging to four species with varying levels of uncertainty (i.e. complexity) in social tolerance (the higher the degree of tolerance, the higher the degree of uncertainty): two intolerant species, Japanese and rhesus macaques, and two tolerant species, Tonkean and crested macaques. We recorded the vocalizations emitted by adult females in affiliative, agonistic and neutral contexts. We analysed several acoustic variables: call duration, entropy, time and frequency energy quantiles. The results showed that tolerant macaques displayed higher levels of vocal diversity and flexibility than intolerant macaques in situations with a greater number of options and consequences, i.e. in agonistic and affiliative contexts. We found no significant differences between tolerant and intolerant macaques in the neutral context where individuals are not directly involved in social interaction. This shows that species experiencing more uncertain social interactions displayed greater vocal diversity and flexibility, which supports the social complexity hypothesis.

Brevity is not a universal in animal communication: evidence for compression depends on the unit of analysis in small ape vocalizations

Evidence for compression, or minimization of code length, has been found across biological systems from genomes to human language and music. Two linguistic laws-Menzerath's Law (which states that longer sequences consist of shorter constituents) and Zipf's Law of abbreviation (a negative relationship between signal length and frequency of use)-are predictions of compression. It has been proposed that compression is a universal in animal communication, but there have been mixed results, particularly in reference to Zipf's Law of abbreviation. Like songbirds, male gibbons (Hylobates muelleri) engage in long solo bouts with unique combinations of notes which combine into phrases. We found strong support for Menzerath's Law as the longer a phrase, the shorter the notes. To identify phrase types, we used state-of-the-art affinity propagation clustering, and were able to predict phrase types using support vector machines with a mean accuracy of 74%. Based on unsupervised phrase type classification, we did not find support for Zipf's Law of abbreviation. Our results indicate that adherence to linguistic laws in male gibbon solos depends on the unit of analysis. We conclude that principles of compression are applicable outside of human language, but may act differently across levels of organization in biological systems.

Classification of Florida manatee (Trichechus manatus latirostris) vocalizations

The vocal repertoire for the Florida manatee is quantitatively categorized from a sample of 1114 calls recorded from 3 different manatee habitats in Florida. First, manatee vocalizations were categorized into five call categories based on visual inspection of spectrograms and following descriptions provided in previous studies. Second, based on measurements of 17 acoustic parameters, the subjective classification scheme was validated using classification and regression trees (CARTs) and model-based cluster analysis paired with silhouette coefficients. CART analysis revealed that these five broad call categories can be successfully distinguished based on correct classification scores of 41.6%-62.5%. Silhouette coefficients determined that the manatee vocal repertoire is highly intergraded. This study supports and expands upon existing subjective categorization schemes by providing a quantifiable methodology for describing the Florida manatees' vocal repertoire. These findings contribute to the increasing number of studies suggesting many animal species vocal repertoires contain graded call types.

Acoustic structure of forest elephant rumbles: a test of the ambiguity reduction hypothesis

Quantitative assessments of the structure of vocalizations are a fundamental prerequisite to understand a species' vocal communication system and, more broadly, the selective pressures shaping vocal repertoires. For example, to reduce ambiguity in signal interpretation in the absence of auxiliary visual cues, species in densely vegetated habitats should exhibit more discrete vocal signals than species in open habitats. To test this "ambiguity reduction hypothesis", we conducted the first quantitative assessment of the rumble vocalizations of the forest elephant. Based on 686 forest elephant rumbles recorded with autonomous acoustic recording units at four sites across Central Africa, we used model-based cluster analyses paired with subsequent evaluation of cluster-discreteness and discriminant function analyses to quantify the structure of rumbles based on 23 source- and filter-related acoustic parameters. Model-based cluster analyses suggest that rumbles can be classified into five to eight types. Similar to previous findings in savannah elephants and contrary to the ambiguity reduction hypothesis, average silhouette coefficients below 0.34 indicated that these rumble types were highly intergraded. However, discriminant function analyses predicted rumble types with at least 75% accuracy whereby the location of the minimum fundamental frequency, middle slope and peak frequency contributed most to separation between types. In line with an increasing number of studies highlighting that a distinction between discrete and graded repertoires may have little biological significance, we propose that ambiguity reduction may take place through the evolution of perceptual and cognitive mechanisms, rather than acting on vocal production.

Finding Meanings in Low Dimensional Structures: Stochastic Neighbor Embedding Applied to the Analysis of Indri indri Vocal Repertoire

Simple Summary

The description of the vocal repertoire represents a critical step before deepening other aspects of animal behaviour. Repertoires may contain both discrete vocalizations—acoustically distinct and distinguishable from each other—or graded ones, with a less rigid acoustic structure. The gradation level is one of the causes that make repertoires challenging to be objectively quantified. Indeed, the higher the level of gradation in a system, the higher the complexity in grouping its components. A large sample of Indri indri calls was divided into ten putative categories from the acoustic similarity among them. We extracted frequency and duration parameters and then performed two different analyses that were able to group the calls accordingly to the a priori categories, indicating the presence of ten robust vocal classes. The analyses also showed a neat grouping of discrete vocalizations and a weaker classification of graded ones.

Abstract

Although there is a growing number of researches focusing on acoustic communication, the lack of shared analytic approaches leads to inconsistency among studies. Here, we introduced a computational method used to examine 3360 calls recorded from wild indris (Indri indri) from 2005–2018. We split each sound into ten portions of equal length and, from each portion we extracted spectral coefficients, considering frequency values up to 15,000 Hz. We submitted the set of acoustic features first to a t-distributed stochastic neighbor embedding algorithm, then to a hard-clustering procedure using a k-means algorithm. The t-distributed stochastic neighbor embedding (t-SNE) mapping indicated the presence of eight different groups, consistent with the acoustic structure of the a priori identification of calls, while the cluster analysis revealed that an overlay between distinct call types might exist. Our results indicated that the t-distributed stochastic neighbor embedding (t-SNE), successfully been employed in several studies, showed a good performance also in the analysis of indris’ repertoire and may open new perspectives towards the achievement of shared methodical techniques for the comparison of animal vocal repertoires.

Baboon vocal repertoires and the evolution of primate vocal diversity

A remarkable and derived trait of humans is the faculty for language, and considerable research effort has been devoted to understanding the evolution of speech. In contrast to spoken language, which constitutes a (learned) symbolic communication system, the acoustic structure of nonhuman primate vocalizations is largely genetically fixed. Yet, appreciable differences between different genera and species may exist. Environmental conditions, sexual selection, and characteristics of the social system have been invoked to explain these differences. Here, we studied the acoustic variation of call types and vocal repertoires in the genus Papio. Because the genus comprises both stable groups as well as multi-level societies, and reveals striking variation in the degree of aggressiveness from south to north, it constitutes a promising model to assess the link between social system characteristics and vocal communication. We found that, the vocal repertoires of the different species were composed of the same general call types. A quantitative analysis of the acoustic features of the grunts and loud calls of chacma (Papio ursinus), olive (P. anubis), and Guinea (P. papio) baboons showed subtle acoustic differences within call types, however. Social system characteristics did not map onto acoustic variation. We found no correlation between the structure of grunts and geographic distance; the same was true for female loud calls. Only for male loud calls from three populations, call structure varied with geographic distance. Our findings corroborate the view that the structure of nonhuman primate vocalizations is highly conserved, despite the differences in social systems. Apparently, variation in rate and intensity of occurrence of signals, probably due to different behavioral dispositions in species, are sufficient to allow for plasticity at the level of the social relationships, mating patterns, and social organization.

Cooperative breeding influences the number and type of vocalizations in avian lineages

Although communicative complexity is often predicted to correlate with social complexity in animal societies, few studies have employed large-scale comparative analyses to test whether socially complex species have more complex systems of communication. I tested this social complexity hypothesis in birds (Class: Aves) using the large amount of natural history information that describes both vocal repertoire and social system in these species. To do so, I marshalled data from primary and secondary records of avian vocal repertoires (n = 253), and for each of the species in the dataset I recorded the reported repertoire size and associated species information. Using phylogenetic comparative methods, I found that cooperative breeding was a strong and repeatable predictor of vocal repertoire size, while other social variables, e.g. group size and group stability, had little or no influence on repertoire size. Importantly, repertoire sizes expanded concurrently with the evolution of cooperative breeding, suggesting a direct link between these two traits. Cooperatively breeding species devoted significantly more of their repertoire to contact calls and alarm calls. Overall, these results therefore lend support to the hypothesis that social complexity via behavioural coordination leads to increases in vocal complexity.

Acoustic variation of spider monkeys' contact calls (whinnies) is related to distance between vocalizing individuals and immediate caller behavior

Due to several factors such as ecological conditions, group size, and social organization, primates frequently spend time out of visual contact with individuals of their own group. Through the use of long-distance vocalizations, often termed "contact calls," primates are able to maintain contact with out-of-sight individuals. Contact calls have been shown to be individually distinct, and reverberation and attenuation provide information about caller distance. It is less clear, however, whether callers actively change the structure of contact calls depending on the distance to the presumed listeners. We studied this question in spider monkeys (Ateles geoffroyi), a species with complex spatial dynamics (fission-fusion society) that produces highly frequency modulated contact calls, denominated "whinnies." We determined the acoustic characteristics of 566 whinnies recorded from 35 free-ranging spider monkeys that belong to a community located in Mexico, and used cluster analyses, discriminant function analyses, and generalized linear mixed models to assess if they varied in relation to the presumed distance to the listener. Whinnies could be grouped into five subtypes. Since the lowest frequency subtype was mainly produced by spider monkeys that exchanged whinnies at longer distances, and lower frequency calls propagate across longer distances, our results suggest that whinnies vary in order to enhance vocal contact between individuals separated by different distances. Our results also revealed that whinnies convey potential information about caller immediate behaviors and corroborated that these calls are individually distinct. Overall, our results suggest that whinny acoustic variation facilitates the maintenance of vocal contact between individuals living in a society with complex spatial dynamics.

Nonlinguistic vocalizations from online amateur videos for emotion research: A validated corpus

This study introduces a corpus of 260 naturalistic human nonlinguistic vocalizations representing nine emotions: amusement, anger, disgust, effort, fear, joy, pain, pleasure, and sadness. The recognition accuracy in a rating task varied greatly per emotion, from <40% for joy and pain, to >70% for amusement, pleasure, fear, and sadness. In contrast, the raters' linguistic-cultural group had no effect on recognition accuracy: The predominantly English-language corpus was classified with similar accuracies by participants from Brazil, Russia, Sweden, and the UK/USA. Supervised random forest models classified the sounds as accurately as the human raters. The best acoustic predictors of emotion were pitch, harmonicity, and the spacing and regularity of syllables. This corpus of ecologically valid emotional vocalizations can be filtered to include only sounds with high recognition rates, in order to study reactions to emotional stimuli of known perceptual types (reception side), or can be used in its entirety to study the association between affective states and vocal expressions (production side).

Human Non-linguistic Vocal Repertoire: Call Types and Their Meaning

Recent research on human nonverbal vocalizations has led to considerable progress in our understanding of vocal communication of emotion. However, in contrast to studies of animal vocalizations, this research has focused mainly on the emotional interpretation of such signals. The repertoire of human nonverbal vocalizations as acoustic types, and the mapping between acoustic and emotional categories, thus remain underexplored. In a cross-linguistic naming task (Experiment 1), verbal categorization of 132 authentic (non-acted) human vocalizations by English-, Swedish- and Russian-speaking participants revealed the same major acoustic types: laugh, cry, scream, moan, and possibly roar and sigh. The association between call type and perceived emotion was systematic but non-redundant: listeners associated every call type with a limited, but in some cases relatively wide, range of emotions. The speed and consistency of naming the call type predicted the speed and consistency of inferring the caller’s emotion, suggesting that acoustic and emotional categorizations are closely related. However, participants preferred to name the call type before naming the emotion. Furthermore, nonverbal categorization of the same stimuli in a triad classification task (Experiment 2) was more compatible with classification by call type than by emotion, indicating the former’s greater perceptual salience. These results suggest that acoustic categorization may precede attribution of emotion, highlighting the need to distinguish between the overt form of nonverbal signals and their interpretation by the perceiver. Both within- and between-call acoustic variation can then be modeled explicitly, bringing research on human nonverbal vocalizations more in line with the work on animal communication.

Quantifying group specificity of animal vocalizations without specific sender information

Recordings of animal vocalization can lack information about sender and context. This is often the case in studies on marine mammals or in the increasing number of automated bioacoustics monitorings. Here, we develop a framework to estimate group specificity without specific sender information. We introduce and apply a bag-of-calls-and-coefficients approach (BOCCA) to study ensembles of cepstral coefficients calculated from vocalization signals recorded from a given animal group. Comparing distributions of such ensembles of coefficients by computing relative entropies reveals group specific differences. Applying the BOCCA to ensembles of calls recorded from group of long-finned pilot whales in northern Norway, we find that differences of vocalizations within social groups of pilot whales (Globicephala melas) are significantly lower than intergroup differences.

Nocturnal "humming" vocalizations: adding a piece to the puzzle of giraffe vocal communication

BACKGROUND

Recent research reveals that giraffes (Giraffa camelopardalis sp.) exhibit a socially structured, fission-fusion system. In other species possessing this kind of society, information exchange is important and vocal communication is usually well developed. But is this true for giraffes? Giraffes are known to produce sounds, but there is no evidence that they use vocalizations for communication. Reports on giraffe vocalizations are mainly anecdotal and the missing acoustic descriptions make it difficult to establish a call nomenclature. Despite inconclusive evidence to date, it is widely assumed that giraffes produce infrasonic vocalizations similar to elephants. In order to initiate a more detailed investigation of the vocal communication in giraffes, we collected data of captive individuals during day and night. We particularly focussed on detecting tonal, infrasonic or sustained vocalizations.

FINDINGS

We collected over 947 h of audio material in three European zoos and quantified the spectral and temporal components of acoustic signals to obtain an accurate set of acoustic parameters. Besides the known burst, snorts and grunts, we detected harmonic, sustained and frequency-modulated "humming" vocalizations during night recordings. None of the recorded vocalizations were within the infrasonic range.

CONCLUSIONS

These results show that giraffes do produce vocalizations, which, based on their acoustic structure, might have the potential to function as communicative signals to convey information about the physical and motivational attributes of the caller. The data further reveal that the assumption of infrasonic communication in giraffes needs to be considered with caution and requires further investigations in future studies.