Food Calls in Common Marmosets, Callithrix jacchus, and Evidence That One Is Functionally Referential

From emotional signals to symbols

The quest for the origins of language is a diverse enterprise, where research from a variety of disciplines brings area-specific ideas and area-specific terminology to bear. This variety often results in misunderstandings and misconceptions about communication in various species. In the present paper, we argue for focus on emotional systems as the primary motivators for social signals in animals in general. This focus can help resolve discrepancies of interpretation among different areas of inquiry and can illuminate distinctions among different social signals as well as their phylogenetic origins in animals and especially in humans. We advocate, following Jaak Panksepp, a view wherein the Seeking System, the endogenous tendency to search and explore, is the most fundamental emotional motivation. The Seeking System forms the basis for flexible, voluntary, and exploratory control of motor systems and makes much of learning possible. The relative lack of vocal learning and expression in nonhuman primates contrasted with extensive vocal learning and expression in humans began, we propose, with the evolution in ancient hominins of a necessary foundation for the many subsequent capabilities required for language. That foundation was, according to the reasoning, naturally selected in the form of neurological connections between the Seeking System and mechanisms of glottal/phonatory control. The new connections allowed ancient hominins to develop flexible, endogenous vocal fitness signals produced at very high rates and including large numbers of discrete syllables, recombinable to form syllable combinations with many prosodic variations. The increasing sociality of hominins supported evolution of massive expansion in the utilization of these flexible vocal forms to allow development of words and sentences.

Development of a Marmoset Apparatus for Automated Pulling (MarmoAAP) to Study Cooperative Behaviors

In recent years, the field of neuroscience has increasingly recognized the importance of studying animal behaviors in naturalistic environments to gain deeper insights into ethologically relevant behavioral processes and neural mechanisms. The common marmoset (Callithrix jacchus), due to its small size, prosocial nature, and genetic proximity to humans, has emerged as a pivotal model toward this effort. However, traditional research methodologies often fail to fully capture the nuances of marmoset social interactions and cooperative behaviors. To address this critical gap, we developed the Marmoset Apparatus for Automated Pulling (MarmoAAP), a novel behavioral apparatus designed for studying cooperative behaviors in common marmosets. MarmoAAP addresses the limitations of traditional behavioral research methods by enabling high-throughput, detailed behavior outputs that can be integrated with video and audio recordings, allowing for more nuanced and comprehensive analyses even in a naturalistic setting. We also highlight the flexibility of MarmoAAP in task parameter manipulation which accommodates a wide range of behaviors and individual animal capabilities. Furthermore, MarmoAAP provides a platform to perform investigations of neural activity underlying naturalistic social behaviors. MarmoAAP is a versatile and robust tool for advancing our understanding of primate behavior and related cognitive processes. This new apparatus bridges the gap between ethologically relevant animal behavior studies and neural investigations, paving the way for future research in cognitive and social neuroscience using marmosets as a model organism.

Representation of conspecific vocalizations in amygdala of awake marmosets

Human speech and animal vocalizations are important for social communication and animal survival. Neurons in the auditory pathway are responsive to a range of sounds, from elementary sound features to complex acoustic sounds. For social communication, responses to distinct patterns of vocalization are usually highly specific to an individual conspecific call, in some species. This includes the specificity of sound patterns and embedded biological information. We conducted single-unit recordings in the amygdala of awake marmosets and presented calls used in marmoset communication, calls of other species and calls from specific marmoset individuals. We found that some neurons (47/262) in the amygdala distinguished 'Phee' calls from vocalizations of other animals and other types of marmoset vocalizations. Interestingly, a subset of Phee-responsive neurons (22/47) also exhibited selectivity to one out of the three Phees from two different 'caller' marmosets. Our findings suggest that, while it has traditionally been considered the key structure in the limbic system, the amygdala also represents a critical stage of socially relevant auditory perceptual processing.

Who is calling? Optimizing source identification from marmoset vocalizations with hierarchical machine learning classifiers

With their highly social nature and complex vocal communication system, marmosets are important models for comparative studies of vocal communication and, eventually, language evolution. However, our knowledge about marmoset vocalizations predominantly originates from playback studies or vocal interactions between dyads, and there is a need to move towards studying group-level communication dynamics. Efficient source identification from marmoset vocalizations is essential for this challenge, and machine learning algorithms (MLAs) can aid it. Here we built a pipeline capable of plentiful feature extraction, meaningful feature selection, and supervised classification of vocalizations of up to 18 marmosets. We optimized the classifier by building a hierarchical MLA that first learned to determine the sex of the source, narrowed down the possible source individuals based on their sex and then determined the source identity. We were able to correctly identify the source individual with high precisions (87.21%-94.42%, depending on call type, and up to 97.79% after the removal of twins from the dataset). We also examine the robustness of identification across varying sample sizes. Our pipeline is a promising tool not only for source identification from marmoset vocalizations but also for analysing vocalizations of other species.

A convergent interaction engine: vocal communication among marmoset monkeys

To understand the primate origins of the human interaction engine, it is worthwhile to focus not only on great apes but also on callitrichid monkeys (marmosets and tamarins). Like humans, but unlike great apes, callitrichids are cooperative breeders, and thus habitually engage in coordinated joint actions, for instance when an infant is handed over from one group member to another. We first explore the hypothesis that these habitual cooperative interactions, the marmoset interactional ethology, are supported by the same key elements as found in the human interaction engine: mutual gaze (during joint action), turn-taking, volubility, as well as group-wide prosociality and trust. Marmosets show clear evidence of these features. We next examine the prediction that, if such an interaction engine can indeed give rise to more flexible communication, callitrichids may also possess elaborate communicative skills. A review of marmoset vocal communication confirms unusual abilities in these small primates: high volubility and large vocal repertoires, vocal learning and babbling in immatures, and voluntary usage and control. We end by discussing how the adoption of cooperative breeding during human evolution may have catalysed language evolution by adding these convergent consequences to the great ape-like cognitive system of our hominin ancestors. This article is part of the theme issue 'Revisiting the human 'interaction engine': comparative approaches to social action coordination'.

Trade-offs between vocal accommodation and individual recognisability in common marmoset vocalizations

Recent studies find increasing evidence for vocal accommodation in nonhuman primates, indicating that this form of vocal learning is more prevalent than previously thought. Convergent vocal accommodation (i.e. becoming more similar to partners) indicates social closeness. At the same time, however, becoming too similar may compromise individual recognisability. This is especially problematic if individual recognisability is an important part of the call function, like in long-distance contact calls. In contrast, in calls with a different function, the trade-off between signalling social closeness and individual recognisability might be less severe. We therefore hypothesized that the extent and consequences of accommodation depend on the function of a given call, and expected (1) more accommodation in calls for which individual identity is less crucial and (2) that individual identity is less compromised in calls that serve mainly to transmit identity compared to calls where individual recognisability is less important. We quantified vocal accommodation in three call types over the process of pair formation in common marmoset monkeys (Callithrix jacchus, n = 20). These three call types have different functions and vary with the degree to which they refer to individual identity of the caller. In accordance with our predictions, we found that animals converged most in close contact calls (trill calls), but less in calls where individual identity is more essential (phee- and food calls). In two out of three call types, the amount of accommodation was predicted by the initial vocal distance. Moreover, accommodation led to a drop in statistical individual recognisability in trill calls, but not in phee calls and food calls. Overall, our study shows that patterns of vocal accommodation vary between call types with different functions, suggestive of trade-offs between signalling social closeness and individual recognisability in marmoset vocalizations.

Cognitive Components of Vocal Communication: A Case Study

Communication among nonhuman animals is often presented as rigid and inflexible, reflecting emotional states rather than having any cognitive basis. Using the world's smallest monkey, the pygmy marmoset (Cebuella pygmaea), with the smallest absolute brain size amongst simian primates as a case study, I review the role of cognition in the development and usage of vocalizations in pygmy marmosets and present new data on the instrumental use of babbling and of food associated vocalizations. Pygmy marmosets have several contact calls that differ in the psychoacoustic properties for sound localization as well as the distance at which they carry through the rainforest. Marmosets use these calls strategically based on distance from neighbors. Marmosets alter spectral and temporal aspects of call structure when exposed to new groups and when newly mated. They display population specific vocal dialects. Young pygmy marmosets engage in extensive babbling behavior rewarded by parents that helps the young develop adult vocal structures, but older monkeys also use babbling instrumentally in conflict situations. Specific food referential calls generally relate to food preferences, but food calls are suppressed in the presence of animate prey. Unmated animals systematically combine a long distance call with food calls as though advertising for mates. Taken together, these examples show that even small brained primates use their vocal signals flexibly and strategically in response to a variety of environmental and social conditions.